Protocol. This trial protocol has been provided by the authors to give readers additional information about their work.

Transcription

1 Protocol This trial protocol has been provided by the authors to give readers additional information about their work. Protocol for: Galiè N, Barberà JA, Frost AE, et al. Initial use of ambrisentan plus tadalafil in pulmonary arterial hypertension. N Engl J Med 2015;373: DOI: /NEJMoa

2 This supplement contains the following items: 1. (a) Original protocol, (b) final protocol with amendments. 2. Reporting analysis plan.

3 CONFIDENTIAL GM2008/00365/00 AMB Division: Worldwide Development Retention Category: GRS019 Information Type: Clinical Protocol Protocol Number: AMB (GSK); GS-US (Gilead); IND No Title: AMBITION: A Randomised, Multicenter Study of First-Line Ambrisentan and Tadalafil Combination Therapy in Subjects with Pulmonary Arterial Hypertension Compound Number: Development Phase Effective Date: Description: Subject: GSK III/IV A randomised multicenter study comparing the effect on time to clinical failure of initial combination therapy (ambrisentan and tadalafil) and initial monotherapy (ambrisentan or tadalafil) in subjects with pulmonary arterial hypertension. Pulmonary Arterial hypertension; Combination Therapy Author(s): (GSK Medical Affairs) (GSK Medical Affairs) (GSK Statistics) (GSK Clinical Operations) (Gilead Sciences Clinical Research) (Gilead Sciences Clinical Research) (Gilead Sciences Clinical Research) (Gilead Sciences Clinical Research) 1

4 CONFIDENTIAL GM2008/00365/00 AMB SPONSOR SIGNATORY: The study will be sponsored by two companies. Gilead act as regulatory sponsor in the United States (US) and GSK in the Rest of the World (RoW). For the purposes of this protocol, the term sponsor shall refer to Gilead where the study is conducted in the USA and GSK where the study is conducted anywhere else. [GSK] Vice President Global Clinical CV MDC Date MD, FACC, FAHA [Gilead] Vice President Clinical Research Head, CV/Metabolic Clinical Development Date, MD, PhD [Gilead] Vice President, Biometrics Date, PhD, MD [Gilead] Senior Vice President, Development Operations Date 2

5 SPONSORS INFORMATION PAGE: GSK CONFIDENTIAL GM2008/00365/00 AMB Clinical Study Identifier: Sponsor Address and Contact Details AMB GlaxoSmithKline Iron Bridge Road Stockley Park West, Uxbridge, Middlesex, UB11 1BU, UK Telephone: In some countries, the clinical trial sponsor may be the local GlaxoSmithKline affiliate company (or designee). Where applicable, the details of the Sponsor and contact person will be provided to the relevant regulatory authority as part of the clinical trial submission. Sponsor Medical Monitor Contact Information: Dr. Telephone: Sponsor Serious Adverse Events (SAE) Contact Information: Regulatory Agency Identifying Number(s):

6 SPONSORS INFORMATION PAGE: GILEAD CONFIDENTIAL GM2008/00365/00 AMB Clinical Study Identifier: Sponsor Address and Contact Details: GS-US Gilead Sciences, Inc. 333 Lakeside Drive Foster City, CA Telephone: Sponsor Medical Monitor Contact Information: Dr. Telephone: IND Number:

7 CONFIDENTIAL GM2008/00365/00 AMB INVESTIGATOR PROTOCOL AGREEMENT PAGE Required Standard Wording: I confirm agreement to conduct the study in compliance with the protocol. I acknowledge that I am responsible for overall study conduct. I agree to personally conduct or supervise the described clinical study. I agree to ensure that all associates, colleagues and employees assisting in the conduct of the study are informed about their obligations. Mechanisms are in place to ensure that site staff receives the appropriate information throughout the study. Investigator Name: Investigator Signature Date 5

8 CONFIDENTIAL GM2008/00365/00 AMB TABLE OF CONTENTS PAGE LIST OF ABBREVIATIONS... 9 PROTOCOL SUMMARY INTRODUCTION Background Pulmonary Arterial Hypertension Current Treatment Options Ambrisentan Clinical Efficacy Safety Tadalafil Clinical efficacy Safety Combination of ambrisentan and tadalafil Rationale Rationale for ambrisentan peak-trough 6MWD assessment OBJECTIVE(S) INVESTIGATIONAL PLAN Study Design SUBJECT SELECTION AND WITHDRAWAL CRITERIA Number of Subjects Inclusion Criteria Exclusion Criteria Withdrawal Criteria STUDY TREATMENTS Investigational Products Ambrisentan or placebo Provision, handling and storage Dosing Tadalafil or placebo Provision, handling and storage Dosing Down titration of Investigational Product Treatment following clinical failure event Blinded Combination Therapy Treatment Assignment Blinding Product Accountability Treatment Compliance Concomitant Medications and Non-Drug Therapies Permitted Medications and Non-Drug Therapies

9 CONFIDENTIAL GM2008/00365/00 AMB Prohibited Medications and Non-Drug Therapies Treatment after the End of the Study Treatment of Investigational Product Overdose Ambrisentan Tadalafil STUDY ASSESSMENTS AND PROCEDURES Critical Baseline Assessments Efficacy Primary Endpoint Time to Clinical Failure Secondary Endpoints Minute Walk Distance and Borg Dyspnea Index Borg Dyspnea Index N-Terminal pro-b-type Natriuretic Peptide WHO Functional Class assessment Exploratory Endpoints Peak-trough assessment of 6MWD Safety Electrocardiogram Vital Signs and Body Weight and Height Clinical Laboratory Tests Safety Tests Pharmacogenetics Management, Monitoring and Follow-up of Serum Aminotransferase Abnormalities Liver event follow up assessments Adverse Events Definition of an AE Definition of a SAE Laboratory and Other Safety Assessment Abnormalities Reported as AEs and SAEs Pregnancy Time Period and Frequency of Detecting AEs and SAEs Prompt Reporting of Serious Adverse Events and Other Events to Sponsor Regulatory reporting requirements for SAEs Health Outcomes Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) Health Related Quality of Life and Quality of Life Measure SF-36 Health Survey DATA MANAGEMENT DATA ANALYSIS AND STATISTICAL CONSIDERATIONS Hypotheses Study Design Considerations Sample Size Assumptions Sample Size Sensitivity

10 CONFIDENTIAL GM2008/00365/00 AMB Sample Size Re-estimation Data Analysis Considerations Analysis Populations Analysis Data Sets Treatment Comparisons Primary Comparisons of Interest Other Comparisons of Interest Interim Analysis Key Elements of Analysis Plan Efficacy Analyses Primary Analysis Secondary Analyses Exploratory Analyses Safety Analyses Health Outcomes STUDY CONDUCT CONSIDERATIONS Posting of Information on Clinicaltrials.gov Regulatory and Ethical Considerations, Including the Informed Consent Process Quality Control (Study Monitoring) Quality Assurance Study and Site Closure Records Retention Provision of Study Results and Information to Investigators Independent Data Monitoring Committee (IDMC) REFERENCES APPENDICES Appendix 1: List of Highly Effective Methods for Avoidance of Pregnancy in Women of Childbearing Potential Appendix 2: 6 Minute Walk Distance Test Appendix 3: Borg Dyspnea Index Appendix 4: Pharmacogenetic Research Appendix 5: CAMPHOR Quality of Life Questionnaire Appendix 6: SF-36 Quality of Life Questionnaire Appendix 7: Country Specific Requirements

11 CONFIDENTIAL GM2008/00365/00 AMB LIST OF ABBREVIATIONS ABS ADME AE ALT AST AUC Bid big ET-1 BDI BMI BNP BSEP BUN CAMPHOR CCB CEC C max CONMED CRF CRO CT CTD CYP DMC DSPH E2 ECE-1 ECG EEG EMA EOS Ambrisentan absorption, distribution, metabolism, and elimination adverse event alanine aminotransferase aspartate aminotransferase area under the plasma concentration-time curve twice daily proendothelin-1 Borg dyspnea index body mass index B-type natriuretic peptide bile salt export pump blood urea nitrogen Cambridge Pulmonary Hypertension Outcome Review calcium channel blockers Clinical Endpoint Committee maximum plasma concentration concomitant medication case report form clinical research organization computed axial tomography scan connective tissue disease cytochrome P450 Data Monitoring Committee Drug Safety and Public Health estradiol endothelin converting enzyme electrocardiogram electroencephalogram European Medicines Authority End of Study 9

12 ERA ET-1 ET A ET B EU CONFIDENTIAL GM2008/00365/00 AMB endothelin receptor antagonist endothelin-1 endothelin receptor type A endothelin receptor type B European Union FAV Final Assessment Visit FAV (see section 6 Table 2) FC FDA FEV 1 HPAH FSH GCP GGT hcg HIV HMG-CoA HRQL IB ICF ICH IDMC IEC INR IP ipah ITT IUD IVRS IRB iv LFT LOCF Functional Class Food and Drug Administration forced expiratory volume in 1 second Hereditary pulmonary arterial hypertension follicle stimulating hormone Good Clinical Practice gamma glutamyl transferase human chorionic gonadotropin Human Immunodeficiency Virus 5-hydroxy-3-methylglutaryl-coenzyme A Health Related Quality of Life Investigator s Brochure Informed Consent Form International Conference on Harmonisation Independent Data Monitoring Committee Independent Ethics Committee international normalized ratio Investigational product idiopathic pulmonary arterial hypertension intent to treat intrauterine device Interactive Voice Response System Institutional Review Board intravenous liver function test last-observation-carried-forward 10

13 LVEDP MedDRA m mg min mmhg mpap CONFIDENTIAL GM2008/00365/00 AMB left ventricle end diastolic pressure Medical Dictionary for Regulatory Activities meter milligram minute millimetres mercury mean pulmonary artery pressure MRP2 multidrug resistance-associated protein 2 NDA NO NT-pro-BNP NTCP NYHA OATP od PAH PAH-CTD PAP PBO PCWP New Drug Application nitric oxide N-terminal pro-b-type natriuretic peptide sodium-taurocholate co-transporter protein New York Heart Association organic anion transporter polypeptide once daily pulmonary arterial hypertension pulmonary arterial hypertension associated with connective tissue disease pulmonary artery pressure Placebo pulmonary capillary wedge pressure PDE-5 phosphodiesterase type 5 PGI 2 P-gp PH po PP PT PVOD PVR QoL RAP RHC prostacyclin P-glycoprotein pulmonary hypertension by mouth, orally Per Protocol prothrombin time Pulmonary veno-occlusive disease pulmonary vascular resistance Quality of Life right atrial pressure right heart catheterization 11

14 CONFIDENTIAL GM2008/00365/00 AMB RoW SAE SaO 2 SF-36 SOC SPC SUSAR TAD tid TLC TtCW TtCF UGT ULN US WHO 6MWD Rest of World serious adverse event arterial oxygen saturation SF-36 Health Survey System Organ Class Summary of Product Characteristics Suspected Unexpected Serious Adverse Reaction Tadalafil three times daily total lung capacity Time to Clinical Worsening Time to Clinical Failure uridine glucuronosyltransferase upper limit of normal United States World Health Organization 6-minute walk distance 12

15 CONFIDENTIAL GM2008/00365/00 AMB Trademark Information Trademarks of the GlaxoSmithKline group of companies FLOLAN Trademarks not owned by the GlaxoSmithKline group of companies VOLIBRIS TRACLEER LETAIRIS CIALIS THELIN LEVITRA VENTAVIS REMODULIN ADCIRCA 13

16 CONFIDENTIAL GM2008/00365/00 AMB PROTOCOL SUMMARY Rationale Pulmonary arterial hypertension (PAH) is a devastating, life threatening disease. A number of treatments are available, but no single drug has been demonstrated to be consistently effective in treating all patients with PAH. Combining drug products with different mechanisms of action is an evolving strategy for the treatment of PAH. This approach is supported by data from several studies in PAH demonstrating the synergistic benefits of combining two or more drugs [Humbert M et al. Eur Respir J 2004;24(3): ; Hoeper MM et al. Eur Respir J 2003;22(2): ; Ghofrani HA et al. Ann Intern Med 2002;136(7): ; Hoeper MM et al. Eur Respir J 2005; 26(5): ; Simonneau G et al. Ann Intern Med 2008; 149: ]. All of the larger clinical studies completed to date have evaluated combination therapy as an add-on approach: the addition of a second PAH therapy to a subject s monotherapy regimen. It is unknown if PAH patients may have greater improvements for sustained periods of time if combination PAH treatment is initiated as first-line therapy, rather than delaying the addition of the second therapy until the PAH status of the patient has reached a plateau or subsequently deteriorated while receiving monotherapy. Objective(s) The primary objective of this study is to compare the two treatment strategies; first-line combination therapy (ambrisentan and tadalafil) versus first-line monotherapy (ambrisentan or tadalafil) in subjects with PAH. This will be assessed by time to the first clinical failure event. The secondary objectives of this study are to compare the change in other clinical measures of PAH after initiating first-line combination therapy or first-line monotherapy, in subjects with PAH. The safety and tolerability of first-line combination therapy will be compared to first-line monotherapy. In addition, the effect of ambrisentan on exercise capacity at both peak and trough plasma concentrations will be assessed in subjects with pulmonary arterial hypertension (PAH). Study Design This Phase III/IV, randomised, double-blind study will compare the safety and efficacy of first-line combination therapy (ambrisentan and tadalafil) to first-line monotherapy (ambrisentan or tadalafil) in subjects with WHO/NYHA functional class II and III PAH. 14

17 CONFIDENTIAL GM2008/00365/00 AMB Enrolment of 352 subjects is planned with 176 subjects in the combination arm and 176 in the monotherapy arm. Subjects must have a confirmed diagnosis of PAH with documented mean pulmonary arterial pressure (mpap) 25 mmhg, pulmonary vascular resistance (PVR) 240 dyne sec/cm 5, and pulmonary capillary wedge pressure (PCWP) or left ventricle end diastolic pressure (LVEDP) of 15 mmhg based on a right heart catheterization (RHC) prior to Screening. Subjects must not have previously received chronic PAH therapy (i.e., prostanoids, ERAs, or PDE-5 inhibitors) within 4 weeks prior to Screening. Subjects must have a 6-minute walk distance (6MWD) of at least 125m and no more than 500m at the screening visit to be eligible for randomisation. Eligible subjects will be stratified based on the underlying aetiology of PAH (IPAH/HPAH and Non-IPAH) and WHO Functional Class (II and III). Subjects will be randomised 2:1:1 to either the combination therapy arm (ambrisentan and tadalafil) or to the monotherapy arm (ambrisentan and placebo or tadalafil and placebo). Subjects will also be randomly assigned in a 1:1 allocation within each treatment group to peak or trough 6MWD assessment at week 16. The target doses of the study medication will be 10mg ambrisentan (ABS) once daily and 40mg tadalafil (TAD) once daily. Combination therapy arm: Subjects randomised to combination treatment will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see section for details) Monotherapy arm: ambrisentan group: Subjects randomised to ambrisentan monotherapy will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter Two tablets of TAD-matching placebo 15

18 CONFIDENTIAL GM2008/00365/00 AMB Monotherapy arm: tadalafil group: Subjects randomised to tadalafil monotherapy will receive both Two tablets of ABS-matching placebo One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see section for details). Subjects will be assessed for efficacy and safety at Screening, Randomisation, Weeks 4, 8, 16, 24, and every 12 weeks thereafter. Subjects will also have monthly liver function and pregnancy safety assessments, which may be performed by a local phlebotomy laboratory or at the Investigator clinic (or by the central lab with Sponsor permission). The primary endpoint of the study is time to the first clinical failure event. Criterion for clinical failure will be adjudicated by a blinded study-specific Clinical Endpoint Committee (CEC). Other than the protocol defined dose adjustments of Investigational Product, PAH therapies may not be added unless a subject has experienced a clinical failure event based on the given criteria. Additional PAH therapies (other than ambrisentan and tadalafil) that are necessary for treatment will not be supplied by the sponsor. All subjects are eligible to receive a minimum of 24 weeks of therapy. At the time that 82 adjudicated primary endpoint events have occurred, i.e. 82 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety assessment will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed. An Independent Data Monitoring Committee (IDMC) will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will also review formal interim analysis and recommend continuation or early termination of the study based on stopping criteria defined in the IDMC charter. 16

19 Study Endpoints/Assessments CONFIDENTIAL GM2008/00365/00 AMB The primary efficacy endpoint is the time (days) to the first clinical failure event of PAH. Time to clinical failure (TtCF) is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non-elective hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III or IV symptoms for 6 months Time to clinical worsening (death, hospitalisation for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Medical Interventions for PAH During the Study: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary endpoint event and subjects may continue in the study following initiation. If a subject experiences an event of clinical failure the following medical interventions may also be utilized. The clinical failure event necessary for medical intervention is based on the opinion of the investigator and does not require prior adjudication from the clinical endpoint committee. The investigator may choose to request blinded combination therapy titration to ensure that the subject is receiving ambrisentan/tadalafil combination therapy The investigator may choose to add prostanoid therapy (parenteral or nonparenteral) (see section 5.1.4). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first 17

20 CONFIDENTIAL GM2008/00365/00 AMB intervention (see section ) the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy Initiate parenteral prostanoid therapy Request blinded ambrisentan/tadalafil combination therapy. Secondary endpoints include: Change from baseline 6MWD measured at week 24 Percentage of subjects with satisfactory clinical response measured at Week 24, defined as: o 10% improvement in 6MWD compared to baseline o Improvement or maintenance of WHO class I or II symptoms o No events of clinical worsening prior to or at the week 24 visit Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP) Change from baseline measured at week 24 in WHO Functional Class Change from baseline measured at week 24 in BDI immediately following exercise Exploratory endpoints include: An assessment of the effects of ambrisentan on 6MWD at peak and trough plasma concentrations, which includes: o The ambrisentan placebo-corrected mean change from baseline 6MWD following 16 weeks of treatment (i.e., tadalafil/ambrisentan combination therapy versus tadalafil/placebo monotherapy) o The ratio of the placebo-corrected mean change from baseline trough 6MWD at week 16 to the placebo-corrected mean change from baseline peak 6MWD at week 16. Health outcomes endpoints include Change from baseline to week 24 in Quality of Life: as measured by the Short Form 36 Health Survey (SF-36) & the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) (selected countries where language translation is validated) Safety endpoints include Incidence and severity of adverse events Vital signs (i.e. blood pressure and heart rate). Laboratory test: clinical chemistry, haematology, testicular function (males only) 18

21 CONFIDENTIAL GM2008/00365/00 AMB The intent to treat (ITT) population will include all subjects randomised to treatment who receive at least one dose of Investigational Product, analyzed according to randomised treatment group. This will be the primary analysis population for assessing efficacy and safety. A per protocol (PP) population will also be defined and used to alternatively describe the primary and the most important secondary and tertiary efficacy endpoints. The PP population will consist of the subset of subjects contained in the ITT without any major protocol violation. If the PP population is greater than 85% or less than 50% of the ITT population, a PP analysis will not be performed. The CEC will provide adjudication of primary and secondary clinical events reported during the study. The members of this committee will be blinded to treatment assignment and investigator. This event-driven study requires 82 subjects with a clinical failure event in order to have 90% power with an overall type I error rate (alpha level) of 5%, accounting for two interim analyses as detailed in Section Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of 10% per year, i.e. an absolute difference of 10%. This leads to an approximate hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial. Alpha spending associated with the stopping criteria are described in Section The study will require 352 subjects (176 per treatment group) assuming a recruitment period of 65 weeks (based on an average recruitment rate of 5.4 subjects per week) and a total study duration of 129 weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 10% per treatment group at one year has been considered for this calculation. The primary comparison will be between the combination therapy arm and the pooled monotherapy arm. For the primary analysis, time to the first clinical failure event of PAH, the log rank test will be used to compare the two treatment arms. Time to clinical failure will be displayed by treatment group as a Kaplan-Meier event-free curve from randomisation up to the Final Assessment Visit (FAV). Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using a Cox proportional hazards regression model. The two interim analyses are planned after 50% and 75% of the primary endpoints events have been reached. Statistical stopping rules are in place to stop the study early for either superiority or futility. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints. Inferences on the superiority of combination versus monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint is statistically significant will 19

22 CONFIDENTIAL GM2008/00365/00 AMB inferences on the Week 24 change from baseline 6MWD endpoint worsening be evaluated. Only if the Week 24 change from baseline 6MWD endpoint is found to be significant will inferences on the percentage of subjects with satisfactory response at Week 24 endpoint be evaluated. This gate keeping approach, described above, will be implemented for all secondary endpoints in the order listed above. With this strategy, all tests will be based on a two-side p-value at a 5% significance level. 20

23 CONFIDENTIAL GM2008/00365/00 AMB INTRODUCTION 1.1. Background Pulmonary Arterial Hypertension PAH consists of a group of diseases characterized by a progressive increase of pulmonary vascular resistance (PVR) and pulmonary artery pressure (PAP). Group 1 PAH includes idiopathic pulmonary arterial hypertension (ipah), hereditary pulmonary arterial hypertension (hpah), and PAH associated with various conditions such as connective tissue disease (CTD), congenital systemic-to-pulmonary shunts, portal hypertension, anorexigen use, and human immunodeficiency virus (HIV) infection [Galie et al. Eur Respir J 2009; 34(6): ]. All of these disorders share similar proliferative and obstructive changes of the pulmonary microcirculation, including plexiform lesions [Galie et al. Eur Respir J 2009; 34(6): ]. The vascular damage that defines PAH is characterized by a mean pulmonary artery pressure (mpap) 25 mmhg at rest and a pulmonary capillary wedge pressure (PCWP) 15 mmhg. Patients with PAH typically develop significant increases in PVR and sustained elevations in PAP, which ultimately lead to right ventricular failure and death [Rubin LJ. Chest 1993; 104(1): ; Galie et al. Eur Respir J 2009; 34(6): ]. Patients diagnosed with PAH have a poor prognosis and equally compromised quality of life, with a mean life expectancy of 2 to 5 years from the time of diagnosis if untreated [D Alonzo GE et al. Ann Intern Med 1991; 115(5):343-9]. PAH medicines that block or reduce the vasoconstriction and hypertrophy associated with the narrowing of the pulmonary arteries may ameliorate the cycle of constriction and improve pulmonary vascular blood flow Current Treatment Options Diuretics and anticoagulants have been widely used in the management of PAH, yet response varies [Humbert M, et al. N Engl J Med 2004; 351(14): ]. Calcium channel blockers (CCBs) have shown improved survival in vasoreactive IPAH patients, yet the relatively low incidence of vasoreactivity make CCBs useful only in a minority of the population with PAH [Humbert M, et al. N Engl J Med 2004; 351(14): ; Rich S et al. N Engl J Med 1992; 327(2):76-81]. Three signalling pathways involved in the pathogenesis of PAH have been targeted for therapeutic intervention [Humbert M, et al. N Engl J Med 2004; 351(14): ]: the camp-dependent prostacyclin (PGI 2 ) pathway, the cgmp-dependent nitric oxide (NO) pathway, and the phospholipase-c-dependent endothelin pathway. These pathways are targeted by the following classes of PAH medicines: prostacyclin derivatives (Flolan (epoprostenol), Ventavis (iloprost) and Remodulin (terprostinil)), phosphodiesterase type-5 (PDE-5) inhibitors (Revatio (sildenafil) and Adcirca (tadalafil)), and endothelin receptor antagonists (ERAs) (Volibris [EU]/Letairis [US] (ambrisentan), Tracleer (bosentan) and Thelin (sitaxentan)). 21

24 CONFIDENTIAL GM2008/00365/00 AMB Despite the number of treatments available, PAH remains a fatal condition. There is high clinical interest in combining treatments targeting different pathophysiological pathways. The first signalling pathway is the camp-dependent PGI 2 pathway. Prostanoids such as IV epoprostenol, and other PGI 2 derivatives target this pathway. These medicinal products mimic the effects of PGI 2, an endogenous prostaglandin produced by the vascular endothelium, to stimulate camp-dependent vasodilation of the pulmonary arterial bed and inhibit platelet aggregation [Humbert M et al. J Am Coll Cardiol 2004; 43(12 Suppl S):13S-24S]. The second signalling pathway is the cgmp dependent nitric oxide (NO) pathway. Tadalafil is the second PDE-5 inhibitor approved in the EU and USA, along with a number of other countries, for the treatment of PAH, that targets the NO pathway. Through inhibition of PDE-5, tadalafil increases cytoplasmic cgmp concentrations and enhances NO-mediated vasodilation of the vasculature [Forgue ST et al. Br J Clin Pharmacol. 2006; 61:280-8]. Tadalafil 40mg once daily demonstrated significant benefit compared to placebo over 12 weeks of treatment and is the licensed dose [Galie et al. Circulation 2009; 119(22): ]. The third targeted pathway is the phospholipase-c-dependent endothelin pathway. Endothelin-1 (ET-1) is the primary member of a family of potent vasoconstrictor peptides, which are known to play an essential role in mammalian cardiovascular physiology [Rubanyi GM, et al. Pharmacol Rev 1994; 46(3): ]. Two receptor subtypes, endothelin receptor type A (ET A ) and endothelin receptor type B (ET B ), mediate the effects of ET-1. In humans, the ET A receptor is preferentially expressed in vascular smooth muscle cells and is primarily responsible for the vasoconstrictive, as well as the mitogenic effects of ET-1 [Davenport AP, et al. Br J Pharmacol 1995; 114(6): ; Kirchengast M, et al. Proc Soc Exp Biol Med 1999; 21(4): ]. In contrast, ET B receptors are found mainly in the vascular endothelium, and their activation results in vasodilation via the production of NO and prostacyclin [Verhaar MC et al. Circulation 1998; 97(8): ; Hirata Y et al. J Clin Invest 1993; 91(4): ]. The ET B receptor is also involved in the regulation of circulating concentrations of ET-1, through effects on endothelin converting enzyme (ECE-1) expression, and the synthesis and reuptake of ET-1 by endothelial cells [Fukuroda T et al. Biochem Biophys Res Commun 1994; 199(3): ; Ozaki S et al. Biochem Biophys Res Commun 1995; 209(2): ]. It has therefore been suggested that ET-1 plays a critical role in the pathogenesis and progression of PAH. Ambrisentan is the second FDA-approved ERA in the US and third in the EU, and is approved in a number of other countries, for the treatment of PAH. Ambrisentan was the first, and is the first non-sulfonamide ERA approved in all regions. It works by targeting the ET A receptor, blocking the deleterious effects of endothelin-1. Ambrisentan 5mg and 10mg demonstrated significant benefit compared to placebo over 12 weeks, and are the licensed doses in most countries [Ambrisentan EU Summary of Product Characteristics 2009; Letairis US Prescribing Information 2009]. 22

25 CONFIDENTIAL GM2008/00365/00 AMB Ambrisentan Ambrisentan [(+)-(2S)-2-[(4,6-dimethylpyrimidin-2-yl)oxy]-3-methoxy-3,3- diphenylpropanoic acid] is an orally active Endothelin Receptor Antagonist (ERA) that is selective for the ET A receptor. Ambrisentan (Volibris, Letairis ). Ambrisentan is the latest ERA approved in the US, EU and many other countries internationally Clinical Efficacy In two Phase 3, placebo-controlled studies (AMB-320 [ARIES-1] and AMB-321 [ARIES-2]), ambrisentan demonstrated statistically significant improvement in 6MWD [Galie N, et al. Circulation 2008; 117: ]. An increase in 6MWD was observed after 4 weeks of treatment with ambrisentan, with a dose-response observed after 12 weeks of treatment. Results from AMB-321 demonstrated that 5 mg and 2.5 mg orally, by mouth once daily of ambrisentan improved the placebo-corrected 6MWD by 59 meters (p <0.001) and 32 meters (p = 0.022), respectively. Similarly, results from AMB-320 demonstrated that 10 mg and 5 mg orally, by mouth once daily of ambrisentan improved the placebo-corrected 6MWD by 51 meters (p 0.001) and 31 meters (p = 0.008), respectively. Time to clinical worsening, an indicator of disease progression, was a key secondary endpoint in the two Phase 3, placebo-controlled studies. The individual studies were not statistically powered to examine secondary endpoints such as time to clinical worsening; therefore, to increase the power to observe a treatment effect, a combined analysis of the two Phase 3 studies was pre-specified to examine secondary endpoints. The log-rank test for the comparison of the combined ambrisentan group versus placebo demonstrated that a significant delay in the time to clinical worsening of PAH was observed for subjects receiving ambrisentan (p <0.001). Furthermore, the hazard ratio was 0.29 (95% CI: 0.14 to 0.59), indicating a 71% reduction in the probability of clinical worsening over the 12- week treatment period for subjects receiving ambrisentan compared to placebo. The conclusions of the combined analysis were supported by similar trends in the individual studies. In AMB-320, a 2-fold increase in the number of subjects with an event of clinical worsening was observed in the placebo group compared to each of the ambrisentan dose groups; however, the log-rank comparison of the combined ambrisentan group versus placebo did not demonstrate a statistically significant difference in the time to clinical worsening of PAH (p = 0.214). In AMB-321, a 4-fold increase in the number of subjects with an event of clinical worsening was observed in the placebo group compared to each of the ambrisentan dose groups. The log-rank test demonstrated a significant delay in time to clinical worsening of PAH for the comparison of the combined ambrisentan group versus placebo (p <0.001). Clinically relevant or statistically significant improvements were observed in the individual studies for the secondary endpoints of WHO functional class, the SF-36 health survey, and BDI. In the pre-specified combined analysis, all secondary endpoints were statistically significant. 23

26 CONFIDENTIAL GM2008/00365/00 AMB In addition, statistically significant improvements from baseline with ambrisentan treatment were observed for several haemodynamics parameters (cardiac index, mpap, and PVR) in a Phase 2 dose-controlled study (AMB-220), consistent with an overall improvement in disease state. [Galie et al. JACC 2005;46(3):529-35] Plasma B-type natriuretic peptide (BNP), a hormone secreted primarily from the cardiac ventricles, has been proposed as a prognostic factor of mortality in patients with PAH [Leuchte HH et al. Chest 2005; 128(4): ; Nagaya N et al. Circulation 2000; 102(8): ; Park MH et al. Congest Heart Fail 2004; 10(5): ]. The final step of BNP synthesis consists of a high molecular weight precursor, probnp cleaved into biologically inactive N-terminal segment (NT-proBNP) and the proper low molecular weight BNP. NT-proBNP has a longer half-life and a better stability both in circulating blood and after sampling. Reductions in BNP/NT-proBNP have been shown to parallel improvements in haemodynamics, 6MWD, and WHO functional class in patients with PAH [Leuchte HH et al. J Am Coll Cardiol 2004; 43(5): ]. Substantial (p <0.05) and clinically relevant decreases (29% to 45%) in BNP/NT-proBNP compared to placebo were observed in the Phase 3 pivotal studies, and decreases in BNP/NT-proBNP correlated with and were predictive of improvements in 6MWD. [Galie N, et al. Circulation 2009;117: ] Data from the long-term follow-up of the subjects who were treated with ambrisentan in the Phase 3, placebo-controlled studies and their open-label extension (N = 383) has been reported. After 2 years of ambrisentan exposure, the mean change from baseline in 6MWD was improved for the 5-mg (+23m; 95% confidence interval: 9 to 38m) and 10- mg (+28m; 95% confidence interval: 11 to 45m) groups. Estimates of survival and freedom from clinical worsening for the combined dose group were 94% and 83%, respectively, at 1 year and 88% and 72%, respectively, at 2 years. The annualized risk of aminotransferase abnormalities >3x the upper limit of normal was ~2% per year; most of these events were mild and did not lead to discontinuation of drug. Without a control group this data should be interpreted with caution [Oudiz R, et al. JACC 2009; 54: ] Safety Safety data for ambrisentan were obtained from the two Phase 3, placebo-controlled studies and four non-placebo-controlled studies in a total of 483 patients with PAH who were treated with doses of 1, 2.5, 5, or 10 mg once daily in the phase 2 (AMB-220, AMB-222) and 3 (AMB-320 and AMB-321) studies. This includes subjects exposed in the long term studies (AMB-220-E, AMB-222, AMB 320/321-E) following rerandomisation of placebo subjects from the parent studies. As of 16 July 2008, a total of 483 subjects with PAH have received ambrisentan in Phase 2 or Phase 3 clinical studies (excluding Study AMB-323), with a mean exposure of weeks (2 years) and a maximum exposure of 230.1weeks (4.3 years). 24

27 CONFIDENTIAL GM2008/00365/00 AMB The adverse drug reactions (ADR) identified from 12 week placebo controlled clinical trial data are: BLOOD AND LYMPHATIC SYSTEM DISORDERS Common: Anemia (decreases in haemoglobin and/or haematocrit) IMMUNE SYSTEM DISORDERS Uncommon: Hypersensitivity (e.g. angioedema, rash) NERVOUS SYSTEM DISORDERS Common: Headache CARDIAC DISORDERS Common: Palpitations VASCULAR DISORDERS Common: Flushing RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS Common: Nasal congestion, sinusitis, nasopharyngitis The incidence of nasal congestion was dose-related during ambrisentan therapy. GASTROINTESTINAL DISORDERS Common: Abdominal pain, constipation GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS Common: Fluid retention, Peripheral edema With longer observation in uncontrolled studies (mean observation of 79 weeks), the safety profile was similar to that observed in the short term studies. Peripheral oedema is a known class effect of ERAs, and is also a clinical consequence of PAH and worsening PAH. Among subjects receiving ambrisentan (but not placebo), the incidence and severity of peripheral oedema appeared greater for the 65 year old subgroup (28.6%; 16/56) compared to the <65 year old subgroup (14.1%; 29/205). Hepatic enzyme elevations have been observed with ERAs. Therefore, evaluation of hepatic function prior to initiation of ambrisentan and during treatment is recommended. In clinical trials, ambrisentan had a low overall incidence and a low incidence of clinically significant liver transaminase abnormalities, even with long-term treatment. The hepatic safety profile for ambrisentan observed in clinical practice has appeared consistent with that seen in the clinical studies and has not suggested an increased risk of hepatic events with ambrisentan compared to background hepatic events in patients with PAH. Teratogenicity is a class effect of ERAs and animal studies in rats and rabbits have shown that ambrisentan is teratogenic. Thus, use of ambrisentan is contraindicated in 25

28 CONFIDENTIAL GM2008/00365/00 AMB pregnancy. Data on teratogenic effects in humans are very limited; no congenital abnormalities or teratogenic effects have been observed in the fetuses of women who have become pregnant while using ambrisentan. Use of reliable contraception and pregnancy testing during treatment is recommended in women of child bearing potential. Additionally, these women should be advised to contact their physician immediately if they become pregnant or suspect they may be pregnant. The development of testicular tubular atrophy in male animals has been linked to the chronic administration of ERAs, including ambrisentan. However, the effect on male human fertility is not known. To date, postmarketing safety surveillance has shown that the nature and severity of AEs reported with ambrisentan have been generally consistent with the known safety profile of ambrisentan. However, several additional adverse reactions have been identified in the post-approval period which has led to amendments to the core safety profile for ambrisentan. Post-marketing reports of fluid retention occurring within weeks after starting Volibris/Letairis have been received and, in some cases, have required intervention with a diuretic or hospitalization for fluid management or decompensated heart failure. Prescribers have been advised that if patients have pre-existing fluid overload, this should be managed as clinically appropriate prior to starting ambrisentan. Additionally, prescribers have been informed of cases of worsening dyspnea of unclear etiology which have been reported shortly after starting ambrisentan therapy, and nausea and vomiting have been added to the undesirable effects section. Further, an ERA class effect has also been added to the prescribing information related to the possibility that patients with pulmonary veno-occlusive disease could develop acute pulmonary oedema during initiation of therapy with vasodilating agents such as an endothelin receptor antagonist. For full details on ambrisentan including contraindications, warnings and precautions and drug-drug interactions refer to the prescribing information. See section for details of permitted concomitant medications in which caution is advised 1.3. Tadalafil Tadalafil is an orally administered selective inhibitor of the enzyme phosphodiesterase type 5 (PDE5), the primary cyclic guanosine monophosphate (cgmp)-hydrolyzing enzyme in smooth muscle [Adcirca US Prescribing Information 2009]. Tadalafil (Adcirca ) 40-mg once daily dosing has been approved for the treatment of pulmonary arterial hypertension (PAH) in the US, EU, Japan and Canada Clinical efficacy A randomised, double-blind, 16 week placebo-controlled study was conducted in 405 patients with pulmonary arterial hypertension [Adcirca US Prescribing Information 2009; Galie N et al. Circulation 2009; 119(22): ], defined as a resting mean pulmonary artery pressure (mpap) 25 mmhg, pulmonary capillary wedge pressure (PCWP) 15 mmhg, and pulmonary vascular resistance (PVR) 3 Wood units via right heart catheterization. Allowed background therapy included bosentan (maintenance dosing up to 125 mg twice daily) and chronic anticoagulation. The use of prostacyclin or analogue, 26

29 CONFIDENTIAL GM2008/00365/00 AMB L-arginine, phosphodiesterase inhibitor, or other chronic PAH medications were not permitted. Subjects were randomly assigned to 1 of 5 treatment groups (tadalafil 2.5, 10, 20, 40 mg, or placebo) in a 1:1:1:1:1 ratio. Subjects had to be at least 12 years of age and had a diagnosis of PAH that was idiopathic, related to collagen vascular disease, anorexigen use, human immunodeficiency virus (HIV) infection, associated with an atrial-septal defect, or associated with surgical repair of a congenital systemic-to-pulmonary shunt of at least 1 year in duration (for example, ventricular septal defect, patent ductus arteriosus). Patients with a history of left-sided heart disease, severe renal insufficiency, or pulmonary hypertension related to conditions other than specified in the inclusion criteria were not eligible for enrolment. The mean age of all subjects was 54 years (range years) with the majority of subjects being Caucasian (81%) and female (78%). PAH aetiologies were predominantly idiopathic PAH (61%) and related to collagen vascular disease (24%). More than half (53%) of the subjects in the study were receiving concomitant bosentan therapy. The majority of subjects had a World Health Organization (WHO) Functional Class III (65%) or II (32%). The mean baseline 6MWD was 343 meters. Of the 405 subjects, 341 completed the study. The primary efficacy endpoint was the change from baseline at week 16 in 6MWD. In the Adcirca 40 mg treatment group, the placebo-adjusted mean change increase in 6MWD was 33 meters (95% C.I meters; p=0.0004). The improvement in 6MWD was apparent at 8 weeks of treatment and then maintained at week 12 and week 16. Placebo-adjusted changes in 6MWD at 16 weeks were evaluated in subgroups. In patients taking only Adcirca 40 mg (i.e., without concomitant bosentan), the placebo-adjusted mean change in 6MWD was 44 meters. In patients taking Adcirca 40 mg and concomitant bosentan therapy, the placebo adjusted mean change in 6MWD was 23 meters. There was less clinical worsening (defined as death, lung transplantation, atrial septostomy, hospitalization because of worsening PAH, initiation of new PAH therapy [prostacyclin or analogue, endothelin receptor antagonist, PDE5 inhibitor], or worsening WHO functional class) in the Adcirca 40 mg group compared to the placebo group and the groups that used lower doses of Adcirca. Patients (N=357) from the placebo-controlled study entered a long-term extension study. Of these, 311 patients have been treated with tadalafil for at least 6 months and 182 for 1 year (median exposure 356 days; range 2 days to 415 days). The survival rate in the extension study was 96.5 per 100 patient years. Without a control group, these data must be interpreted cautiously Safety Tadalafil was administered to 398 patients with PAH during clinical trials worldwide [Adcirca US Prescribing Information 2009]. In trials of Adcirca, a total of 311 and

30 CONFIDENTIAL GM2008/00365/00 AMB subjects have been treated for at least 182 days and 360 days, respectively. The overall rates of discontinuation because of an adverse event (AE) in the placebo-controlled trial were 9% for Adcirca 40 mg and 15% for placebo. The rates of discontinuation because of AEs, other than those related to worsening of PAH, in patients treated with Adcirca 40 mg was 4% compared to 5% in placebo-treated patients. In the placebo-controlled study, the most common AEs were generally transient and mild to moderate in intensity. Table 2 presents treatment-emergent adverse events reported by 9% of patients in the Adcirca 40 mg group and occurring more frequently than with placebo [Adcirca US Prescribing Information 2009]. TABLE 2: Treatment-Emergent Adverse Events Reported by 9% of Patients in Adcirca and More Frequent than Placebo by 2% (16 week study) Event Placebo (%) N = 82 Adcirca 20 mg Adcirca 40 mg (%) N=79 (%) N=82 Headache Myalgia Nasopharyngitis Flushing Respiratory Tract Infection (Upper and Lower) Pain in Extremity Nausea Back Pain Dyspepsia Nasal Congestion (Including sinus congestion) Adverse Reactions and Post-marketing Experience from Adcirca USPI: The following adverse reactions have been identified during post-approval use of tadalafil. These events have been chosen for inclusion either because of their seriousness, reporting frequency, lack of clear alternative causation, or a combination of these factors. Because these reactions are reported voluntarily from a population of uncertain size, it is not always possible to estimate reliably their frequency or establish a causal relationship to drug exposure. The list does not include adverse events that are reported from clinical trials and that are listed above in this section. Cardiovascular and cerebrovascular Serious cardiovascular events, including myocardial infarction, sudden cardiac death, stroke, chest pain, palpitations, and tachycardia, have been reported post-marketing in temporal association with the use of tadalafil. Most, but not all, of these patients had pre-existing cardiovascular risk factors. Many of these events were reported to occur during or shortly after sexual activity, and a few were reported to occur shortly after the use of tadalafil without sexual activity. Others were reported to have occurred hours to days after the use of tadalafil and sexual activity. It is not possible to determine whether these events are related directly to tadalafil, to sexual activity, to the patient s underlying cardiovascular disease, to a combination of these factors, or to other factors. 28

31 CONFIDENTIAL GM2008/00365/00 AMB Body as a whole Hypersensitivity reactions including urticaria, Stevens Johnson syndrome, and exfoliative dermatitis Nervous Migraine, seizure and seizure recurrence, and transient global amnesia Ophthalmologic Visual field defect, retinal vein occlusion, and retinal artery occlusion Non arteritic anterior ischemic optic neuropathy (NAION), a cause of decreased vision including permanent loss of vision, has been reported rarely postmarketing in temporal association with the use of PDE5 inhibitors, including tadalafil. Most, but not all, of these patients had underlying anatomic or vascular risk factors for development of NAION, including but not necessarily limited to: low cup to disc ratio ( crowded disc ), age over 50, diabetes, hypertension, coronary artery disease, hyperlipidemia, and smoking. It is not possible to determine whether these events are related directly to the use of PDE5 inhibitors, to the patient s underlying vascular risk factors or anatomical defects, to a combination of these factors, or to other factors. Otologic Cases of sudden decrease or loss of hearing have been reported postmarketing in temporal association with the use of PDE5 inhibitors, including tadalafil. In some of the cases, medical conditions and other factors were reported that may have also played a role in the otologic adverse events. In many cases, medical follow-up information was limited. It is not possible to determine whether these reported events are related directly to the use of tadalafil, to the patient s underlying risk factors for hearing loss, a combination of these factors, or to other factors. Urogenital Priapism. In clinical studies, tadalafil was shown to augment the hypotensive effects of nitrates. This is thought to result from the combined effects of nitrates and tadalafil on the nitric oxide/cgmp pathway. Therefore, administration of Adcirca to patients who are using any form of organic nitrate is contraindicated. For full details on tadalafil including contraindications, warnings and precautions and drug-drug interactions refer to the prescribing information. See section for details of permitted concomitant medications in which caution is advised 1.4. Combination of ambrisentan and tadalafil A Phase 1 study in 26 healthy subjects demonstrated no clinically relevant pharmacokinetic interaction between tadalafil and ambrisentan [Spence R et al. J Pharm Sci. 2009;98(12): ] Single-dose PK of ambrisentan (10 mg) and its metabolite, 4-hydroxymethyl ambrisentan, were determined in the absence and presence of multiple doses of tadalafil (40 mg once daily). Similarly, single-dose PK of tadalafil (40 mg) were evaluated in the absence and presence of multiple doses of ambrisentan (10 mg once daily). In the presence of tadalafil, ambrisentan maximum plasma concentration (Cmax) was similar (105.0% [90% CI: %]) and systemic exposure (AUC 0 1 ) was slightly decreased (87.5% [ %]), compared with ambrisentan alone. Similar changes were observed with 4-hydroxymethyl ambrisentan. Tadalafil C max (100.6% [

32 CONFIDENTIAL GM2008/00365/00 AMB %]) and AUC 0 1 (100.2% [ %]) showed no difference in the absence and presence of ambrisentan. The safety profile of the drugs combined was similar to that of either drug alone. Overall, there was a trend for the mean SBP and DBP in the sitting, supine, and standing positions to decrease over the course of both Treatment Regimens. The potential for additive effects on blood pressure when ambrisentan and tadalafil are administered in combination in the PAH population will be monitored during the trial. No animal or in vitro studies have been undertaken with ambrisentan in combination with tadalafil. Both ambrisentan and tadalafil are indicated individually for the treatment of patients with pulmonary arterial hypertension. Pharmacologically ambrisentan is a highly specific and potent endothelin Type-A receptor antagonist whilst tadalafil is a specific and potent inhibitor of cgmp specific phosphodiesterase type 5 (PDE5). The possibility for additive toxicity for AEs which are associated with both drugs e.g. nasal congestion/upper respiratory infections, flushing, nausea, headache cannot be excluded. From an analysis of pharmacological activity and the known toxicological findings of each of these compounds there is no evidence to suggest that there would be any synergistic toxicity or exacerbation of the known toxicological profile of either individual if they were administered to patients in combination. However, the development of testicular tubular atrophy in male animals has been linked to the chronic administration of ERAs, including ambrisentan. In dogs given tadalafil daily for 6 to 12 months at doses of 25 mg/kg/day (resulting in at least a 3-fold greater exposure [range ] than seen in humans given a single 20 mg dose) and above, there was regression of the seminiferous tubular epithelium that resulted in a decrease in spermatogenesis in some dogs. In the AMBITION study, clinical laboratory assessments of testicular function will be collected and male subjects will be informed of the potential risks Rationale Pulmonary arterial hypertension (PAH) is a devastating, life threatening disease. A number of treatments are available but no single drug has been demonstrated to be consistently effective in treating all patients with PAH. Combining medicines with different mechanisms of action is an evolving strategy for the treatment of PAH [Galie N et al. Eur Respir J 2009; 34(6): ]. This approach is supported by data from several exploratory studies in PAH demonstrating the synergistic benefits of combining two or more drugs [Humbert M et al. Eur Respir J 2004; 24(3): ; Hoeper MM et al. Eur Respir J 2003; 22(2): ; Ghofrani HA et al. Ann Intern Med 2002; 136(7): ; Hoeper MM et al. Eur Respir J 2005; 26(5): ; Simonneau G et al. Ann Intern Med 2008; 149: ]. Ambrisentan (5 and 10 mg once daily) is an ETA-selective ERA that is approved by the FDA and EMA for the treatment of PAH (ipah and PAH-CTD) in patients with WHO class II or III symptoms [Ambrisentan EU Summary of Product Characteristics 2009; Letairis US Prescribing Information]. 30

33 CONFIDENTIAL GM2008/00365/00 AMB Tadalafil (40 mg once daily) is a PDE-5 inhibitor approved in numerous countries for the treatment of PAH (ipah and PAH -CTD) in patients with WHO class II and III symptoms [Adcirca (tadalafil) Summary of Product Characteristics 2009; Adcirca US Prescribing Information 2009]. Because ambrisentan and tadalafil are both orally administered once a day, have different mechanisms of action targeting different intracellular pathways, and have not demonstrated any clinically relevant pharmacokinetic concerns when coadministered, an ambrisentan/tadalafil combination therapy is a rational treatment strategy for patients with PAH. All of the larger clinical studies completed to date have evaluated combination therapy as an add-on approach: the addition of a second PAH therapy to a subject s monotherapy regimen. It is unknown if PAH patients may have greater improvements for sustained periods of time if combination PAH treatment is initiated as first-line therapy, rather than delaying the addition of the second therapy until the PAH status of the patient has reached a plateau or subsequently deteriorated while receiving monotherapy. The primary objective of this study is to compare the difference between two treatment strategies; first-line combination therapy (ambrisentan and tadalafil) versus first-line monotherapy (ambrisentan or tadalafil) in subjects with PAH. This will be assessed by time to the first clinical failure event Rationale for ambrisentan peak-trough 6MWD assessment The use of ambrisentan as a once daily, oral therapy for the treatment of PAH is supported by its mean terminal half-life of 12.9 to 17.9 hours in PAH patients [Ambrisentan Summary of Product Characteristics 2009; Letairis US Prescribing Information] and the consistent and clinically relevant efficacy observed in the placebocontrolled studies (see section 1.2.1). However, the mean plasma concentration of ambrisentan varies considerably during the 24-hour dosing interval (mean trough concentration at steady-state is approximately 15% of the mean peak concentration at steady-state). The relationships between ambrisentan plasma concentrations, endothelin receptor occupancy in the pulmonary vasculature, and clinical efficacy outcomes, have not been clearly established. However, in prior clinical studies, the timing of the 6MWD assessment relative to ambrisentan dosing was not specified or collected; therefore, the potential for variability of this assessment over the 24-hour dosing interval has not been evaluated. This study will evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations. To demonstrate an ambrisentan drug treatment effect, the change from baseline to Week 16 in 6MWD for the ambrisentan/tadalafil combination group and the tadalafil monotherapy group will be compared (ABS+TAD vs. Placebo+TAD). The potential for differences in treatment effect over the 24-hour dosing interval will subsequently be explored by comparing the 31

34 CONFIDENTIAL GM2008/00365/00 AMB ratio of placebo-corrected 6MWD response at peak and trough ambrisentan plasma concentrations. 2. OBJECTIVE(S) The primary objective of this study is to compare the difference between two treatment strategies; first-line combination therapy (with ambrisentan 10mg once daily and tadalafil 40mg once daily) vs. monotherapy (with ambrisentan 10mg once daily or tadalafil 40mg once daily) in subjects with pulmonary arterial hypertension (PAH). This will be assessed by time to first clinical failure (TtCF) event. TtCF is a variation on time to clinical worsening (TtCW), adding an additional component of unsatisfactory long-term clinical response. This additional component is proposed based on the recent guidelines for the clinical treatment of PAH [Galie N et al. Eur Respir J 2009]. These guidelines recommend a target based treatment approach, based on parameters with established importance for assessing disease severity, stability and prognosis in pulmonary arterial hypertension. The patients status is divided into three categories: stable and satisfactory, stable and not satisfactory, unstable and deteriorating. The aim of treatment is to move all patients to the stable and satisfactory category. Therefore current treatment guidelines do not deem it acceptable to wait for a patient to experience a significant deterioration prior to medical intervention. As this is a treatment strategy study, this unsatisfactory long-term response component is intended to better reflect the principle of target based treatment approach by counting unsatisfactory response as clinical failure. This allows physicians to intervene prior to a patient experiencing significant deterioration and better reflects current clinical practice [Galie N et al. Eur Respir J 2009] The secondary objectives of this study are to compare the change in other clinical measures of PAH after initiating combination therapy (with ambrisentan and tadalafil) or monotherapy (with ambrisentan or tadalafil) in subjects with PAH. The safety and tolerability of combination therapy (with ambrisentan and tadalafil) will be compared to monotherapy (with ambrisentan or tadalafil). In addition, the effect of ambrisentan on exercise capacity at both peak and trough plasma concentrations will be assessed. The study endpoints are described in section

35

36 CONFIDENTIAL GM2008/00365/00 AMB One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (Creatinine Clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see section for details). Monotherapy arm: ambrisentan group: Subjects randomised to ambrisentan monotherapy will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter Two tablets of TAD-matching placebo Monotherapy arm: tadalafil group: Subjects randomised to tadalafil monotherapy will receive both Two tablets of ABS-matching placebo One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and, two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (Creatinine Clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see section for details). As there is limited data on combining these two drugs in patients a cautious approach has been taken with a titration of ambrisentan (combination group and AMB monotherapy group) and tadalafil (combination group only). In addition the uptitration time-points have been separated to avoid dose increases at the same time. The tadalafil monotherapy group will also start on 20mg od to provide a true comparison with the combination group and allow subjects with mild to moderate renal impairment to follow dosing recommendations. It should be noted that since the recommended dose for tadalafil for subjects with PAH without renal insufficiency is 40 mg, it is possible that some subjects in the trial may be under dosed on tadalafil for the first 4 weeks of the trial. Should tolerability issues be experienced the investigator may separate dosing of Investigational Products in to morning and evening (i.e., ambrisentan in the morning and 34

37 CONFIDENTIAL GM2008/00365/00 AMB tadalafil in the evening). If tolerability issues are still experienced the investigator may request a down titration of ambrisentan from 10mg to 5mg once daily (it should be noted that subjects in TAD monotherapy group will not have a change in dose). Therefore all subjects should be closely monitored for continued lack of tolerance to IP). If tolerability issues are still experienced the investigator may then also request a down titration of tadalafil from 40mg once daily to 20mg once daily (requires approval of the Sponsor Medical Monitor). See section for details. Subjects will be assessed for efficacy and safety at Screening, Randomisation, Weeks 4, 8, 16, 24, and every 12 weeks thereafter. In between clinic visits subjects will have monthly laboratory safety assessments. These may be performed by a local phlebotomy laboratory or at the Investigator clinic (or by the central lab with Sponsor permission). The primary efficacy endpoint is the time to the first clinical failure event of PAH. Time to clinical failure is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) o Non-elective hospitalization for worsening PAH o Lung or heart/lung transplant o Atrial septostomy o Initiation of parenteral prostanoid therapy Disease progression (adjudicated) o >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) o Receiving randomised treatment for at least 6 months o A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days o Sustained WHO class III or IV symptoms for 6 months Time to clinical worsening (death, hospitalization for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Medical Interventions for PAH During the Study: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary endpoint event and subjects may continue in the study following initiation. If a subject experiences an event of clinical failure the following medical interventions may also be utilized. The clinical failure event necessary for medical intervention is 35

38 CONFIDENTIAL GM2008/00365/00 AMB based on the opinion of the investigator and does not require prior adjudication from the clinical endpoint committee. The investigator may choose to request blinded combination therapy to ensure that the subject is receiving ambrisentan/tadalafil combination therapy The investigator may choose to add prostanoid therapy (parenteral or nonparenteral) (see section 5.1.4). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first intervention (see section ) the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy Initiate parenteral prostanoid therapy Request blinded ambrisentan/tadalafil combination therapy. Prostanoids will not be supplied by the sponsor. The investigator is required to complete a Change of Randomised Treatment Visit to assess safety and efficacy. Any additional PAH therapies necessary for treatment or changes in dose outside the protocol design will not be supplied by the sponsor. Prior to any change in PAH treatment, a subject is required to complete a Change of Randomised Treatment Visit to assess safety and efficacy. Criterion for clinical failure events will be reviewed and adjudicated by a blinded studyspecific Clinical Endpoint Committee. Secondary endpoints include: Change from baseline 6MWD measured at week 24 Percentage of subjects with satisfactory clinical response measured at Week 24, defined as: o 10% improvement in 6MWD compared to baseline o Improvement or maintenance of WHO class I or II symptoms o No events of clinical worsening prior to or at the week 24 visit Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP) Change from baseline measured at week 24 in WHO Functional Class Change from baseline measured at week 24 in BDI immediately following exercise Exploratory endpoints include: An assessment of the effects of ambrisentan on 6MWD at peak and trough plasma concentrations, which includes: 36

39 CONFIDENTIAL GM2008/00365/00 AMB o The ambrisentan placebo-corrected mean change from baseline 6MWD following 16 weeks of treatment (i.e., tadalafil/ambrisentan combination therapy versus tadalafil/placebo monotherapy) o The ratio of the placebo-corrected mean change from baseline trough 6MWD at week 16 to the placebo-corrected mean change from baseline peak 6MWD at week 16. Health outcomes endpoints include Change from baseline to week 24 in Quality of Life: as measured by the Short Form 36 Health Survey (SF-36) & the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) (selected countries where language translation is validated) Safety endpoints include Incidence and severity of adverse events Vital signs (i.e. blood pressure and heart rate). Laboratory test: clinical chemistry, haematology, testicular function (males only) All subjects are eligible to receive a minimum of 24 weeks of therapy. When 82 adjudicated primary endpoint events are projected to have occurred, i.e. 82 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety follow-up by phone will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed (see figure 2). An IDMC will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will also review formal interim analysis and recommend continuation or early termination of the study based on stopping criteria defined in the IDMC charter. 37

40

41 CONFIDENTIAL GM2008/00365/00 AMB i. connective tissue disease (e.g., limited scleroderma, diffuse scleroderma, mixed CTD, systemic lupus erythematosus, or overlap syndrome) ii. iii. iv. drugs or toxins HIV infection congenital heart defects repaired greater than 1 year prior to screening (i.e., atrial septal defects, ventricular septal defects, and patent ductus arteriosus) NB: subjects with portopulmonary hypertension and PVOD are NOT eligible for the study 4. Subject must have a current diagnosis of being in WHO Functional Class II or III. 5. Subject with a diagnosis of HIV must have stable disease status. For this study, stable HIV status is defined as: i. No addition of medications for treatment of HIV for at least 8 weeks prior to screening ii. iii. No active opportunistic infection during the Screening Period No hospitalizations due to HIV for at least 4 weeks prior to screening 6. Subject must meet all of the following haemodynamic criteria by means of a RHC prior to screening: i. mpap of 25 mmhg ii. iii. PVR 240 dyne sec/cm5 PCWP or LVEDP of 15 mmhg 7. Subject must meet all of the following pulmonary function tests completed no more than 24 weeks before the Screening visit: i. Total lung capacity (TLC) 60% of predicted normal and ii. Forced expiratory volume in one second (FEV1) 55% of predicted normal 8. Subject must walk a distance of 125m and 500m at the screening visit 9. Subject, with or without supplemental oxygen, must have a resting arterial oxygen saturation (SaO2) 88% as measured by pulse oximetry at the Screening Visit. Exercise Programmes 10. Subject has not enrolled in an exercise training program for pulmonary rehabilitation within 12 weeks prior to the Screening Visit and must agree not to enroll in an exercise training program for pulmonary rehabilitation during the Screening Period and the first 24 weeks of the study. Subjects enrolled in an exercise program for pulmonary rehabilitation 12 weeks prior to screening may enter the study if they agree to maintain their current level of rehabilitation for the first 24 weeks of the study. 39

42 General CONFIDENTIAL GM2008/00365/00 AMB Female subject of childbearing potential must agree to use 2 reliable methods of contraception from the Screening Visit until study completion and for at least 30 days following the last dose of Investigational Product (reliable methods of contraception are described in Appendix 1) 12. Subject must agree not to participate in a clinical study involving another investigational drug or device throughout this study 13. Subject must be competent to understand the information given in the Institutional Review Board (IRB) or Independent Ethics Committee (IEC) approved ICF and must sign the form prior to the initiation of any study procedures Specific information regarding warnings, precautions, contraindications, adverse events, and other pertinent information on the investigational product that may impact subject eligibility is provided in the IB and product label. French subjects: In France, a subject will be eligible for inclusion in this study only if either affiliated to or a beneficiary of a social security category Exclusion Criteria Subjects meeting any of the following criteria must not be enrolled in the study: PAH Treatments 1. Subject received previous PAH therapy (PDE5i, ERA, chronic prostanoid*) within 4 weeks prior to the screening visit *Chronic prostanoid use is considered >7 days of treatment 2. Subject received ERA treatment (e.g., bosentan or sitaxentan) or PDE5i treatment (e.g. Sildenafil) at any time AND discontinued due to tolerance issues other than those associated with liver function abnormalities 3. Subjects who have previously discontinued ambrisentan or tadalafil in either another clinical study or commercial product (Volibris/Letairis or Adcirca) for safety or tolerability reasons. 4. Subject has a known hypersensitivity to the Investigational Products, the metabolites, or formulation excipients Other Therapies 5. Subject receiving intravenous inotropes within 2 weeks prior to the Screening Visit (e.g. dopamine, dobutamine) 6. Subject is receiving treatment with protease inhibitors, systemic ketoconazole, or systemic itraconazole 7. Subject is receiving treatment with a potent inducer of CYP3A4 (e.g. rifampicin) 40

43 CONFIDENTIAL GM2008/00365/00 AMB Subjects receiving Calcium Channel Blockers or HMG-CoA reductase inhibitors (i.e., statins) on an unstable dose 4 weeks prior to the Screening Visit (to be eligible subjects must not have changed their dose <4 weeks prior to the screening visit) 9. Subject has a history of angina pectoris or other condition that was treated with long or short-acting nitrates <12 weeks of screening Laboratory Values at Screening 10. Subject has a serum ALT or AST lab value that is > 2xULN at the Screening Visit 11. Subject has serum bilirubin lab value that is >1.5xULN at the screening visit 12. Subject has severe renal impairment (creatinine clearance <30 ml/min) at the Screening Visit Medical History/Current Medical Conditions Liver 13. Subject has severe hepatic impairment (Child-Pugh class C with or without cirrhosis) at the Screening Visit Haematology and bleeding disorders 14. Subject has clinically significant anaemia in the opinion of the investigator 15. Subjects with bleeding disorders or significant active peptic ulceration in the opinion of the investigator Cardiovascular 16. Subject has uncontrolled hypertension ( 180/110 mmhg) at screening 17. Subject has severe hypotension (<90/50 mmhg) at screening 18. Subject has had an acute myocardial infarction within the last 90 days prior to screening 19. Subject has, in the opinion of the investigator, clinically significant aortic or mitral valve disease; pericardial constriction; restrictive or congestive cardiomyopathy; lifethreatening cardiac arrhythmias; significant left ventricular dysfunction; left ventricular outflow obstruction; symptomatic coronary artery disease; autonomic hypotension; fluid depletion. Ophthalmic 20. Subject has a past medical history of NAION 21. Subject has a hereditary degenerative retinal disorder (e.g. retinitis pigmentosa) General Medical Conditions 22. Subject has clinically significant fluid retention in the opinion of the investigator 23. Subject with cardiovascular, liver, renal, haematologic, gastrointestinal, immunologic, endocrine, metabolic, or central nervous system disease that, in the opinion of the Investigator, may adversely affect the safety of the subject and/or 41

44 CONFIDENTIAL GM2008/00365/00 AMB efficacy of the investigational product or severely limit the lifespan of the subject other than the condition being studied 24. Subject has a history of malignancies within the past 5 years, except for a subject with localized, non-metastatic basal cell carcinoma of the skin, in situ carcinoma of the cervix, or prostate cancer who is not currently or expected, during the study, to undergo radiation therapy, chemotherapy, and/or surgical intervention, or to initiate hormonal treatment General Criteria 25. Female subject who is pregnant or breastfeeding 26. Subject has demonstrated noncompliance with previous medical regimens 27. Subject has a recent (within 1 year) history of abusing alcohol or illicit drugs 28. Subject has participated in a clinical study involving another investigational drug or device within 4 weeks before the Screening Visit Subjects who fail inclusion/exclusion criteria may be re-screened once Withdrawal Criteria If a subject prematurely and permanently discontinues Investigational Product administration, a Change of Randomised Treatment Visit must be performed. If a subject discontinues Investigational Product prematurely, the subject should be encouraged to continue all study Clinic Visits according to the planned schedule of events, after completing the Change of Randomised Treatment Visit. After discontinuing Investigational Product, if a subject continues to perform Clinic Visits, and later prematurely discontinues from the study the subject should perform an End of Study Visit. Every effort will be made by the investigator to follow the subject to assess primary endpoint events. All subjects have the right to withdraw formal consent without prejudice at any time during the study. If a subject withdraws formal consent, the Investigator should make a reasonable effort to determine the cause for withdrawal of consent. Possible reasons for subject discontinuation from IP include, but are not limited to, the following: Development of an AE where continuation of the subject s participation in the study is thought by the Investigator to be inappropriate Subject meets liver stopping criteria Subject begins treatment with a prohibited concomitant therapy Lost to follow-up 42

45 CONFIDENTIAL GM2008/00365/00 AMB Possible reasons for subject withdrawal from the study include, but are not limited to, the following: Subject withdraws consent/subject requests to discontinue for any reason Subject noncompliance (e.g., refusal to complete study procedures) Discretion of the Investigator Discontinuation of the study at the request of the Sponsor, a regulatory agency, or an IRB/IEC 5. STUDY TREATMENTS 5.1. Investigational Products Investigational products are ambrisentan, in the form of 5mg tablets or matching placebo and tadalafil, in the form of 20mg tablets, or matching placebo Ambrisentan or placebo Provision, handling and storage The sponsor will provide ambrisentan in the form of round, white, film-coated, immediate-release tablets, containing 5mg ambrisentan. Matching placebo will also be provided and will be identical in appearance to ambrisentan 5mg tablets. Investigational Product will be packaged in ambrisentan monthly dosing packs containing a total of 2 x 35 tablets of ambrisentan and/or ambrisentan-matching placebo in blister strips. The packaged material will be labelled and the contents of the label will be in accordance with all applicable regulatory requirements. Investigational Product will be stored at a temperature up to 30 C (59-86 F). The Investigational Product distributor will send the initial shipment of Investigational Product to site and Investigational Product re-supply will occur automatically via site inventory monitoring in IVRS. Under normal conditions of handling and administration, investigational product is not expected to pose significant safety risks to site staff. A Material Safety Data Sheet (MSDS) describing the occupational hazards and recommended handling precautions will be provided to site staff if required by local laws or will otherwise be available from Sponsor upon request. Investigational product must be stored in a secure area under the appropriate physical conditions for the product. Access to and administration of the investigational product will be limited to the investigator and authorized site staff. Investigational product must be dispensed or administered only to subjects enrolled in the study and in accordance with the protocol. 43

46 CONFIDENTIAL GM2008/00365/00 AMB Subjects should be instructed to bring their used and unused Investigational Product and containers with them to every visit to assess compliance and drug accountability. Any unused Investigational Product will be returned by the Investigator to the Sponsor Dosing Subjects should be recommended to take Investigational Product once daily preferably in the morning with or without food throughout the study unless otherwise instructed. The target dose of ambrisentan is 10mg once daily in the form of 2 x 5mg tablets, or matching placebo, and will follow the dose schedule below The initial dose of ambrisentan will be 5mg once daily in the form of 1x5mg ambrisentan tablet (or ambrisentan-matching placebo) and 1 x ambrisentanmatching placebo tablet. The subject will take this dose for the first 8 weeks of treatment. At the 8 week visit the subject will be up-titrated to the target dose of 10mg ambrisentan once daily, in the form of 2 x 5mg ambrisentan tablets (or ambrisentan-matching placebo) If the investigator judges that the subject has not tolerated 5mg of ambrisentan well, he may request that the subject remain on 5mg. The subject will continue to receive ambrisentan 5mg once daily in the form of 1 x 5mg ambrisentan tablet (or ambrisentan-matching placebo) and 1 x ambrisentanmatching placebo tablet Following week 8: Down-titration from ambrisentan 10mg to ambrisentan 5mg may be requested should the subject not tolerate the investigational product. See section for details. At the Investigator s discretion, chronic parenteral prostanoid therapy may be initiated at any time throughout the study (which meets the criterion for clinical failure). Other PAH therapies may not be added unless a subject has experienced a clinical failure event Tadalafil or placebo Provision, handling and storage The sponsor will provide tadalafil in the form of a large almond, dark yellow tablet, containing 20 mg tadalafil. Matching placebo will also be provided and will be identical in appearance to tadalafil 20mg tablets. Investigational Product will be packaged in tadalafil monthly dosing packs containing 2 x 35 tablets of tadalafil and/or tadalafil-matching placebo in blister strips. 44

47 CONFIDENTIAL GM2008/00365/00 AMB The packaged material will be labelled and the contents of the label will be in accordance with all applicable regulatory requirements. Investigational Product will be stored at room temperature (15 C to 30 C (59-86 F)). The Investigational Product distributor will send the initial shipment of Investigational Product and Investigational Product re-supply will occur automatically via site inventory monitoring in IVRS. Under normal conditions of handling and administration, investigational product is not expected to pose significant safety risks to site staff. A Material Safety Data Sheet (MSDS) describing the occupational hazards and recommended handling precautions will be provided to site staff if required by local laws or will otherwise be available from Sponsor upon request. Investigational product must be stored in a secure area under the appropriate physical conditions for the product. Access to and administration of the investigational product will be limited to the investigator and authorized site staff. Investigational product must be dispensed or administered only to subjects enrolled in the study and in accordance with the protocol. Subjects should be instructed to bring their used and unused Investigational Product and containers with them to every visit to assess compliance and drug accountability. Any unused Investigational Product will be returned by the Investigator to the Sponsor Dosing Subjects should be recommended to take Investigational Product once daily preferably in the morning, with or without food, throughout the study unless otherwise instructed. The target dose of tadalafil is 40mg once daily (as this is the licensed dose) in the form of 2 x 20mg tablets, or matching placebo and will be initiated at 20mg od and titrated to 40mg od at week 4. Those subjects with mild to moderate impaired renal function (creatinine clearance >30mL/min and <80 ml/min) should be carefully assess for tolerability prior to any dose up titration as tadalafil exposure has been shown to double in these subjects. The investigator may decide that the subject remain on 20mg od or uptitrated to 40mg od. It should be noted that since the recommended dose for tadalafil for subjects with PAH without renal insufficiency is 40 mg, it is possible that some subjects in the trial may be under dosed on tadalafil for the first 4 weeks of the trial. The dosing will follow the dose schedule below: Combination Therapy Group The dose of tadalafil will be initiated at 20mg and titrated to 40mg after 4 weeks. This rationale for this is to try to minimise any potential tolerability issues, particularly those that may be due to a synergistic effect of the 2 investigational products 45

48 CONFIDENTIAL GM2008/00365/00 AMB The initial dose of tadalafil will be 20mg once daily in the form of 1 x 20mg tablet and 1x tadalafil-matching placebo tablet. The subject will take this dose for the first 4 weeks of treatment. The 20mg od dose may not provide subjects will optimal efficacy. At the 4 week visit the subject will be automatically up-titrated to the target dose of 40mg tadalafil once daily, in the form of 2 x 20mg tadalafil tablets Subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80 ml/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on the 20mg od dose or up-titrate to 40mg od. Tadalafil Monotherapy Group The initial dose of tadalafil will be 20mg once daily in the form of 1 x 20mg tablet and 1x tadalafil-matching placebo tablet. The subject will take this dose for the first 4 weeks of treatment. The 20mg od dose may not provide subjects will optimal efficacy for the first 4 week. The rationale for starting with 20mg is to provide a better comparator to the combination therapy group as well as enabling subjects with mild to moderate renal impairment to follow the recommended dose regimen. At the 4 week visit the subject will be up-titrated to the target dose of 40mg tadalafil once daily, in the form of 2 x 20mg tadalafil tablets Those subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on the 20mg od dose or up-titrate to 40mg od Down-titration from tadalafil 40mg to tadalafil 20mg may be requested following consultation with the Sponsor Medical Monitor should the subject not tolerate the study medication. See section for details. At the Investigator s discretion, chronic parenteral prostanoid therapy may be initiated at any time throughout the study (which meets the criterion for clinical failure). Other PAH therapies may not be added unless a subject has experienced a clinical failure event Down titration of Investigational Product Should the subject not tolerate Investigational Product the following algorithm should be followed: 46

49 CONFIDENTIAL GM2008/00365/00 AMB Separate dosing of Investigational Products in to morning and evening, i.e., ambrisentan in the morning and tadalafil in the evening. DO NOT split the doses (e.g. ABS 5mg AM and ABS 5mg PM is not permitted). 2. Request a blinded down-titration of ambrisentan from 10mg once daily to 5mg once daily using the Interactive Voice Response System (IVRS) (see section 5.2). It should be noted that subjects in TAD monotherapy group will not have a change in dose. Therefore all subjects should be closely monitored for continued lack of tolerance to IP. 3. If the subject continues to experience tolerability issues the investigator may seek permission from the Sponsor medical monitor to down-titrate the subject to tadalafil 20mg once daily. If permission is given the investigator may request a blinded down-titration using the Interactive Voice Response System (IVRS) (see section 5.2). The investigator should complete a Change of Randomised Treatment visit prior to any changes in study medication. If tolerance issues persist the investigator should consider withdrawing the subject from Investigational Product. A Change of Randomised Treatment Visit should be performed Treatment following clinical failure event Following the declaration of a clinical failure event by an Investigator, based on the criteria given, the investigator may do either of the following: Request the subject receive blinded combination therapy. It should be noted that only those subjects on monotherapy will receive a change in their treatment regimen (or those in whom events occur in the first 8 weeks). Subjects in the combination arm will remain on the same combination therapy. Therefore if this is chosen it is important to re-assess the subjects status. Non-parenteral prostanoids may be given following a clinical failure event. These will not be provided by the sponsor. Both of these options may be utilised following a TtCF event but may not be initiated at the same time. Should blinded combination therapy be requested first, non-parenteral prostanoids may not be given until after stage 3 blinded combination therapy has been initiated, unless subject remains on current stage due to tolerability reasons (see section ). Should non-parenteral prostanoids be initiated first, blinded combination therapy may not be requested until the next clinic visit. Parenteral prostanoids may be given at any point during the study. Prior to any change in PAH treatment, a subject is required to complete a Change of Treatment Visit to assess safety and efficacy. 47

50 CONFIDENTIAL GM2008/00365/00 AMB All events will be retrospectively adjudicated by a clinical endpoint committee Blinded Combination Therapy Should the investigator choose the subject receive combination therapy following an event it will be provided in the following format to ensure that the blind is maintained (investigator will know that the subject is on combination therapy, but not what group the subject was randomised to): Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg (event prior to week 4) ABS 5mg + TAD 40mg (TAD optimised) ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 5mg + TAD 40mg (event prior to week 8) ABS 10mg + TAD 40mg (ABS dose optimised) ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg (event post week 8) ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo (event prior to week 8) ABS 10mg + TAD PBO (ABS optimised) ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD Placebo (event post week 8) ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg Monotherapy arm: tadalafil group ABS Placebo + TAD 40mg (event post week 4) ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg For those Subjects in whom the investigator has requested either no uptitration in Investigational Product at week 8 or a down titration in Investigational Product based on tolerability reasons, the following format will apply Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg Monotherapy arm: ambrisentan group 48

51 CONFIDENTIAL GM2008/00365/00 AMB ABS 5mg + TAD Placebo ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS 5mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg Combination Therapy will be fully blinded. Investigators are required to complete a Change of Randomised Treatment visit prior to a subject receiving blinded combination therapy and prior to progressing to the next treatment Stage. In addition subjects should continue to follow the 4 weekly safety visit schedule throughout the treatment Stages Subjects will initially receive Medical Intervention Stage 1, then Stage 2, then Stage 3, upon request of the investigator Subjects must be assessed in the clinic no more than 8 weeks following initiation of Stage 1 (by completing a Change of Randomised Treatment Visit) If the investigator judges that the subject has not sufficiently tolerated any of the treatment Stages, the Investigator may request that the subject remain on their current Stage. Subjects in whom the investigator has requested either no uptitration in study medication or a down titration in study medication based on tolerability reasons may be re-uptitrated at any point at the request of the investigator. This uptitration will follow the format in the first Medical Intervention table above Treatment Assignment Subjects will be assigned to study treatment in accordance with the randomisation schedule. The randomisation code will be generated using the Sponsor randomisation system Randall. Randomisation will be performed by an IVRS which will direct the investigator to assign the appropriate treatment to each subject. Eligible subjects will be stratified based on the underlying aetiology of PAH (IPAH/HPAH and Non-IPAH) and WHO Functional Class (II and III). Subjects will be randomised 2:1:1 to either the combination therapy arm (ambrisentan and tadalafil) or to the monotherapy arm (ambrisentan group or tadalafil group). Subjects will also be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group Blinding Blinding will be accomplished by providing ambrisentan or placebo in tablets that are visually indistinguishable and provided in numbered containers and by providing tadalafil or placebo in tablets that are visually indistinguishable and provided in numbered containers. Only the number of the containers to be administered to a given study subject, and not the identity of the Investigational Product, will be provided to sites. 49

52 CONFIDENTIAL GM2008/00365/00 AMB Please refer to section 3.1 (Study Design) and section 5 (Investigational Product) for details. Unblinding of randomised treatment will occur for all subjects after the database has been locked (Database Lock 1). The investigator or treating physician may request unblinding of a subject s treatment assignment only in the case of an emergency, when knowledge of the study treatment is essential for the appropriate clinical management or welfare of the subject. The investigator must first discuss options with the Sponsor Medical Monitor or appropriate Sponsor study personnel before unblinding the subject s treatment assignment. If this is impractical, the investigator must notify the Sponsor as soon as possible, but without revealing the treatment assignment of the unblinded subject, unless that information is important for the safety of subjects currently in the study. The date and reason for the unblinding must be recorded in the appropriate data collection tool. If the blind is broken by the investigator, the subject will be permanently discontinued from IP and a Change of Randomised Treatment visit will be completed. The subject will continue to be followed until the end of the study. Sponsor s Central Safety Department staff may unblind the treatment assignment for any subject with an SAE. If the SAE requires that an expedited regulatory report be sent to one or more regulatory agencies, a copy of the report, identifying the subject s treatment assignment, may be sent to clinical investigators in accordance with local regulations and/or Sponsor policy Product Accountability In accordance with local regulatory requirements, the investigator, designated site staff, or head of the medical institution (where applicable) must document the amount of Sponsor investigational product dispensed and/or administered to study subjects, the amount returned by study subjects, and the amount received from and returned to Sponsor, when applicable. Product accountability records must be maintained throughout the course of the study Treatment Compliance Subjects should be instructed to bring their used and unused Investigational Product containers with them to every visit to assess compliance and drug accountability Concomitant Medications and Non-Drug Therapies Permitted Medications and Non-Drug Therapies Standard medical treatment(s) being taken by the subject upon study entry may be maintained throughout the study. 50

53 CONFIDENTIAL GM2008/00365/00 AMB Cyclosporine A: Steady-state co-administration of ambrisentan and cyclosporine A resulted in a 2-fold increase in ambrisentan exposure in healthy volunteers. This may be due to the inhibition by cyclosporine A of transporters and metabolic enzymes involved in the pharmacokinetics of ambrisentan. Therefore the dose of ambrisentan should be limited to 5 mg once daily when co-administered with cyclosporine A. Multiple doses of ambrisentan had no effect on cyclosporine A exposure, and no dose adjustment of cyclosporine A is warranted. Alpha blockers: Tadalafil and alpha adrenergic blocking agents are vasodilators with blood pressure-lowering effects. When vasodilators are used in combination, an additive effect on blood pressure may be anticipated. In some patients, concomitant use of these two drug classes can lower blood pressure significantly which may lead to symptomatic hypotension (e.g., fainting). Although no clinically significant effect on blood pressure was observed in healthy volunteers when tadalafil was co-administered with either alfuzosin or tamsulosin, symptomatic augmentation of the blood pressure lowering effect of doxazosin was reported, including syncope, when tadalafil was co-administered. Given the possibility that the combination of these two classes of drugs may lead to additive effects on blood pressure, caution should be exercised when deciding to coadminister tadalafil with alpha blockers. Alcohol: Three clinical trials were conducted in healthy volunteers to study the effects of coadministration of tadalafil and alcohol, both mild systemic vasodilators. Some subjects in these trials experienced clinically significant decreases in blood pressure, postural dizziness, and orthostatic hypotension when alcohol (0.7g/kg which is roughly equivalent to 6 ounces of 80-proof vodka in an 80-kg male consumed within 10 minutes) was administered in conjunction with either 10mg or 20 mg of tadalafil. Four ounces of 80- proof vodka administered within 10 minutes in conjunction with 20 mg tadalafil was not associated with orthostatic hypotension or potentiation of the hypotensive effects of alcohol. Therefore, it is recommended that subjects participating in the AMBITION trial to limit alcohol consumption to less than 5 units (per Adcirca USPI) Prohibited Medications and Non-Drug Therapies Drugs prohibited while receiving investigational product: Other ERAs, such as Bosentan (Tracleer) and Sitaxentan (Thelin) Commercial ambrisentan (Volibris or Letairis) Inhaled nitric oxide Intravenous inotropes (e.g. dopamine, dobutamine) PDE-5i such as sildenafil (Revatio or Viagra) and vardenafil (Levitra) Commercial tadalafil (Adcirca or Cialis) Any other investigational therapy All forms of prostanoids are prohibited unless a subject has met the definition of clinical failure (Section 3.1) 51

54 Nitrates CONFIDENTIAL GM2008/00365/00 AMB Potent inhibitors of CYP3A4 (e.g. protease inhibitors, systemic ketoconazole, or systemic itraconazole) Potent inducers of CYP3A4 (e.g. rifampicin) 5.7. Treatment after the End of the Study Following the first database lock, subjects will be requested to return for an End of Study visit where they will have their final assessment done and be unblinded to study treatment. The Investigator should then treat the subjects according to best standard of care available to the investigator. Subjects will be followed for 30 days post their last dose of study medication. The sponsor will no longer supply ambrisentan or tadalafil after the end of the study except in those countries where a marketing license is yet to be granted Treatment of Investigational Product Overdose Ambrisentan There is no experience in PAH patients of ambrisentan at daily doses greater than 10mg. In healthy volunteers, single doses of 50mg and 100mg (5 to 10 times the maximum recommended dose) were associated with headache, flushing, dizziness, nausea and nasal congestion. Due to the mechanism of action, an overdose of ambrisentan could potentially result in hypotension. In the case of pronounced hypotension, active cardiovascular support may be required. No specific antidote is available. Advice on overdose is available in the product label Tadalafil Single doses of up to 500 mg of tadalafil have been given to healthy subjects, and multiple daily doses up to 100 mg have been given to patients with erectile dysfunction. Adverse events were similar to those seen at lower doses. In cases of overdose, standard supportive measures should be adopted as required. Haemodialysis contributes negligibly to tadalafil elimination. No specific antidote is available. Advice on overdose is available in the product label. 52

55 CONFIDENTIAL GM2008/00365/00 AMB STUDY ASSESSMENTS AND PROCEDURES Procedures Screen Baseline Week (Wk)/Visit Wk Wk -4 0 Visit Windows Wk 4 Wk 8 Wk 16 Wk week intervals Treatment FAV 3 EOS Visit 4 Safety Visits / telephone contact 4 Weekly 1 Unscheduled/ Change of Tx 2 Follow Up Call 13 All visit windows are +/- 3 days with the exception of screen baseline. If eligible, subjects should be randomised within 4 weeks of the screening visit. Written Informed X Consent Subject X Demography Medical History X Disease History X Therapy History X Inclusion/Exclusion X Criteria Efficacy Assessments 6 MWD 5 X X X X X X X X X X BDI Score 5 X X X X X X X X X X WHO FC assessment X X X X X X X X X X Clinical Failure X X X X X X X X X Health Outcome Assessments SF-36 X X X X X X X 14 Camphor X X X X X X X 14 Safety Assessments Concomitant Medication X X X X X X X X X X X Physical X X X Examination Vital Signs 6 X X X X X X X X X X 12-lead ECG X X X Pulse Oximetry X Pulmonary Function Test 7 X Adverse Events X X X X X X X X X X X Serious Adverse X X X X X X X X X X X X Events Laboratory Assessments Haematology X X X X X X X X X X Chemistry (inc. X X X X X X X X X X X LFT s) 8 Testicular Function 9 X X X X X 12 X X X Urinalysis X X X Pregnancy Test 10 S U S S S S S S S S S NT-pro-BNP X X X X X X X X X Coagulation Pharmacogenetics (optional) X X 53

56 CONFIDENTIAL GM2008/00365/00 AMB Investigational product (IP) Dispense IP X X X X X X X X 11 Assess IP X X X X X X X X compliance IVRS X X X X X X X X X X 11 Peak-Trough X 1. Safety Visits will be required every 4 weeks after week 4 unless a clinic visit is scheduled. A phone call is required to assess concomitant medications and adverse events (including serious). See the Clinical Laboratory Test section for details of the laboratory tests required. If any additional procedures are performed at the safety visit, an unscheduled visit should be completed. 2. Unscheduled / Change of randomised treatment visit. Also refer to foot note 12. If a subject permanently withdraws from investigational product and from the study at the same time, the End of Study Visit should be completed. 3. Final Assessment Visit 4. End of Study. (This visit should also be completed by subjects who prematurely withdraw from the study) 5. Subjects will take their investigational products dose at least 16 hours prior to the 6MWD and BDI assessments 6. Includes weight, blood pressure and heart rate plus height (height at screening only). 7. Pulmonary function test may also be carried out at screening if not done within 24 weeks prior. 8. Liver Function Test (LFT) should be measured and Investigational Product dose adjusted / stopped as specified in section and Testicular function (Total testosterone, sex hormone binding globulin (SHBG needed to calculate free testosterone), FSH, LH, and inhibin B. For males only. 10. Pregnancy test: urine = U, serum = S. For women of child bearing potential only 11. If applicable 12. Alternate visits only (i.e. every 24 weeks) 13. Follow up telephone contact 30 days after last dose of Investigational Product 14. Only if change of randomised treatment See appendix 2 for instructions on how to conduct the 6 minute walk test. (2 tests [i.e. 1 at screening and 1 at randomisation] should be performed during the screening period such that 2 consecutive tests do not differ by more than 10%) 6.1. Critical Baseline Assessments All ongoing conditions and relevant/significant medical history (including all major hospitalizations and surgeries) will be recorded at the Screening Visit. Symptoms related to PAH and/or the underlying aetiology of the disease need not be listed on the medical history form; however, worsening of any symptoms during the course of this study must be captured as an AE. Review historical diagnostic tests, such as Right Heart Catheter (RHC); chest X-ray; echocardiogram; ventilation/perfusion scan, computed axial tomography (CT) or spiral CT, and pulmonary arteriogram (if applicable), to ensure that none of the exclusion criteria for PAH aetiology are met. Prior PAH treatment history will also be assessed. Historical haemodynamic data (e.g., mpap, RAP, PVR, and cardiac index) based on the most recent RHC prior to starting this study will be collected. 54

57 CONFIDENTIAL GM2008/00365/00 AMB Efficacy Primary Endpoint Time to Clinical Failure The primary efficacy endpoint is the time to the first clinical failure event of PAH. Time to clinical failure is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) i. Non-elective hospitalization for worsening PAH ii. Lung or heart/lung transplant iii. Atrial septostomy iv. Initiation of parenteral prostanoid therapy Disease progression (adjudicated) i. >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) i. Receiving randomised treatment for at least 6 months ii. A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days iii. Sustained WHO class III or IV symptoms for 6 months Time to clinical worsening (death, hospitalization for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Medical Interventions for PAH During the Study: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary endpoint event and subjects may continue in the study following initiation. If a subject experiences an event of clinical failure the following medical interventions may also be utilized. The clinical failure event necessary for medical intervention is based on the opinion of the investigator and does not require prior adjudication from the clinical endpoint committee. The investigator may choose to request blinded combination therapy to ensure that the subject is receiving ambrisentan/tadalafil combination therapy The investigator may choose to add prostanoid therapy (parenteral or nonparenteral) (see section 5.1.4). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first 55

58 CONFIDENTIAL GM2008/00365/00 AMB intervention (see section ) the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy Initiate parenteral prostanoid therapy Request blinded ambrisentan/tadalafil combination therapy. The CEC will review the case (blinded to treatment and investigator) and decide if the event meets the definition of clinical failure of PAH. The CEC Charter provides full details. All events of clinical failure of PAH, as well as confirmation that a specific event has not occurred during this time period, will be collected on a clinical failure CRF throughout the study. If a subject prematurely discontinues from the study, an assessment will be made at the time of discontinuation from the study Secondary Endpoints Minute Walk Distance and Borg Dyspnea Index See Appendix 2. The 6MWD will be assessed at each clinic visit. Screening period (i.e. screening visit and baseline visit): Two 6MWD tests should be performed during the screening period, the first 6MWD test at the screening visit, and the second 6MWD test at the baseline visit. The tests should not vary by >10%. A maximum of 4 tests may be conducted during the Screening Period per subject to generate two consecutive tests that meet the criteria below. A maximum of two tests can be performed per day and tests must be separated by at least two hours. If two consecutive tests that vary by <10% cannot be performed the subject should not be entered into the study. If they do not vary by >10% the average value of the two consecutive tests should be used as the baseline value Borg Dyspnea Index See Appendix 3. The Borg Dyspnea Index score will be collected at each clinic visit. It should be performed immediately after the 6MWD test. Screening period (i.e. screening visit and baseline visit): The Borg Dyspnea Index should be performed straight after every 6MWD test. A maximum of 4 tests may be conducted during the Screening Period per subject to generate two consecutive tests that meet the criteria in section The Borg score following the screening 6MWD test should be 56

59 CONFIDENTIAL GM2008/00365/00 AMB recorded as the screening value, and the average value of the two scores following the two consecutive 6MWD tests should be used as the baseline Borg value N-Terminal pro-b-type Natriuretic Peptide Blood samples for determination of NT-proBNP plasma concentrations will be collected at each clinic visit from randomisation onwards and analysed via a central laboratory. Investigators will remain blinded to NT-proBNP results during the study WHO Functional Class assessment WHO functional class will be assessed at each clinic visit. Every effort will be made to have the same Investigator assess the subject at each study visit Exploratory Endpoints Peak-trough assessment of 6MWD At the Randomisation visit (Week 0), all subjects will be randomly assigned to perform the Week 16 6MWD assessment at either peak or trough ambrisentan concentrations. Prior to arriving at the investigative site for the Week 16 visit, subjects will not administer their daily dose of either Investigational Products (ABS or TAD). The site will dispense blinded Investigational Product (ambrisentan/ambrisentan matching placebo) as directed by the IVRS for the peak-trough assessment. Subjects in the ambrisentan monotherapy arm and the combination arm will receive either ambrisentan followed by placebo (peak assessment) or placebo followed by ambrisentan (trough assessment). Subjects randomised to the tadalafil monotherapy arm will receive placebo followed by placebo to maintain the blind. Subjects will take the first dose of ambrisentan Investigational Product (placebo or ambrisentan) and the 6MWD will be assessed 2 hours (+/- 30 minutes) after Investigational Product administration (approximate C max ). Borg dyspnea index (BDI) will be assessed immediately following this test. After the remaining clinical assessments have been completed, subjects will then take the second dose of ambrisentan Investigational Product to receive either their daily dose of ambrisentan or placebo to maintain the blind. Subjects will also take their daily dose of tadalafil Investigational Product at this time Safety Electrocardiogram A 12-lead ECG will be performed at Screening, Week 24 and Final Assessment Visit. Any clinically relevant ECG findings will be captured in the CRF. Any clinically relevant changes since the Screening Visit will be captured as an AE in the CRF. 57

60 Vital Signs and Body Weight and Height CONFIDENTIAL GM2008/00365/00 AMB Vital signs (including heart rate and blood pressure) will be collected at each clinic visit. One measurement of blood pressure and heart rate after the subject has been sitting quietly for at least 5 minutes should be taken. This same procedure should be followed throughout the study. Oxygen saturation will be collected at screening only. All measures of blood pressure will be performed using standard sphygmomanometry. If possible, the same sphygmomanometer and arm should be used. In addition, body weight will be collected at each clinic visit. Height will be collected at screening only Clinical Laboratory Tests Safety Tests The following tests are required at all clinic visits (including FAV, EOS visit, Unscheduled/Change of Randomised Treatment Visit (if applicable)) and will be performed by a central laboratory: Chemistry: serum ALT, AST, alkaline phosphatase, GGT, total bilirubin, creatinine, egfr, amylase, blood urea nitrogen (BUN), sodium, potassium, chloride, bicarbonate, calcium, total cholesterol, uric acid, glucose, total protein and albumin. Haematology: haemoglobin, haematocrit, red cell count, red cell indices, white blood cell count (total and differential) and platelet count Pregnancy*: Serum pregnancy test at all visits, except baseline when a urine pregnancy test will be performed The following tests are required at all safety visits and will be performed by a local phlebotomy laboratory or at the Investigator clinic (or the central lab with Sponsor permission): Liver function: serum ALT, AST, alkaline phosphatase, GGT, total bilirubin Pregnancy* : serum *In female subjects of child bearing potential The following test will be performed in males only at baseline, weeks 4, 16, 24, subsequently at alternate 12 week visits (i.e. every 24 weeks), FAV and EOS visits as well as Unscheduled/Change of Randomised Treatment Visit (if applicable) and will be performed by a central laboratory: 58

61 CONFIDENTIAL GM2008/00365/00 AMB Testicular Function: Total testosterone, Sex hormone binding globulin (SHBG needed to calculate free testosterone), FSH, LH and inhibin B The following test will be performed at baseline, week 24, Final Assessment Visit and will be performed by a central laboratory: Total urinalysis: including a microscopic examination of the urine sediment Additional laboratory tests may be performed by a local laboratory or the central laboratory if clinically relevant abnormal values are obtained at any time during the course of the study. Local laboratory reports must be immediately reviewed by the Investigator. Laboratory normal ranges will be collected Pharmacogenetics A pharmacogenetics sample will be collected ideally at baseline (or after baseline if the sample cannot be collected then). See appendix 4 for details Collection, processing, labelling, and shipping of the pharmacogenetic samples will follow central laboratory guidelines Management, Monitoring and Follow-up of Serum Aminotransferase Abnormalities All subjects who are receiving investigational product will have blood drawn every 4 weeks to assess hepatic function. Assessments for hepatic function include serum aminotransferases (alanine aminotransferase, ALT or aspartate aminotransferase, AST), alkaline phosphatase, gamma glutamyl transferase (GGT), and total bilirubin concentrations. In order to ensure the safety of all subjects, the results of these tests must be reviewed by the Investigator immediately upon receipt. An increase in serum ALT or AST concentration >3xULN must be confirmed by a second test (>3xULN) that is collected no more than 7 days after receipt of the initial lab report and is analyzed by the central laboratory. The requirements presented in Table 1 must be applied to the review and management of elevated serum aminotransferases in all subjects receiving investigational product; however, investigators may elect to interrupt or discontinue investigative product(s) at a lower threshold based on their assessment of the patient s clinical status. Table 1 Management, Monitoring and Follow-up of Subjects Experiencing Serum Aminotransferase Elevations 59

62 CONFIDENTIAL GM2008/00365/00 AMB Serum Aminotransferase Elevations ALT>3xULN and bilirubin 2XULN or if ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. ALT/AST >3x and 5xULN ALT/AST >5x and 8xULN ALT/AST >8xULN Actions to be Taken Stop investigational product treatment immediately. Report the liver event to the Medical Monitor within 24 hours of learning of its occurrence and report as an SAE. Make every reasonable attempt to have subjects return to clinic within 24 hours for repeat liver chemistries, liver event follow up assessments, and close monitoring Complete liver event CRF within 24 hours. After discontinuation of investigational product, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. A specialist or hepatology consultation is recommended investigational product may not be reintroduced. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. If confirmed, interrupt therapy. Continue to monitor serum aminotransferase concentrations every 2 weeks or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of investigational product may be considered by the Investigator after consultation with the Medical Monitor. Stop investigational product treatment immediately. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. Continue to monitor serum aminotransferase concentrations every 2 weeks, or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of investigational product may be considered by the Investigator after consultation with the Medical Monitor. Stop investigational product treatment immediately. Report the liver event to the Medical Monitor and within 24 hours of learning of its occurrence. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. 60

63 CONFIDENTIAL GM2008/00365/00 AMB After discontinuation of investigational product, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. investigational product may not be reintroduced. *If investigational product is reintroduced, serum aminotransferase concentrations should be checked within 3 days, and then every 2 weeks or more frequently if clinically indicated. Once investigator, in consultation with Medical Monitor, feels that subject is clinically stable, liver safety testing may be resumed at standard protocol defined intervals. If at any time these subjects experience serum aminotransferase elevations, proceed as described above. Subjects withdrawn from the investigational product may remain in the study for followup and collection of data for outcome Liver event follow up assessments For liver events, make every attempt to carry out the liver event follow up assessments described below: Viral hepatitis serology including: Hepatitis A IgM antibody; Hepatitis B surface antigen and Hepatitis B Core Antibody (IgM); Hepatitis C RNA; Cytomegalovirus IgM antibody; Epstein-Barr viral capsid antigen IgM antibody (or if unavailable, obtain heterophile antibody or monospot testing); Hepatitis E IgM antibody (if subject resides outside the US or Canada, or has travelled outside US or Canada in past 3 months); Blood sample for PK analysis, obtained within 24 hours of last dose. Record the date/time of the PK blood sample draw and the date/time of the last dose of investigational product(s) prior to blood sample draw on the CRF. If the date or time of the last dose is unclear, provide the subject s best approximation. If the date/time of the last dose cannot be approximated OR a PK sample cannot be collected in the time period indicated above, do not obtain a PK sample. Instructions for sample handling and shipping are in the SPM. Serum creatine phosphokinase (CPK) and lactate dehydrogenase (LDH). Fractionate bilirubin, if total bilirubin 2xULN.. Serum bilirubin fractionation should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick. 61

64 CONFIDENTIAL GM2008/00365/00 AMB Obtain complete blood count with differential to assess eosinophilia Record the appearance or worsening of clinical symptoms of hepatitis or hypersensitivity, such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever rash or eosinophilia as relevant on the AE report form Record use of concomitant medications, acetaminophen, herbal remedies, other over the counter medications, or putative hepatotoxins, on the concomitant medications report form. Record alcohol use on the liver event alcohol intake case report form Complete the liver imaging and/or liver biopsy CRFs if these tests are performed The following are required for subjects with ALT 3xULN and bilirubin 2xULN (>35% direct) but are optional for other abnormal liver chemistries: Anti-nuclear antibody, anti-smooth muscle antibody, and Type 1 anti-liver kidney microsomal antibodies. Liver imaging (ultrasound, magnetic resonance, or computerized tomography) to evaluate liver disease Adverse Events The investigator or site staff will be responsible for detecting, documenting and reporting events that meet the definition of an AE or SAE Definition of an AE Any untoward medical occurrence in a patient or clinical investigation subject, temporally associated with the use of a medicinal product, whether or not considered related to the medicinal product. Note: An AE can therefore be any unfavourable and unintended sign (including an abnormal laboratory finding), symptom, or disease (new or exacerbated) temporally associated with the use of a medicinal product. For marketed medicinal products, this also includes failure to produce expected benefits (i.e., lack of efficacy), abuse or misuse. Events meeting the definition of an AE include: Exacerbation of a chronic or intermittent pre-existing condition including either an increase in frequency and/or intensity of the condition New conditions detected or diagnosed after investigational product administration even though it may have been present prior to the start of the study Signs, symptoms, or the clinical sequelae of a suspected interaction Signs, symptoms, or the clinical sequelae of a suspected overdose of either investigational product or a concomitant medication (overdose per se will not be reported as an AE/SAE). 62

65 CONFIDENTIAL GM2008/00365/00 AMB Lack of efficacy or failure of expected pharmacological action per se will not be reported as an AE or SAE. However, the signs and symptoms and/or clinical sequelae resulting from lack of efficacy will be reported if they fulfil the definition of an AE or SAE. Events that do not meet the definition of an AE include: Medical or surgical procedure (e.g., endoscopy, appendectomy); the condition that leads to the procedure is an AE Situations where an untoward medical occurrence did not occur (social and/or convenience admission to a hospital) Anticipated day-to-day fluctuations of pre-existing disease(s) or condition(s) present or detected at the start of the study that do not worsen The disease/disorder being studied, or expected progression, signs, or symptoms of the disease/disorder being studied, unless more severe than expected for the subject s condition Definition of a SAE A serious adverse event is any untoward medical occurrence that, at any dose: a. Results in death b. Is life-threatening NOTE: The term 'life-threatening' in the definition of 'serious' refers to an event in which the subject was at risk of death at the time of the event. It does not refer to an event, which hypothetically might have caused death, if it were more severe. c. Requires hospitalization or prolongation of existing hospitalization NOTE: In general, hospitalization signifies that the subject has been detained (usually involving at least an overnight stay) at the hospital or emergency ward for observation and/or treatment that would not have been appropriate in the physician s office or out-patient setting. Complications that occur during hospitalization are AEs. If a complication prolongs hospitalization or fulfils any other serious criteria, the event is serious. When in doubt as to whether hospitalization occurred or was necessary, the AE should be considered serious. Hospitalization for elective treatment of a pre-existing condition that did not worsen from baseline is not considered an AE. d. Results in disability/incapacity, or NOTE: The term disability means a substantial disruption of a person s ability to conduct normal life functions. This definition is not intended to include experiences of relatively minor medical significance such as uncomplicated headache, nausea, vomiting, diarrhoea, influenza, and accidental trauma (e.g. sprained ankle) which 63

66 CONFIDENTIAL GM2008/00365/00 AMB may interfere or prevent everyday life functions but do not constitute a substantial disruption. e. Is a congenital anomaly/birth defect f. Medical or scientific judgment should be exercised in deciding whether reporting is appropriate in other situations, such as important medical events that may not be immediately life-threatening or result in death or hospitalization but may jeopardize the subject or may require medical or surgical intervention to prevent one of the other outcomes listed in the above definition. These should also be considered serious. Examples of such events are invasive or malignant cancers, intensive treatment in an emergency room or at home for allergic bronchospasm, blood dyscrasias or convulsions that do not result in hospitalization, or development of drug dependency or drug abuse. g. All events of possible drug-induced liver injury with hyperbilirubinaemia defined as ALT 3xULN and bilirubin 2xULN (>35% direct) (or ALT 3xULN and INR>1.5, if INR measured) termed Hy s Law events (INR measurement is not required and the threshold value stated will not apply to patients receiving anticoagulants). NOTE: bilirubin fractionation is performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If testing is unavailable and a subject meets the criterion of total bilirubin 2xULN, then the event is still reported as an SAE. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury Laboratory and Other Safety Assessment Abnormalities Reported as AEs and SAEs Any abnormal laboratory test results (haematology, clinical chemistry, or urinalysis) or other safety assessments (e.g., ECGs, radiological scans, vital signs measurements), including those that worsen from baseline, and felt to be clinically significant in the medical and scientific judgement of the investigator are to be recorded as AEs or SAEs. However, any clinically significant safety assessments that are associated with the underlying disease, unless judged by the investigator to be more severe than expected for the subject s condition, are not to be reported as AEs or SAEs. All events of possible drug-induced liver injury with hyperbilirubinaemia (defined as ALT 3xULN plus bilirubin 2xULN and/or INR>1.5) or Hy s Law events, require immediate Investigational Product cessation and reporting as an SAE (see section g above). NOTE: bilirubin fractionation is performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. 64

67 CONFIDENTIAL GM2008/00365/00 AMB If testing is unavailable and a subject meets the criterion of total bilirubin 2.0 xuln, then the event is still reported as an SAE. If INR is obtained, include values on the SAE form. INR elevations>1.5 suggest severe liver injury Pregnancy Subjects who become pregnant during the study must discontinue Investigational Product immediately. The Investigator should counsel the subject regarding the possible effects of prior Investigational Product exposure on the foetus (See SmPC and USPI) and the need to inform the study site of the outcome of the pregnancy. Subjects should be instructed to notify the Investigator if they become pregnant at any time during the study, or if they become pregnant within 30 days of last Investigational Product dose. Any pregnancy that occurs during study participation must be reported using a clinical trial pregnancy form. To ensure subject safety, each pregnancy must be reported to Sponsor within 24 hours of learning of its occurrence. The pregnancy must be followed up to determine outcome (including premature termination) and status of mother and child. Pregnancy complications and elective terminations for medical reasons must be reported as an AE or SAE. Spontaneous abortions must be reported as an SAE. Any SAE occurring in association with a pregnancy, brought to the investigator s attention after the subject has completed the study and considered by the investigator as possibly related to the investigational product, must be promptly reported to Sponsor Time Period and Frequency of Detecting AEs and SAEs AEs will be collected from the start of investigational product and until 30 days post the last dose of investigational product. AE s occurring between screening and randomisation will be recorded as medical history. All deaths, regardless of cause or relationship, must be reported for subjects on study and for all deaths occurring within 30 days of last investigational product dose or within 30 days of last study evaluation, whichever is longer. SAEs will be collected over the same time period as stated above for AEs. However, any SAEs assessed as related to study participation (e.g., investigational product, protocolmandated procedures, invasive tests, or change in existing therapy) or related to a Sponsor concomitant medication, will be recorded from the time a subject consents to participate in the study up to and including any follow up contact. All SAEs will be reported to Sponsor within 24 hours, as indicated in Section

68 CONFIDENTIAL GM2008/00365/00 AMB Prompt Reporting of Serious Adverse Events and Other Events to Sponsor SAEs, pregnancies, and liver function abnormalities meeting pre-defined criteria will be reported promptly by the investigator to Sponsor as described in the following table once the investigator determines that the event meets the protocol definition for that event. Initial Reports Follow-up Information on a Previous Report Type of Event Time Frame Documents Time Frame Documents All SAEs 24 hours SAE data 24 hours Updated SAE collection tool Pregnancy 24 hours Pregnancy Notification Form Liver chemistry abnormalities ALT 3xULN and 24 hours* SAE data Bilirubin 2xULN collection tool. (>35% direct) (or **Liver Event ALT 3xULN and Case Report Form INR>1.5, if INR (CRF) and liver measured)*** imaging and/or biopsy CRFs if ALT 5xULN; ALT 3xULN with hepatitis or rash or 3xULN 4 weeks ALT 3xULN and <5xULN and biliribin <2xULN data collection tool 24 hours Pregnancy Follow up Form 24 hours Updated SAE data collection tool. **Updated Liver Event CRF applicable 24 hours* **Liver Event CRF 24 hours **Updated Liver Event CRF 24 hours* **Liver Event CRF does not need completing unless elevations persist for 4 weeks or subject cannot be monitored weekly for 4 weeks *Sponsor to be notified at onset of liver chemistry elevations to discuss subject safety. ** Liver event documents should be completed as soon as possible *** INR measurement is not required; if measured, the threshold value stated will not apply to patients receiving anticoagulants. The method of detecting, recording, evaluating and follow-up of AEs and SAEs plus procedures for completing and transmitting SAE reports to Sponsor are provided in the SPM. Procedures for post-study AEs/SAEs are provided in the SPM. 66

69 CONFIDENTIAL GM2008/00365/00 AMB Regulatory reporting requirements for SAEs Prompt notification of SAEs by the investigator to the Sponsor is essential so that legal obligations and ethical responsibilities towards the safety of subjects are met. The sponsor has a legal responsibility to notify both the local regulatory authority and other regulatory agencies about the safety of a product under clinical investigation. The sponsor will comply with country specific regulatory requirements relating to safety reporting to the regulatory authority, Institutional Review Board (IRB)/Independent Ethics Committee (IEC) and investigators. Investigator safety reports are prepared for suspected unexpected serious adverse reactions according to local regulatory requirements and the sponsor policy and are forwarded to investigators as necessary. An investigator who receives an investigator safety report describing a SAE(s) or other specific safety information (e.g., summary or listing of SAEs) from the sponsor will file it with the IB and will notify the IRB/IEC, if appropriate according to local requirements Health Outcomes Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) Health Related Quality of Life and Quality of Life Measure The CAMPHOR [McKenna SP et al. Quality of Life Research 2006;15: ] is a pulmonary hypertension specific instrument, has been developed and validated to facilitate the measurement of HRQL (symptoms and functioning) and QoL in this patient group. It is provided in Appendix 5. It is the only PH disease specific tool available. Subjects will complete the CAMPHOR questionnaire at Randomisation (Week 0), Weeks 16, 24, at Clinic Visits every 12 weeks subsequent to Week 24, the Change of Randomised Treatment Visit (if applicable), and the FAV and EOS visit. The CAMPHOR will only be used in countries where a validated translation is available. The CAMPHOR consists of three sections. The first two address HRQL issues. 1. The first section is the symptom (impairment) score. It contains 25 negatively weighted items consisting of three sub-scales relating to energy, breathlessness and mood. Each item is in the form of a simple statement to which patients indicate whether or not it is true for them at that moment. This two-point response system was selected as it presents the minimum burden on respondents and is the simplest in relation to ease of scoring and equivalence of translation. 67

70 CONFIDENTIAL GM2008/00365/00 AMB The second section is an Activity (disability) score. This consists of a 15 item scale, each of which relates to activities described by patients as being affected by PAH. The patients rate themselves as being able to perform each activity. 3. The third section is Quality of Life and contains 25 negatively weighted items. A change score of 2 points is meaningful, this is the MID (Minimal Important Difference). Further details of the CAMPHOR will be detailed in the Reporting Analysis Plan (RAP) SF-36 Health Survey The SF-36 Health Survey [Ware, JE et al. SF-36 Health Survey: Manual and Interpretation Guide, Lincoln, RI: QualityMetric Inc, 2000] is provided (the US Version) in Appendix 6. The SF-26v2 Health Survey will be used. The SF-36 Health Survey asks 36 questions to measure functional health and well-being from the patient's point of view. It s called a generic health survey because it can be used across age (18 and older), disease, and treatment group, as opposed to a disease-specific health survey which focuses on a particular condition or disease. Whilst not specifically validated in PAH, it has been used in a number of PAH trials. Subjects will complete the SF-36 Health Survey at Randomisation (Week 0), Weeks 16, 24, at Clinic Visits every 12 weeks subsequent to Week 24, the Change of Randomised Treatment Visit (if applicable), and the FAV and EOS visit. The SF-36 Health Survey consists of 8 health domains (Physical Functioning, Role- Physical, Bodily Pain, General Health, Vitality, Social Functioning, Role-Emotional, and Mental Health) as well as 2 summary measures (Physical Health and Mental Health). Further details of the SF-36 will be detailed in the RAP and the scoring will follow the algorithm of Ware, Kosinski and Gandek (2000). 7. DATA MANAGEMENT For this study subject data will be entered into electronic case report forms (ecrfs), transmitted electronically to the Sponsor and designees, and combined with data provided from other sources in a validated data system. Management of clinical data will be performed in accordance with the applicable sponsor standards and data cleaning procedures to ensure the integrity of the data, e.g., removing errors and inconsistencies in the data. Adverse events and concomitant medications terms will be coded using MedDRA and an internal validated medication dictionary, GSKDrug. ecrfs (including queries and audit trails) will be retained by Sponsor, and copies will be sent to the investigator to maintain as the investigator copy. In all cases, 68

71 CONFIDENTIAL GM2008/00365/00 AMB subject initials will not be collected or transmitted to Sponsor according to Sponsor policy. 8. DATA ANALYSIS AND STATISTICAL CONSIDERATIONS This section will cover the statistical considerations and data analyses in order to assess the primary objective of the study. Further details will be presented in the Reporting Analysis Plan (RAP) Hypotheses The study is designed to determine the efficacy of a treatment strategy of combination therapy compared to monotherapy in subjects with PAH. The null hypothesis tested for the primary endpoint of efficacy is that there is no difference in the time to clinical failure of PAH when treated with monotherapy (ambrisentan or tadalafil) compared to combination therapy (ambrisentan and tadalafil) (i.e., H 0 : HR = 1). The two-sided alternative hypothesis is that there is a difference (i.e., H 1 : HR 1, a two-sided test) Study Design Considerations Sample Size Assumptions This event-driven study requires 82 subjects with a clinical failure event in order to have 90% power with an overall type I error rate (alpha level) of 5%, accounting for two interim analyses as detailed in Section Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of 10% per year, i.e. an absolute difference of 10%. This leads to an approximate hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial. The alpha spending associated with the stopping criteria are described in Section Assuming a recruitment period of 65 weeks (based on a recruitment rate of 5.4 subjects per week) and a total study duration of 129 weeks the study requires 352 subjects (176 per treatment group). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 10% per treatment group at one year has been considered for this calculation. The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have 75% power for each comparison. For this calculation, it has been assumed that there are 55 clinical failure events for 264 subjects (176 on combination and 88 on each monotherapy), no interim analyses and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been assumed. 69

72 CONFIDENTIAL GM2008/00365/00 AMB Calculations were performed in EAST version 5.2, Cytel Inc, using simulated power. Central randomisation is planned stratified by aetiology of PAH (IPAH/HPAH and Non- IPAH) and WHO Functional Class (II and III). Additionally, subjects will be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 176 ABS/TAD subjects and 88 PBO/TAD subjects yields approximately 86% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. Treatment effect and standard deviation were based on data from Phase 2 and Phase 3 clinical studies of ambrisentan. Sample size and final alpha level were calculated using nquery Advisor Version 6.0. Statistical Solutions, Cork, Ireland. Although the sample size was calculated using a 2-sided t-test, the specified analysis for this endpoint uses a Wilcoxon rank sum test; therefore, the actual power may vary slightly Sample Size Sensitivity The sample size required to achieve a given number of necessary primary endpoint events depends on the basic assumptions as outlined in section The power (based on simulation) resulting from an observed 82 events for various reductions in risk different from the assumed 53% are shown in the table below. Table 1 Assumptions for Power Calculation Reduction in Risk Hazard Ratio Approximate Power 40% % 45% % 50% % 53% % 60% % 65% 0.35 >99% Assuming 352 subjects, 82 events, alpha of 0.05, 10% yearly dropout rate per group, and varying 70

73 CONFIDENTIAL GM2008/00365/00 AMB hazard ratios. If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at 352, the study duration may differ from the current planned 129 weeks (see Table 2). Table 2. Sample Size Sensitivity for 90% Power and 52.8% Reduction in Risk and Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n=352) 25% 106 weeks 20% 129 weeks 15% 171 weeks 10% 268 weeks Assuming 352 subjects, 82 events, alpha of 0.05, power of approximately 90%, 10% yearly combination therapy event rate, 10% yearly dropout rate per group and varying Monotherapy yearly event rates If the recruitment rate differs from the assumed recruitment rate of 5.4 subjects per week, the study duration may differ from the current planned 129 weeks (see Table 3). Table 3. Sample Size Sensitivity for 90% Power, 52.8% Reduction in Risk, 352 Subjects, and Varying Recruitment Rates Recruitment Period (Weeks) Recruitment Rate (pts/week) Follow-up Period after Last Subject Enrolled (Weeks) Total Study Duration (Weeks)*

74 CONFIDENTIAL GM2008/00365/00 AMB Assuming 352 subjects total, 82 events, alpha of 0.05; power of approximately 90%, 10% yearly combination therapy event rate, 20% yearly monotherapy event rate,10% yearly dropout rate per group, and varying recruitment rates All calculations were performed using EAST versions 5.2, Cytel Inc, using simulated power Sample Size Re-estimation The event rate will be monitored using data blinded with respect to treatment assignment. Should the observed event rate be lower than anticipated or discontinuation rates (i.e., withdrawal from study or discontinuation of IP) higher than expected, then the sponsors and the Steering Committee will use blinded data to assess the value of increasing study size or length of duration Data Analysis Considerations Analysis Populations The populations to be used in the analysis of this study are: ITT- All Randomised Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. PP-The per-protocol population will consist of the subset of subjects in the ITT without any major protocol violation. If the PP population is greater than 85% of the ITT population or less than 50% of the ITT population, a PP analysis will not be performed Analysis Data Sets All subjects included into this study will be included in the analyses according to the analysis populations defined above in For time to event analyses, all lost to follow-up subjects will be censored at their last know date in the study. For other endpoints, LOCF will be used to account for missing data. 72

75 Treatment Comparisons Primary Comparisons of Interest CONFIDENTIAL GM2008/00365/00 AMB The primary comparison is of the hazard rates of clinical failure between the combination therapy arm (ambrisentan and tadalafil) and the pooled monotherapy arm (ambrisentan and placebo plus tadalafil and placebo). The ITT population will be utilized for this comparison. The comparison will be made at a value close to 5% significance level, adjusting for interim analyses using the alpha-spending approach detailed in Section Other Comparisons of Interest The primary comparison will be between the combination and pooled monotherapy treatment arms. If the combination treatment arm is demonstrated as statistically significant to the pooled monotherapy arm, comparisons of combination treatment versus each individual monotherapy may be performed. Alpha adjustments to accommodate the two monotherapy treatment comparisons will be implemented with a Hochberg approach. It is acknowledged that these comparisons are underpowered. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints. Inferences on the superiority of combination versus pooled monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint is statistically significant will inferences on the Week 24 change from baseline 6MWD endpoint worsening be evaluated. Only if the Week 24 change from baseline 6MWD endpoint is found to be significant will inferences on the percentage of subjects with satisfactory response at Week 24 endpoint be evaluated. This gate keeping approach, described above, will be implemented for all secondary endpoints in the order listed. With this approach, no adjustment to the significance level at each step is needed. The primary, secondary and exploratory analyses may be repeated for the PP population Interim Analysis The IDMC will monitor progress of the study, review safety data to protect subject safety and assess efficacy data against the pre-specified stopping rules identified in the protocol. An external statistician will produce the summaries and analyses required by the IDMC. Further details will be specified in the IDMC charter. Only the IDMC will be authorised to review unblinded interim efficacy and safety analyses during the study. Based on the interim analyses, the IDMC will give 73

76 CONFIDENTIAL GM2008/00365/00 AMB recommendations to the Sponsors and Chair of the Steering Committee regarding whether the trial should be stopped. Safety data will be monitored throughout the study on an ongoing basis. Formal interim analyses of efficacy data will be performed when approximately 50% and 75% of the expected number of primary endpoint events have been reached. Statistical stopping rules for superiority and futility based on O Brien Fleming will then be applied. Table 4. p-values at each Interim Analysis Interim 1 Interim 2 Approximate Percentage of Events 50% 75% Average Time of Analysis (weeks) Probability of stopping for Superiority / Futility 25% / 1% 46% / 4% Efficacy p-value for superiority Futility p-value (non-binding) As the alpha spend associated with each interim is relatively small the final analysis will be performed at alpha= If the study is to be terminated early as a result of an interim analysis, a further additional sensitivity analysis will be performed including all subjects and blinded data collected, i.e. including additional blinded data that did not contribute to the interim analysis. Calculations were performed in EAST version 5.2, Cytel Inc, using simulated power. Further details of the IDMC will be provided in the IDMC Charter Key Elements of Analysis Plan The analyses planned in this protocol will be expanded in the RAP. This will be completed prior to unblinding of the study. Any deviations from the analyses described in this protocol will be described in the RAP and the final study report(s) as appropriate Efficacy Analyses Primary Analysis Time to clinical failure will be displayed as Kaplan-Meier event-free curves from randomisation up to the final visit. Events which occur after the final visit will not be 74

77 CONFIDENTIAL GM2008/00365/00 AMB used in the primary analysis. Differences between the curves will be tested for significance by the stratified log-rank statistic. The stratifying variables are the aetiology of PAH and WHO functional class. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using a Cox proportional hazards regression model. In case data for some subjects is not available following loss to follow up, their event times will be treated as censored for statistical analyses. The first event that occurs will be counted for subjects experiencing more than one component of the primary (composite) endpoint during the study. In such cases, all events will be counted separately for the analysis of the individual components. The primary analysis will include all events (target of 82 events) up to the end of the Final Assessment Visit of the main study phase. Any events occurring after this timepoint will be reported as a sensitivity analysis or as a separate summary Secondary Analyses The analyses of the secondary efficacy endpoints will compare combination therapy to the pooled monotherapy arm consisting of ambrisentan or tadalafil. A multiplicity adjustment will be applied for the secondary endpoints as detailed in Section Individual components of the primary endpoint, time to clinical worsening and death will be analysed as for the primary endpoint. Percentage of subjects with satisfactory clinical response will be summarised and response (yes, no) will be analysed as a binary endpoint using logistic regression, with PAH aetiology and WHO functional class as covariates. Descriptive statistics for each endpoint will be presented for the change from baseline and percentage change from baseline, where appropriate, by treatment group for each visit. Continuous data will be analysed using an analysis of covariance model with terms for treatment, aetiology and baseline measurement (continuous covariate). The point estimates and corresponding 95% confidence intervals for treatment differences will be calculated. Where the data is not normally distributed appropriate methods will be used and detailed in the Reporting and Analysis plan. The 6MWD test will be analysed using the Wilcoxon rank sum test stratified by PAH aetiology and WHO Functional Class. Change from baseline in WHO functional class will be tested using a Wilcoxon rank sum test, stratified by PAH aetiology and WHO functional class at baseline, between treatments. An analysis on change from baseline categorization (+2, +1, 0, -1, -2) and change from baseline categorization using the following three categories: Improved, No Change, and Deteriorated will also be performed. 75

78 Change from baseline in BDI will be analysed as for 6MWD. CONFIDENTIAL GM2008/00365/00 AMB Change from baseline at week 24 in N-terminal pro-b-type natriuretic peptide (NTproBNP) will be summarised and analysed as the geometric mean and the geometric mean ratio Exploratory Analyses Peak-trough assessment of 6MWD To assess the treatment effect of ambrisentan (assay sensitivity), the change from baseline to Week 16 for 6MWD will be analyzed for the ambrisentan/tadalafil combination group versus the tadalafil monotherapy group. This will be analysed as for 6MWD as described in the secondary analyses section. The estimate of peak-trough 6MWD effect will be determined by the ratio of the placebo-corrected mean change from baseline 6MWD measured at trough ambrisentan plasma concentrations following 16 weeks of treatment divided by the placebocorrected mean change from baseline 6MWD measured at peak ambrisentan plasma concentrations following 16 weeks of treatment. i.e. 6MWD 6MWD trough peak - - 6MWD 6MWD placebo placebo, where 6MWD trough = mean change from baseline 6MWD for the subjects in the ambrisentan/tadalafil combination group whose 6MWD was assessed at trough ambrisentan plasma concentrations 6MWD peak = mean change from baseline 6MWD for the subjects in the ambrisentan/tadalafil combination group whose 6MWD was assessed at peak ambrisentan plasma concentrations 6MWD placebo = mean change from baseline 6MWD for the tadalafil monotherapy group. Baseline 6MWD is defined as the average of the two 6MWDs performed at randomisation. In addition to reporting the above ratio, an interval estimate for the ratio using Hinkley s Method (Hinkley, 1969) will be provided. Full details of the analysis of all endpoints will be provided in the Reporting and Analysis Plan (RAP). The RAP will become part of the study documentation and will be finalised before the treatment code is unblinded and the database closed. 76

79 CONFIDENTIAL GM2008/00365/00 AMB Safety Analyses Adverse Events AE summaries will be performed using the preferred terms and system organ class assigned by the Medical Dictionary for Regulatory Activities (MedDRA). The proportion of subjects reporting each type of AE will be tabulated for each treatment group according to the treatment received at the time of the AE. AEs may also be tabulated separately for those occurring prior to a clinical failure event by randomised treatment and for those AEs events occurring after a clinical failure event according to actual treatment received. The following summaries of AEs will be provided: All AEs Treatment-related AEs AEs leading to withdrawal Serious AEs Details for all AEs will be listed. AEs of special interest will be summarised. Of special interest are Liver Events, Anaemia, Hypersensitivity, Hypotension and Fluid Retention. Exposure The number of days of exposure to study medication will be summarized by treatment group. Subject exposure may also be categorised and the number and percentage of subjects in each category presented for each treatment group. Vital Signs All vital sign data including systolic and diastolic blood pressure, heart rate, and BMI will be summarised by visit and treatment group. Summary statistics for values and change from baseline at measured time points will be provided by treatment group. For systolic and diastolic blood pressure and heart rate, the number and percentage of subjects with values of clinical concern will be summarized by treatment group. Clinical concern criteria will be specified in the RAP. Laboratory Evaluations Laboratory measurements will be summarized by visit and treatment group. Summary statistics for values and change from baseline will be provided by treatment group. In addition, the number and percentage of subjects with values of clinical concern will be summarized by treatment group. Clinical concern criteria will be specified in the RAP. Of special interest are some liver function laboratory parameters and their corresponding change from baseline. 77

80 CONFIDENTIAL GM2008/00365/00 AMB Health Outcomes SF-36 and CAMPHOR, in selected countries, will be summarised by treatment group. Each section of the CAMPHOR will be summarised separately. Further details will be described in the Reporting and Analysis Plan. 9. STUDY CONDUCT CONSIDERATIONS 9.1. Posting of Information on Clinicaltrials.gov Study information from this protocol will be posted on clinicaltrials.gov before enrolment of subjects begins Regulatory and Ethical Considerations, Including the Informed Consent Process Prior to initiation of a study site, the Sponsors (e.g., Gilead in the USA and GSK in the Rest of World) will obtain approval from the appropriate Regulatory Agency to conduct the study in accordance with applicable country-specific regulatory requirements. The study will be conducted in accordance with all applicable regulatory requirements, The study will be conducted in accordance with Good Clinical Practice (GCP), all applicable subject privacy requirements, and the ethical principles that are outlined in the Declaration of Helsinki 2000, as amended in 2002, 2004 and 2008 including, but not limited to: Institutional Review Board (IRB)/Independent Ethics Committee (IEC) review and approval of study protocol and any subsequent amendments. Subject informed consent. Investigator reporting requirements. The Sponsor will provide full details of the above procedures, either verbally, in writing, or both. Written informed consent must be obtained from each subject prior to participation in the study Quality Control (Study Monitoring) In accordance with applicable regulations, GCP, and the Sponsor procedures, monitors will contact the site prior to the start of the study to review with the site staff the protocol, study requirements, and their responsibilities to satisfy regulatory, ethical, and Sponsor requirements. When reviewing data collection procedures, the discussion will include identification, agreement and documentation of data items for which the CRF will serve as the source document. 78

81 The sponsor will monitor the study to ensure that the: Data are authentic, accurate, and complete. Safety and rights of subjects are being protected. CONFIDENTIAL GM2008/00365/00 AMB Study is conducted in accordance with the currently approved protocol and any other study agreements, GCP, and all applicable regulatory requirements. The investigator and the head of the medical institution (where applicable) agrees to allow the monitor direct access to all relevant documents Quality Assurance To ensure compliance with GCP and all applicable regulatory requirements, the Sponsor may conduct a quality assurance audit of the site records, and the regulatory agencies may conduct a regulatory inspection at any time during or after completion of the study. In the event of an audit or inspection, the investigator (and institution) must agree to grant the auditor(s) and inspector(s) direct access to all relevant documents and to allocate their time and the time of their staff to discuss any findings/relevant issues Study and Site Closure The study will be considered over after the database lock 2. Upon completion or termination of the study, the Sponsor monitor will conduct site closure activities with the investigator or site staff (as appropriate), in accordance with applicable regulations, GCP, and the sponsor or protocol specific Standard Operating Procedures. The Sponsor reserves the right to temporarily suspend or terminate the study at any time for reasons including (but not limited to) safety issues, ethical issues, or severe noncompliance. If the Sponsor determines that such action is required, the Sponsor will discuss the reasons for taking such action with the investigator or head of the medical institution (where applicable). When feasible, the Sponsor will provide advance notice to the investigator or head of the medical institution of the impending action. If a study is suspended or terminated for safety reasons, the Sponsor will promptly inform all investigators, heads of the medical institutions (where applicable),and/or institutions conducting the study. The Sponsor will also promptly inform the relevant regulatory authorities of the suspension/termination along with the reasons for such action. Where required by applicable regulations, the investigator or head of the medical institution must inform the IRB/IEC promptly and provide the reason(s) for the suspension/termination. 79

82 CONFIDENTIAL GM2008/00365/00 AMB Records Retention Following closure of the study, the investigator or head of the medical institution (where applicable) must maintain all site study records (except for those required by local regulations to be maintained elsewhere) in a safe and secure location. The records must be easily accessible when needed (e.g., for a Sponsor audit or regulatory inspection) and must be available for review in conjunction with assessment of the facility, supporting systems, and relevant site staff. Where permitted by local laws/regulations or institutional policy, some or all of the records may be maintained in a format other than hard copy (e.g., microfiche, scanned, electronic); however, caution must be exercised before such action is taken. The investigator must ensure that all reproductions are legible and are a true and accurate copy of the original. In addition, they must meet accessibility and retrieval standards, including regeneration of a hard copy, if required. The investigator must also ensure that an acceptable back-up of the reproductions exists and that there is an acceptable quality control procedure in place for creating the reproductions. The Sponsor will inform the investigator of the time period for retaining the site records in order to comply with all applicable regulatory requirements. The minimum retention time will meet the strictest standard applicable to a particular site, as dictated by local laws/regulations, the Sponsor or study-specific standard operating procedures, and/or institutional requirements. The investigator must notify the Sponsor of any changes in the archival arrangements, including, but not limited to archival of records at an off-site facility or transfer of ownership of the records in the event that the investigator is no longer associated with the site Provision of Study Results and Information to Investigators Where required by applicable regulatory requirements, an investigator signatory will be identified for the approval of the clinical study report. The investigator will be provided reasonable access to statistical tables, figures, and relevant reports and will have the opportunity to review the complete study results at a Sponsor site or other mutuallyagreeable location. The Sponsor will also provide the investigator with the full summary of the study results. The investigator is encouraged to share the summary results with the study subjects, as appropriate. The Sponsor will provide the investigator with the randomisation codes for their site only after completion of the full statistical analysis. The results summary will be posted to the Clinical Study Register at the time of the first regulatory approval or within 12 months of any decision to terminate development. In addition, a manuscript will be submitted to a peer-reviewed journal for publication within 80

83 CONFIDENTIAL GM2008/00365/00 AMB months of the first approval or within 12 months of any decision to terminate development. When manuscript publication in a peer-reviewed journal is not feasible, further study information will be posted to the Sponsor Clinical Study Register to supplement the results summary Independent Data Monitoring Committee (IDMC) An IDMC will be utilized in this study to ensure external objective medical and/or statistical review of safety and/or efficacy issues in order to protect the ethical and safety interests of subjects and to protect the scientific validity of the study. The schedule of the analysis plan for IDMC review is described in the charter, which is available upon request. 81

84 CONFIDENTIAL GM2008/00365/00 AMB REFERENCES Ambrisentan EU Summary of Product Characteristics 2009 Ambrisentan US Prescribing Information 2009 Adcirca (tadalafil) US Prescribing Information 2009 Adcirca (tadalafil) Summary of Product Characteristics 2009 D'Alonzo GE, Barst RJ, Ayres SM, Bergofsky EH, Brundage BH, Detre KM et al. Survival in patients with primary pulmonary hypertension. Results from a national prospective registry. Ann Intern Med 1991; 115(5): Davenport AP, O'Reilly G, Kuc RE. Endothelin ETA and ETB mrna and receptors expressed by smooth muscle in the human vasculature: majority of the ETA sub-type. Br J Pharmacol 1995; 114(6): Forgue ST, Patterson BE, Bedding AW, Payne CD, Phillips DL, Wrishko RE, Mitchell MI. Br J Clin Pharmacol. Tadalafil pharmacokinetics in healthy subjects. 2006; 61: Fukuroda T, Fujikawa T, Ozaki S, Ishikawa K, Yano M, Nishikibe M. Clearance of circulating endothelin-1 by ETB receptors in rats. Biochem Biophys Res Commun 1994; 199(3): Galie N, Badesch D, Oudiz R, Simonneau G, McGoon MD, Keogh AM et al. Ambrisentan therapy for pulmonary arterial hypertension. J Am Coll Cardiol 2005; 46(3): Galiè N, Olschewski H, Oudiz RJ, Torres F, Frost A, Ghofrani HA, et al. Ambrisentan in Pulmonary Arterial Hypertension, Randomized, Double-Blind, Placebo-Controlled, Multicenter, Efficacy Studies (ARIES) Group. Ambrisentan for the treatment of pulmonary arterial hypertension: results of the ambrisentan in pulmonary arterial hypertension, randomized, double-blind, placebo-controlled, multicenter, efficacy (ARIES) study 1 and 2. Circulation 2008; 117: Galiè N, Brundage BH, Ghofrani HA, Oudiz RJ, Simonneau G, Safdar Z, et al. Pulmonary Arterial Hypertension and Response to Tadalafil (PHIRST) Study Group. Tadalafil therapy for pulmonary arterial hypertension. Circulation 2009; 119(22): Galiè N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, Barbera JA, et al. Task Force for Diagnosis and Treatment of Pulmonary Hypertension of European Society of Cardiology (ESC); European Respiratory Society (ERS); International Society of Heart and Lung Transplantation (ISHLT). Guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Respir J 2009; 34(6):

85 CONFIDENTIAL GM2008/00365/00 AMB Ghofrani HA, Wiedemann R, Rose F, Olschewski H, Schermuly RT, Weissmann N et al. Combination therapy with oral sildenafil and inhaled iloprost for severe pulmonary hypertension. Ann Intern Med 2002; 136(7): Hirata Y, Emori T, Eguchi S, Kanno K, Imai T, Ohta K et al. Endothelin receptor subtype B mediates synthesis of nitric oxide by cultured bovine endothelial cells. J Clin Invest 1993; 91(4): Hoeper MM, Taha N, Bekjarova A, Gatzke R, Spiekerkoetter E. Bosentan treatment in patients with primary pulmonary hypertension receiving nonparenteral prostanoids. Eur Respir J 2003; 22(2): Hoeper MM, Markevych I, Spiekerkoetter E, Welte T, Niedermeyer J. Goal-oriented treatment and combination therapy for pulmonary arterial hypertension. Eur Respir J 2005; 26(5): Humbert M, Barst RJ, Robbins IM, Channick RN, Galie N, Boonstra A et al. Combination of bosentan with epoprostenol in pulmonary arterial hypertension: BREATHE-2. Eur Respir J 2004; 24(3): Humbert M, Morrell NW, Archer SL, Stenmark KR, MacLean MR, Lang IM et al. Cellular and molecular pathobiology of pulmonary arterial hypertension. J Am Coll Cardiol 2004; 43(12 Suppl S):13S-24S. Humbert M, Sitbon O, Simmonneau G. Treatment of pulmonary arterial hypertension. N Engl J Med 2004; 351(14): Kirchengast M, Munter K. Endothelin-1 and endothelin receptor antagonists in cardiovascular remodeling. Proc Soc Exp Biol Med 1999; 221(4): Leuchte HH, Holzapfel M, Baumgartner RA, Ding I, Neurohr C, Vogeser M et al. Clinical significance of brain natriuretic peptide in primary pulmonary hypertension. J Am Coll Cardiol 2004; 43(5): Leuchte HH, Holzapfel M, Baumgartner RA, Neurohr C, Vogeser M, Behr J. Characterization of brain natriuretic peptide in long-term follow-up of pulmonary arterial hypertension. Chest 2005; 128(4): Letairis US Prescribing Information 2009 McKenna SP et al. The Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR): A measure of health-related quality of life and quality of life for patients with pulmonary hypertension. Quality of Life Research 2006; 15: Nagaya N, Nishikimi T, Uematsu M, Satoh T, Kyotani S, Sakamaki F et al. Plasma brain natriuretic peptide as a prognostic indicator in patients with primary pulmonary hypertension. Circulation 2000; 102(8):

86 CONFIDENTIAL GM2008/00365/00 AMB Oudiz RJ, Galiè N, Olschewski H, Torres F, Frost A, Ghofrani HA, et al. ARIES Study Group. Long-term ambrisentan therapy for the treatment of pulmonary arterial hypertension. JACC 2009; 54: Ozaki S, Ohwaki K, Ihara M, Fukuroda T, Ishikawa K, Yano M. ETB-mediated regulation of extracellular levels of endothelin-1 in cultured human endothelial cells. Biochem Biophys Res Commun 1995; 209(2): Park MH, Scott RL, Uber PA, Ventura HO, Mehra MR. Usefulness of B-type natriuretic peptide as a predictor of treatment outcome in pulmonary arterial hypertension. Congest Heart Fail 2004; 10(5): Simonneau G, Rubin LJ, Galiè N, Barst RJ, Fleming TR, Frost AE et al. Addition of sildenafil to long-term intravenous epoprostenol therapy in patients with pulmonary arterial hypertension: a randomized trial. Ann Intern Med 2008; 149: Spence R, Mandagere A, Harrison B, Dufton C, Boinpally R. No clinically relevant pharmacokinetic and safety interactions of ambrisentan in combination with tadalafil in healthy volunteers. J Pharm Sci 2009; 98(12): Rich S, Kaufmann E, Levy PS. The effect of high doses of calcium-channel blockers on survival in primary pulmonary hypertension. N Engl J Med 1992; 327(2): Rubanyi GM, Polokoff MA. Endothelins: molecular biology, biochemistry, pharmacology, physiology, and pathophysiology. Pharmacol Rev 1994; 46(3): Rubin LJ. Primary pulmonary hypertension. Chest 1993; 104(1): Verhaar MC, Strachan FE, Newby DE, Cruden NL, Koomans HA, Rabelink TJ et al. Endothelin-A receptor antagonist-mediated vasodilatation is attenuated by inhibition of nitric oxide synthesis and by endothelin-b receptor blockade. Circulation 1998; 97(8): Ware, JE; Kosinski, M; Gandek, B. SF-36 Health Survey: Manual and Interpretation Guide, Lincoln, RI: QualityMetric Inc,

87 CONFIDENTIAL GM2008/00365/00 AMB APPENDICES Appendix 1: List of Highly Effective Methods for Avoidance of Pregnancy in Women of Childbearing Potential The following is a list of the highly effective methods for avoiding pregnancy (i.e., have a failure rate of less than 1% per year). Abstinence [Hatcher, 2004] Combination Oral Contraceptive [Hatcher, 2004] Injectable progestogen [Hatcher, 2004] Implants of levonorgestrel [Hatcher, 2004] Estrogenic vaginal ring [Hatcher, 2004] Percutaneous contraceptive patches [Hatcher, 2004] Intrauterine device (IUD) or intrauterine system (IUS) that meets the SOP effectiveness criteria as stated in the product label [Hatcher, 2004] Male partner sterilization (vasectomy with documentation of azoospermia) prior to the female subject's entry into the study, and this male is the sole partner for that subject [Hatcher, 2004]. For this definition, documented refers to the outcome of the investigator's/designee s medical examination of the subject or review of the subject's medical history for study eligibility, as obtained via a verbal interview with the subject or from the subject s medical records. Double barrier method: condom and an occlusive cap (diaphragm or cervical/vault caps) with a vaginal spermicidal agent (foam/gel/film/cream/suppository) [Trussell] Nonoxynol-9 is the critical component in most spermicides, and is regarded as an acceptable spermicidal agent. Concern has been raised that nonoxynol-9 damages the epithelial lining of the vagina, and exposure may facilitate transmission of viruses, particularly human immunodeficiency virus (HIV). The World Health Organization (WHO) conducted a technical consultation in October 2001 and concluded that the increased risk for such transmission was low to minimal [WHO/CONRAD, 2003]. References 1. Hatcher RA, Trussell J, Stewart F, Nelson AL, Cates W, Guest F, Kowal DD, editors. Contraceptive Technology. New York: Ardent Media, 2004: 226. Table 9-2, % of Women Experiencing an Unintended Pregnancy During the First Year of Typical Use and the First Year of Perfect Use of Contraception and the Percentage Continuing Use at the End of the First Year. United States, column entitled, % of Women Experiencing an Unintended Pregnancy Within the First Year of Use. Perfect Use. 85

88 CONFIDENTIAL GM2008/00365/00 AMB Trussell J. Personal communication WHO/CONRAD Technical consultation on nonoxynol-9. WHO, Geneva, 9-10 October Summary Report. World Health Organization,

89 CONFIDENTIAL GM2008/00365/00 AMB Appendix 2: 6 Minute Walk Distance Test The 6-minute walk test will be conducted according to the American Thoracic Society guidelines [ in accordance with local standard operating procedures. Please refer to the following guidelines for conducting the 6-minute walk test. The Test Location Select a quiet, enclosed corridor, free of distractions and drafts. Depending on your site s standard operating procedures, please mark out a 30 m or 100 ft course. The start and end of the course must be visibly marked by placing tape on the floor, using chairs etc. The course should be sub-divided into 3m or 10ft sections using a method unnoticeable to the subject. The Day Before the Test Tell the subject the time their test will take place. Give the subject the following instructions: Take a light meal 3 to 4 hours before the test. Take nothing by mouth (po) except clear liquids for 1 hour before the test. Do not take heart medication within 2 hours of the test, including shortacting nitrates. Do not smoke for at least 2 hours before the test. Prior to all clinic visits, withhold the daily dose of Investigational Product such that the last dose of Investigational Product is administered more than 16 hours prior to the 6MWD and BDI assessments Immediately Before the Test The test should be performed on the subject s "usual" oxygen. If the subject usually breathes room air, they should perform the test on room air. If the subject normally requires supplemental oxygen with ambulation, they should perform the test with supplemental oxygen. Subjects who require supplemental oxygen should receive the same oxygen flow rate as baseline during all subsequent 6-minute walk tests. Subjects who add supplemental oxygen after initiation of therapy should walk without oxygen whenever possible so data is comparable to baseline conditions. If the oxygen flow rate must be increased during subsequent visits due to worsening gas exchange, the 6-minute walk test may be conducted at the increased oxygen flow rate and this should be noted in the CRF. 87

90 CONFIDENTIAL GM2008/00365/00 AMB If the subject reduces or discontinues supplemental oxygen use during the study, they should still perform the 6-minute walk test with the same flow rate of oxygen that was used during previous tests. Ask the subject to sit quietly for 10 minutes. Read the following instructions verbatim to the subject: The object of this test is to walk as far as possible for 6 minutes. You will walk back and forth in this hallway. Six minutes is a long time to walk, so you will be exerting yourself. You will probably get out of breath or become exhausted. You are permitted to slow down, to stop, and to rest as necessary. You may lean against the wall while resting, but resume walking as soon as you are able. Are you ready to do that? Repeat the entire set of instructions if the subject does not seem to understand. Repeat the sentence: Remember that the object is to walk AS FAR AS POSSIBLE for 6 minutes, but don t run or jog. Instruct the subject to start by saying: Start now, or whenever you are ready. During the Test: Do not influence the walking pace of the subject. Use an even tone of voice when using the standard phrases of encouragement. Encourage the subject every 60 seconds (1 minute) during the test but do not convey information about their performance. Examples of appropriate encouragement are, You are doing well. or Keep up the good work. Only speak to the subject at the 60-second points. If the subject is slowing down or expresses that he/she wants to stop say: You can lean against the wall if you would like; then continue walking whenever you feel able. Note the number of laps of the corridor walked by the subject but do not convey this information to the subject. At 6 minutes, tell the subject to stop and not to move until the distance walked has been measured. After the Test 88

91 CONFIDENTIAL GM2008/00365/00 AMB If the subject walked for less than 6 minutes (i.e., stops prior to 6 minutes and DOES NOT start again), record the time walked as well as the distance. When the distance has been measured tell the subject to sit down and observe him/her for at least 10 minutes. The distance covered during any walking test must not be revealed to the subject at any time during the study. Parameters to be recorded: Distance walked, to the nearest meter or foot Time walked, to the nearest second (if subject walked for less than 6 minutes) Comments if subject was unable to walk for 6 minutes The BDI score immediately following completion of the 6-minute walk test (Appendix 3) 89

92 11.3. Appendix 3: Borg Dyspnea Index CONFIDENTIAL GM2008/00365/00 AMB The BDI score immediately following completion of the 6-minute walk test (See appendix 2) Borg CR10 scale G. Borg, 1998, 2007 English 90

93 CONFIDENTIAL GM2008/00365/00 AMB Instruction. Use this rating scale to report how strong your perception is. It can be exertion, pain or something else. First look at the verbal expressions. Start with them and then the numbers. Of these ten (10) or Extremely strong, Maximal is a very important intensity level. This is the most intense perception or feeling you have ever had. If your experience or feeling is Very weak, you should say 1, if it is Moderate, say 3. Note that Moderate is 3 and thus weaker than Medium, Mean or Middle. If the experience is Strong or Heavy (it feels Difficult ) say 5. Note that Strong is about half of Maximal. If your feeling is Very strong, choose a number from 6 to 8. If your perception or feeling is stronger than 10, - Extremely strong, Maximal you can use a larger number, e.g. 12 or still higher (that s why Absolute maximum is marked with a dot ). It's very important that you report what you actually experience or feel, not what you think you should report. Be as spontaneous and honest as possible and try to avoid underor overestimating. Look at the verbal descriptors and then choose a number. When rating exertion give a number that corresponds to how hard and strenuous you perceive the work to be. The perception of exertion is mainly felt as strain and fatigue in your muscles and as breathlessness or any aches. 0 Nothing at all, means that you don t feel any exertion whatsoever, no muscle fatigue, no breathlessness or difficulties breathing. 1 Very weak means a very light exertion. As taking a shorter walk at your own pace. 3 Moderate is somewhat but not especially hard. It feels good and not difficult to go on. 5 Strong. The work is hard and tiring, but continuing isn't terribly difficult. The effort and exertion is about half as intense as Maximal. 7 Very strong is quite strenuous. You can still go on, but you really have to push yourself and you are very tired. 10 Extremely strong Maximal is an extremely strenuous level. For most people this is the most strenuous exertion they have ever experienced previously in their lives. Is Absolute maximum for example 12 or even more. Any questions? Borg CR10 scale G. Borg, 1998, 2007 English 91

94 CONFIDENTIAL GM2008/00365/00 AMB Appendix 4: Pharmacogenetic Research Pharmacogenetics Background Pharmacogenetics (PGx) is the study of variability in drug response due to hereditary factors in different populations. There is increasing evidence that an individual's genetic composition (i.e., genotype) may impact the pharmacokinetics (absorption, distribution, metabolism, elimination), pharmacodynamics (relationship between concentrations and pharmacologic effects or the time course of pharmacologic effects) and/or clinical outcome (in terms of efficacy and/or safety and tolerability). Some reported examples of PGx analysis include: Drug Disease Gene Outcome Abacavir HIV [Hetherington, 2002; Mallal, 2002] HLA B*5701 Individuals with HLA-B*5701 variant may be at increased risk for experiencing hypersensitivity to abacavir. Clinical assays are available for HLA-B*5701 but none has been validated. HLA-B*5701 screening would supplement but never replace abacavir clinical risk management strategies aimed at minimising rare but serious outcomes associated with Warfarin Irinotecan Cardiovascular [Neergard, 2006; Wilke, 2005] Cancer [FDA News Release, 2005] CYP2C9 UGT1A1 abacavir hypersensitivity. Serious adverse events (SAEs) experienced by some patients on warfarin may be explained by variations in the CYP2C9 gene that influences the degree of anticoagulation achieved. Variations in the UGT1A1 gene can influence a patient s ability to break down irinotecan, which can lead to increased blood levels of the drug and a higher risk of side effects. A dose of irinotecan that is safe for one patient with a particular UGT1A1 gene variation, might be too high for another patient without this variation, raising the risk of certain sideeffects. A genetic blood test (Invader UGT1A1 molecular assay) is available that can detect variations in the gene. A key component to successful PGx research is the collection of samples during the conduct of clinical studies. Collection of whole blood samples, even when no a priori hypothesis has been identified, may enable PGx analysis to be conducted if at any time it appears that there is a potential unexpected or unexplained variation in handling or response to ambrisentan or tadalafil. 92

95 Pharmacogenetic Research Objectives CONFIDENTIAL GM2008/00365/00 AMB The objective of the PGx research (if there is a potential unexpected or unexplained variation) is to investigate a possible genetic relationship to handling or response to ambrisentan or tadalafil. If at any time it appears there is potential variability in response in this clinical study or in a series of clinical studies with ambrisentan or tadalafil that may be attributable to genetic variations of subjects, the following objectives may be investigated: Relationship between genetic variants and the pharmacokinetics and/or pharmacodynamics of ambrisentan/tadalafil Relationship between genetic variants and safety and/or tolerability of ambrisentan/tadalafil Relationship between genetic variants and efficacy of ambrisentan/tadalafil Study Population Any subject who has given informed consent to participate in the clinical study, has met all the entry criteria for the clinical study, and receives investigational product may take part in the PGx research. Any subject who has received an allogeneic bone marrow transplant must be excluded from the PGx research. Subject participation in the PGx research is voluntary and refusal to participate will not indicate withdrawal from the clinical study. Refusal to participate will involve no penalty or loss of benefits to which the subject would otherwise be entitled. Study Assessments and Procedures In addition to any blood samples taken for the clinical study, a whole blood sample (~10ml) will be collected for the PGx research using a tube containing EDTA. It is recommended that the blood sample be taken at the first opportunity after a subject has been randomised and provided informed consent for PGx research, but may be taken at any time while the subject is participating in the clinical study. The PGx sample is labelled (or coded ) with a study specific number that can be traced or linked back to the subject by the investigator or site staff. Coded samples do not carry personal identifiers (such as name or social security number). The blood sample is taken on a single occasion unless a duplicate sample is required due to inability to utilize the original sample. The DNA extracted from the blood sample may be subjected to sample quality control analysis. This analysis will involve the genotyping of several genetic markers to confirm the integrity of individual samples. If inconsistencies are noted in the analysis, then those samples may be destroyed. The need to conduct PGx analysis may be identified after a study (or a set of studies) of ambrisentan or tadalafil has been completed and the study data reviewed. 93

96 CONFIDENTIAL GM2008/00365/00 AMB In some cases, the samples may not be studied. e.g., no questions are raised about how people respond to ambrisentan or tadalafil. Samples will be stored securely and may be kept for up to 15 years after the last subject completes the study or the Sponsor may destroy the samples sooner. The Sponsor or those working with the Sponsor (for example, other researchers) will use samples collected from the study for the purpose stated in this protocol and in the informed consent form. Subjects can request their sample to be destroyed at any time. Subject Withdrawal from Study If a subject who has consented to participate in PGx research and has a sample taken for PGx research withdraws from the clinical study for any reason other than lost to followup, the subject will be given the following options: The sample is retained for PGx research Any PGx sample is destroyed. If a subject withdraws consent from the PGx research or requests sample destruction for any reason, the investigator must complete the appropriate documentation to request sample destruction within the timeframe specified by the Sponsor and maintain the documentation in the site study records. If the sample has already been processed, it will be destroyed after all steps are complete. The Sponsor will ensure that any data related to the sample will not be analysed. The sample will be destroyed after processing is complete. Screen and Baseline Failures If a blood sample for PGx research has been collected and it is determined that the subject does not meet the entry criteria for participation in the clinical study, then the investigator must complete the appropriate documentation to request sample destruction within 5 days. The sample will be destroyed and documentation sent to the site within 30 working days of receipt of the request for destruction. All documents pertaining to sample destruction must be maintained in the site study records. Pharmacogenetics Analyses 1. Specific sections of DNA may be selected from areas of the genome (e.g., candidate genes) known to encode the drug target, drug metabolizing enzymes, areas associated with mechanisms underlying adverse events, and those linked to study disease and, thus, linked to drug response. The candidate genes that may be investigated in this study include the following: the Sponsor Absorption, Distribution, Metabolism and Excretion genes. These play a central role in drug pharmacokinetics and pharmacodynamics. In addition, continuing research may identify other enzymes, transporters, proteins or receptors that may be involved in 94

97 CONFIDENTIAL GM2008/00365/00 AMB response to ambrisentan or tadalafil. The genes that may code for these proteins may also be studied. 2. By evaluating large numbers of polymorphic markers (e.g., single nucleotide polymorphisms or SNPs) throughout the genome, sets of markers may be identified that correspond to differential drug response. The results of PGx investigations will be reported either as part of the main clinical study report or as a separate report. All endpoints of interest from all comparisons will be descriptively and/or graphically summarised as appropriate to the data. In all cases, appropriate statistical methods will be used to analyse the genetic markers in the context of other clinical data. Statistical methods may include, but are not limited to Hardy- Weinberg Equilibrium testing, Comparison of Demographic and Baseline Characteristics by Genotype, Evaluation of Genotypic Effects, Evaluation of Treatment by Genotype and Gene-Gene Interaction, Linkage Disequilibrium, Multiple Comparison and Multiplicity and/or Power and Sample Size Considerations. Detailed description of the analyses to be conducted will be documented in the Pharmacogenetics Reporting and Analysis Plan. 3. Genome-wide scans involving a large number of polymorphic markers (e.g., single nucleotide polymorphisms) located throughout the genome. This approach is often employed when potential genetic effects are not well understood. Informed Consent Subjects who do not wish to participate in the PGx research may still participate in the clinical study. PGx informed consent must be obtained prior to any blood being taken for PGx research. Provision of Study Results and Confidentiality of Subject s PGx Data Sponsor may summarize the cumulative PGx research results in the clinical study report. In general, the Sponsor does not inform the investigator, subject, or anyone else (e.g., family members, study investigators, primary care physicians, insurers, or employers) of the PGx research results that are not known to be relevant to the subject s medical care at the time of the study, because the information generated from PGx studies is preliminary in nature, and the significance and scientific validity of the results are undetermined at such an early stage of research, under any circumstances unless required by law. References Hetherington S, Hughes AR, Mosteller M, Shortino D, Baker KL, Spreen W, Lai E, Davies K, Handley A, Dow DJ, Fling ME, Stocum M, Bowman C, Thurmond LM, Roses AD. Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet. 2002; 359: Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, Sayer D, Castley A, Mamotte C, Maxwell D, James I. Association between presence of HLA-B*5701, HLA-DR7, and 95

98 CONFIDENTIAL GM2008/00365/00 AMB HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet. 2002; 359: Neergard. Reducing the risk of blood thinners. Associated press, September U.S. Food and Drug Administration, FDA Clears Genetic Test That Advances Personalized Medicine Test Helps Determine Safety of Drug Therapy 22 August 2005, Wilke RA, Musana AK, Weber WW. Cytochrome P450 gene-based drug prescribing, and factors impacting translation into routine clinical practice. Personalized Med 2005; 2:

99

100 CONFIDENTIAL GM2008/00365/00 AMB Symptoms Please read each statement carefully and decide if it applies to you today 1. My stamina levels are low 2. I have to rest during the day 3. I feel worn out 4. I get tired very quickly 5. I m tired all the time 6. I feel very weak Yes No Yes No Yes No Yes No Yes No Yes No 7. I feel completely exhausted Yes No 8. I want to sit down all the time Yes No 9. I quickly run out of energy Yes No Please turn the page 98

101 CONFIDENTIAL GM2008/00365/00 AMB Everything is an effort 11. I get out of breath when I stand up 12. I get out of breath when I talk Yes No Yes No Yes No Please read each statement carefully and decide if it applies to you today 13. I get out of breath when I walk 14. I get out of breath when I bend over 15. I get out of breath when I go up one step Yes No Yes No Yes No 16. I get out of breath when I walk up a slight incline 17. I get out of breath without doing anything 18. I get out of breath climbing a flight of stairs Yes No Yes No Yes No 99

102 CONFIDENTIAL GM2008/00365/00 AMB I have mood swings 20. I get very down 21. I rarely feel happy Yes No Yes No Yes No Please read each statement carefully and decide if it applies to you today 22. I ve forgotten what it s like to enjoy myself 23. I feel hopeless 24. My condition gets me down 25. I often feel anxious Yes No Yes No Yes No Yes No Please turn the page 100

103 CONFIDENTIAL GM2008/00365/00 AMB Activities Please put a check in the box under the response which best describes your abilities today. For each of the 15 statements please provide one response only. Please check your response based on your abilities without the use of equipment (for example a cane or walker) or assistance. Note for daytime oxygen users: Please check your response based on your abilities with your normal oxygen use Please check only 1 box for each of the following statements Can do on my own without difficulty Can do on my own with difficulty Cannot do on my own 1. Cut your toenails 4. Walk around inside the house (not including climbing stairs) 2. Bathe yourself 3. Get dressed 5. Walk short distances on level ground 6. Walk longer distances on level ground 7. Walk up a slight incline 8. Climb a flight of stairs 9. Bend down to pick up objects off the floor 10. Stand for a short time 11. Stand for a long time 12. Lift heavy objects 13. Carry heavy objects 14. Do light work around the house or yard 15. Do heavy work around the house or yard Quality of Life Please turn the page 101

104 CONFIDENTIAL GM2008/00365/00 AMB Please read each statement carefully and put a check next to the response that applies best to you today 1. I have to speak softly True False 2. I can't be away from home 3. I've lost interest in food 4. I don t have the energy for my close relationships 5. Walking for pleasure is out of the question 6. My condition puts a strain on my close relationships True False True False True False True False True False 7. I feel very isolated 8. I can t do things on the spur of the moment True False True False True 9. I feel vulnerable when I m on my own False Please turn the page 102

105 CONFIDENTIAL GM2008/00365/00 AMB It feels like my body has let me down 11. I feel like I m not in control of my life 12. I feel dependent on other people True False True False True False Please remember to check only one of the alternative responses for each of the statements 13..Sometimes it takes too much effort to speak 14. I feel like I m a burden to other people 15. Travelling distances is difficult 16. I don't like to be seen like this 17. I feel like I'm losing my purpose in life 18. I worry that I m neglecting the people close to me True False True False True False True False True False True False 103

106 CONFIDENTIAL GM2008/00365/00 AMB Please read each statement carefully and decide if it applies to you today 19. I feel guilty asking for help 20. My condition limits the places I can go 21. I don t like having to rely on other people True False True False True False 22. I don t want to talk to anybody 23. I feel like I let people down 24. I m reluctant to leave the house 25. I m unable to participate in activities with my family and friends True False True False True False True False Thank you for taking the time to fill out this questionnaire. Please check all the pages to make sure that you have responded to each statement. 104

107 CONFIDENTIAL GM2008/00365/00 AMB Appendix 6: SF-36 Quality of Life Questionnaire 105

108 CONFIDENTIAL GM2008/00365/00 AMB

109 CONFIDENTIAL GM2008/00365/00 AMB

110 CONFIDENTIAL GM2008/00365/00 AMB

111 CONFIDENTIAL GM2008/00365/00 AMB

112 CONFIDENTIAL GM2008/00365/00 AMB

113 CONFIDENTIAL GM2008/00365/00 AMB Appendix 7: Country Specific Requirements 111

114

115 GM2008/00365/04 CONFIDENTIAL The GlaxoSmithKline group of companies AMB112565/GS-US Revision Chronology: GM2008/00365/ APR-15 Original GM2008/00365/ AUG-25 Amendment No1 Clarify primary End Point, fourth component unsatisfactory long-term clinical response ; removal of reference to functional class IV to clarify that subjects should not, and do not, need to be in WHO FC IV for 6 months to qualify as an event. To clarify that 2 reliable methods of contraception are only required if the subject is sexually active and to clarify that specific types of IUD are the only required method of contraception. To highlight that the screening and baseline 6MWD Test values must not vary by more than 10%. Add new exclusion criterion that prohibits concomitant use of cyclosporine A Clarification of blinded combination therapy options following TtCF event. Amendments to the blinded combination therapy section to outline all possible starting dose combinations including provision for initial dose of ABS 10mg and TAD 20mg at stage 1 blinded combination therapy for subjects with mild to moderate renal impairment. Clarify when, following TtCF, an investigator may request a subject reuptitrate their dose of IP Specify that text messaging may be used (if chosen by the centre/subject) to notify subjects of visit dates. 2

116 GM2008/00365/04 CONFIDENTIAL The GlaxoSmithKline group of companies AMB112565/GS-US Removal of a redundant cross reference to a separate footnote in footnote 2 of the time and events schedule. GM2008/00365/ NOV-17 Amendment No: 2 Correct an error stating that randomization for the peak-trough assessment will occur at the week 0 visit it will in fact occur at the week 16 visit. The definition of liver events that potentially meet the criteria for Hy s law cases was amended from ALT value from 3xULN to >3xULN and bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. Further, text related to fractionation of bilirubin as part of Hy s law cases was removed to simplify the definition of these events. Additionally, text was added to Section to clarify that symptomatic elevations >3xULN are considered protocol defined SAEs. Amendments were made to Section to clarify reporting timelines for liver events. Clarification that the Final Assessment Visits are to be scheduled when it is projected that 82 primary events will have occurred. Make Appendix 2 consistent with rest of protocol with removal of text stating the requirement that subject should withhold investigational product prior to the clinic visit and reference to withholding nitrates as these are prohibited by protocol. Change in Sponsor [GSK] Medical Monitor and associated contact information. 3

117 GM2008/00365/04 CONFIDENTIAL The GlaxoSmithKline group of companies AMB112565/GS-US Clarification to wording of primary endpoint; second component to allow any hospitalisation due to worsening PAH to be included; and fourth component, through removal of contradictory wording. Amendments to Protocol Summary, Section Amend study inclusion/exclusion criteria to reduce the likelihood of subjects with PH due to covert diastolic dysfunction (WHO Group 2 PH) being enrolled in the study by introducing more stringent haemodynamic criteria excluding subjects with multiple left heart disease risk factors and requiring a documented ventilation/perfusion (V/Q) scan to rule out thromboembolic PH. Clarification to wording about use of cyclosporine; ophthalmic formulation is permitted. Amendments to Section 4.3 [Exclusion Criteria] and Section [Prohibited Medications and Non-Drug Therapies] United States Subjects: Change to allow right heart catheterization procedure under limited circumstances during Screening Period in the United States. Amendment to Section 4.2 [Inclusion Criteria] Study Assessments and Procedures table: Clarification to footnotes, relocation of footnote 5 to only be required for the week 16 visit instead of all visits. This corrects contradictory text. Change to allow use of local lab tests at screening to assess eligibility for randomisation. Amendments to Section 4.3 [Exclusion Criteria] and Section [Safety Tests]. Clarify that the start point for reporting of death is from the subject signing 4

118 GM2008/00365/04 CONFIDENTIAL The GlaxoSmithKline group of companies AMB112565/GS-US informed consent. Increase sample size, recruitment and study duration and removal of interim analyses, to increase power for comparison of combination therapy to individual monotherapy arms. Update to and ordering of the hierarchical testing of secondary endpoints. Adds an as treated population for the reporting of safety data. GM2008/00365/ MAY-30 Amendment No.: 03 Change to Sponsor (GSK) Medical Monitor. Change in primary analysis population to a modified ITT based on inclusion criteria in amendment 02. Increase in study sample size and duration to retain power for the mitt. Change to Exclusion criteria to clarify the previous use of PAH specific medications. Change in exclusion criteria and the previous use of nitrates in line with tadalafil product labeling Changes to the liver stopping criteria to allow the stopping of AMB/PBO only, not both IPs, if chosen by the investigator, in the event of a liver enzyme elevation. Relaxation of the Clinic Visit Schedule and visit assessments for subjects discontinuing Investigation Product, to enhance subject retention and aid collection of outcome data for all subjects in the trial. 5

119 GM2008/00365/04 CONFIDENTIAL The GlaxoSmithKline group of companies AMB112565/GS-US GM2008/00365/ MAR-19 Amendment No.: 04 Increase in anticipated sample size and revision to wording to highlight that the study will continue to recruit up to 24 weeks prior to the predicted final event. Change to wording of the Time to Clinical Failure End Point to clarify that all subjects having received one dose of IP may experience component 4 of TTCF unsatisfactory long-term clinical response, they do not have to have received randomised treatment for 6 months. This is in the spirit of the ITT principle. 6

120 GM2008/00365/04 CONFIDENTIAL The study will be sponsored by two companies. Gilead act as regulatory sponsor in the United States (US) and GSK in the Rest of the World (RoW). For the purposes of this protocol, the term sponsor shall refer to Gilead where the study is conducted in the USA and GSK where the study is conducted anywhere else. Sponsor Signatory: Signature: Date:, MD Vice President, Head Unit Physician MPC Therapy Area, MD, PhD [Gilead] Vice President, Biometrics Signature/date available upon request, PhD, MD [Gilead] Senior Vice President, Development Operations Signature/date available upon request, MD, PhD [Gilead] Senior Vice President, Cardiovascular Therapeutics Signature/date available upon request 7

121 GM2008/00365/04 CONFIDENTIAL SPONSORS INFORMATION PAGE: GSK Clinical Study Identifier: Sponsor Address and Contact Details AMB GlaxoSmithKline Iron Bridge Road Stockley Park West, Uxbridge, Middlesex, UB11 1BU, UK Telephone: In some countries, the clinical trial sponsor may be the local GlaxoSmithKline affiliate company (or designee). Where applicable, the details of the Sponsor and contact person will be provided to the relevant regulatory authority as part of the clinical trial submission. Sponsor Medical Monitor Contact Information: Dr Brian Sanderson Telephone: Sponsor Serious Adverse Events (SAE) Contact Information: As above Regulatory Agency Identifying Number(s):

122

123 GM2008/00365/04 CONFIDENTIAL INVESTIGATOR PROTOCOL AGREEMENT PAGE For protocol number I confirm agreement to conduct the study in compliance with the protocol, as amended by this protocol amendment. I acknowledge that I am responsible for overall study conduct. I agree to personally conduct or supervise the described clinical study. I agree to ensure that all associates, colleagues and employees assisting in the conduct of the study are informed about their obligations. Mechanisms are in place to ensure that site staff receives the appropriate information throughout the study. Investigator Name: Investigator Signature Date 10

124 GM2008/00365/04 CONFIDENTIAL TABLE OF CONTENTS PAGE LIST OF ABBREVIATIONS PROTOCOL SUMMARY INTRODUCTION Background Pulmonary Arterial Hypertension Current Treatment Options Ambrisentan Clinical Efficacy Safety Tadalafil Clinical efficacy Safety Combination of ambrisentan and tadalafil Rationale Rationale for ambrisentan peak-trough 6MWD assessment OBJECTIVE(S) INVESTIGATIONAL PLAN Study Design SUBJECT SELECTION AND WITHDRAWAL CRITERIA Number of Subjects Inclusion Criteria Exclusion Criteria Withdrawal Criteria STUDY TREATMENTS Investigational Products Ambrisentan or placebo Provision, handling and storage Dosing Tadalafil or placebo Provision, handling and storage Dosing Down titration of Investigational Product Treatment following clinical failure event Blinded Combination Therapy Treatment Assignment Blinding Product Accountability Treatment Compliance Concomitant Medications and Non-Drug Therapies Permitted Medications and Non-Drug Therapies Prohibited Medications and Non-Drug Therapies

125 GM2008/00365/04 CONFIDENTIAL 5.7. Treatment after the End of the Study Treatment of Investigational Product Overdose Ambrisentan Tadalafil STUDY ASSESSMENTS AND PROCEDURES Critical Baseline Assessments Efficacy Primary Endpoint Time to Clinical Failure Secondary Endpoints Minute Walk Distance and Borg Dyspnea Index Borg Dyspnea Index N-Terminal pro-b-type Natriuretic Peptide WHO Functional Class assessment Exploratory Endpoints Peak-trough assessment of 6MWD Safety Electrocardiogram Vital Signs and Body Weight and Height Clinical Laboratory Tests Safety Tests Pharmacogenetics Management, Monitoring and Follow-up of Serum Aminotransferase Abnormalities Liver event follow up assessments Adverse Events Definition of an AE Definition of a SAE Laboratory and Other Safety Assessment Abnormalities Reported as AEs and SAEs Pregnancy Time Period and Frequency of Detecting AEs and SAEs Prompt Reporting of Serious Adverse Events and Other Events to Sponsor Regulatory reporting requirements for SAEs Health Outcomes Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) Health Related Quality of Life and Quality of Life Measure SF-36 Health Survey DATA MANAGEMENT DATA ANALYSIS AND STATISTICAL CONSIDERATIONS Hypotheses Study Design Considerations Sample Size Assumptions Sample Size Sensitivity Sample Size Re-estimation Data Analysis Considerations

126 GM2008/00365/04 CONFIDENTIAL Analysis Population Analysis Data Sets Treatment Comparisons Primary Comparisons of Interest Other Comparisons of Interest Interim Analysis Key Elements of Analysis Plan Efficacy Analyses Primary Analysis Secondary Analyses Exploratory Analyses Safety Analyses Health Outcomes STUDY CONDUCT CONSIDERATIONS Posting of Information on Clinicaltrials.gov Regulatory and Ethical Considerations, Including the Informed Consent Process Quality Control (Study Monitoring) Quality Assurance Study and Site Closure Records Retention Provision of Study Results and Information to Investigators Independent Data Monitoring Committee (IDMC) REFERENCES APPENDICES Appendix 1: List of Highly Effective Methods for Avoidance of Pregnancy in Women of Childbearing Potential Appendix 2: 6 Minute Walk Distance Test Appendix 3: Borg Dyspnea Index Appendix 4: Pharmacogenetic Research Appendix 5: CAMPHOR Quality of Life Questionnaire Appendix 6: SF-36 Quality of Life Questionnaire Appendix 7: Country Specific Requirements Appendix 8: Protocol Amendment 01 Changes Appendix 9: Protocol Amendment 02 Changes Appendix 10: Protocol Amendment 03 Changes Appendix 11: Protocol Amendment 04 Changes

127 GM2008/00365/04 LIST OF ABBREVIATIONS CONFIDENTIAL ABS ADME AE ALT AST AUC Bid big ET-1 BDI BMI BNP BSEP BUN CAMPHOR CCB CEC C max CONMED CRF CRO CT CTD CYP DMC DSPH E2 ECE-1 ECG EEG EMA EOS ERA ET-1 ET A ET B EU FAV FC FDA FEV 1 HPAH FSH GCP GGT Ambrisentan absorption, distribution, metabolism, and elimination adverse event alanine aminotransferase aspartate aminotransferase area under the plasma concentration-time curve twice daily proendothelin-1 Borg dyspnea index body mass index B-type natriuretic peptide bile salt export pump blood urea nitrogen Cambridge Pulmonary Hypertension Outcome Review calcium channel blockers Clinical Endpoint Committee maximum plasma concentration concomitant medication case report form clinical research organization computed axial tomography scan connective tissue disease cytochrome P450 Data Monitoring Committee Drug Safety and Public Health estradiol endothelin converting enzyme electrocardiogram electroencephalogram European Medicines Authority End of Study endothelin receptor antagonist endothelin-1 endothelin receptor type A endothelin receptor type B European Union Final Assessment Visit FAV Functional Class Food and Drug Administration forced expiratory volume in 1 second Hereditary pulmonary arterial hypertension follicle stimulating hormone Good Clinical Practice gamma glutamyl transferase 14

128 GM2008/00365/04 CONFIDENTIAL hcg human chorionic gonadotropin HIV Human Immunodeficiency Virus HMG-CoA 5-hydroxy-3-methylglutaryl-coenzyme A HRQL Health Related Quality of Life IB Investigator s Brochure ICF Informed Consent Form ICH International Conference on Harmonisation IDMC Independent Data Monitoring Committee IEC Independent Ethics Committee INR international normalized ratio IP Investigational product ipah idiopathic pulmonary arterial hypertension ITT intent to treat IUD intrauterine device IVRS Interactive Voice Response System IRB Institutional Review Board iv intravenous LFT liver function test LOCF last-observation-carried-forward LVEDP left ventricle end diastolic pressure MedDRA Medical Dictionary for Regulatory Activities m meter mg milligram min minute mitt Modified Intent To Treat mmhg millimetres mercury mpap mean pulmonary artery pressure MRP2 multidrug resistance-associated protein 2 NDA New Drug Application NO nitric oxide NT-pro-BNP N-terminal pro-b-type natriuretic peptide NTCP sodium-taurocholate co-transporter protein NYHA New York Heart Association OATP organic anion transporter polypeptide od once daily PAH pulmonary arterial hypertension PAH-CTD pulmonary arterial hypertension associated with connective tissue disease PAP pulmonary artery pressure PBO Placebo PCWP pulmonary capillary wedge pressure PDE-5 phosphodiesterase type 5 PGI 2 prostacyclin P-gp P-glycoprotein PH pulmonary hypertension po by mouth, orally 15

129 GM2008/00365/04 CONFIDENTIAL PP PT PVOD PVR QoL RAP RHC RoW SAE SaO 2 SF-36 SOC SPC SUSAR TAD tid TLC TtCW TtCF UGT ULN US WHO 6MWD Per Protocol prothrombin time Pulmonary veno-occlusive disease pulmonary vascular resistance Quality of Life right atrial pressure right heart catheterization Rest of World serious adverse event arterial oxygen saturation SF-36 Health Survey System Organ Class Summary of Product Characteristics Suspected Unexpected Serious Adverse Reaction Tadalafil three times daily total lung capacity Time to Clinical Worsening Time to Clinical Failure uridine glucuronosyltransferase upper limit of normal United States World Health Organization 6-minute walk distance Trademark Information Trademarks of the GlaxoSmithKline group of companies FLOLAN Trademarks not owned by the GlaxoSmithKline group of companies Adcirca Borg CR10 scale Cialis Letairis Levitra nquery Advisor Remodulin SF-36 Thelin Tracleer Ventavis Volibris 16

130 GM2008/00365/04 PROTOCOL SUMMARY Rationale CONFIDENTIAL Pulmonary arterial hypertension (PAH) is a devastating, life threatening disease. A number of treatments are available, but no single drug has been demonstrated to be consistently effective in treating all patients with PAH. Combining drug products with different mechanisms of action is an evolving strategy for the treatment of PAH. This approach is supported by data from several studies in PAH demonstrating the synergistic benefits of combining two or more drugs [Humbert, 2004a; Hoeper, 2003; Ghofrani, 2002; Hoeper, 2005; Simonneau, 2008]. All of the larger clinical studies completed to date have evaluated combination therapy as an add-on approach: the addition of a second PAH therapy to a subject s monotherapy regimen. It is unknown if PAH patients may have greater improvements for sustained periods of time if combination PAH treatment is initiated as first-line therapy, rather than delaying the addition of the second therapy until the PAH status of the patient has reached a plateau or subsequently deteriorated while receiving monotherapy. Objective(s) The primary objective of this study is to compare the two treatment strategies; first-line combination therapy (ambrisentan and tadalafil) versus first-line monotherapy (ambrisentan or tadalafil) in subjects with PAH. This will be assessed by time to the first clinical failure event. The secondary objectives of this study are to compare the change in other clinical measures of PAH after initiating first-line combination therapy or first-line monotherapy, in subjects with PAH. The safety and tolerability of first-line combination therapy will be compared to first-line monotherapy. In addition, the effect of ambrisentan on exercise capacity at both peak and trough plasma concentrations will be assessed in subjects with pulmonary arterial hypertension (PAH). Study Design This Phase III/IV, randomised, double-blind study will compare the safety and efficacy of first-line combination therapy (ambrisentan and tadalafil) to first-line monotherapy (ambrisentan or tadalafil) in subjects with WHO/NYHA functional class II and III PAH. 17

131 GM2008/00365/04 CONFIDENTIAL The study is event driven and therefore enrolment and study duration are dependent on study event rate. Enrolment of subjects will continue up until 24 weeks prior to the anticipated 105 th first event in the mitt population. Currently 614 total subjects is planned to provide 520 modified ITT subjects who meet the PAH Diagnosis and classification eligibility criteria defined in protocol amendment 02 with 260 subjects in the combination arm and 260 in the monotherapy arm (130 subjects receiving ambrisentan and 130 subjects receiving tadalafil). Subjects must have a confirmed diagnosis of PAH with documented mean pulmonary arterial pressure (mpap) 25 mmhg, pulmonary vascular resistance (PVR) 300 dyne sec/cm 5, and pulmonary capillary wedge pressure (PCWP) or left ventricle end diastolic pressure (LVEDP) of 12 mmhg if PVR 300 to <500 dyne sec/cm 5, or PCWP/LVEDP 15 mmhg if PVR 500 dyne sec/cm 5, based on a right heart catheterization (RHC) prior to Screening. Subjects must not have 3 or more of the following left ventricular disease/dysfunction risk factors: I. Body Mass Index (BMI) 30 II. History of Essential Hypertension III. Diabetes Mellitus any type IV. Historical evidence of significant coronary disease established by any one of: history of myocardial infarction history of percutaneous intervention angiographic evidence of CAD (>50% stenosis in at least one vessel), either by invasive angiography or by CT Angiography positive stress test with imaging (either pharmacologic or with exercise treadmill) previous coronary artery surgery chronic stable angina Subjects are required to have a negative V/Q scan. If results of the V/Q scan are not definitive, evidence of a pulmonary arteriogram (negative for CTEPH) is required prior to screening. Subjects must not have previously received chronic PAH therapy (i.e., prostanoids, ERAs, or PDE-5 inhibitors) within 4 weeks prior to Screening. Subjects who previously received PAH therapy for less than 14 days must not have received any PAH therapy within 7 days prior to Screening Visit. Subjects must have a 6-minute walk distance (6MWD) of at least 125m and no more than 500m at the screening visit, and the screening and baseline 6MWD tests must not vary by greater than 10%, to be eligible for randomisation. 18

132 GM2008/00365/04 CONFIDENTIAL Eligible subjects will be stratified based on the underlying aetiology of PAH (IPAH/HPAH and Non-IPAH) and WHO Functional Class (II and III). Subjects will be randomised 2:1:1 to either the combination therapy arm (ambrisentan and tadalafil) or to the monotherapy arm (ambrisentan and placebo or tadalafil and placebo). Subjects will also be randomly assigned in a 1:1 allocation within each treatment group to peak or trough 6MWD assessment at week 16. The target doses of the study medication will be 10mg ambrisentan (ABS) once daily and 40mg tadalafil (TAD) once daily. Combination therapy arm: Subjects randomised to combination treatment will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see Section for details) Monotherapy arm: ambrisentan group: Subjects randomised to ambrisentan monotherapy will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter Two tablets of TAD-matching placebo Monotherapy arm: tadalafil group: Subjects randomised to tadalafil monotherapy will receive both Two tablets of ABS-matching placebo One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see Section for details). 19

133 GM2008/00365/04 CONFIDENTIAL Subjects will be assessed for efficacy and safety at Screening, Randomisation, Weeks 4, 8, 16, 24, and every 12 weeks thereafter. Subjects will also have monthly liver function and pregnancy safety assessments, which may be performed by a local phlebotomy laboratory or at the Investigator clinic (or by the central lab with Sponsor permission). The primary endpoint of the study is time to the first clinical failure event. Criterion for clinical failure will be adjudicated by a blinded study-specific Clinical Endpoint Committee (CEC). Other than the protocol defined dose adjustments of Investigational Product, PAH therapies may not be added unless a subject has experienced a clinical failure event based on the given criteria. Additional PAH therapies (other than ambrisentan and tadalafil) that are necessary for treatment will not be supplied by the sponsor. All subjects are eligible to receive a minimum of 24 weeks of therapy. At the time that 105 adjudicated primary endpoint events are projected to have occurred, i.e. 105 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety assessment will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed. An Independent Data Monitoring Committee (IDMC) will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will recommend continuation or early termination of the study based on the criteria defined in the IDMC charter. Study Endpoints/Assessments The primary efficacy endpoint is the time (days) to the first clinical failure event of PAH. Time to clinical failure (TtCF) is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) 20

134 GM2008/00365/04 CONFIDENTIAL >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO class III symptoms assessed at two clinic visits separated by 6 months Time to clinical worsening (death, hospitalisation for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Medical Interventions for PAH During the Study: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary endpoint event and subjects may continue in the study following initiation. If a subject experiences an event of clinical failure the following medical interventions may also be utilized. The clinical failure event necessary for medical intervention is based on the opinion of the investigator and does not require prior adjudication from the clinical endpoint committee. The investigator may choose to request blinded combination therapy titration to ensure that the subject is receiving ambrisentan/tadalafil combination therapy The investigator may choose to add prostanoid therapy (parenteral or nonparenteral) (see Section 5.1.4). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first intervention (see Section ) the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy Initiate parenteral prostanoid therapy Request blinded ambrisentan/tadalafil combination therapy. Secondary endpoints include: Change from baseline 6MWD measured at week 24 Percentage of subjects with satisfactory clinical response measured at week 24, defined as: 10% improvement in 6MWD compared to baseline 21

135 GM2008/00365/04 CONFIDENTIAL Improvement to or maintenance of WHO class I or II symptoms No events of clinical worsening prior to or at the week 24 visit Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP) Change from baseline measured at week 24 in WHO Functional Class Change from baseline measured at week 24 in BDI immediately following exercise Exploratory endpoints include: An assessment of the effects of ambrisentan on 6MWD at peak and trough plasma concentrations, which includes: The ambrisentan placebo-corrected mean change from baseline 6MWD following 16 weeks of treatment (i.e., tadalafil/ambrisentan combination therapy versus tadalafil/placebo monotherapy) The ratio of the placebo-corrected mean change from baseline trough 6MWD at week 16 to the placebo-corrected mean change from baseline peak 6MWD at week 16. Health outcomes endpoints include Change from baseline to week 24 in Quality of Life: as measured by the Short Form 36 Health Survey (SF-36) & the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) (selected countries where language translation is validated) Safety endpoints include Incidence and severity of adverse events Vital signs (i.e. blood pressure and heart rate). Laboratory test: clinical chemistry, haematology, testicular function (males only) The modified intent to treat (mitt) population will include all subjects randomised to treatment who receive at least one dose of Investigational Product and who meet the PAH diagnosis and classification eligibility criteria defined in Protocol Amendment No. 2, analyzed according to randomised treatment group. This will be the primary analysis population for assessing efficacy. A per protocol (PP) population will also be defined and used to alternatively describe the primary and the most important secondary efficacy endpoints. The PP population will consist of the subset of subjects contained in the mitt without any major protocol violation. If the PP population is greater than 85% or less than 50% of the mitt population, a PP analysis will not be performed. Additional sensitivity analyses of all randomised and dosed patients (ITT) and non-mitt patients will also be performed for the primary endpoint. Safety data will be summarised for the mitt, full ITT and non-mitt patients. 22

136 GM2008/00365/04 CONFIDENTIAL The CEC will provide adjudication of primary and secondary clinical events reported during the study. The members of this committee will be blinded to treatment assignment and investigator. This event-driven study requires 105 subjects from the mitt population with a clinical failure event in order to have approximately 97% power for the comparison of combination therapy with pooled monotherapy, and 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and were originally projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. Following a blinded review of event rate after approximately 2 years of recruitment, the number of estimated adjudicated events (adjudicated events plus events pending adjudication adjusted for concordance) was approximately 77% of the predicted overall event rate of 15%. The overall event rate was re-estimated to be approximately 12%, and a combination event rate of 8% and a monotherapy event rate of 16% were assumed to maintain the reduction in the hazard of 53% (see Section 8.2.3). Enrolment of subjects will continue up until 24 weeks prior to the anticipated 105 th first event in the mitt population (up to a maximum of 680 subjects). Based on the current blinded overall event rate estimate the study will require approximately 614 subjects to obtain 520 subjects in the mitt population (260 subjects in the combination arm and 260 in the monotherapy arm [130 subjects receiving ambrisentan and 130 subjects receiving tadalafil]) assuming a recruitment period of 148 weeks (based on an average recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 175 weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 5% per treatment group at one year has been considered for this calculation. For the primary analysis, time to the first clinical failure event of PAH, the log rank test will be used to compare the two treatment arms. Time to clinical failure will be displayed by treatment group as a Kaplan-Meier event-free curve from randomisation up to the Final Assessment Visit (FAV). Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using a Cox proportional hazards regression model. The primary comparison will be between the combination therapy arm and the pooled monotherapy arms. Comparison of the combination therapy to the individual monotherapy arms is a secondary comparison. These comparisons will only be performed if the comparison of the combination vs. pooled monotherapy is significant. 23

137 GM2008/00365/04 CONFIDENTIAL To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints and treatment comparisons. Inferences on the superiority of combination versus pooled monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance for combination therapy vs pooled monotherapy therapy. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint combination vs pooled comparison is statistically significant will inferences on the first secondary endpoint be evaluated. Only if the first secondary endpoint combination vs pooled comparison is found to be significant will inferences on the subsequent secondary endpoint be evaluated. This gate keeping approach will be implemented for all secondary endpoints in a pre-defined order, and all tests will be performed at a 5% significance level. Additionally, for each endpoint, if the combination treatment arm is demonstrated as statistically significantly superior to the pooled monotherapy arm using the gate keeping approach described above, comparisons of combination treatment versus each individual monotherapy treatment will be performed at a 5% significance level. 24

138 GM2008/00365/04 1. INTRODUCTION 1.1. Background CONFIDENTIAL Pulmonary Arterial Hypertension PAH consists of a group of diseases characterized by a progressive increase of pulmonary vascular resistance (PVR) and pulmonary artery pressure (PAP). Group 1 PAH includes idiopathic pulmonary arterial hypertension (ipah), hereditary pulmonary arterial hypertension (hpah), and PAH associated with various conditions such as connective tissue disease (CTD), congenital systemic-to-pulmonary shunts, portal hypertension, anorexigen use, and human immunodeficiency virus (HIV) infection [Galiè, 2009b]. All of these disorders share similar proliferative and obstructive changes of the pulmonary microcirculation, including plexiform lesions [Galiè, 2009b]. The vascular damage that defines PAH is characterized by a mean pulmonary artery pressure (mpap) 25 mmhg at rest and a pulmonary capillary wedge pressure (PCWP) 15 mmhg. Patients with PAH typically develop significant increases in PVR and sustained elevations in PAP, which ultimately lead to right ventricular failure and death [Rubin, 1993; Galiè, 2009b]. Patients diagnosed with PAH have a poor prognosis and equally compromised quality of life, with a mean life expectancy of 2 to 5 years from the time of diagnosis if untreated [D'Alonzo, 1991]. PAH medicines that block or reduce the vasoconstriction and hypertrophy associated with the narrowing of the pulmonary arteries may ameliorate the cycle of constriction and improve pulmonary vascular blood flow Current Treatment Options Diuretics and anticoagulants have been widely used in the management of PAH, yet response varies [Humbert, 2004c]. Calcium channel blockers (CCBs) have shown improved survival in vasoreactive IPAH patients, yet the relatively low incidence of vasoreactivity make CCBs useful only in a minority of the population with PAH [Humbert, 2004c; Rich, 1992]. Three signalling pathways involved in the pathogenesis of PAH have been targeted for therapeutic intervention [Humbert, 2004c]: the camp-dependent prostacyclin (PGI 2 ) pathway, the cgmp-dependent nitric oxide (NO) pathway, and the phospholipase-c-dependent endothelin pathway. These pathways are targeted by the following classes of PAH medicines: prostacyclin derivatives (FLOLAN (epoprostenol), Ventavis (iloprost) and Remodulin (treprostinil)), phosphodiesterase type-5 (PDE-5) inhibitors (Revatio (sildenafil) and Adcirca (tadalafil)), and endothelin receptor antagonists (ERAs) (Volibris [EU]/Letairis [US] (ambrisentan), Tracleer (bosentan) and Thelin (sitaxentan)). Despite the number of treatments available, PAH remains a fatal condition. There is high clinical interest in combining treatments targeting different pathophysiological pathways. 25

139 GM2008/00365/04 CONFIDENTIAL The first signalling pathway is the camp-dependent PGI 2 pathway. Prostanoids such as IV epoprostenol, and other PGI 2 derivatives target this pathway. These medicinal products mimic the effects of PGI 2, an endogenous prostaglandin produced by the vascular endothelium, to stimulate camp-dependent vasodilation of the pulmonary arterial bed and inhibit platelet aggregation [Humbert, 2004b]. The second signalling pathway is the cgmp dependent nitric oxide (NO) pathway. Tadalafil is the second PDE-5 inhibitor approved in the EU and USA, along with a number of other countries, for the treatment of PAH, that targets the NO pathway. Through inhibition of PDE-5, tadalafil increases cytoplasmic cgmp concentrations and enhances NO-mediated vasodilation of the vasculature [Forgue, 2006]. Tadalafil 40mg once daily demonstrated significant benefit compared to placebo over 12 weeks of treatment and is the licensed dose [Galiè, 2009a]. The third targeted pathway is the phospholipase-c-dependent endothelin pathway. Endothelin-1 (ET-1) is the primary member of a family of potent vasoconstrictor peptides, which are known to play an essential role in mammalian cardiovascular physiology [Rubanyi, 1994]. Two receptor subtypes, endothelin receptor type A (ET A ) and endothelin receptor type B (ET B ), mediate the effects of ET-1. In humans, the ET A receptor is preferentially expressed in vascular smooth muscle cells and is primarily responsible for the vasoconstrictive, as well as the mitogenic effects of ET-1 [Davenport, 1995; Kirchengast, 1999]. In contrast, ET B receptors are found mainly in the vascular endothelium, and their activation results in vasodilation via the production of NO and prostacyclin [Verhaar, 1998; Hirata, 1993]. The ET B receptor is also involved in the regulation of circulating concentrations of ET-1, through effects on endothelin converting enzyme (ECE-1) expression, and the synthesis and reuptake of ET-1 by endothelial cells [Fukuroda, 1994; Ozaki, 1995]. It has therefore been suggested that ET- 1 plays a critical role in the pathogenesis and progression of PAH. Ambrisentan is the second FDA-approved ERA in the US and third in the EU, and is approved in a number of other countries, for the treatment of PAH. Ambrisentan was the first, and is the only non-sulfonamide ERA approved in all regions. It works by targeting the ET A receptor, blocking the deleterious effects of endothelin-1. Ambrisentan 5mg and 10mg demonstrated significant benefit compared to placebo over 12 weeks, and are the licensed doses in most countries [Ambrisentan EU Summary of Product Characteristics, 2010; Ambrisentan US Prescribing Information, 2009] Ambrisentan Ambrisentan [(+)-(2S)-2-[(4,6-dimethylpyrimidin-2-yl)oxy]-3-methoxy-3,3- diphenylpropanoic acid] is an orally active Endothelin Receptor Antagonist (ERA) that is selective for the ET A receptor. Ambrisentan (Volibris, Letairis) is the latest ERA approved in the US, EU and many other countries internationally. 26

140 GM2008/00365/ Clinical Efficacy CONFIDENTIAL In two Phase 3, placebo-controlled studies (AMB-320 [ARIES-1] and AMB-321 [ARIES-2]), ambrisentan demonstrated statistically significant improvement in 6MWD [Galiè, 2008]. An increase in 6MWD was observed after 4 weeks of treatment with ambrisentan, with a dose-response observed after 12 weeks of treatment. Results from AMB-321 demonstrated that 5 mg and 2.5 mg orally, by mouth once daily of ambrisentan improved the placebo-corrected 6MWD by 59 meters (p <0.001) and 32 meters (p = 0.022), respectively. Similarly, results from AMB-320 demonstrated that 10 mg and 5 mg orally, by mouth once daily of ambrisentan improved the placebo-corrected 6MWD by 51 meters (p <0.001) and 31 meters (p = 0.008), respectively. Time to clinical worsening, an indicator of disease progression, was a key secondary endpoint in the two Phase 3, placebo-controlled studies. The individual studies were not statistically powered to examine secondary endpoints such as time to clinical worsening; therefore, to increase the power to observe a treatment effect, a combined analysis of the two Phase 3 studies was pre-specified to examine secondary endpoints. The log-rank test for the comparison of the combined ambrisentan group versus placebo demonstrated that a significant delay in the time to clinical worsening of PAH was observed for subjects receiving ambrisentan (p <0.001). Furthermore, the hazard ratio was 0.29 (95% CI: 0.14 to 0.59), indicating a 71% reduction in the probability of clinical worsening over the 12- week treatment period for subjects receiving ambrisentan compared to placebo. The conclusions of the combined analysis were supported by similar trends in the individual studies. In AMB-320, a 2-fold increase in the number of subjects with an event of clinical worsening was observed in the placebo group compared to each of the ambrisentan dose groups; however, the log-rank comparison of the combined ambrisentan group versus placebo did not demonstrate a statistically significant difference in the time to clinical worsening of PAH (p = 0.214). In AMB-321, a 4-fold increase in the number of subjects with an event of clinical worsening was observed in the placebo group compared to each of the ambrisentan dose groups. The log-rank test demonstrated a significant delay in time to clinical worsening of PAH for the comparison of the combined ambrisentan group versus placebo (p <0.001). Clinically relevant or statistically significant improvements were observed in the individual studies for the secondary endpoints of WHO functional class, the SF-36 health survey, and BDI. In the pre-specified combined analysis, all secondary endpoints were statistically significant. In addition, statistically significant improvements from baseline with ambrisentan treatment were observed for several haemodynamics parameters (cardiac index, mpap, and PVR) in a Phase 2 dose-controlled study (AMB-220), consistent with an overall improvement in disease state. [Galie, 2005] 27

141 GM2008/00365/04 CONFIDENTIAL Plasma B-type natriuretic peptide (BNP), a hormone secreted primarily from the cardiac ventricles, has been proposed as a prognostic factor of mortality in patients with PAH [Leuchte, 2005; Nagaya, 2000; Park, 2004]. The final step of BNP synthesis consists of a high molecular weight precursor, probnp cleaved into biologically inactive N-terminal segment (NT-proBNP) and the proper low molecular weight BNP. NT-proBNP has a longer half-life and a better stability both in circulating blood and after sampling. Reductions in BNP/NT-proBNP have been shown to parallel improvements in haemodynamics, 6MWD, and WHO functional class in patients with PAH [Leuchte, 2004]. Substantial (p <0.05) and clinically relevant decreases (29% to 45%) in BNP/NT-proBNP compared to placebo were observed in the Phase 3 pivotal studies, and decreases in BNP/NT-proBNP correlated with and were predictive of improvements in 6MWD. [Galiè, 2009a] Data from the long-term follow-up of the subjects who were treated with ambrisentan in the Phase 3, placebo-controlled studies and their open-label extension (N = 383) has been reported. After 2 years of ambrisentan exposure, the mean change from baseline in 6MWD was improved for the 5mg (+23m; 95% confidence interval: 9 to 38m) and 10mg (+28m; 95% confidence interval: 11 to 45m) groups. Estimates of survival and freedom from clinical worsening for the combined dose group were 94% and 83%, respectively, at 1 year and 88% and 72%, respectively, at 2 years. The annualized risk of aminotransferase abnormalities >3x the upper limit of normal was ~2% per year; most of these events were mild and did not lead to discontinuation of drug. Without a control group this data should be interpreted with caution [Oudiz, 2009] Safety Safety data for ambrisentan were obtained from the two Phase 3, placebo-controlled studies and four non-placebo-controlled studies in a total of 483 patients with PAH who were treated with doses of 1, 2.5, 5, or 10 mg once daily in the phase 2 (AMB-220, AMB-222) and 3 (AMB-320 and AMB-321) studies. This includes subjects exposed in the long term studies (AMB-220-E, AMB-222, AMB 320/321-E) following rerandomisation of placebo subjects from the parent studies. As of 16 July 2008, a total of 483 subjects with PAH have received ambrisentan in Phase 2 or Phase 3 clinical studies (excluding Study AMB-323), with a mean exposure of ± 55.4 weeks (2 years) and a maximum exposure of 230.1weeks (4.3 years). The adverse drug reactions (ADR) identified from 12 week placebo controlled clinical trial data are: BLOOD AND LYMPHATIC SYSTEM DISORDERS Common: Anemia (decreases in haemoglobin and/or haematocrit) IMMUNE SYSTEM DISORDERS Uncommon: Hypersensitivity (e.g. angioedema, rash) NERVOUS SYSTEM DISORDERS Common: Headache 28

142 GM2008/00365/04 CONFIDENTIAL CARDIAC DISORDERS Common: Palpitations VASCULAR DISORDERS Common: Flushing RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS Common: Nasal congestion, sinusitis, nasopharyngitis The incidence of nasal congestion was dose-related during ambrisentan therapy. GASTROINTESTINAL DISORDERS Common: Abdominal pain, constipation GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS Common: Fluid retention, Peripheral edema With longer observation in uncontrolled studies (mean observation of 79 weeks), the safety profile was similar to that observed in the short term studies. Peripheral oedema is a known class effect of ERAs, and is also a clinical consequence of PAH and worsening PAH. Among subjects receiving ambrisentan (but not placebo), the incidence and severity of peripheral oedema appeared greater for the 65 year old subgroup (28.6%; 16/56) compared to the <65 year old subgroup (14.1%; 29/205). Hepatic enzyme elevations have been observed with ERAs. Therefore, evaluation of hepatic function prior to initiation of ambrisentan and during treatment is recommended. In clinical trials, ambrisentan had a low overall incidence and a low incidence of clinically significant liver transaminase abnormalities, even with long-term treatment. The hepatic safety profile for ambrisentan observed in clinical practice has appeared consistent with that seen in the clinical studies and has not suggested an increased risk of hepatic events with ambrisentan compared to background hepatic events in patients with PAH. Teratogenicity is a class effect of ERAs and animal studies in rats and rabbits have shown that ambrisentan is teratogenic. Thus, use of ambrisentan is contraindicated in pregnancy. Data on teratogenic effects in humans are very limited; no congenital abnormalities or teratogenic effects have been observed in the fetuses of women who have become pregnant while using ambrisentan. Use of reliable contraception and pregnancy testing during treatment is recommended in women of child bearing potential. Additionally, these women should be advised to contact their physician immediately if they become pregnant or suspect they may be pregnant. The development of testicular tubular atrophy in male animals has been linked to the chronic administration of ERAs, including ambrisentan. However, the effect on male human fertility is not known. To date, postmarketing safety surveillance has shown that the nature and severity of AEs reported with ambrisentan have been generally consistent with the known safety profile of ambrisentan. However, several additional adverse reactions have been identified in the post-approval period which has led to amendments to the core safety profile for 29

143 GM2008/00365/04 CONFIDENTIAL ambrisentan. Post-marketing reports of fluid retention occurring within weeks after starting Volibris/Letairis have been received and, in some cases, have required intervention with a diuretic or hospitalization for fluid management or decompensated heart failure. Prescribers have been advised that if patients have pre-existing fluid overload, this should be managed as clinically appropriate prior to starting ambrisentan. Additionally, prescribers have been informed of cases of worsening dyspnea of unclear etiology which have been reported shortly after starting ambrisentan therapy, and nausea and vomiting have been added to the undesirable effects section. Further, an ERA class effect has also been added to the prescribing information related to the possibility that patients with pulmonary veno-occlusive disease could develop acute pulmonary oedema during initiation of therapy with vasodilating agents such as an endothelin receptor antagonist. For full details on ambrisentan including contraindications, warnings and precautions and drug-drug interactions refer to the prescribing information. See Section for details of permitted concomitant medications in which caution is advised Tadalafil Tadalafil is an orally administered selective inhibitor of the enzyme phosphodiesterase type 5 (PDE5), the primary cyclic guanosine monophosphate (cgmp)-hydrolyzing enzyme in smooth muscle [Adcirca US Prescribing Information, 2009]. Tadalafil (Adcirca) 40mg once daily dosing has been approved for the treatment of pulmonary arterial hypertension (PAH) in the US, EU, Japan and Canada Clinical efficacy A randomised, double-blind, 16 week placebo-controlled study was conducted in 405 patients with pulmonary arterial hypertension [Adcirca US Prescribing Information, 2009; Galiè, 2009a], defined as a resting mean pulmonary artery pressure (mpap) 25 mmhg, pulmonary capillary wedge pressure (PCWP) 15 mmhg, and pulmonary vascular resistance (PVR) 3 Wood units via right heart catheterization. Allowed background therapy included bosentan (maintenance dosing up to 125mg twice daily) and chronic anticoagulation. The use of prostacyclin or analogue, L-arginine, phosphodiesterase inhibitor, or other chronic PAH medications were not permitted. Subjects were randomly assigned to 1 of 5 treatment groups (tadalafil 2.5, 10, 20, 40 mg, or placebo) in a 1:1:1:1:1 ratio. Subjects had to be at least 12 years of age and had a diagnosis of PAH that was idiopathic, related to collagen vascular disease, anorexigen use, human immunodeficiency virus (HIV) infection, associated with an atrial-septal defect, or associated with surgical repair of a congenital systemic-to-pulmonary shunt of at least 1 year in duration (for example, ventricular septal defect, patent ductus arteriosus). Patients with a history of left-sided heart disease, severe renal insufficiency, or pulmonary hypertension related to conditions other than specified in the inclusion criteria were not eligible for enrolment. The mean age of all subjects was 54 years (range years) with the majority of subjects being Caucasian (81%) and female (78%). PAH aetiologies were predominantly idiopathic PAH (61%) and related to collagen vascular disease (24%). More than half 30

144 GM2008/00365/04 CONFIDENTIAL (53%) of the subjects in the study were receiving concomitant bosentan therapy. The majority of subjects had a World Health Organization (WHO) Functional Class III (65%) or II (32%). The mean baseline 6MWD was 343 meters. Of the 405 subjects, 341 completed the study. The primary efficacy endpoint was the change from baseline at week 16 in 6MWD. In the Adcirca 40mg treatment group, the placebo-adjusted mean change increase in 6MWD was 33 meters (95% C.I meters; p=0.0004). The improvement in 6MWD was apparent at 8 weeks of treatment and then maintained at week 12 and week 16. Placebo-adjusted changes in 6MWD at 16 weeks were evaluated in subgroups. In patients taking only Adcirca 40mg (i.e., without concomitant bosentan), the placebo-adjusted mean change in 6MWD was 44 meters. In patients taking Adcirca 40mg and concomitant bosentan therapy, the placebo adjusted mean change in 6MWD was 23 meters. There was less clinical worsening (defined as death, lung transplantation, atrial septostomy, hospitalization because of worsening PAH, initiation of new PAH therapy [prostacyclin or analogue, endothelin receptor antagonist, PDE5 inhibitor], or worsening WHO functional class) in the Adcirca 40mg group compared to the placebo group and the groups that used lower doses of Adcirca. Patients (N=357) from the placebo-controlled study entered a long-term extension study. Of these, 311 patients have been treated with tadalafil for at least 6 months and 182 for 1 year (median exposure 356 days; range 2 days to 415 days). The survival rate in the extension study was 96.5 per 100 patient years. Without a control group, these data must be interpreted cautiously Safety Tadalafil was administered to 398 patients with PAH during clinical trials worldwide [Adcirca US Prescribing Information, 2009]. In trials of Adcirca, a total of 311 and 251 subjects have been treated for at least 182 days and 360 days, respectively. The overall rates of discontinuation because of an adverse event (AE) in the placebo-controlled trial were 9% for Adcirca 40mg and 15% for placebo. The rates of discontinuation because of AEs, other than those related to worsening of PAH, in patients treated with Adcirca 40mg was 4% compared to 5% in placebo-treated patients. In the placebo-controlled study, the most common AEs were generally transient and mild to moderate in intensity. Table 1 presents treatment-emergent adverse events reported by 9% of patients in the Adcirca 40mg group and occurring more frequently than with placebo [Adcirca US Prescribing Information, 2009]. 31

145 GM2008/00365/04 CONFIDENTIAL Table 1 Treatment-Emergent Adverse Events Reported by >= 9% of Patients in Adcirca and More Frequent than Placebo by 2% (16 week study) Event Placebo (%) N = 82 Adcirca 20mg (%) N=82 Adcirca 40mg (%) N=79 Headache Myalgia Nasopharyngitis Flushing Respiratory Tract Infection (Upper and Lower) Pain in Extremity Nausea Back Pain Dyspepsia Nasal Congestion (Including sinus congestion) Adverse Reactions and Post-marketing Experience from Adcirca USPI: The following adverse reactions have been identified during post-approval use of tadalafil. These events have been chosen for inclusion either because of their seriousness, reporting frequency, lack of clear alternative causation, or a combination of these factors. Because these reactions are reported voluntarily from a population of uncertain size, it is not always possible to estimate reliably their frequency or establish a causal relationship to drug exposure. The list does not include adverse events that are reported from clinical trials and that are listed above in this section. Cardiovascular and cerebrovascular Serious cardiovascular events, including myocardial infarction, sudden cardiac death, stroke, chest pain, palpitations, and tachycardia, have been reported post-marketing in temporal association with the use of tadalafil. Most, but not all, of these patients had pre-existing cardiovascular risk factors. Many of these events were reported to occur during or shortly after sexual activity, and a few were reported to occur shortly after the use of tadalafil without sexual activity. Others were reported to have occurred hours to days after the use of tadalafil and sexual activity. It is not possible to determine whether these events are related directly to tadalafil, to sexual activity, to the patient s underlying cardiovascular disease, to a combination of these factors, or to other factors. Body as a whole Hypersensitivity reactions including urticaria, Stevens Johnson syndrome, and exfoliative dermatitis Nervous Migraine, seizure and seizure recurrence, and transient global amnesia Ophthalmologic Visual field defect, retinal vein occlusion, and retinal artery occlusion Non arteritic anterior ischemic optic neuropathy (NAION), a cause of decreased vision including permanent loss of vision, has been reported rarely postmarketing in temporal association with the use of PDE5 inhibitors, including tadalafil. Most, but not all, of these 32

146 GM2008/00365/04 CONFIDENTIAL patients had underlying anatomic or vascular risk factors for development of NAION, including but not necessarily limited to: low cup to disc ratio ( crowded disc ), age over 50, diabetes, hypertension, coronary artery disease, hyperlipidemia, and smoking. It is not possible to determine whether these events are related directly to the use of PDE5 inhibitors, to the patient s underlying vascular risk factors or anatomical defects, to a combination of these factors, or to other factors. Otologic Cases of sudden decrease or loss of hearing have been reported postmarketing in temporal association with the use of PDE5 inhibitors, including tadalafil. In some of the cases, medical conditions and other factors were reported that may have also played a role in the otologic adverse events. In many cases, medical follow-up information was limited. It is not possible to determine whether these reported events are related directly to the use of tadalafil, to the patient s underlying risk factors for hearing loss, a combination of these factors, or to other factors. Urogenital Priapism. In clinical studies, tadalafil was shown to augment the hypotensive effects of nitrates. This is thought to result from the combined effects of nitrates and tadalafil on the nitric oxide/cgmp pathway. Therefore, administration of Adcirca to patients who are using any form of organic nitrate is contraindicated. For full details on tadalafil including contraindications, warnings and precautions and drug-drug interactions refer to the prescribing information. See Section for details of permitted concomitant medications in which caution is advised Combination of ambrisentan and tadalafil A Phase 1 study in 26 healthy subjects demonstrated no clinically relevant pharmacokinetic interaction between tadalafil and ambrisentan [Spence, 2009]. Single-dose PK of ambrisentan (10mg) and its metabolite, 4-hydroxymethyl ambrisentan, were determined in the absence and presence of multiple doses of tadalafil (40mg once daily). Similarly, single-dose PK of tadalafil (40mg) were evaluated in the absence and presence of multiple doses of ambrisentan (10mg once daily). In the presence of tadalafil, ambrisentan maximum plasma concentration (Cmax) was similar (105.0% [90% CI: %]) and systemic exposure (AUC 0 1 ) was slightly decreased (87.5% [ %]), compared with ambrisentan alone. Similar changes were observed with 4- hydroxymethyl ambrisentan. Tadalafil C max (100.6% [ %]) and AUC 0 1 (100.2% [ %]) showed no difference in the absence and presence of ambrisentan. The safety profile of the drugs combined was similar to that of either drug alone. Overall, there was a trend for the mean SBP and DBP in the sitting, supine, and standing positions to decrease over the course of both Treatment Regimens. The potential for additive effects on blood pressure when ambrisentan and tadalafil are administered in combination in the PAH population will be monitored during the trial. 33

147 GM2008/00365/04 CONFIDENTIAL No animal or in vitro studies have been undertaken with ambrisentan in combination with tadalafil. Both ambrisentan and tadalafil are indicated individually for the treatment of patients with pulmonary arterial hypertension. Pharmacologically ambrisentan is a highly specific and potent endothelin Type-A receptor antagonist whilst tadalafil is a specific and potent inhibitor of cgmp specific phosphodiesterase type 5 (PDE5). The possibility for additive toxicity for AEs which are associated with both drugs e.g. nasal congestion/upper respiratory infections, flushing, nausea, headache cannot be excluded. From an analysis of pharmacological activity and the known toxicological findings of each of these compounds there is no evidence to suggest that there would be any synergistic toxicity or exacerbation of the known toxicological profile of either individual compound if they were administered to patients in combination. However, the development of testicular tubular atrophy in male animals has been linked to the chronic administration of ERAs, including ambrisentan. In dogs given tadalafil daily for 6 to 12 months at doses of 25mg/kg/day (resulting in at least a 3-fold greater exposure [range ] than seen in humans given a single 20mg dose) and above, there was regression of the seminiferous tubular epithelium that resulted in a decrease in spermatogenesis in some dogs. In the AMBITION study, clinical laboratory assessments of testicular function will be collected and male subjects will be informed of the potential risks Rationale Pulmonary arterial hypertension (PAH) is a devastating, life threatening disease. A number of treatments are available but no single drug has been demonstrated to be consistently effective in treating all patients with PAH. Combining medicines with different mechanisms of action is an evolving strategy for the treatment of PAH [Galiè, 2009b]. This approach is supported by data from several exploratory studies in PAH demonstrating the synergistic benefits of combining two or more drugs [Humbert, 2004a; Hoeper, 2003; Ghofrani, 2002; Hoeper, 2005; Simonneau, 2008]. Ambrisentan (5 and 10mg once daily) is an ETA-selective ERA that is approved by the FDA and EMA for the treatment of PAH (ipah and PAH-CTD) in patients with WHO class II or III symptoms [Ambrisentan EU Summary of Product Characteristics, 2010; Ambrisentan US Prescribing Information, 2009]. Tadalafil (40mg once daily) is a PDE-5 inhibitor approved in numerous countries for the treatment of PAH (ipah and PAH -CTD) in patients with WHO class II and III symptoms [Adcirca Summary of Product Characteristics, 2010; Adcirca US Prescribing Information, 2009]. Because ambrisentan and tadalafil are both orally administered once a day, have different mechanisms of action targeting different intracellular pathways, and have not demonstrated any clinically relevant pharmacokinetic concerns when coadministered, an ambrisentan/tadalafil combination therapy is a rational treatment strategy for patients with PAH. 34

148 GM2008/00365/04 CONFIDENTIAL All of the larger clinical studies completed to date have evaluated combination therapy as an add-on approach: the addition of a second PAH therapy to a subject s monotherapy regimen. It is unknown if PAH patients may have greater improvements for sustained periods of time if combination PAH treatment is initiated as first-line therapy, rather than delaying the addition of the second therapy until the PAH status of the patient has reached a plateau or subsequently deteriorated while receiving monotherapy. The primary objective of this study is to compare the difference between two treatment strategies; first-line combination therapy (ambrisentan and tadalafil) versus first-line monotherapy (ambrisentan or tadalafil) in subjects with PAH. This will be assessed by time to the first clinical failure event Rationale for ambrisentan peak-trough 6MWD assessment The use of ambrisentan as a once daily, oral therapy for the treatment of PAH is supported by its mean terminal half-life of 12.9 to 17.9 hours in PAH patients [Ambrisentan Summary of Product Characteristics, 2010; Ambrisentan US Prescribing Information, 2009] and the consistent and clinically relevant efficacy observed in the placebo-controlled studies (see Section 1.2.1). However, the mean plasma concentration of ambrisentan varies considerably during the 24-hour dosing interval (mean trough concentration at steady-state is approximately 15% of the mean peak concentration at steady-state). The relationships between ambrisentan plasma concentrations, endothelin receptor occupancy in the pulmonary vasculature, and clinical efficacy outcomes, have not been clearly established. However, in prior clinical studies, the timing of the 6MWD assessment relative to ambrisentan dosing was not specified or collected; therefore, the potential for variability of this assessment over the 24-hour dosing interval has not been evaluated. This study will evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations. To demonstrate an ambrisentan drug treatment effect, the change from baseline to week 16 in 6MWD for the ambrisentan/tadalafil combination group and the tadalafil monotherapy group will be compared (ABS+TAD vs. Placebo+TAD). The potential for differences in treatment effect over the 24-hour dosing interval will subsequently be explored by comparing the ratio of placebo-corrected 6MWD response at peak and trough ambrisentan plasma concentrations. 2. OBJECTIVE(S) The primary objective of this study is to compare the difference between two treatment strategies; first-line combination therapy (with ambrisentan 10mg once daily and tadalafil 40mg once daily) vs. monotherapy (with ambrisentan 10mg once daily or tadalafil 40mg once daily) in subjects with pulmonary arterial hypertension (PAH). This will be assessed by time to first clinical failure (TtCF) event. TtCF is a variation on time to clinical worsening (TtCW), adding an additional component of unsatisfactory long-term clinical response. This additional component is proposed based on the recent guidelines for the clinical treatment of PAH [Galiè, 2009b]. 35

149 GM2008/00365/04 CONFIDENTIAL These guidelines recommend a target based treatment approach, based on parameters with established importance for assessing disease severity, stability and prognosis in pulmonary arterial hypertension. The patients status is divided into three categories: stable and satisfactory, stable and not satisfactory, unstable and deteriorating. The aim of treatment is to move all patients to the stable and satisfactory category. Therefore current treatment guidelines do not deem it acceptable to wait for a patient to experience a significant deterioration prior to medical intervention. As this is a treatment strategy study, this unsatisfactory long-term response component is intended to better reflect the principle of target based treatment approach by counting unsatisfactory response as clinical failure. This allows physicians to intervene prior to a patient experiencing significant deterioration and better reflects current clinical practice [Galiè, 2009b]. The secondary objectives of this study are to compare the change in other clinical measures of PAH after initiating combination therapy (with ambrisentan and tadalafil) or monotherapy (with ambrisentan or tadalafil) in subjects with PAH. The safety and tolerability of combination therapy (with ambrisentan and tadalafil) will be compared to monotherapy (with ambrisentan or tadalafil). In addition, the effect of ambrisentan on exercise capacity at both peak and trough plasma concentrations will be assessed. The study endpoints are described in Section INVESTIGATIONAL PLAN 3.1. Study Design Figure 1 Study Schematic This Phase III/IV, randomised, double-blind study will compare the safety and efficacy of initiating combination therapy (with ambrisentan and tadalafil) or monotherapy (with ambrisentan or tadalafil) as first-line therapy in subjects with WHO FC II and III PAH. 36

150 GM2008/00365/04 CONFIDENTIAL This is an event driven trial. The duration of the study and sample size will be determined by the rate of first occurrence of time to clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. Recruitment will continue up to 24 weeks prior to the predicted 105 th event in the mitt population, and is anticipated to be 614 subjects based on the current blinded event rate estimate. Should the event rate estimate increase over projections, recruitment will be stopped prior to 614 subjects, or should it reduce, may continue past 614 subjects, up to a maximum of 680. The study will be terminated when it is projected that approximately 105 events in the mitt population have occurred. Based on current assumptions, the median treatment duration is anticipated to be 2.1 years with the total study duration estimated at 3.4 years. Eligible subjects will be stratified based on the underlying aetiology of PAH (IPAH/HPAH and Non-IPAH) and WHO Functional Class (II and III). Subject will be randomised 2:1:1 to either the combination therapy arm (ambrisentan and tadalafil) or to the monotherapy arm (ambrisentan and placebo or tadalafil and placebo). Additionally, subjects will also be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. The target doses of the study medication will be 10mg ambrisentan (ABS) once daily and 40mg tadalafil (TAD) once daily. Combination therapy arm: Subjects randomised to combination treatment will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (Creatinine Clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see Section for details). Monotherapy arm: ambrisentan group: Subjects randomised to ambrisentan monotherapy will receive both One tablet of 5mg ABS and one tablet of ABS-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg ABS (10mg once daily), thereafter Two tablets of TAD-matching placebo Monotherapy arm: tadalafil group: Subjects randomised to tadalafil monotherapy will receive both Two tablets of ABS-matching placebo 37

151 GM2008/00365/04 CONFIDENTIAL One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and, two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (Creatinine Clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on tadalafil 20mg od or up-titrate to tadalafil 40mg od (see Section for details). As there is limited data on combining these two drugs in patients a cautious approach has been taken with a titration of ambrisentan and tadalafil. In addition the uptitration timepoints have been separated to avoid dose increases at the same time. The tadalafil monotherapy group will also start on 20mg od to provide a true comparison with the combination group and allow subjects with mild to moderate renal impairment to follow dosing recommendations. It should be noted that since the recommended dose for tadalafil for subjects with PAH without renal insufficiency is 40mg, it is possible that some subjects in the trial may be under dosed on tadalafil for the first 4 weeks of the trial. Should tolerability issues be experienced the investigator may separate dosing of Investigational Products in to morning and evening (i.e., ambrisentan in the morning and tadalafil in the evening). If tolerability issues are still experienced the investigator may request a down titration of ambrisentan from 10mg to 5mg once daily (it should be noted that subjects in TAD monotherapy group will not have a change in dose). Therefore all subjects should be closely monitored for continued lack of tolerance to IP). If tolerability issues are still experienced the investigator may then also request a down titration of tadalafil from 40mg once daily to 20mg once daily (requires approval of the Sponsor Medical Monitor). See Section for details. Subjects will be assessed for efficacy and safety at Screening, Randomisation, Weeks 4, 8, 16, 24, and every 12 weeks thereafter. In between clinic visits subjects will have monthly laboratory safety assessments. These may be performed by a local phlebotomy laboratory or at the Investigator clinic (or by the central lab with Sponsor permission). The primary efficacy endpoint is the time to the first clinical failure event of PAH. Time to clinical failure is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) 38

152 GM2008/00365/04 CONFIDENTIAL >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO class III symptoms assessed at two clinic visits separated by 6 months Time to clinical worsening (death, hospitalization for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Medical Interventions for PAH During the Study: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary endpoint event and subjects may continue in the study following initiation. If a subject experiences an event of clinical failure the following medical interventions may also be utilized. The clinical failure event necessary for medical intervention is based on the opinion of the investigator and does not require prior adjudication from the clinical endpoint committee. The investigator may choose to request blinded combination therapy to ensure that the subject is receiving ambrisentan/tadalafil combination therapy The investigator may choose to add prostanoid therapy (parenteral or nonparenteral) (see Section 5.1.4). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first intervention (see Section ) the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy Initiate parenteral prostanoid therapy Request blinded ambrisentan/tadalafil combination therapy. Prostanoids will not be supplied by the sponsor. The investigator is required to complete a Change of Randomised Treatment Visit to assess safety and efficacy. Any additional PAH therapies necessary for treatment or changes in dose outside the protocol design will not be supplied by the sponsor. Prior to any change in PAH treatment, a subject is required to complete a Change of Randomised Treatment Visit to assess safety and efficacy. 39

153 GM2008/00365/04 CONFIDENTIAL Criteria for clinical failure events will be reviewed and adjudicated by a blinded studyspecific Clinical Endpoint Committee. Secondary endpoints include: Change from baseline 6MWD measured at week 24 Percentage of subjects with satisfactory clinical response measured at week 24, defined as: 10% improvement in 6MWD compared to baseline Improvement to or maintenance of WHO class I or II symptoms No events of clinical worsening prior to or at the week 24 visit Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP) Change from baseline measured at week 24 in WHO Functional Class Change from baseline measured at week 24 in BDI immediately following exercise Exploratory endpoints include: An assessment of the effects of ambrisentan on 6MWD at peak and trough plasma concentrations, which includes: The ambrisentan placebo-corrected mean change from baseline 6MWD following 16 weeks of treatment (i.e., tadalafil/ambrisentan combination therapy versus tadalafil/placebo monotherapy) The ratio of the placebo-corrected mean change from baseline trough 6MWD at week 16 to the placebo-corrected mean change from baseline peak 6MWD at week 16. Health outcomes endpoints include: Change from baseline to week 24 in Quality of Life: as measured by the Short Form 36 Health Survey (SF-36) & the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) (selected countries where language translation is validated) Safety endpoints include: Incidence and severity of adverse events Vital signs (i.e. blood pressure and heart rate). Laboratory test: clinical chemistry, haematology, testicular function (males only) 40

154 GM2008/00365/04 CONFIDENTIAL All subjects are eligible to receive a minimum of 24 weeks of therapy. When 105 adjudicated primary endpoint events are projected to have occurred, i.e. 105 mitt subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety follow-up by phone will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed (see Figure 2). An IDMC will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will recommend continuation or early termination of the study based on the criteria defined in the IDMC charter. Figure 2 End of Study Strategy Including Unblinding Following Database Lock Figure 2: Study Visit Strategy Primary analysis Sensitivity analysis Projected 105 events Database Lock 1 (triggers unblinding of subjects) Database Lock 2 *Subjects with events who do not return within 28 days will not have their events counted in the primary analysis, but will in the sensitivity analsysis. Subjects to return within 28* days to clinic Last subject Final Assessment Visit [all events to FAS visit included in primary analysis ] 4 weeks (standard timelines for ecrf between last subject last visit and database lock) Blinded Study Drug Subjects to return to clinic within 4* weeks for unblinding following database lock Last subject End of Study visit. Subject returns to clinic and is unblinded. Investigator prescribes standard care 4 weeks (30 day safety follow up after last dose of study medication) Last subject follow up phone call for safety 4 weeks (standard timelines for ecrf between last subject last visit and database lock) Standard of Care 41

155 GM2008/00365/04 CONFIDENTIAL 4. SUBJECT SELECTION AND WITHDRAWAL CRITERIA 4.1. Number of Subjects Planned enrolment is approximately 614 subjects to provide 520 mitt subjects with PAH i.e. meeting the inclusion/exclusion criteria defined in Protocol Amendment No.2 (260 subjects in the combination arm and 260 in the monotherapy arm [130 subjects receiving ambrisentan and 130 subjects receiving tadalafil]). Since this is an event-driven study, the final number of subjects will depend on the rate of clinical failure events, and may be increased to a maximum of 680 subjects. Approximately 150 investigational sites world-wide will be selected for participation in the study Inclusion Criteria Subjects eligible for enrolment in the study must meet all of the following criteria: Demographics 1. Subject must be between 18 and 75 years of age, inclusive, at the Screening Visit 2. Subject must weigh 40 kg at the Screening Visit PAH Diagnosis and Classification 3. Subjects must have a diagnosis of PAH due to the following: a. idiopathic or heritable PAH b. PAH associated with: i. connective tissue disease (e.g., limited scleroderma, diffuse scleroderma, mixed CTD, systemic lupus erythematosus, or overlap syndrome) ii. drugs or toxins iii. HIV infection iv. congenital heart defects repaired greater than 1 year prior to screening (i.e., atrial septal defects, ventricular septal defects, and patent ductus arteriosus) NB: subjects with portopulmonary hypertension and PVOD are NOT eligible for the study NB: Subjects must not have 3 or more of the following left ventricular disease/dysfunction risk factors: i. Body Mass Index (BMI) 30 ii. History of Essential Hypertension iii. Diabetes Mellitus any type iv. Historical evidence of significant coronary disease established by any one of: 42

156 GM2008/00365/04 CONFIDENTIAL history of myocardial infarction history of percutaneous intervention angiographic evidence of CAD (>50% stenosis in at least one vessel), either by invasive angiography or by CT Angiography positive stress test with imaging (either pharmacologic or with exercise) previous coronary artery surgery chronic stable angina 4. Subject must have a current diagnosis of being in WHO Functional Class II or III. 5. Subject with a diagnosis of HIV must have stable disease status. For this study, stable HIV status is defined as: i. No addition of medications for treatment of HIV for at least 8 weeks prior to screening ii. iii. No active opportunistic infection during the Screening Period No hospitalizations due to HIV for at least 4 weeks prior to screening 6. Subject must meet all of the following haemodynamic criteria by means of a RHC prior to screening: i. mpap of 25 mmhg ii. PVR 300 dynes/sec/cm5 iii. PCWP or LVEDP of 12 mmhg if PVR 300 to <500 dyne sec/cm 5, or PCWP/LVEDP 15 mmhg if PVR 500 dynes/sec/cm 5 U.S. Specific Text: Subjects in the United States who have had previous RHC and have evidence to suggest the presence of PH but require an RHC during the Screening Period to confirm the diagnosis of PH (as defined above) may be permitted to receive a RHC as a study procedure, provided approval from the medical monitor is obtained in advance. Approval will be based upon the following criteria: The existing RHC data in the subject's medical record is deemed to be inadequate or non-interpretable per the investigator's judgment An adequate RHC record could not be obtained or was not accessible from the subject's medical record Subject has met all other screening requirements prior to undergoing RHC The site/subject are able to perform the RHC in advance and within the screening period Documentation of approval will be maintained in the source documents and by the Sponsor 7. Subject must meet all of the following pulmonary function tests completed no more than 24 weeks before the Screening visit: 43

157 GM2008/00365/04 CONFIDENTIAL i. Total lung capacity (TLC) 60% of predicted normal and ii. Forced expiratory volume in one second (FEV1) 55% of predicted normal Subjects are required to have a documented negative V/Q scan or pulmonary arteriogram confirming the absence of CTEPH prior to screening. 8. Subject must walk a distance of 125m and 500m at the screening visit. In addition the screening and baseline 6MWD tests must not vary by greater than 10% (see Section for further details) 9. Subject, with or without supplemental oxygen, must have a resting arterial oxygen saturation (SaO2) 88% as measured by pulse oximetry at the Screening Visit. Exercise Programmes 10. Subject has not enrolled in an exercise training program for pulmonary rehabilitation within 12 weeks prior to the Screening Visit and must agree not to enroll in an exercise training program for pulmonary rehabilitation during the Screening Period and the first 24 weeks of the study. Subjects enrolled in an exercise program for pulmonary rehabilitation 12 weeks prior to screening may enter the study if they agree to maintain their current level of rehabilitation for the first 24 weeks of the study. General 11. Female subject of childbearing potential, if sexually active, must agree to use 2 reliable methods of contraception from the Screening Visit until study completion and for at least 30 days following the last dose of Investigational Product (reliable methods of contraception are described in Appendix 1). Subjects who have had a Copper T 380A IUD or LNg 20 IUD inserted are not required to use additional methods of contraception 12. Subject must agree not to participate in a clinical study involving another investigational drug or device throughout this study 13. Subject must be competent to understand the information given in the Institutional Review Board (IRB) or Independent Ethics Committee (IEC) approved ICF and must sign the form prior to the initiation of any study procedures Specific information regarding warnings, precautions, contraindications, adverse events, and other pertinent information on the investigational product that may impact subject eligibility is provided in the IB and product label. French subjects: In France, a subject will be eligible for inclusion in this study only if either affiliated to or a beneficiary of a social security category Exclusion Criteria Subjects meeting any of the following criteria must not be enrolled in the study (please note that screening visit laboratory tests may be performed by local or central laboratory): 44

158 GM2008/00365/04 PAH Treatments CONFIDENTIAL 1. Subject received previous PAH therapy continuously for 14 days or more (PDE5i, ERA, prostanoid) prior to the screening visit. Subjects who previously received PAH therapy for less than 14 days must not have received any PAH therapy within 7 days prior to Screening Visit. 2. Subject received ERA treatment (e.g., bosentan or sitaxentan) or PDE5i treatment (e.g. Sildenafil) at any time AND discontinued due to tolerance issues other than those associated with liver function abnormalities 3. Subjects who have previously discontinued ambrisentan or tadalafil in either another clinical study or commercial product (Volibris/Letairis or Adcirca) for safety or tolerability reasons. 4. Subject has a known hypersensitivity to the Investigational Products, the metabolites, or formulation excipients Other Therapies 5. Subject receiving intravenous inotropes within 2 weeks prior to the Screening Visit (e.g. dopamine, dobutamine) 6. Subject is receiving treatment with a potent inhibitor of CYP3A4 (e.g. protease inhibitors, systemic ketoconazole, or systemic itraconazole) 7. Subject is receiving treatment with a potent inducer of CYP3A4 (e.g. rifampicin) 8. Subject is receiving treatment with cyclosporine A (except ophthalmic formulation) 9. Subjects receiving Calcium Channel Blockers or HMG-CoA reductase inhibitors (i.e., statins) on an unstable dose 4 weeks prior to the Screening Visit (to be eligible subjects must not have changed their dose <4 weeks prior to the screening visit) 10. Subject has a history of angina pectoris that was treated with long or short-acting nitrates <12 weeks of screening or nitrate use for any other condition within 48 hours of screening Laboratory Values at Screening 11. Subject has a serum ALT or AST lab value that is > 2xULN at the Screening Visit 12. Subject has serum bilirubin lab value that is >1.5xULN at the screening visit 13. Subject has severe renal impairment (creatinine clearance <30 ml/min) at the Screening Visit Medical History/Current Medical Conditions Liver 14. Subject has severe hepatic impairment (Child-Pugh class C with or without cirrhosis) at the Screening Visit 45

159 GM2008/00365/04 Haematology and bleeding disorders CONFIDENTIAL 15. Subject has clinically significant anaemia in the opinion of the investigator 16. Subjects with bleeding disorders or significant active peptic ulceration in the opinion of the investigator Cardiovascular 17. Subject has uncontrolled hypertension ( 180/110 mmhg) at screening 18. Subject has severe hypotension (<90/50 mmhg) at screening 19. Subject has had an acute myocardial infarction within the last 90 days prior to screening 20. Subject has, in the opinion of the investigator, clinically significant aortic or mitral valve disease; pericardial constriction; restrictive or congestive cardiomyopathy; lifethreatening cardiac arrhythmias; significant left ventricular dysfunction; left ventricular outflow obstruction; symptomatic coronary artery disease; autonomic hypotension; fluid depletion. Ophthalmic 21. Subject has a past medical history of NAION 22. Subject has a hereditary degenerative retinal disorder (e.g. retinitis pigmentosa) General Medical Conditions 23. Subject has clinically significant fluid retention in the opinion of the investigator 24. Subject with cardiovascular, liver, renal, haematologic, gastrointestinal, immunologic, endocrine, metabolic, or central nervous system disease that, in the opinion of the Investigator, may adversely affect the safety of the subject and/or efficacy of the investigational product or severely limit the lifespan of the subject other than the condition being studied 25. Subject has a history of malignancies within the past 5 years, except for a subject with localized, non-metastatic basal cell carcinoma of the skin, in situ carcinoma of the cervix, or prostate cancer who is not currently or expected, during the study, to undergo radiation therapy, chemotherapy, and/or surgical intervention, or to initiate hormonal treatment General Criteria 26. Female subject who is pregnant or breastfeeding 27. Subject has demonstrated noncompliance with previous medical regimens 28. Subject has a recent (within 1 year) history of abusing alcohol or illicit drugs 29. Subject has participated in a clinical study involving another investigational drug or device within 4 weeks before the Screening Visit 46

160 GM2008/00365/04 CONFIDENTIAL Subjects who fail inclusion/exclusion criteria may be re-screened once Withdrawal Criteria If a subject prematurely and permanently discontinues Investigational Product administration, a Change of Randomised Treatment Visit must be performed. If a subject discontinues Investigational Product prematurely, the subject should be encouraged to continue all study Clinic Visits according to the planned schedule of events, after completing the Change of Randomised Treatment Visit. If, following discontinuation of Investigation Product, a subject does not wish to continue with the burden of all Clinic Visits and assessments, it is acceptable to continue in the study with a reduced clinic visit schedule and set of assessments. The subject may perform 24 weekly clinic visits and only the key information required to assess the TTCF primary EP (TTCF, 6MWD, WHO FC), safety (AEs. SAE s) and concomitant medications need to be collected. The safety follow-up call 30 days after discontinuation of IP should still be completed. After discontinuing Investigational Product, if a subject continues to perform Clinic Visits, and later prematurely discontinues from the study the subject should perform an End of Study Visit. Every effort will be made by the investigator to follow the subject to assess primary endpoint events. All subjects have the right to withdraw formal consent without prejudice at any time during the study. If a subject withdraws formal consent, the Investigator should make a reasonable effort to determine the cause for withdrawal of consent. Possible reasons for subject discontinuation from IP include, but are not limited to, the following: Development of an AE where continuation of the subject s participation in the study is thought by the Investigator to be inappropriate Subject meets liver stopping criteria and discontinues both Investigational products Subject begins treatment with a prohibited concomitant therapy Lost to follow-up Possible reasons for subject withdrawal from the study include, but are not limited to, the following: Subject withdraws consent/subject requests to discontinue for any reason Subject noncompliance (e.g., refusal to complete study procedures) Discretion of the Investigator Discontinuation of the study at the request of the Sponsor, a regulatory agency, or an IRB/IEC 47

161 GM2008/00365/04 CONFIDENTIAL 5. STUDY TREATMENTS 5.1. Investigational Products Investigational products are ambrisentan, in the form of 5mg tablets or matching placebo and tadalafil, in the form of 20mg tablets, or matching placebo Ambrisentan or placebo Provision, handling and storage The sponsor will provide ambrisentan in the form of round, white, film-coated, immediate-release tablets, containing 5mg ambrisentan. Matching placebo will also be provided and will be identical in appearance to ambrisentan 5mg tablets. Investigational Product will be packaged in ambrisentan monthly dosing packs containing a total of 2 x 35 tablets of ambrisentan and/or ambrisentan-matching placebo in blister strips. The packaged material will be labelled and the contents of the label will be in accordance with all applicable regulatory requirements. Investigational Product will be stored at a temperature up to 30 ο C (59-86 ο F), see Investigational Product label. The Investigational Product distributor will send the initial shipment of Investigational Product to site and Investigational Product re-supply will occur automatically via site inventory monitoring in IVRS. Under normal conditions of handling and administration, investigational product is not expected to pose significant safety risks to site staff. A Material Safety Data Sheet (MSDS) describing the occupational hazards and recommended handling precautions will be provided to site staff if required by local laws or will otherwise be available from Sponsor upon request. Investigational product must be stored in a secure area under the appropriate physical conditions for the product. Access to and administration of the investigational product will be limited to the investigator and authorized site staff. Investigational product must be dispensed or administered only to subjects enrolled in the study and in accordance with the protocol. Subjects should be instructed to bring their used and unused Investigational Product and blister cards with them to every visit to assess compliance and drug accountability. Any unused Investigational Product will be returned by the Investigator to the Sponsor. 48

162 GM2008/00365/ Dosing CONFIDENTIAL Subjects should be recommended to take Investigational Product once daily preferably in the morning with or without food throughout the study unless otherwise instructed. The target dose of ambrisentan is 10mg once daily in the form of 2 x 5mg tablets, or matching placebo, and will follow the dose schedule below - The initial dose of ambrisentan will be 5mg once daily in the form of 1x5mg ambrisentan tablet (or ambrisentan-matching placebo) and 1 x ambrisentanmatching placebo tablet. The subject will take this dose for the first 8 weeks of treatment. - At the 8 week visit the subject will be up-titrated to the target dose of 10mg ambrisentan once daily, in the form of 2 x 5mg ambrisentan tablets (or ambrisentan-matching placebo) - If the investigator judges that the subject has not tolerated 5mg of ambrisentan well, he may request that the subject remain on 5mg. The subject will continue to receive ambrisentan 5mg once daily in the form of 1 x 5mg ambrisentan tablet (or ambrisentan-matching placebo) and 1 x ambrisentan-matching placebo tablet Following week 8: Down-titration from ambrisentan 10mg to ambrisentan 5mg may be requested should the subject not tolerate the investigational product. See Section for details. At the Investigator s discretion, chronic parenteral prostanoid therapy may be initiated at any time throughout the study (which meets the criterion for clinical failure). Other PAH therapies may not be added unless a subject has experienced a clinical failure event Tadalafil or placebo Provision, handling and storage The sponsor will provide tadalafil in the form of a large almond, dark yellow tablet, containing 20mg tadalafil. Matching placebo will also be provided and will be identical in appearance to tadalafil 20mg tablets. Investigational Product will be packaged in tadalafil monthly dosing packs containing 2 x 35 tablets of tadalafil and/or tadalafil-matching placebo in blister strips. The packaged material will be labelled and the contents of the label will be in accordance with all applicable regulatory requirements. Investigational Product will be stored at room temperature (15 ο C to 30 ο C (59-86 ο F)). 49

163 GM2008/00365/04 CONFIDENTIAL The Investigational Product distributor will send the initial shipment of Investigational Product and Investigational Product re-supply will occur automatically via site inventory monitoring in IVRS. Under normal conditions of handling and administration, investigational product is not expected to pose significant safety risks to site staff. A Material Safety Data Sheet (MSDS) describing the occupational hazards and recommended handling precautions will be provided to site staff if required by local laws or will otherwise be available from Sponsor upon request. Investigational product must be stored in a secure area under the appropriate physical conditions for the product. Access to and administration of the investigational product will be limited to the investigator and authorized site staff. Investigational product must be dispensed or administered only to subjects enrolled in the study and in accordance with the protocol. Subjects should be instructed to bring their used and unused Investigational Product and blister cards with them to every visit to assess compliance and drug accountability. Any unused Investigational Product will be returned by the Investigator to the Sponsor Dosing Subjects should be recommended to take Investigational Product once daily preferably in the morning, with or without food, throughout the study unless otherwise instructed. The target dose of tadalafil is 40mg once daily (as this is the licensed dose) in the form of 2 x 20mg tablets, or matching placebo and will be initiated at 20mg od and titrated to 40mg od at week 4. Those subjects with mild to moderate impaired renal function (creatinine clearance >30mL/min and <80 ml/min) should be carefully assessed for tolerability prior to any dose up titration as tadalafil exposure has been shown to double in these subjects. The investigator may decide that the subject remain on 20mg od or uptitrated to 40mg od. It should be noted that since the recommended dose for tadalafil for subjects with PAH without renal insufficiency is 40mg, it is possible that some subjects in the trial may be under dosed on tadalafil for the first 4 weeks of the trial. The dosing will follow the dose schedule below: Combination Therapy Group The dose of tadalafil will be initiated at 20mg and titrated to 40mg after 4 weeks. This rationale for this is to try to minimise any potential tolerability issues, particularly those that may be due to a synergistic effect of the 2 investigational products - The initial dose of tadalafil will be 20mg once daily in the form of 1 x 20mg tablet and 1x tadalafil-matching placebo tablet. The subject will take this dose for the first 4 weeks of treatment. The 20mg od dose may not provide subjects will optimal efficacy. 50

164 GM2008/00365/04 CONFIDENTIAL - At the 4 week visit the subject will be automatically up-titrated to the target dose of 40mg tadalafil once daily, in the form of 2 x 20mg tadalafil tablets - Subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80 ml/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on the 20mg od dose or up-titrate to 40mg od. Tadalafil Monotherapy Group - The initial dose of tadalafil will be 20mg once daily in the form of 1 x 20mg tablet and 1x tadalafil-matching placebo tablet. The subject will take this dose for the first 4 weeks of treatment. The 20mg od dose may not provide subjects will optimal efficacy for the first 4 weeks. The rationale for starting with 20mg is to provide a better comparator to the combination therapy group as well as enabling subjects with mild to moderate renal impairment to follow the recommended dose regimen. - At the 4 week visit the subject will be up-titrated to the target dose of 40mg tadalafil once daily, in the form of 2 x 20mg tadalafil tablets - Those subjects with mild to moderate renal impairment (creatinine clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk:benefit decision made by the investigator whether the subject remain on the 20mg od dose or up-titrate to 40mg od Down-titration from tadalafil 40mg to tadalafil 20mg may be requested following consultation with the Sponsor Medical Monitor should the subject not tolerate the study medication. See Section for details. At the Investigator s discretion, chronic parenteral prostanoid therapy may be initiated at any time throughout the study (which meets the criterion for clinical failure). Other PAH therapies may not be added unless a subject has experienced a clinical failure event Down titration of Investigational Product Should the subject not tolerate Investigational Product the following algorithm should be followed: 1. Separate dosing of Investigational Products in to morning and evening, i.e., ambrisentan in the morning and tadalafil in the evening. DO NOT split the doses (e.g. ABS 5mg AM and ABS 5mg PM is not permitted). 2. Request a blinded down-titration of ambrisentan from 10mg once daily to 5mg once daily using the Interactive Voice Response System (IVRS) (see Section 5.2). It should be noted that subjects in TAD monotherapy group will not have a change in dose. Therefore all subjects should be closely monitored for continued lack of tolerance to IP. 51

165 GM2008/00365/04 CONFIDENTIAL 3. If the subject continues to experience tolerability issues the investigator may seek permission from the Sponsor medical monitor to down-titrate the subject to tadalafil 20mg once daily. If permission is given the investigator may request a blinded down-titration using the Interactive Voice Response System (IVRS) (see Section 5.2). The investigator should complete a Change of Randomised Treatment visit prior to any changes in study medication. If tolerance issues persist the investigator should consider withdrawing the subject from Investigational Product. A Change of Randomised Treatment Visit should be performed Treatment following clinical failure event Following the declaration of a clinical failure event by an Investigator, based on the criteria given, the investigator may do either of the following: Request the subject receive blinded combination therapy. It should be noted that only those subjects on monotherapy will receive an additional treatment. Subjects in the combination arm will remain on the same combination therapy although they may up-titrate their doses of IP to maximum permitted doses if not previously receiving these doses (10mg AMB and/or 40mg TAD), see Section for further details. Therefore if this is chosen it is important to re-assess the subjects status. Non-parenteral prostanoids may be given following a clinical failure event. These will not be provided by the sponsor. Both of these options may be utilised following a TtCF event but may not be initiated at the same time. Should blinded combination therapy be requested first, non-parenteral prostanoids may not be given until after stage 3 blinded combination therapy has been initiated, unless subject remains on current stage due to tolerability reasons (see Section ). Should non-parenteral prostanoids be initiated first, blinded combination therapy may not be requested until the next clinic visit. Parenteral prostanoids may be given at any point during the study. Prior to any change in PAH treatment, a subject is required to complete a Change of Treatment Visit to assess safety and efficacy. All events will be retrospectively adjudicated by a clinical endpoint committee Blinded Combination Therapy Combination Therapy will be fully blinded. Investigators are required to complete a Change of Randomised Treatment visit prior to a subject receiving blinded combination therapy and prior to progressing to the next treatment Stage. In addition subjects should continue to follow the 4 weekly safety visit schedule throughout the treatment Stages 52

166 GM2008/00365/04 CONFIDENTIAL Subjects will initially receive Medical Intervention Stage 1, then Stage 2, then Stage 3, upon request of the investigator Subjects must be assessed in the clinic no more than 8 weeks following initiation of Stage 1 or following progression to the next treatment stage (by completing a Change of Randomised Treatment Visit) If the investigator judges that the subject has not sufficiently tolerated any of the treatment Stages, the Investigator may request that the subject remain on their current Stage. During this post-clinical failure blinded combination therapy, down titration of study drug is not permitted. Subjects who either had no uptitration in study medication or had a down titration in study medication based on tolerability reasons prior to initiation of blinded combination therapy may be re-uptitrated as part of the post-clinical failure blinded combination therapy at the request of the investigator. This uptitration will follow the format in the first Medical Intervention table below. The Investigator will be asked via the IVRS prior to the initiation of blinded combination treatment if the subject should receive full dose of blinded combination therapy or if the subject should stay on any previously requested lower dose of IP. Should the investigator choose the subject receive combination therapy following an event it will be provided in the following example formats to ensure that the blind is maintained (investigator will know that the subject is on combination therapy, but not what group the subject was randomised to): Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg (TAD optimised) ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg (ABS dose optimised) ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo ABS 10mg + TAD PBO ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg (ABS optimised) ABS 10mg + TAD Placebo ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS Placebo + TAD 40mg ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg (TAD optimised) ABS Placebo + TAD 40mg ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg The following example formats will apply in: a. subjects in whom the investigator has requested no uptitration in tadalafil due to renal impairment 53

167 GM2008/00365/04 CONFIDENTIAL b. subjects in whom the investigator has requested no uptitration in ambrisentan at week 8 due to tolerability reasons, or a down titration in either Investigational Product based on tolerability reasons. Exception: Subjects receiving placebo will receive full dose of the corresponding IP (see table above). E.g. Prior to clinical failure the investigator requests down titration of ambrisentan in a subject receiving tadalafil monotherapy (i.e. subject continues to receive ambrisentan placebo). Following clinical failure the investigator requests blinded combination treatment and the subject is up-titrated to 5mg then 10mg of ambrisentan). Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg [prior to week 8 and renal impairment] ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD Placebo ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg [renal impairment] ABS 5mg + TAD Placebo ABS 10mg + TAD PBO ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg [prior to week 8 and renal impairment] ABS 10mg + TAD Placebo ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg [renal impairment] Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS 5mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg 5.2. Treatment Assignment Subjects will be assigned to study treatment in accordance with the randomisation schedule. The randomisation code will be generated using the Sponsor s randomisation system Randall. Randomisation will be performed by an IVRS which will direct the investigator to assign the appropriate treatment to each subject. Eligible subjects will be stratified based on the underlying aetiology of PAH (IPAH/HPAH and Non-IPAH) and WHO Functional Class (II and III). Subjects will be randomised 2:1:1 to either the combination therapy arm (ambrisentan and tadalafil) or to the monotherapy arm (ambrisentan group or tadalafil group). Subjects will also be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. 54

168 GM2008/00365/ Blinding CONFIDENTIAL Blinding will be accomplished by providing ambrisentan or placebo in tablets that are visually indistinguishable and provided in numbered containers and by providing tadalafil or placebo in tablets that are visually indistinguishable and provided in numbered containers. Only the number of the containers to be administered to a given study subject, and not the identity of the Investigational Product, will be provided to sites. Please refer to Section 3.1 (Study Design) and Section 5 (Investigational Product) for details. Unblinding of randomised treatment will occur for all subjects after the database has been locked (Database Lock 1). The investigator or treating physician may request unblinding of a subject s treatment assignment only in the case of an emergency, when knowledge of the study treatment is essential for the appropriate clinical management or welfare of the subject. The investigator must first discuss options with the Sponsor Medical Monitor or appropriate Sponsor study personnel before unblinding the subject s treatment assignment. If this is impractical, the investigator must notify the Sponsor as soon as possible, but without revealing the treatment assignment of the unblinded subject, unless that information is important for the safety of subjects currently in the study. The date and reason for the unblinding must be recorded in the appropriate data collection tool. If the blind is broken by the investigator, the subject will be permanently discontinued from IP and a Change of Randomised Treatment visit will be completed. The subject will continue to be followed until the end of the study. Sponsor s Central Safety Department staff may unblind the treatment assignment for any subject with an SAE. If the SAE requires that an expedited regulatory report be sent to one or more regulatory agencies, a copy of the report, identifying the subject s treatment assignment, may be sent to clinical investigators in accordance with local regulations and/or Sponsor policy Product Accountability In accordance with local regulatory requirements, the investigator, designated site staff, or head of the medical institution (where applicable) must document the amount of Sponsor investigational product dispensed and/or administered to study subjects, the amount returned by study subjects, and the amount received from and returned to Sponsor, when applicable. Product accountability records must be maintained throughout the course of the study Treatment Compliance Subjects should be instructed to bring their used and unused Investigational Product containers with them to every visit to assess compliance and drug accountability. 55

169 GM2008/00365/04 CONFIDENTIAL 5.6. Concomitant Medications and Non-Drug Therapies Permitted Medications and Non-Drug Therapies Standard medical treatment(s) being taken by the subject upon study entry may be maintained throughout the study. Alpha blockers: Tadalafil and alpha adrenergic blocking agents are vasodilators with blood pressure-lowering effects. When vasodilators are used in combination, an additive effect on blood pressure may be anticipated. In some patients, concomitant use of these two drug classes can lower blood pressure significantly which may lead to symptomatic hypotension (e.g., fainting). Although no clinically significant effect on blood pressure was observed in healthy volunteers when tadalafil was co-administered with either alfuzosin or tamsulosin, symptomatic augmentation of the blood pressure lowering effect of doxazosin was reported, including syncope, when tadalafil was co-administered. Given the possibility that the combination of these two classes of drugs may lead to additive effects on blood pressure, caution should be exercised when deciding to coadminister tadalafil with alpha blockers. Alcohol: Three clinical trials were conducted in healthy volunteers to study the effects of coadministration of tadalafil and alcohol, both mild systemic vasodilators. Some subjects in these trials experienced clinically significant decreases in blood pressure, postural dizziness, and orthostatic hypotension when alcohol (0.7g/kg which is roughly equivalent to 6 ounces of 80-proof vodka in an 80-kg male consumed within 10 minutes) was administered in conjunction with either 10mg or 20mg of tadalafil. Four ounces of 80- proof vodka administered within 10 minutes in conjunction with 20mg tadalafil was not associated with orthostatic hypotension or potentiation of the hypotensive effects of alcohol. Therefore, it is recommended that subjects participating in the AMBITION trial to limit alcohol consumption to less than 5 units (per Adcirca USPI) Prohibited Medications and Non-Drug Therapies Drugs prohibited while receiving investigational product: Other ERAs, such as Bosentan (Tracleer) and Sitaxentan (Thelin) Commercial ambrisentan (Volibris or Letairis) Inhaled nitric oxide Intravenous inotropes (e.g. dopamine, dobutamine) PDE-5i such as sildenafil (Revatio or Viagra) and vardenafil (Levitra) Commercial tadalafil (Adcirca or Cialis) Any other investigational therapy All forms of prostanoids are prohibited unless a subject has met the definition of clinical failure (Section 3.1) Nitrates 56

170 GM2008/00365/04 CONFIDENTIAL Potent inhibitors of CYP3A4 (e.g. protease inhibitors, systemic ketoconazole, or systemic itraconazole) Potent inducers of CYP3A4 (e.g. rifampicin) Cyclosporine A (except ophthalmic formulation) 5.7. Treatment after the End of the Study Following the first database lock, subjects will be requested to return for an End of Study visit where they will have their final assessment done and be unblinded to study treatment. The Investigator should then treat the subjects according to best standard of care available to the investigator. Subjects will be followed for 30 days post their last dose of study medication. The sponsor will no longer supply ambrisentan or tadalafil after the end of the study except in those countries where a marketing license is yet to be granted Treatment of Investigational Product Overdose Ambrisentan There is no experience in PAH patients of ambrisentan at daily doses greater than 10mg. In healthy volunteers, single doses of 50mg and 100mg (5 to 10 times the maximum recommended dose) were associated with headache, flushing, dizziness, nausea and nasal congestion. Due to the mechanism of action, an overdose of ambrisentan could potentially result in hypotension. In the case of pronounced hypotension, active cardiovascular support may be required. No specific antidote is available. Advice on overdose is available in the product label Tadalafil Single doses of up to 500mg of tadalafil have been given to healthy subjects, and multiple daily doses up to 100mg have been given to patients with erectile dysfunction. Adverse events were similar to those seen at lower doses. In cases of overdose, standard supportive measures should be adopted as required. Haemodialysis contributes negligibly to tadalafil elimination. No specific antidote is available. Advice on overdose is available in the product label. 57

171 GM2008/00365/04 CONFIDENTIAL 6. STUDY ASSESSMENTS AND PROCEDURES Procedures Week (Wk)/Visit Screen Base line Wk Wk -4 0 Wk Wk Wk Wk 12-week intervals Treatment FAV 3 EOS Visit 4 Safety Visits / telephone contact 4 Weekly 1 Unscheduled/ Follow Change of Tx 2 Up Call 13 Visit Windows All visit windows are ± 3 days with the exception of screen baseline. If eligible, subjects should be randomised within 4 weeks of the screening visit. Written Informed X Consent Subject Demography X Medical History X Disease History X Therapy History X Inclusion/Exclusion X Criteria Efficacy Assessments 6 MWD X X X X X 5 X X X X X BDI Score X X X X X 5 X X X X X WHO FC assessment X X X X X X X X X X Clinical Failure X X X X X X X X X Health Outcome Assessments SF-36 X X X X X X X 14 CAMPHOR X X X X X X X 14 Safety Assessments Concomitant Medication X X X X X X X X X X X Physical X X X Examination Vital Signs 6 X X X X X X X X X X 12-lead ECG X X X Pulse Oximetry X Pulmonary Function Test 7 X Adverse Events X X X X X X X X X X X Serious Adverse X X X X X X X X X X X X Events Laboratory Assessments Haematology X X X X X X X X X X Chemistry X X X X X X X X X X X (inc. LFT s) 8 Testicular Function 9 X X X X X 12 X X X Urinalysis X X X Pregnancy Test 10 S U S S S S S S S S S NT-pro-BNP X X X X X X X X X Coagulation X 58

172 GM2008/00365/04 CONFIDENTIAL Procedures Week (Wk)/Visit Screen Base line Wk Wk -4 0 Wk Wk Wk Wk 12-week intervals Treatment FAV 3 EOS Visit 4 Safety Visits / telephone contact 4 Weekly 1 Unscheduled/ Follow Change of Tx 2 Up Call 13 Pharmacogenetics X (optional) Investigational product (IP) Dispense IP X X X X X X X X 11 Assess IP X X X X X X X X compliance IVRS X X X X X X X X X X 11 Peak-Trough X 1. Safety Visits will be required every 4 weeks after week 4 unless a clinic visit is scheduled. A phone call is required to assess concomitant medications and adverse events (including serious). See the Clinical Laboratory Test section for details of the laboratory tests required. If any additional procedures are performed at the safety visit, an unscheduled visit should be completed. 2. Unscheduled / Change of randomised treatment visit. If a subject permanently withdraws from investigational product and from the study at the same time, the End of Study Visit should be completed. 3. Final Assessment Visit 4. End of Study. (This visit should also be completed by subjects who prematurely withdraw from the study) 5. Subjects will take their investigational products dose at least 16 hours prior to the 6MWD and BDI assessments 6. Includes weight, blood pressure and heart rate plus height (height at screening only). 7. Pulmonary function test may also be carried out at screening if not done within 24 weeks prior. 8. Liver Function Test (LFT) should be measured and Investigational Product dose adjusted / stopped as specified in Section and Section Testicular function (Total testosterone, sex hormone binding globulin (SHBG needed to calculate free testosterone), FSH, LH, and inhibin B. For males only. Every effort should be made to collect blood samples for testicular function between 8am and 1pm 10. Pregnancy test: urine = U, serum = S. For women of child bearing potential only 11. If applicable 12. Alternate visits only (i.e. every 24 weeks) 13. Follow up telephone contact 30 days after last dose of Investigational Product 14. Only if change of randomised treatment See Appendix 2 for instructions on how to conduct the 6 minute walk test. (2 tests [i.e. 1 at screening and 1 at randomisation] should be performed during the screening period such that 2 consecutive tests do not differ by more than 10%) Subjects will have the option to receive study related SMS/text messages including visit reminders and a reminder not to take investigational product less than 16 hours prior to the week 16 visit Critical Baseline Assessments All ongoing conditions and relevant/significant medical history (including all major hospitalizations and surgeries) will be recorded at the Screening Visit. Symptoms related to PAH and/or the underlying aetiology of the disease need not be listed on the medical history form; however, worsening of any symptoms during the course of this study must be captured as an AE. Review historical diagnostic tests, such as Right Heart Catheter (RHC); chest X-ray; echocardiogram; ventilation/perfusion scan, computed axial 59

173 GM2008/00365/04 CONFIDENTIAL tomography (CT) or spiral CT, and pulmonary arteriogram (if applicable), to ensure that none of the exclusion criteria for PAH aetiology are met. Prior PAH treatment history will also be assessed. Historical haemodynamic data (e.g., mpap, RAP, PVR, and cardiac index) based on the most recent RHC prior to starting this study will be collected Efficacy Primary Endpoint Time to Clinical Failure The primary efficacy endpoint is the time to the first clinical failure event of PAH. Time to clinical failure is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO class III symptoms assessed at two clinic visits separated by 6 months Time to clinical worsening (death, hospitalization for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Medical Interventions for PAH During the Study: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary endpoint event and subjects may continue in the study following initiation. 60

174 GM2008/00365/04 CONFIDENTIAL If a subject experiences an event of clinical failure the following medical interventions may also be utilized. The clinical failure event necessary for medical intervention is based on the opinion of the investigator and does not require prior adjudication from the clinical endpoint committee. The investigator may choose to request blinded combination therapy to ensure that the subject is receiving ambrisentan/tadalafil combination therapy The investigator may choose to add prostanoid therapy (parenteral or nonparenteral) (see Section 5.1.4). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first intervention (see Section ) the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy Initiate parenteral prostanoid therapy Request blinded ambrisentan/tadalafil combination therapy. The CEC will review the case (blinded to treatment and investigator) and decide if the event meets the definition of clinical failure of PAH. The CEC Charter provides full details. All events of clinical failure of PAH, as well as confirmation that a specific event has not occurred during this time period, will be collected on a clinical failure CRF throughout the study. If a subject prematurely discontinues from the study, an assessment will be made at the time of discontinuation from the study Secondary Endpoints Minute Walk Distance and Borg Dyspnea Index See Appendix 2. The 6MWD will be assessed at each clinic visit. Screening period (i.e. screening visit and baseline visit): Two 6MWD tests should be performed during the screening period, the first 6MWD test at the screening visit, and the second 6MWD test at the baseline visit. The tests should not vary by >10%. A maximum of 4 tests may be conducted during the Screening Period per subject to generate two consecutive tests that meet the criteria below. A maximum of two tests can be performed per day and tests must be separated by at least two hours. If two consecutive tests that vary by <10% cannot be performed the subject should not be entered into the study. If they do not vary by >10% the average value of the two consecutive tests should be used as the baseline value. 61

175 GM2008/00365/ Borg Dyspnea Index See Appendix 3. CONFIDENTIAL The Borg Dyspnea Index score will be collected at each clinic visit. It should be performed immediately after the 6MWD test. Screening period (i.e. screening visit and baseline visit): The Borg Dyspnea Index should be performed straight after every 6MWD test. A maximum of 4 tests may be conducted during the Screening Period per subject to generate two consecutive tests that meet the criteria in Section The Borg score following the screening 6MWD test should be recorded as the screening value, and the average value of the two scores following the two consecutive 6MWD tests should be used as the baseline Borg value N-Terminal pro-b-type Natriuretic Peptide Blood samples for determination of NT-proBNP plasma concentrations will be collected at each clinic visit from randomisation onwards and analysed via a central laboratory. Investigators will remain blinded to NT-proBNP results during the study WHO Functional Class assessment WHO functional class will be assessed at each clinic visit. Every effort will be made to have the same Investigator assess the subject at each study visit Exploratory Endpoints Peak-trough assessment of 6MWD All subjects will be randomly assigned to perform the Week 16 6MWD assessment at either peak or trough ambrisentan concentrations. Prior to arriving at the investigative site for the Week 16 visit, subjects will not administer their daily dose of either Investigational Products (ABS or TAD). The site will dispense blinded Investigational Product (ambrisentan/ambrisentan matching placebo) as directed by the IVRS for the peak-trough assessment. Subjects in the ambrisentan monotherapy arm and the combination arm will receive either ambrisentan followed by placebo (peak assessment) or placebo followed by ambrisentan (trough assessment). Subjects randomised to the tadalafil monotherapy arm will receive placebo followed by placebo to maintain the blind. Subjects will take the first dose of ambrisentan Investigational Product (placebo or ambrisentan) and the 6MWD will be assessed 2 hours (± 30 minutes) after Investigational Product administration (approximate C max ). Borg dyspnea index (BDI) will be assessed immediately following this test. After the remaining clinical assessments have been completed, subjects will then take the second dose of ambrisentan Investigational Product to receive either their daily dose of ambrisentan or placebo to maintain the blind. Subjects will also take their daily dose of tadalafil Investigational Product at this time. 62

176 GM2008/00365/ Safety Electrocardiogram CONFIDENTIAL A 12-lead ECG will be performed at Screening, Week 24 and Final Assessment Visit. Any clinically relevant ECG findings will be captured in the CRF. Any clinically relevant changes since the Screening Visit will be captured as an AE in the CRF Vital Signs and Body Weight and Height Vital signs (including heart rate and blood pressure) will be collected at each clinic visit. One measurement of blood pressure and heart rate after the subject has been sitting quietly for at least 5 minutes should be taken. This same procedure should be followed throughout the study. Oxygen saturation will be collected at screening only. All measures of blood pressure will be performed using standard sphygmomanometry. If possible, the same sphygmomanometer and arm should be used. In addition, body weight will be collected at each clinic visit. Height will be collected at screening only Clinical Laboratory Tests Safety Tests The following tests are required at all clinic visits (including FAV, EOS visit, Unscheduled/Change of Randomised Treatment Visit (if applicable)) and will be performed by a central laboratory, except for the screening visit, where a local or central laboratory may be used: Chemistry: serum ALT, AST, alkaline phosphatase, GGT, total bilirubin, creatinine, egfr, amylase, blood urea nitrogen (BUN), sodium, potassium, chloride, bicarbonate, calcium, total cholesterol, uric acid, glucose, total protein and albumin. Haematology: haemoglobin, haematocrit, red cell count, red cell indices, white blood cell count (total and differential) and platelet count At baseline only coagulation will also be performed: Prothrombin Time with INR (PT) and activated Partial Thromboplastin Time (aptt). Pregnancy*: Serum pregnancy test at all visits, except baseline when a urine pregnancy test will be performed The following tests are required at all safety visits and will be performed by a local phlebotomy laboratory or at the Investigator clinic (or the central lab with Sponsor permission): 63

177 GM2008/00365/04 CONFIDENTIAL Liver function: serum ALT, AST, alkaline phosphatase, GGT, total bilirubin Pregnancy* : serum *In female subjects of child bearing potential The following test will be performed in males only at baseline, weeks 4, 16, 24, subsequently at alternate 12 week visits (i.e. every 24 weeks), FAV and EOS visits as well as Unscheduled/Change of Randomised Treatment Visit (if applicable) and will be performed by a central laboratory. Every effort should be made to collect blood samples for testicular function between 8am and 1pm: Testicular Function: Total testosterone, Sex hormone binding globulin (SHBG needed to calculate free testosterone), FSH, LH and inhibin B The following test will be performed at baseline, week 24, Final Assessment Visit and will be performed by a central laboratory: Total urinalysis: including a microscopic examination of the urine sediment Additional laboratory tests may be performed by a local laboratory or the central laboratory if clinically relevant abnormal values are obtained at any time during the course of the study. Local laboratory reports must be immediately reviewed by the Investigator. Laboratory normal ranges will be collected Pharmacogenetics A pharmacogenetics sample will be collected ideally at baseline (or after baseline if the sample cannot be collected then). See Appendix 4 for details Collection, processing, labelling, and shipping of the pharmacogenetic samples will follow central laboratory guidelines Management, Monitoring and Follow-up of Serum Aminotransferase Abnormalities All subjects who are receiving investigational product will have blood drawn every 4 weeks to assess hepatic function. Assessments for hepatic function include serum aminotransferases (alanine aminotransferase, ALT or aspartate aminotransferase, AST), alkaline phosphatase, gamma glutamyl transferase (GGT), and total bilirubin concentrations. In order to ensure the safety of all subjects, the results of these tests must be reviewed by the Investigator immediately upon receipt. An increase in serum ALT or AST concentration >3xULN must be confirmed by a second test (>3xULN) that is collected no more than 7 days after receipt of the initial lab report and is analyzed by the central laboratory. The requirements presented in Table 2 must be applied to the review and management of elevated serum aminotransferases in all subjects receiving investigational product; however, investigators may elect to interrupt or discontinue investigative product(s) at a lower threshold based on their assessment of the patient s clinical status. 64

178 GM2008/00365/04 CONFIDENTIAL Table 2 Management, Monitoring and Follow-up of Subjects Experiencing Serum Aminotransferase Elevations Serum Aminotransferase Elevations ALT>3xULN and bilirubin >2xULN or if ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. ALT/AST >3x and 5xULN ALT/AST >5x and 8xULN Actions to be Taken Stop AMB/PBO treatment immediately. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Report the liver event to the Medical Monitor within 24 hours of learning of its occurrence and report as an SAE. Make every reasonable attempt to have subjects return to clinic within 24 hours for repeat liver chemistries, liver event follow up assessments, and close monitoring. Complete liver event CRF within 24 hours. After discontinuation of AMB/PBO, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. A specialist or hepatology consultation is recommended AMB/PBO may not be reintroduced. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. If confirmed, interrupt AMB/PBO therapy. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Continue to monitor serum aminotransferase concentrations every 2 weeks or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of AMB/PBO therapy may be considered by the Investigator after consultation with the Medical Monitor. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. Stop AMB/PBO treatment immediately. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. 65

179 GM2008/00365/04 CONFIDENTIAL Serum Aminotransferase Actions to be Taken Elevations Continue to monitor serum aminotransferase concentrations every 2 weeks, or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of investigational product may be considered by the Investigator after consultation with the Medical Monitor. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. ALT/AST >8xULN Stop AMB/PBO treatment immediately. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo and subjects should therefore be carefully monitored subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Report the liver event to the Medical Monitor and within 24 hours of learning of its occurrence. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. After discontinuation of AMB/PBO, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. AMB/PBO therapy may not be reintroduced. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. *If AMB/PBO therapy is reintroduced, serum aminotransferase concentrations should be checked within 3 days, and then every 2 weeks or more frequently if clinically indicated. Once investigator, in consultation with Medical Monitor, feels that subject is clinically stable, liver safety testing may be resumed at standard protocol defined intervals. If at any time these subjects experience serum aminotransferase elevations, proceed as described above. Subjects withdrawn from the investigational product may remain in the study for followup and collection of data for outcome Liver event follow up assessments For liver events, make every attempt to carry out the liver event follow up assessments described below: Viral hepatitis serology including: Hepatitis A IgM antibody; Hepatitis B surface antigen and Hepatitis B Core Antibody (IgM); Hepatitis C RNA; 66

180 GM2008/00365/04 CONFIDENTIAL Cytomegalovirus IgM antibody; Epstein-Barr viral capsid antigen IgM antibody (or if unavailable, obtain heterophile antibody or monospot testing); Hepatitis E IgM antibody (if subject resides outside the US or Canada, or has travelled outside US or Canada in past 3 months); Blood sample for PK analysis, obtained within 24 hours of last dose. Record the date/time of the PK blood sample draw and the date/time of the last dose of investigational product(s) prior to blood sample draw on the CRF. If the date or time of the last dose is unclear, provide the subject s best approximation. If the date/time of the last dose cannot be approximated OR a PK sample cannot be collected in the time period indicated above, do not obtain a PK sample. Instructions for sample handling and shipping are in the SPM. Serum creatine phosphokinase (CPK) and lactate dehydrogenase (LDH). Fractionate bilirubin, if total bilirubin >2xULN. Serum bilirubin fractionation should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick. Obtain complete blood count with differential to assess eosinophilia Record the appearance or worsening of clinical symptoms of hepatitis or hypersensitivity, such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia as relevant on the AE report form Record use of concomitant medications, acetaminophen, herbal remedies, other over the counter medications, or putative hepatotoxins, on the concomitant medications report form. Record alcohol use on the liver event alcohol intake case report form Complete the liver imaging and/or liver biopsy CRFs if these tests are performed The following are required for subjects with ALT >3xULN and bilirubin >2xULN but are optional for other abnormal liver chemistries: Anti-nuclear antibody, anti-smooth muscle antibody, and Type 1 anti-liver kidney microsomal antibodies. Liver imaging (ultrasound, magnetic resonance, or computerized tomography) to evaluate liver disease Adverse Events The investigator or site staff will be responsible for detecting, documenting and reporting events that meet the definition of an AE or SAE. 67

181 GM2008/00365/ Definition of an AE CONFIDENTIAL Any untoward medical occurrence in a patient or clinical investigation subject, temporally associated with the use of a medicinal product, whether or not considered related to the medicinal product. Note: An AE can therefore be any unfavourable and unintended sign (including an abnormal laboratory finding), symptom, or disease (new or exacerbated) temporally associated with the use of a medicinal product. For marketed medicinal products, this also includes failure to produce expected benefits (i.e., lack of efficacy), abuse or misuse. Events meeting the definition of an AE include: Exacerbation of a chronic or intermittent pre-existing condition including either an increase in frequency and/or intensity of the condition New conditions detected or diagnosed after investigational product administration even though it may have been present prior to the start of the study Signs, symptoms, or the clinical sequelae of a suspected interaction Signs, symptoms, or the clinical sequelae of a suspected overdose of either investigational product or a concomitant medication (overdose per se will not be reported as an AE/SAE). Lack of efficacy or failure of expected pharmacological action per se will not be reported as an AE or SAE. However, the signs and symptoms and/or clinical sequelae resulting from lack of efficacy will be reported if they fulfil the definition of an AE or SAE. Events that do not meet the definition of an AE include: Medical or surgical procedure (e.g., endoscopy, appendectomy); the condition that leads to the procedure is an AE Situations where an untoward medical occurrence did not occur (social and/or convenience admission to a hospital) Anticipated day-to-day fluctuations of pre-existing disease(s) or condition(s) present or detected at the start of the study that do not worsen The disease/disorder being studied, or expected progression, signs, or symptoms of the disease/disorder being studied, unless more severe than expected for the subject s condition Definition of a SAE A serious adverse event is any untoward medical occurrence that, at any dose: a. Results in death b. Is life-threatening 68

182 GM2008/00365/04 CONFIDENTIAL NOTE: The term 'life-threatening' in the definition of 'serious' refers to an event in which the subject was at risk of death at the time of the event. It does not refer to an event, which hypothetically might have caused death, if it were more severe. c. Requires hospitalization or prolongation of existing hospitalization NOTE: In general, hospitalization signifies that the subject has been detained (usually involving at least an overnight stay) at the hospital or emergency ward for observation and/or treatment that would not have been appropriate in the physician s office or out-patient setting. Complications that occur during hospitalization are AEs. If a complication prolongs hospitalization or fulfils any other serious criteria, the event is serious. When in doubt as to whether hospitalization occurred or was necessary, the AE should be considered serious. Hospitalization for elective treatment of a pre-existing condition that did not worsen from baseline is not considered an AE. d. Results in disability/incapacity, or NOTE: The term disability means a substantial disruption of a person s ability to conduct normal life functions. This definition is not intended to include experiences of relatively minor medical significance such as uncomplicated headache, nausea, vomiting, diarrhoea, influenza, and accidental trauma (e.g. sprained ankle) which may interfere or prevent everyday life functions but do not constitute a substantial disruption. e. Is a congenital anomaly/birth defect f. Medical or scientific judgment should be exercised in deciding whether reporting is appropriate in other situations, such as important medical events that may not be immediately life-threatening or result in death or hospitalization but may jeopardize the subject or may require medical or surgical intervention to prevent one of the other outcomes listed in the above definition. These should also be considered serious. Examples of such events are invasive or malignant cancers, intensive treatment in an emergency room or at home for allergic bronchospasm, blood dyscrasias or convulsions that do not result in hospitalization, or development of drug dependency or drug abuse. g. All events of possible drug-induced liver injury with hyperbilirubinaemia defined as ALT >3xULN and bilirubin >2xULN (or ALT >3xULN and INR>1.5, if INR measured) termed Hy s Law events (INR measurement is not required and the threshold value stated will not apply to patients receiving anticoagulants). For the purposes of this protocol, ALT >3xULN associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia will be considered a protocol defined SAE. h. NOTE: bilirubin fractionation should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury. 69

183 GM2008/00365/04 CONFIDENTIAL Laboratory and Other Safety Assessment Abnormalities Reported as AEs and SAEs Any abnormal laboratory test results (haematology, clinical chemistry, or urinalysis) or other safety assessments (e.g., ECGs, radiological scans, vital signs measurements), including those that worsen from baseline, and felt to be clinically significant in the medical and scientific judgement of the investigator are to be recorded as AEs or SAEs. However, any clinically significant safety assessments that are associated with the underlying disease, unless judged by the investigator to be more severe than expected for the subject s condition, are not to be reported as AEs or SAEs. All events of possible drug-induced liver injury with hyperbilirubinaemia (defined as ALT >3xULN plus bilirubin >2xULN and/or INR>1.5) or Hy s Law events, require immediate AMB/PBO cessation and reporting as an SAE (see Section above). NOTE: bilirubin fractionation should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury Pregnancy Subjects who become pregnant during the study must discontinue Investigational Product immediately. The Investigator should counsel the subject regarding the possible effects of prior Investigational Product exposure on the foetus (See SmPC and USPI) and the need to inform the study site of the outcome of the pregnancy. Subjects should be instructed to notify the Investigator if they become pregnant at any time during the study, or if they become pregnant within 30 days of last Investigational Product dose. Any pregnancy that occurs during study participation must be reported using a clinical trial pregnancy form. To ensure subject safety, each pregnancy must be reported to Sponsor within 24 hours of learning of its occurrence. The pregnancy must be followed up to determine outcome (including premature termination) and status of mother and child. Pregnancy complications and elective terminations for medical reasons must be reported as an AE or SAE. Spontaneous abortions must be reported as an SAE. Any SAE occurring in association with a pregnancy, brought to the investigator s attention after the subject has completed the study and considered by the investigator as possibly related to the investigational product, must be promptly reported to Sponsor Time Period and Frequency of Detecting AEs and SAEs AEs will be collected from the start of investigational product and until 30 days post the last dose of investigational product. AEs occurring between screening and randomisation will be recorded as medical history. 70

184 GM2008/00365/04 CONFIDENTIAL All deaths, regardless of cause or relationship, must be reported for subjects on study (defined as the point from which the subject is consented into the study) and for all deaths occurring within 30 days of last investigational product dose or within 30 days of last study evaluation, whichever is longer. SAEs will be collected over the same time period as stated above for AEs. However, any SAEs assessed as related to study participation (e.g., investigational product, protocolmandated procedures, invasive tests, or change in existing therapy) or related to a Sponsor concomitant medication, will be recorded from the time a subject consents to participate in the study up to and including any follow up contact. All SAEs will be reported to Sponsor within 24 hours, as indicated in Section Prompt Reporting of Serious Adverse Events and Other Events to Sponsor SAEs, pregnancies, and liver function abnormalities meeting pre-defined criteria will be reported promptly by the investigator to Sponsor as described in the following table once the investigator determines that the event meets the protocol definition for that event. Initial Reports Follow-up Information on a Previous Report Type of Event Time Frame Documents Time Frame Documents All SAEs 24 hours SAE data 24 hours Updated SAE collection tool Pregnancy 24 hours Pregnancy Notification Form Liver chemistry abnormalities ALT>3xULN and bilirubin>2xuln (or ALT>3xULN and INR>1.5, if INR measured)*** or ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. ALT/AST>3xULN and bilirubin 2xULN 24 hours* SAE data collection tool. **Liver Event Case Report Form (CRF) and liver imaging and/or biopsy CRFs if applicable data collection tool 24 hours Pregnancy Follow up Form 24 hours Updated SAE data collection tool. **Updated Liver Event CRF 24 hours* **Liver Event CRF 24 hours **Updated Liver Event CRF *Sponsor to be notified at onset of liver chemistry elevations to discuss subject safety. ** Liver event documents should be completed as soon as possible *** INR measurement is not required; if measured, the threshold value stated will not apply to patients receiving anticoagulants. 71

185 GM2008/00365/04 CONFIDENTIAL The method of detecting, recording, evaluating and follow-up of AEs and SAEs plus procedures for completing and transmitting SAE reports to Sponsor are provided in the SPM. Procedures for post-study AEs/SAEs are provided in the SPM Regulatory reporting requirements for SAEs Prompt notification of SAEs by the investigator to the Sponsor is essential so that legal obligations and ethical responsibilities towards the safety of subjects are met. The Sponsor has a legal responsibility to notify both the local regulatory authority and other regulatory agencies about the safety of a product under clinical investigation. The Sponsor will comply with country specific regulatory requirements relating to safety reporting to the regulatory authority, Institutional Review Board (IRB)/Independent Ethics Committee (IEC) and investigators. Investigator safety reports are prepared for suspected unexpected serious adverse reactions according to local regulatory requirements and the sponsor policy and are forwarded to investigators as necessary. An investigator who receives an investigator safety report describing a SAE(s) or other specific safety information (e.g., summary or listing of SAEs) from the Sponsor will file it with the IB and will notify the IRB/IEC, if appropriate according to local requirements Health Outcomes Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) Health Related Quality of Life and Quality of Life Measure The CAMPHOR [McKenna, 2006] is a pulmonary hypertension specific instrument, has been developed and validated to facilitate the measurement of HRQL (symptoms and functioning) and QoL in this patient group. It is provided in Appendix 5. It is the only PH disease specific tool available. Subjects will complete the CAMPHOR questionnaire at Randomisation (Week 0), Weeks 16, 24, at Clinic Visits every 12 weeks subsequent to Week 24, the Change of Randomised Treatment Visit (if applicable), and the FAV and EOS visit. The CAMPHOR will only be used in countries where a validated translation is available. The CAMPHOR consists of three sections. The first two address HRQL issues. 1. The first section is the symptom (impairment) score. It contains 25 negatively weighted items consisting of three sub-scales relating to energy, breathlessness and mood. Each item is in the form of a simple statement to which patients indicate whether or not it is true for them at that moment. This two-point response system was selected as it presents the minimum burden on respondents and is the simplest in relation to ease of scoring and equivalence of translation. 72

186 GM2008/00365/04 CONFIDENTIAL 2. The second section is an Activity (disability) score. This consists of a 15 item scale, each of which relates to activities described by patients as being affected by PAH. The patients rate themselves as being able to perform each activity. 3. The third section is Quality of Life and contains 25 negatively weighted items. A change score of 2 points is meaningful, this is the MID (Minimal Important Difference). Further details of the CAMPHOR will be detailed in the Reporting Analysis Plan (RAP) SF-36 Health Survey The SF-36 Health Survey [Ware, 2000] is provided (the US Version) in Appendix 6. The SF-36v2 Health Survey will be used. The SF-36 Health Survey asks 36 questions to measure functional health and well-being from the patient's point of view. It s called a generic health survey because it can be used across age (18 and older), disease, and treatment group, as opposed to a disease-specific health survey which focuses on a particular condition or disease. Whilst not specifically validated in PAH, it has been used in a number of PAH trials. Subjects will complete the SF-36 Health Survey at Randomisation (Week 0), Weeks 16, 24, at Clinic Visits every 12 weeks subsequent to Week 24, the Change of Randomised Treatment Visit (if applicable), and the FAV and EOS visit. The SF-36 Health Survey consists of 8 health domains (Physical Functioning, Role- Physical, Bodily Pain, General Health, Vitality, Social Functioning, Role-Emotional, and Mental Health) as well as 2 summary measures (Physical Health and Mental Health). Further details of the SF-36 will be detailed in the RAP and the scoring will follow the algorithm of [Ware, 2000]. 7. DATA MANAGEMENT For this study subject data will be entered into electronic case report forms (ecrfs), transmitted electronically to the Sponsor and designees, and combined with data provided from other sources in a validated data system. Management of clinical data will be performed in accordance with the applicable sponsor standards and data cleaning procedures to ensure the integrity of the data, e.g., removing errors and inconsistencies in the data. Adverse events and concomitant medications terms will be coded using MedDRA and an internal validated medication dictionary, GSKDrug. ecrfs (including queries and audit trails) will be retained by Sponsor, and copies will be sent to the investigator to maintain as the investigator copy. In all cases, subject initials will not be collected or transmitted to Sponsor according to Sponsor policy. 73

187 GM2008/00365/04 CONFIDENTIAL 8. DATA ANALYSIS AND STATISTICAL CONSIDERATIONS This section will cover the statistical considerations and data analyses in order to assess the primary objective of the study. Further details will be presented in the Reporting Analysis Plan (RAP) Hypotheses The study is designed to determine the efficacy of a treatment strategy of combination therapy compared to monotherapy in subjects with PAH. The null hypothesis tested for the primary endpoint of efficacy is that there is no difference in the time to clinical failure of PAH when treated with monotherapy (ambrisentan or tadalafil) compared to combination therapy (ambrisentan and tadalafil) (i.e., H 0 : HR = 1). The two-sided alternative hypothesis is that there is a difference (i.e., H 1 : HR 1, a two-sided test) Study Design Considerations Sample Size Assumptions This section presents sample size calculations based on an overall event rate of 15% (a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year). Based on the current blinded event rate estimate, the sample size has been adjusted. Section presents additional sample size calculations based on an overall event rate of 12% (a monotherapy event rate of 16% per year and a combination event rate of approximately 8% per year). This event-driven study requires 105 mitt subjects with an adjudicated clinical failure event in order to have approximate 97% power for the comparison of combination therapy with pooled monotherapy, and approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial as defined in the IDMC Charter. Assuming a recruitment period of 134 weeks (based on a recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 159 weeks the study requires approximately 545 subjects enrolled to obtain 456 mitt subjects (228 subjects in the combination arm and 228 in the monotherapy arm [114 subjects receiving ambrisentan and 114 subjects receiving tadalafil]). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 5% per treatment group at one year has been considered for this calculation. 74

188 GM2008/00365/04 CONFIDENTIAL The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have approximately 85% power for each comparison. For this calculation, it has been assumed that there are approximately 70 clinical failure events for 342 mitt subjects (228 on combination and 114 on each monotherapy) and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been assumed. Calculations were performed in EAST version 5.2, Cytel Inc. Central randomisation is planned stratified by aetiology of PAH (IPAH/HPAH and Non- IPAH) and WHO Functional Class (II and III). Additionally, subjects will be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 228 ABS/TAD subjects and 114 PBO/TAD subjects yields approximately 97% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. The primary analysis for the peak/trough assessment will be in the mitt population; but the analysis may be repeated for the ITT population and nonmitt population. These analyses will be defined further in the RAP prior to unblinding. Treatment effect and standard deviation were based on data from Phase 2 and Phase 3 clinical studies of ambrisentan. Sample size and final alpha level were calculated using nquery Advisor Version 6.0. Statistical Solutions, Cork, Ireland. Although the sample size was calculated using a 2-sided t-test, the specified analysis for this endpoint uses a Wilcoxon rank sum test; therefore, the actual power may vary slightly Sample Size Sensitivity The sample size required to achieve a given number of necessary primary endpoint events depends on the basic assumptions as outlined in Section Power calculations for combination therapy vs the pooled monotherapy resulting from a range of treatment effects and monotherapy event rates are shown in the table below. 75

189 GM2008/00365/04 CONFIDENTIAL Table 3 Power Calculations for Varying Event Rates (Combination Therapy vs Pooled Monotherapy) Monotherapy Event Rate 15% 20% 25% 0.70 (30% ) 40% 50% 58% Hazard Ratio (relative reduction) 0.60 (40% ) 66% 78% 85% 0.50 (50% ) 88% 95% 98% 0.47 (53% ) 92% 97% 99% Assuming: 545 subjects (228 combination therapy; 228 pooled monotherapy [114 ambrisentan and 114 tadalafil]), 134 wks recruitment, 159 wks study duration (~3yrs 2 months) alpha of 0.05, 5% annual dropout rate per group and varying Monotherapy annual event rates and varying treatment effects Simulations were run to determine the empirical probability of different study outcomes, based on the assumptions outlined in Section The results are presented in the table below. Table 4 Empirical Probabilities for Study Outcomes Based on Simulations Outcome Combination therapy better than pooled monotherapy Combination therapy better than an individual monotherapy Combination therapy better than pooled monotherapy and better than each individual monotherapy Empirical Probability 96% 88% 79% Assuming 456 mitt subjects, with 105 events, alpha of 0.05, 5% yearly dropout rate per group, 10% combination therapy yearly event rate, and 20% monotherapy yearly event rate (HR=0.47). Simulations were performed using R software. Note: Simulation results differ slightly from power calculations generated using EAST (large sample theory based method). If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at 456, the study duration may differ from the current planned 159 weeks (see Table 5). 76

190 GM2008/00365/04 CONFIDENTIAL Table 5 Study Duration Sensitivity for Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n=456 mitt) 25% 139 weeks 20% 159 weeks 15% 194 weeks Assuming 456 mitt subjects, with 105 events, alpha of 0.05, power of approximately 97% for combination vs pooled monotherapy, 53% Risk Reduction between groups, 5% yearly dropout rate per group and varying Monotherapy yearly event rates Sample size calculations were performed using EAST versions 5.2, Cytel Inc. Simulations in Table 4 were performed using R software Sample Size Re-estimation The event rate will be monitored using data blinded with respect to treatment assignment. Should the observed event rate be lower than anticipated or discontinuation rates (i.e., withdrawal from study or discontinuation of IP) higher than expected, then the sponsors and the Steering Committee will use blinded data to assess the value of increasing study size or length of duration. Following a blinded review of the overall event rate after approximately 2 years of recruitment, the number of estimated adjudicated events (adjudicated events plus events pending adjudication adjusted for concordance) was approximately 77% of the predicted overall event rate of 15%. The overall event rate was re-estimated to be approximately 12%. Sample size calculations were performed assuming a combination event rate of 8% and a monotherapy event rate of 16% to maintain the original assumed reduction in the hazard of 53%. Assuming a recruitment period of 148 weeks (based on a recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 175 weeks, the study requires approximately 614 subjects enrolled to obtain 520 mitt subjects (260 subjects in the combination arm and 260 in the monotherapy arm [130 subjects receiving ambrisentan and 130 subjects receiving tadalafil]). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 5% per treatment group per year has been considered for this calculation. 77

191 GM2008/00365/04 CONFIDENTIAL Based on the revised sample size of 520 mitt patients, 105 mitt subjects with an adjudicated clinical failure event are required in order to maintain the same approximate 97% power for the comparison of combination therapy with pooled monotherapy, and the same approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Calculations were performed in EAST version 5.2, Cytel Inc. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 260 ABS/TAD subjects and 130 PBO/TAD subjects yields approximately 98% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. As this is an event driven trial, the duration of the study and sample size will be determined by the rate of clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. Recruitment will continue up to 24 weeks prior to the predicted 105 th event in the mitt population, and up to a maximum of 680 subjects Data Analysis Considerations Analysis Population The populations to be used in the analysis of this study are: mitt Population: All randomised subjects who meet the inclusion/exclusion criteria defined in Protocol Amendment No. 2 and who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. ITT- All Randomised Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. PP- The per-protocol population will consist of the subset of subjects in the mitt without any major protocol violation. If the PP population is greater than 85% of the mitt population or less than 50% of the mitt population, a PP analysis will not be performed. Safety- As Treated Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment they received. An As Treated population will be defined to correspond to each of the mitt, ITT and non-mitt populations. Non-mITT Population: All randomised subjects who fail to meet the inclusion/exclusion criteria defined in Protocol Amendment No. 2 and who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. 78

192 GM2008/00365/ Analysis Data Sets CONFIDENTIAL All subjects included into this study will be included in the analyses according to the analysis populations defined above in Section For time to event analyses, all lost to follow-up subjects will be censored at their last known date in the study. For other endpoints, LOCF will be used to account for missing data Treatment Comparisons Primary Comparisons of Interest Since the primary objective of this study is to compare the difference between two treatment strategies (first-line combination therapy (ambrisentan and tadalafil) versus first-line monotherapy (ambrisentan or tadalafil) in subjects with PAH), the primary comparison is of the hazard rates of clinical failure between the combination therapy arm (ambrisentan and tadalafil) and the pooled monotherapy arm (ambrisentan and placebo plus tadalafil and placebo). The mitt population will be utilized for this comparison. The comparison will be made at a 5% significance level Other Comparisons of Interest For each endpoint, the primary comparison will be between the combination therapy arm and the pooled monotherapy arms, and the primary analysis will use the mitt population. Comparison of the combination therapy to the individual monotherapy arms is a secondary comparison. These comparisons will only be performed if the comparison of the combination vs. pooled monotherapy is significant. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints and treatment comparisons. Inferences on the superiority of combination versus pooled monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance for combination therapy vs. pooled monotherapy therapy. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint combination vs. pooled monotherapy comparison is statistically significant will inferences on the first secondary endpoint be evaluated. Only if the first secondary endpoint combination vs. pooled monotherapy comparison is found to be significant will inferences on the subsequent secondary endpoint be evaluated. This gate keeping approach will be implemented for all secondary endpoints in pre-defined order (defined below), and all tests will be performed at a 5% significance level. 1. Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP) 2. Percentage of subjects with satisfactory clinical response measured at week Change from baseline in 6MWD measured at week 24 79

193 GM2008/00365/04 CONFIDENTIAL 4. Change from baseline measured at week 24 in WHO Functional Class 5. Change from baseline measured at week 24 in Borg Dyspnea Index Additionally, for each endpoint, if the combination treatment arm is demonstrated as statistically significantly superior to the pooled monotherapy arm using the gate keeping approach described above, comparisons of combination treatment versus each individual monotherapy treatment will be performed at a 5% significance level. The primary, secondary and exploratory analyses may be repeated for the ITT population, PP population and non-mitt population. These analyses will be defined further in the RAP prior to unblinding Interim Analysis There are no planned interim analyses of efficacy data Key Elements of Analysis Plan The analyses planned in this protocol will be expanded in the RAP. This will be completed prior to unblinding of the study. Any deviations from the analyses described in this protocol will be described in the RAP and the final study report(s) as appropriate Efficacy Analyses Primary Analysis Time to clinical failure will be displayed as Kaplan-Meier event-free curves from randomisation up to the final assessment visit. Events which occur after the final assessment visit will not be used in the primary analysis. Differences between the curves will be tested for significance by the stratified log-rank statistic. The stratifying variables are the aetiology of PAH and WHO functional class. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using a Cox proportional hazards regression model. In case data for some subjects is not available following loss to follow up, their event times will be treated as censored for statistical analyses. The first event that occurs will be counted for subjects experiencing more than one component of the primary (composite) endpoint during the study. In such cases, all events will be counted separately for the analysis of the individual components. The primary analysis will include all adjudicated events in the mitt population (target of 105 events) up to the end of the Final Assessment Visit of the main study phase. Any events occurring after this timepoint will also be adjudicated and be reported as a sensitivity analysis or as a separate summary for that particular endpoint Secondary Analyses The analyses of the secondary efficacy endpoints will compare combination therapy to the pooled monotherapy arm consisting of ambrisentan or tadalafil. 80

194 GM2008/00365/04 CONFIDENTIAL The gatekeeper approach described in Section will be applied for the secondary endpoints to protect against type I error. Individual components of the primary endpoint, time to clinical worsening and death will be analysed as for the primary endpoint. Percentage of subjects with satisfactory clinical response will be summarised and response (yes, no) will be analysed as a binary endpoint using logistic regression, with PAH aetiology and WHO functional class as covariates. Descriptive statistics for each endpoint will be presented for the change from baseline and percentage change from baseline, where appropriate, by treatment group for each visit. Continuous data will be analysed using an analysis of covariance model with terms for treatment, aetiology and baseline measurement (continuous covariate). The point estimates and corresponding 95% confidence intervals for treatment differences will be calculated. Where the data is not normally distributed appropriate methods will be used and detailed in the Reporting and Analysis plan. The 6MWD test will be analysed using the Wilcoxon rank sum test stratified by PAH aetiology and WHO Functional Class. Change from baseline in WHO functional class will be tested using a Wilcoxon rank sum test, stratified by PAH aetiology and WHO functional class at baseline, between treatments. An analysis on change from baseline categorization (+2, +1, 0, -1, -2) and change from baseline categorization using the following three categories: Improved, No Change, and Deteriorated will also be performed. Change from baseline in BDI will be analysed as for 6MWD. Change from baseline at week 24 in N-terminal pro-b-type natriuretic peptide (NTproBNP) will be summarised and analysed as the geometric mean and the geometric mean ratio. A complete list of sensitivity analyses will be defined in the RAP Exploratory Analyses Peak-trough assessment of 6MWD To assess the treatment effect of ambrisentan (assay sensitivity), the change from baseline to Week 16 for 6MWD will be analyzed for the ambrisentan/tadalafil combination group versus the tadalafil monotherapy group. This will be analysed as for 6MWD as described in the secondary analyses section. The estimate of peak-trough 6MWD effect will be determined by the ratio of the placebocorrected mean change from baseline 6MWD measured at trough ambrisentan plasma concentrations following 16 weeks of treatment divided by the placebo-corrected mean change from baseline 6MWD measured at peak ambrisentan plasma concentrations following 16 weeks of treatment. 81

195 GM2008/00365/04 CONFIDENTIAL 6MWD trough = mean change from baseline 6MWD for the subjects in the ambrisentan/tadalafil combination group whose 6MWD was assessed at trough ambrisentan plasma concentrations 6MWD peak = mean change from baseline 6MWD for the subjects in the ambrisentan/tadalafil combination group whose 6MWD was assessed at peak ambrisentan plasma concentrations 6MWD placebo = mean change from baseline 6MWD for the tadalafil monotherapy group. Baseline 6MWD is defined as the average of the two 6MWDs performed at randomisation. In addition to reporting the above ratio, an interval estimate for the ratio using Hinkley s Method [Hinkley, 1969] will be provided. Full details of the analysis of all endpoints will be provided in the Reporting and Analysis Plan (RAP). The RAP will become part of the study documentation and will be finalised before the treatment code is unblinded and the database closed Safety Analyses Adverse Events AE summaries will be performed using the preferred terms and system organ class assigned by the Medical Dictionary for Regulatory Activities (MedDRA). The proportion of subjects reporting each type of AE will be tabulated for each treatment group according to the treatment received at the time of the AE. AEs may also be tabulated separately for those occurring prior to a clinical failure event by randomised treatment and for those AEs events occurring after a clinical failure event according to actual treatment received. The following summaries of AEs will be provided: All AEs Treatment-related AEs AEs leading to withdrawal Serious AEs Details for all AEs will be listed. AEs of special interest will be summarised. Of special interest are Liver Events, Anaemia, Hypersensitivity, Hypotension and Fluid Retention. 82

196 GM2008/00365/04 Exposure CONFIDENTIAL The number of days of exposure to study medication will be summarized by treatment group. Subject exposure may also be categorised and the number and percentage of subjects in each category presented for each treatment group. Vital Signs All vital sign data including systolic and diastolic blood pressure, heart rate, and BMI will be summarised by visit and treatment group. Summary statistics for values and change from baseline at measured time points will be provided by treatment group. For systolic and diastolic blood pressure and heart rate, the number and percentage of subjects with values of clinical concern will be summarized by treatment group. Clinical concern criteria will be specified in the RAP. Laboratory Evaluations Laboratory measurements will be summarized by visit and treatment group. Summary statistics for values and change from baseline will be provided by treatment group. In addition, the number and percentage of subjects with values of clinical concern will be summarized by treatment group. Clinical concern criteria will be specified in the RAP. Of special interest are some liver function laboratory parameters and their corresponding change from baseline Health Outcomes SF-36 and CAMPHOR, in selected countries, will be summarised by treatment group. Each section of the CAMPHOR will be summarised separately. Further details will be described in the Reporting and Analysis Plan. 9. STUDY CONDUCT CONSIDERATIONS 9.1. Posting of Information on Clinicaltrials.gov Study information from this protocol will be posted on clinicaltrials.gov before enrolment of subjects begins Regulatory and Ethical Considerations, Including the Informed Consent Process Prior to initiation of a study site, the Sponsors (e.g., Gilead in the USA and GSK in the Rest of World) will obtain approval from the appropriate Regulatory Agency to conduct the study in accordance with applicable country-specific regulatory requirements. The study will be conducted in accordance with all applicable regulatory requirements, 83

197 GM2008/00365/04 CONFIDENTIAL The study will be conducted in accordance with Good Clinical Practice (GCP), all applicable subject privacy requirements, and the ethical principles that are outlined in the Declaration of Helsinki 2000, as amended in 2002, 2004 and 2008 including, but not limited to: Institutional Review Board (IRB)/Independent Ethics Committee (IEC) review and approval of study protocol and any subsequent amendments. Subject informed consent. Investigator reporting requirements. The Sponsor will provide full details of the above procedures, either verbally, in writing, or both. Written informed consent must be obtained from each subject prior to participation in the study Quality Control (Study Monitoring) In accordance with applicable regulations, GCP, and the Sponsor procedures, monitors will contact the site prior to the start of the study to review with the site staff the protocol, study requirements, and their responsibilities to satisfy regulatory, ethical, and Sponsor requirements. When reviewing data collection procedures, the discussion will include identification, agreement and documentation of data items for which the CRF will serve as the source document. The sponsor will monitor the study to ensure that the: Data are authentic, accurate, and complete. Safety and rights of subjects are being protected. Study is conducted in accordance with the currently approved protocol and any other study agreements, GCP, and all applicable regulatory requirements. The investigator and the head of the medical institution (where applicable) agrees to allow the monitor direct access to all relevant documents Quality Assurance To ensure compliance with GCP and all applicable regulatory requirements, the Sponsor may conduct a quality assurance audit of the site records, and the regulatory agencies may conduct a regulatory inspection at any time during or after completion of the study. In the event of an audit or inspection, the investigator (and institution) must agree to grant the auditor(s) and inspector(s) direct access to all relevant documents and to allocate their time and the time of their staff to discuss any findings/relevant issues Study and Site Closure The study will be considered over after the database lock 2. 84

198 GM2008/00365/04 CONFIDENTIAL Upon completion or termination of the study, the Sponsor monitor will conduct site closure activities with the investigator or site staff (as appropriate), in accordance with applicable regulations, GCP, and the sponsor or protocol specific Standard Operating Procedures. The Sponsor reserves the right to temporarily suspend or terminate the study at any time for reasons including (but not limited to) safety issues, ethical issues, or severe noncompliance. If the Sponsor determines that such action is required, the Sponsor will discuss the reasons for taking such action with the investigator or head of the medical institution (where applicable). When feasible, the Sponsor will provide advance notice to the investigator or head of the medical institution of the impending action. If a study is suspended or terminated for safety reasons, the Sponsor will promptly inform all investigators, heads of the medical institutions (where applicable),and/or institutions conducting the study. The Sponsor will also promptly inform the relevant regulatory authorities of the suspension/termination along with the reasons for such action. Where required by applicable regulations, the investigator or head of the medical institution must inform the IRB/IEC promptly and provide the reason(s) for the suspension/termination Records Retention Following closure of the study, the investigator or head of the medical institution (where applicable) must maintain all site study records (except for those required by local regulations to be maintained elsewhere) in a safe and secure location. The records must be easily accessible when needed (e.g., for a Sponsor audit or regulatory inspection) and must be available for review in conjunction with assessment of the facility, supporting systems, and relevant site staff. Where permitted by local laws/regulations or institutional policy, some or all of the records may be maintained in a format other than hard copy (e.g., microfiche, scanned, electronic); however, caution must be exercised before such action is taken. The investigator must ensure that all reproductions are legible and are a true and accurate copy of the original. In addition, they must meet accessibility and retrieval standards, including regeneration of a hard copy, if required. The investigator must also ensure that an acceptable back-up of the reproductions exists and that there is an acceptable quality control procedure in place for creating the reproductions. The Sponsor will inform the investigator of the time period for retaining the site records in order to comply with all applicable regulatory requirements. The minimum retention time will meet the strictest standard applicable to a particular site, as dictated by local laws/regulations, the Sponsor or study-specific standard operating procedures, and/or institutional requirements. The investigator must notify the Sponsor of any changes in the archival arrangements, including, but not limited to archival of records at an off-site facility or transfer of ownership of the records in the event that the investigator is no longer associated with the site. 85

199 GM2008/00365/04 CONFIDENTIAL 9.7. Provision of Study Results and Information to Investigators Where required by applicable regulatory requirements, an investigator signatory will be identified for the approval of the clinical study report. The investigator will be provided reasonable access to statistical tables, figures, and relevant reports and will have the opportunity to review the complete study results at a Sponsor site or other mutuallyagreeable location. The Sponsor will also provide the investigator with the full summary of the study results. The investigator is encouraged to share the summary results with the study subjects, as appropriate. The Sponsor will provide the investigator with the randomisation codes for their site only after completion of the full statistical analysis. The results summary will be posted to the Clinical Study Register at the time of the first regulatory approval or within 12 months of any decision to terminate development. In addition, a manuscript will be submitted to a peer-reviewed journal for publication within 12 months of the first approval or within 12 months of any decision to terminate development. When manuscript publication in a peer-reviewed journal is not feasible, further study information will be posted to the Sponsor Clinical Study Register to supplement the results summary Independent Data Monitoring Committee (IDMC) An IDMC will be utilized in this study to ensure external objective medical and/or statistical review of safety and/or efficacy issues in order to protect the ethical and safety interests of subjects and to protect the scientific validity of the study. The schedule of the analysis plan for IDMC review is described in the charter, which is available upon request. 86

200 GM2008/00365/ REFERENCES CONFIDENTIAL Adcirca (tadalafil) Summary of Product Characteristics May 2010 Adcirca (tadalafil) US Prescribing Information May 2009 Ambrisentan EU Summary of Product Characteristics March 2010 Ambrisentan US Prescribing Information August 2009 D'Alonzo GE, Barst RJ, Ayres SM, Bergofsky EH, Brundage BH, Detre KM et al. Survival in patients with primary pulmonary hypertension. Results from a national prospective registry. Ann Intern Med 1991; 115(5): Davenport AP, O'Reilly G, Kuc RE. Endothelin ETA and ETB mrna and receptors expressed by smooth muscle in the human vasculature: majority of the ETA sub-type. Br J Pharmacol 1995; 114(6): Forgue ST, Patterson BE, Bedding AW, Payne CD, Phillips DL, Wrishko RE, Mitchell MI. Br J Clin Pharmacol. Tadalafil pharmacokinetics in healthy subjects. 2006; 61: Fukuroda T, Fujikawa T, Ozaki S, Ishikawa K, Yano M, Nishikibe M. Clearance of circulating endothelin-1 by ETB receptors in rats. Biochem Biophys Res Commun 1994; 199(3): Galie N, Badesch D, Oudiz R, Simonneau G, McGoon MD, Keogh AM et al. Ambrisentan therapy for pulmonary arterial hypertension. J Am Coll Cardiol 2005; 46(3): Galiè N, Brundage BH, Ghofrani HA, Oudiz RJ, Simonneau G, Safdar Z, et al. Pulmonary Arterial Hypertension and Response to Tadalafil (PHIRST) Study Group. Tadalafil therapy for pulmonary arterial hypertension. Circulation 2009a; 119(22): Galiè N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, Barbera JA, et al. Task Force for Diagnosis and Treatment of Pulmonary Hypertension of European Society of Cardiology (ESC); European Respiratory Society (ERS); International Society of Heart and Lung Transplantation (ISHLT). Guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Respir J 2009b; 34(6): Galiè N, Olschewski H, Oudiz RJ, Torres F, Frost A, Ghofrani HA, et al. Ambrisentan in Pulmonary Arterial Hypertension, Randomized, Double-Blind, Placebo-Controlled, Multicenter, Efficacy Studies (ARIES) Group. Ambrisentan for the treatment of pulmonary arterial hypertension: results of the ambrisentan in pulmonary arterial hypertension, randomized, double-blind, placebo-controlled, multicenter, efficacy (ARIES) study 1 and 2. Circulation 2008; 117:

201 GM2008/00365/04 CONFIDENTIAL Ghofrani HA, Wiedemann R, Rose F, Olschewski H, Schermuly RT, Weissmann N et al. Combination therapy with oral sildenafil and inhaled iloprost for severe pulmonary hypertension. Ann Intern Med 2002; 136(7): Hinkley, D. V. (December 1969). "On the Ratio of Two Correlated Normal Random Variables". Biometrika 56 (3): Hirata Y, Emori T, Eguchi S, Kanno K, Imai T, Ohta K et al. Endothelin receptor subtype B mediates synthesis of nitric oxide by cultured bovine endothelial cells. J Clin Invest 1993; 91(4): Hoeper MM, Markevych I, Spiekerkoetter E, Welte T, Niedermeyer J. Goal-oriented treatment and combination therapy for pulmonary arterial hypertension. Eur Respir J 2005; 26(5): Hoeper MM, Taha N, Bekjarova A, Gatzke R, Spiekerkoetter E. Bosentan treatment in patients with primary pulmonary hypertension receiving nonparenteral prostanoids. Eur Respir J 2003; 22(2): Humbert M, Barst RJ, Robbins IM, Channick RN, Galie N, Boonstra A et al. Combination of bosentan with epoprostenol in pulmonary arterial hypertension: BREATHE-2. Eur Respir J 2004a; 24(3): Humbert M, Morrell NW, Archer SL, Stenmark KR, MacLean MR, Lang IM et al. Cellular and molecular pathobiology of pulmonary arterial hypertension. J Am Coll Cardiol 2004b; 43(12 Suppl S):13S-24S. Humbert M, Sitbon O, Simmonneau G. Treatment of pulmonary arterial hypertension. N Engl J Med 2004c; 351(14): Kirchengast M, Munter K. Endothelin-1 and endothelin receptor antagonists in cardiovascular remodeling. Proc Soc Exp Biol Med 1999; 221(4): Leuchte HH, Holzapfel M, Baumgartner RA, Ding I, Neurohr C, Vogeser M et al. Clinical significance of brain natriuretic peptide in primary pulmonary hypertension. J Am Coll Cardiol 2004; 43(5): Leuchte HH, Holzapfel M, Baumgartner RA, Neurohr C, Vogeser M, Behr J. Characterization of brain natriuretic peptide in long-term follow-up of pulmonary arterial hypertension. Chest 2005; 128(4): McKenna SP et al. The Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR): A measure of health-related quality of life and quality of life for patients with pulmonary hypertension. Quality of Life Research 2006; 15: Nagaya N, Nishikimi T, Uematsu M, Satoh T, Kyotani S, Sakamaki F et al. Plasma brain natriuretic peptide as a prognostic indicator in patients with primary pulmonary hypertension. Circulation 2000; 102(8):

202 GM2008/00365/04 CONFIDENTIAL Oudiz RJ, Galiè N, Olschewski H, Torres F, Frost A, Ghofrani HA, et al. ARIES Study Group. Long-term ambrisentan therapy for the treatment of pulmonary arterial hypertension. JACC 2009; 54: Ozaki S, Ohwaki K, Ihara M, Fukuroda T, Ishikawa K, Yano M. ETB-mediated regulation of extracellular levels of endothelin-1 in cultured human endothelial cells. Biochem Biophys Res Commun 1995; 209(2): Park MH, Scott RL, Uber PA, Ventura HO, Mehra MR. Usefulness of B-type natriuretic peptide as a predictor of treatment outcome in pulmonary arterial hypertension. Congest Heart Fail 2004; 10(5): Rich S, Kaufmann E, Levy PS. The effect of high doses of calcium-channel blockers on survival in primary pulmonary hypertension. N Engl J Med 1992; 327(2): Rubanyi GM, Polokoff MA. Endothelins: molecular biology, biochemistry, pharmacology, physiology, and pathophysiology. Pharmacol Rev 1994; 46(3): Rubin LJ. Primary pulmonary hypertension. Chest 1993; 104(1): Simonneau G, Rubin LJ, Galiè N, Barst RJ, Fleming TR, Frost AE et al. Addition of sildenafil to long-term intravenous epoprostenol therapy in patients with pulmonary arterial hypertension: a randomized trial. Ann Intern Med 2008; 149: Spence R, Mandagere A, Harrison B, Dufton C, Boinpally R. No clinically relevant pharmacokinetic and safety interactions of ambrisentan in combination with tadalafil in healthy volunteers. J Pharm Sci 2009; 98(12): Verhaar MC, Strachan FE, Newby DE, Cruden NL, Koomans HA, Rabelink TJ et al. Endothelin-A receptor antagonist-mediated vasodilatation is attenuated by inhibition of nitric oxide synthesis and by endothelin-b receptor blockade. Circulation 1998; 97(8): Ware, JE; Kosinski, M; Gandek, B. SF-36 Health Survey: Manual and Interpretation Guide, Lincoln, RI: QualityMetric Inc,

203 GM2008/00365/ APPENDICES CONFIDENTIAL Appendix 1: List of Highly Effective Methods for Avoidance of Pregnancy in Women of Childbearing Potential The following is a list of the highly effective methods for avoiding pregnancy (i.e., have a failure rate of less than 1% per year). Abstinence [Hatcher, 2004] Combination Oral Contraceptive [Hatcher, 2004] Injectable progestogen [Hatcher, 2004] Implants of levonorgestrel [Hatcher, 2004] Estrogenic vaginal ring [Hatcher, 2004] Percutaneous contraceptive patches [Hatcher, 2004] Intrauterine device (IUD) or intrauterine system (IUS) that meets the SOP effectiveness criteria as stated in the product label [Hatcher, 2004] Male partner sterilization (vasectomy with documentation of azoospermia) prior to the female subject's entry into the study, and this male is the sole partner for that subject [Hatcher, 2004]. For this definition, documented refers to the outcome of the investigator's/designee s medical examination of the subject or review of the subject's medical history for study eligibility, as obtained via a verbal interview with the subject or from the subject s medical records. Double barrier method: condom and an occlusive cap (diaphragm or cervical/vault caps) with a vaginal spermicidal agent (foam/gel/film/cream/suppository) [Trussell, 2003]. Nonoxynol-9 is the critical component in most spermicides, and is regarded as an acceptable spermicidal agent. Concern has been raised that nonoxynol-9 damages the epithelial lining of the vagina, and exposure may facilitate transmission of viruses, particularly human immunodeficiency virus (HIV). The World Health Organization (WHO) conducted a technical consultation in October 2001 and concluded that the increased risk for such transmission was low to minimal [Trussell, 2003]. References Hatcher RA, Trussell J, Stewart F, Nelson AL, Cates W, Guest F, Kowal DD, editors. Contraceptive Technology. New York: Ardent Media, 2004: 226. Table 9-2, % of Women Experiencing an Unintended Pregnancy During the First Year of Typical Use and the First Year of Perfect Use of Contraception and the Percentage Continuing Use at the End of the First Year. United States, column entitled, % of Women Experiencing an Unintended Pregnancy Within the First Year of Use. Perfect Use. Trussell J. Personal communication WHO/CONRAD Technical consultation on nonoxynol-9. WHO, Geneva, 9-10 October Summary Report. World Health Organization,

204 GM2008/00365/04 CONFIDENTIAL Appendix 2: 6 Minute Walk Distance Test The 6-minute walk test will be conducted according to the American Thoracic Society guidelines [in accordance with local standard operating procedures]. Please refer to the following guidelines for conducting the 6-minute walk test. The Test Location Select a quiet, enclosed corridor, free of distractions and drafts. Depending on your site s standard operating procedures, please mark out a 6MWT course. See Study Procedures Manual for further guidance. The start and end of the course must be visibly marked by placing tape on the floor, using chairs etc. The course should be sub-divided into 3m or 10ft sections using a method unnoticeable to the subject. The Day Before the Test Tell the subject the time their test will take place. Give the subject the following instructions: - Take a light meal 3 to 4 hours before the test. - Take nothing by mouth (po) except clear liquids for 1 hour before the test. - Do not take heart medication within 2 hours of the test - Do not smoke for at least 2 hours before the test. Immediately Before the Test The test should be performed on the subject s "usual" oxygen. If the subject usually breathes room air, they should perform the test on room air. If the subject normally requires supplemental oxygen with ambulation, they should perform the test with supplemental oxygen. Subjects who require supplemental oxygen should receive the same oxygen flow rate as baseline during all subsequent 6-minute walk tests. Subjects who add supplemental oxygen after initiation of therapy should walk without oxygen whenever possible so data is comparable to baseline conditions. If the oxygen flow rate must be increased during subsequent visits due to worsening gas exchange, the 6-minute walk test may be conducted at the increased oxygen flow rate and this should be noted in the CRF. If the subject reduces or discontinues supplemental oxygen use during the study, they should still perform the 6-minute walk test with the same flow rate of oxygen that was used during previous tests. 91

205 GM2008/00365/04 CONFIDENTIAL Ask the subject to sit quietly for 10 minutes. Read the following instructions verbatim to the subject: The object of this test is to walk as far as possible for 6 minutes. You will walk back and forth in this hallway. Six minutes is a long time to walk, so you will be exerting yourself. You will probably get out of breath or become exhausted. You are permitted to slow down, to stop, and to rest as necessary. You may lean against the wall while resting, but resume walking as soon as you are able. Are you ready to do that? Repeat the entire set of instructions if the subject does not seem to understand. Repeat the sentence: Remember that the object is to walk AS FAR AS POSSIBLE for 6 minutes, but don t run or jog. Instruct the subject to start by saying: Start now, or whenever you are ready. During the Test: Do not influence the walking pace of the subject. Use an even tone of voice when using the standard phrases of encouragement. Encourage the subject every 60 seconds (1 minute) during the test but do not convey information about their performance. Examples of appropriate encouragement are, You are doing well. or Keep up the good work. Only speak to the subject at the 60-second points. If the subject is slowing down or expresses that he/she wants to stop say: You can lean against the wall if you would like; then continue walking whenever you feel able. Note the number of laps of the corridor walked by the subject but do not convey this information to the subject. At 6 minutes, tell the subject to stop and not to move until the distance walked has been measured. After the Test If the subject walked for less than 6 minutes (i.e., stops prior to 6 minutes and DOES NOT start again), record the time walked as well as the distance. When the distance has been measured tell the subject to sit down and observe him/her for at least 10 minutes. 92

206 GM2008/00365/04 CONFIDENTIAL The distance covered during any walking test must not be revealed to the subject at any time during the study. Parameters to be recorded: Distance walked, to the nearest meter or foot Time walked, to the nearest second (if subject walked for less than 6 minutes) Comments if subject was unable to walk for 6 minutes The BDI score immediately following completion of the 6-minute walk test (Appendix 3). 93

207 GM2008/00365/04 CONFIDENTIAL Appendix 3: Borg Dyspnea Index Assess the BDI score immediately following completion of the 6-minute walk test (See Appendix 2) Borg CR10 scale G. Borg, 1998, 2007 English 94

208 GM2008/00365/04 CONFIDENTIAL Instruction. Use this rating scale to report how strong your perception is. It can be exertion, pain or something else. First look at the verbal expressions. Start with them and then the numbers. Of these ten (10) or Extremely strong, Maximal is a very important intensity level. This is the most intense perception or feeling you have ever had. If your experience or feeling is Very weak, you should say 1, if it is Moderate, say 3. Note that Moderate is 3 and thus weaker than Medium, Mean or Middle. If the experience is Strong or Heavy (it feels Difficult ) say 5. Note that Strong is about half of Maximal. If your feeling is Very strong, choose a number from 6 to 8. If your perception or feeling is stronger than 10, - Extremely strong, Maximal you can use a larger number, e.g. 12 or still higher (that s why Absolute maximum is marked with a dot ). It's very important that you report what you actually experience or feel, not what you think you should report. Be as spontaneous and honest as possible and try to avoid underor overestimating. Look at the verbal descriptors and then choose a number. When rating exertion give a number that corresponds to how hard and strenuous you perceive the work to be. The perception of exertion is mainly felt as strain and fatigue in your muscles and as breathlessness or any aches. 0 Nothing at all, means that you don t feel any exertion whatsoever, no muscle fatigue, no breathlessness or difficulties breathing. 1 Very weak means a very light exertion. As taking a shorter walk at your own pace. 3 Moderate is somewhat but not especially hard. It feels good and not difficult to go on. 5 Strong. The work is hard and tiring, but continuing isn't terribly difficult. The effort and exertion is about half as intense as Maximal. 7 Very strong is quite strenuous. You can still go on, but you really have to push yourself and you are very tired. 10 Extremely strong Maximal is an extremely strenuous level. For most people this is the most strenuous exertion they have ever experienced previously in their lives. Is Absolute maximum for example 12 or even more. Any questions? Borg CR10 scale G. Borg, 1998, 2007 English 95

209 GM2008/00365/04 CONFIDENTIAL Appendix 4: Pharmacogenetic Research Pharmacogenetics Background Pharmacogenetics (PGx) is the study of variability in drug response due to hereditary factors in different populations. There is increasing evidence that an individual's genetic composition (i.e., genotype) may impact the pharmacokinetics (absorption, distribution, metabolism, elimination), pharmacodynamics (relationship between concentrations and pharmacologic effects or the time course of pharmacologic effects) and/or clinical outcome (in terms of efficacy and/or safety and tolerability). Some reported examples of PGx analysis include: Drug Disease Gene Outcome Abacavir HIV [Hetherington, 2002; Mallal, 2002] HLA B*5701 Individuals with HLA-B*5701 variant may be at increased risk for experiencing hypersensitivity to abacavir. Clinical assays are available for HLA-B*5701 but none has been validated. HLA-B*5701 screening would supplement but never replace abacavir clinical risk management strategies aimed at minimising rare but serious outcomes associated with Warfarin Irinotecan Cardiovascular [Neergard, 2006; Wilke, 2005] Cancer [FDA News Release, 2005] CYP2C9 UGT1A1 abacavir hypersensitivity. Serious adverse events (SAEs) experienced by some patients on warfarin may be explained by variations in the CYP2C9 gene that influences the degree of anticoagulation achieved. Variations in the UGT1A1 gene can influence a patient s ability to break down irinotecan, which can lead to increased blood levels of the drug and a higher risk of side effects. A dose of irinotecan that is safe for one patient with a particular UGT1A1 gene variation might be too high for another patient without this variation, raising the risk of certain sideeffects. A genetic blood test (Invader UGT1A1 molecular assay) is available that can detect variations in the gene. A key component to successful PGx research is the collection of samples during the conduct of clinical studies. Collection of whole blood samples, even when no a priori hypothesis has been identified, may enable PGx analysis to be conducted if at any time it appears that there is a potential unexpected or unexplained variation in handling or response to ambrisentan or tadalafil. 96

210 GM2008/00365/04 CONFIDENTIAL Pharmacogenetic Research Objectives The objective of the PGx research (if there is a potential unexpected or unexplained variation) is to investigate a possible genetic relationship to handling or response to ambrisentan or tadalafil. If at any time it appears there is potential variability in response in this clinical study or in a series of clinical studies with ambrisentan or tadalafil that may be attributable to genetic variations of subjects, the following objectives may be investigated: Relationship between genetic variants and the pharmacokinetics and/or pharmacodynamics of ambrisentan/tadalafil Relationship between genetic variants and safety and/or tolerability of ambrisentan/tadalafil Relationship between genetic variants and efficacy of ambrisentan/tadalafil Study Population Any subject who has given informed consent to participate in the clinical study, has met all the entry criteria for the clinical study, and receives investigational product may take part in the PGx research. Any subject who has received an allogeneic bone marrow transplant must be excluded from the PGx research. Subject participation in the PGx research is voluntary and refusal to participate will not indicate withdrawal from the clinical study. Refusal to participate will involve no penalty or loss of benefits to which the subject would otherwise be entitled. Study Assessments and Procedures In addition to any blood samples taken for the clinical study, a whole blood sample (~10ml) will be collected for the PGx research using a tube containing EDTA. It is recommended that the blood sample be taken at the first opportunity after a subject has been randomised and provided informed consent for PGx research, but may be taken at any time while the subject is participating in the clinical study. The PGx sample is labelled (or coded ) with a study specific number that can be traced or linked back to the subject by the investigator or site staff. Coded samples do not carry personal identifiers (such as name or social security number). The blood sample is taken on a single occasion unless a duplicate sample is required due to inability to utilize the original sample. The DNA extracted from the blood sample may be subjected to sample quality control analysis. This analysis will involve the genotyping of several genetic markers to confirm the integrity of individual samples. If inconsistencies are noted in the analysis, then those samples may be destroyed. The need to conduct PGx analysis may be identified after a study (or a set of studies) of ambrisentan or tadalafil has been completed and the study data reviewed. 97

211 GM2008/00365/04 CONFIDENTIAL In some cases, the samples may not be studied. e.g., no questions are raised about how people respond to ambrisentan or tadalafil. Samples will be stored securely and may be kept for up to 15 years after the last subject completes the study or the Sponsor may destroy the samples sooner. The Sponsor or those working with the Sponsor (for example, other researchers) will use samples collected from the study for the purpose stated in this protocol and in the informed consent form. Subjects can request their sample to be destroyed at any time. Subject Withdrawal from Study If a subject who has consented to participate in PGx research and has a sample taken for PGx research withdraws from the clinical study for any reason other than lost to followup, the subject will be given the following options: The sample is retained for PGx research Any PGx sample is destroyed. If a subject withdraws consent from the PGx research or requests sample destruction for any reason, the investigator must complete the appropriate documentation to request sample destruction within the timeframe specified by the Sponsor and maintain the documentation in the site study records. If the sample has already been processed, it will be destroyed after all steps are complete. The Sponsor will ensure that any data related to the sample will not be analysed. The sample will be destroyed after processing is complete. Screen and Baseline Failures If a blood sample for PGx research has been collected and it is determined that the subject does not meet the entry criteria for participation in the clinical study, then the investigator must complete the appropriate documentation to request sample destruction within 5 days. The sample will be destroyed and documentation sent to the site within 30 working days of receipt of the request for destruction. All documents pertaining to sample destruction must be maintained in the site study records. Pharmacogenetics Analyses 1. Specific sections of DNA may be selected from areas of the genome (e.g., candidate genes) known to encode the drug target, drug metabolizing enzymes, areas associated with mechanisms underlying adverse events, and those linked to study disease and, thus, linked to drug response. The candidate genes that may be investigated in this study include the following: the Sponsor Absorption, Distribution, Metabolism and Excretion genes. These play a central role in drug pharmacokinetics and pharmacodynamics. In addition, continuing research may identify other enzymes, transporters, proteins or receptors that may be involved in response to ambrisentan or tadalafil. The genes that may code for these proteins may also be studied. 98

212 GM2008/00365/04 CONFIDENTIAL 2. By evaluating large numbers of polymorphic markers (e.g., single nucleotide polymorphisms or SNPs) throughout the genome, sets of markers may be identified that correspond to differential drug response. The results of PGx investigations will be reported either as part of the main clinical study report or as a separate report. All endpoints of interest from all comparisons will be descriptively and/or graphically summarised as appropriate to the data. In all cases, appropriate statistical methods will be used to analyse the genetic markers in the context of other clinical data. Statistical methods may include, but are not limited to Hardy- Weinberg Equilibrium testing, Comparison of Demographic and Baseline Characteristics by Genotype, Evaluation of Genotypic Effects, Evaluation of Treatment by Genotype and Gene-Gene Interaction, Linkage Disequilibrium, Multiple Comparison and Multiplicity and/or Power and Sample Size Considerations. Detailed description of the analyses to be conducted will be documented in the Pharmacogenetics Reporting and Analysis Plan. 3. Genome-wide scans involving a large number of polymorphic markers (e.g., single nucleotide polymorphisms) located throughout the genome. This approach is often employed when potential genetic effects are not well understood. Informed Consent Subjects who do not wish to participate in the PGx research may still participate in the clinical study. PGx informed consent must be obtained prior to any blood being taken for PGx research. Provision of Study Results and Confidentiality of Subject s PGx Data Sponsor may summarize the cumulative PGx research results in the clinical study report. In general, the Sponsor does not inform the investigator, subject, or anyone else (e.g., family members, study investigators, primary care physicians, insurers, or employers) of the PGx research results that are not known to be relevant to the subject s medical care at the time of the study, because the information generated from PGx studies is preliminary in nature, and the significance and scientific validity of the results are undetermined at such an early stage of research, under any circumstances unless required by law. References Hetherington S, Hughes AR, Mosteller M, Shortino D, Baker KL, Spreen W, Lai E, Davies K, Handley A, Dow DJ, Fling ME, Stocum M, Bowman C, Thurmond LM, Roses AD. Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet. 2002; 359: Mallal S, Nolan D, Witt C, Masel G, Martin AM, Moore C, Sayer D, Castley A, Mamotte C, Maxwell D, James I. Association between presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet. 2002; 359: Neergard. Reducing the risk of blood thinners. Associated press, September

213 GM2008/00365/04 CONFIDENTIAL U.S. Food and Drug Administration, FDA Clears Genetic Test That Advances Personalized Medicine Test Helps Determine Safety of Drug Therapy 22 August 2005, Wilke RA, Musana AK, Weber WW. Cytochrome P450 gene-based drug prescribing, and factors impacting translation into routine clinical practice. Personalized Med 2005; 2:

214 GM2008/00365/04 CONFIDENTIAL Appendix 5: CAMPHOR Quality of Life Questionnaire 101

215 GM2008/00365/04 CONFIDENTIAL 102

216 GM2008/00365/04 CONFIDENTIAL 103

217 GM2008/00365/04 CONFIDENTIAL 104

218 GM2008/00365/04 CONFIDENTIAL 105

219 GM2008/00365/04 CONFIDENTIAL 106

220 GM2008/00365/04 CONFIDENTIAL 107

221 GM2008/00365/04 CONFIDENTIAL 108

222 GM2008/00365/04 CONFIDENTIAL Appendix 6: SF-36 Quality of Life Questionnaire 109

223 GM2008/00365/04 CONFIDENTIAL 110

224 GM2008/00365/04 CONFIDENTIAL 111

225 GM2008/00365/04 CONFIDENTIAL 112

226 GM2008/00365/04 CONFIDENTIAL 113

227 GM2008/00365/04 CONFIDENTIAL 114

228 GM2008/00365/04 CONFIDENTIAL Appendix 7: Country Specific Requirements 115

229 GM2008/00365/04 CONFIDENTIAL Appendix 8: Protocol Amendment 01 Changes AMENDMENT 1 Where the Amendment Applies This is a global amendment Deletions are noted using strikethrough text Insertions are noted in Bold Face Summary of Amendment Changes with Rationale The principle aim of this amendment is to clarify the primary endpoint, specifically the 4th component Unsatisfactory long-term clinical response. The reference to functional class IV has been removed to clarify that subjects should not, and do not, need to be in WHO FC IV for 6 months to qualify as an event. The following clarifications are also included: Clarification that the Final Assessment Visits are to be scheduled when it is projected that 82 primary events have occurred. This is to address an inconsistency in the original protocol. Clarify that 2 reliable methods of contraception are only required if the subject is sexually active and that subjects who have had a Copper T 380A IUD or LNg 20 IUD inserted are not required to use additional methods of contraception. Highlight in the inclusion criteria that the screening and baseline 6MWD Test values should not vary by more than 10%, as described later in Section Change status of cyclosporine A from a permitted medication to a prohibited medication. Add criterion for cyclosporine A under the Other Therapies section of the Exclusion Criteria. Provide additional clarification to the blinded combination therapy section including to outline all possible dose combinations including provision for initial dose of ABS 10mg and TAD 20mg at stage 1 blinded combination therapy for subjects with mild to moderate renal impairment. This is to ensure there is a match for all the protocol defined doses prior to clinical failure. This was an oversight in the original protocol. Clarification of when and how investigators may up-titrate blinded combination therapy following a TtCF event to simplify the process. Additionally clarify when, following TtCF, an investigator may request a subject re-uptitrate their dose of IP. This is to simplify the study dosing. Addition of text specifying that SMS/text messaging may be used (if chosen by the centre/subject). This is a service designed to assist the subject to attend all scheduled visits (both clinic visits and safety visits). It was omitted from the original version of the protocol. 116

230 GM2008/00365/04 CONFIDENTIAL Removal of a redundant cross reference to a separate footnote in footnote 2 of the time and events schedule. Correct a protocol inconsistency stating that randomization for the peak-trough assessment will occur at the week 0 visit it will in fact occur at the week 16 visit as described in Section 5.2. The definition of liver events that potentially meet the criteria for Hy s law cases was amended from ALT value from 3xULN to >3xULN and bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. Further, text related to fractionation of bilirubin as part of Hy s law cases was removed to simplify the definition of these events. Additionally, text was added to Section to clarify that symptomatic elevations >3xULN are considered protocol defined SAEs. Amendments were made to Section to clarify reporting timelines for liver events. Clarification that subjects are not required to withhold investigational product on the day of the clinic visit (exception week 16). This is to address an inconsistency in the original protocol. In addition, removal of reference to withholding nitrates 2 hours prior to the 6MWD test as nitrate use is prohibited during the study. List of Specific Changes Protocol Summary; Study Design Section; paragraphs 5 and 15 PREVIOUS TEXT Subjects must have a 6-minute walk distance (6MWD) of at least 125m and no more than 500m at the screening visit to be eligible for randomisation. All subjects are eligible to receive a minimum of 24 weeks of therapy. At the time that 82 adjudicated primary endpoint events have occurred, i.e. 82 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. REVISED TEXT Subjects must have a 6-minute walk distance (6MWD) of at least 125m and no more than 500m at the screening visit, and the screening and baseline 6MWD tests must not vary by greater than 10%, to be eligible for randomisation. All subjects are eligible to receive a minimum of 24 weeks of therapy. At the time that 82 adjudicated primary endpoint events are projected to have occurred, i.e. 82 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. 117

231 GM2008/00365/04 RATIONALE CONFIDENTIAL This is a clarification to reflect the wording in the main body of the protocol. Section 3.1 Study Design; paragraph 8 and Section Primary End point; Time to Clinical Failure; 4 th bullet point unsatisfactory long term clinical response PREVIOUS TEXT Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III or IV symptoms for 6 months REVISED TEXT Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III or IV symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) RATIONALE The essence of the 4 th component of the TtCF-endpoint ( Unsatisfactory long-term clinical response ) and all the clinical situations it aims to capture in terms of the primary endpoint remains completely unchanged. The reason for the clarification is to avoid any misinterpretation of the last sub-bullet that this would endorse or only accept an endpoint being met when a patient remains in WHO-FC class IV for at least 6 months. By dropping or IV this misinterpretation is avoided. Any patients deteriorating into WHO-FC class IV at any point in time during the trial do have all options for immediate intervention and likely will meet other components of the composite primary endpoint anyway. In addition, the new text under the 3 rd sub-bullet in brackets clarifies that any patient remaining in WHO-FC class-iii for at least 6 months or a patient starting in WHO-FC class-iii, improving to class-ii initially but deteriorating back to class-iii at the 6 month visit, would both be considered to meet the criteria for clinical failure, assuming they also meet the other criteria of the unsatisfactory response definition of the primary EP. 118

232 GM2008/00365/04 CONFIDENTIAL Section 4.2, Inclusion Criteria; Criteria 8 and 11 PREVIOUS TEXT 8. Subject must walk a distance of 125m and 500m at the screening visit. 11. Female subject of childbearing potential must agree to use 2 reliable methods of contraception from the Screening Visit until study completion and for at least 30 days following the last dose of Investigational Product (reliable methods of contraception are described in Appendix 1) AMENDED TEXT 8. Subject must walk a distance of 125m and 500m at the screening visit. In addition the screening and baseline 6MWD tests must not vary by greater than 10% (see Section for further details) 11. Female subject of childbearing potential, if sexually active, must agree to use 2 reliable methods of contraception from the screening visit until study completion and for at least 30 days following the last dose of investigational product (reliable methods of contraception are described in Appendix 1). Subjects who have had a Copper T 380A IUD or LNg 20 IUD inserted are not required to use additional methods of contraception. RATIONALE To highlight that the screening and baseline 6MWD Test values must not vary by more than 10%, as outlined in Section To clarify that 2 reliable methods of contraception are only required if the subject is sexually active. In addition to simplify the criteria for those subjects with a Copper T 380A or LNg 20 IUD inserted, allowing this to be the only method of contraception required as it is considered sufficiently effective on its own. Section 4.3, Exclusion Criteria; New Criterion (Number 8) ADDITIONAL TEXT 8. Subject is receiving treatment with cyclosporine A. Note: this will result in renumbering of subsequent items in the exclusion criteria section. RATIONALE In a controlled clinical trial the Sponsor does not want to be allowing concomitant use of medications that are known to increase the mean exposure of AMB by two fold. This aligns with the Sponsor s other ambrisentan PAH protocols, where subjects receiving cyclosporine A are excluded. 119

233 GM2008/00365/04 CONFIDENTIAL Section Treatment following clinical failure event: First bullet PREVIOUS TEXT Following the declaration of a clinical failure event by an Investigator, based on the criteria given, the investigator may do either of the following: Request the subject receive blinded combination therapy. It should be noted that only those subjects on monotherapy will receive a change in their treatment regimen (or those in whom events occur in the first 8 weeks). Subjects in the combination arm will remain on the same combination therapy. Therefore if this is chosen it is important to re-assess the subjects status. REVISED TEXT Following the declaration of a clinical failure event by an Investigator, based on the criteria given, the investigator may do either of the following: RATIONALE Request the subject receive blinded combination therapy. It should be noted that only those subjects on monotherapy will receive an additional change in their treatment regimen (or those in whom events occur in the first 8 weeks). Subjects in the combination arm will remain on the same combination therapy, although they may up-titrate their doses of IP to maximum permitted doses if not previously receiving these doses (10mg AMB and/or 40mg TAD), see Section for further details. Therefore if this is chosen it is important to re-assess the subjects status. To provide additional clarity on the blinded combination treatment option following a clinical failure event. Section Blinded Combination Therapy: Dose tables PREVIOUS TEXT Should the investigator choose the subject receive combination therapy following an event it will be provided in the following format to ensure that the blind is maintained (investigator will know that the subject is on combination therapy, but not what group the subject was randomised to): 120

234 GM2008/00365/04 CONFIDENTIAL Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg (event ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg prior to week 4) (TAD optimised) ABS 5mg + TAD 40mg (event ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg prior to week 8) (ABS dose optimised) ABS 10mg + TAD 40mg (event ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg post week 8) Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo ABS 10mg + TAD PBO ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg (event prior to week 8) (ABS optimised) ABS 10mg + TAD Placebo ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg (event post week 8) Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS Placebo + TAD ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg (event prior to week 4) 40mg ABS Placebo + TAD 40mg (event post week 4) ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg For those Subjects in whom the investigator has requested either no uptitration in Investigational Product at weeks 4 or 8 or a down titration in Investigational Product based on tolerability reasons, the following format will apply: Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS 5mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg Combination Therapy will be fully blinded. Investigators are required to complete a Change of Randomised Treatment visit prior to a subject receiving blinded combination therapy and prior to progressing to the next treatment Stage. In addition subjects should continue to follow the 4 weekly safety visit schedule throughout the treatment Stages Subjects will initially receive Medical Intervention Stage 1, then Stage 2, then Stage 3, upon request of the investigator Subjects must be assessed in the clinic no more than 8 weeks following initiation of Stage 1 (by completing a Change of Randomised Treatment Visit) If the investigator judges that the subject has not sufficiently tolerated any of the treatment Stages, the Investigator may request that the subject remain on their current Stage. 121

235 GM2008/00365/04 CONFIDENTIAL Subjects in whom the investigator has requested either no uptitration in study medication or a down titration in study medication based on tolerability reasons may be re-uptitrated at any point at the request of the investigator. This uptitration will follow the format in the first Medical Intervention table above. REVISED TEXT Combination Therapy will be fully blinded. Investigators are required to complete a Change of Randomised Treatment visit prior to a subject receiving blinded combination therapy and prior to progressing to the next treatment Stage. In addition subjects should continue to follow the 4 weekly safety visit schedule throughout the treatment Stages Subjects will initially receive Medical Intervention Stage 1, then Stage 2, then Stage 3, upon request of the investigator Subjects must be assessed in the clinic no more than 8 weeks following initiation of Stage 1 or following progression to the next treatment stage (by completing a Change of Randomised Treatment Visit) If the investigator judges that the subject has not sufficiently tolerated any of the treatment Stages, the Investigator may request that the subject remain on their current Stage. During this post-clinical failure blinded combination therapy, down titration of study drug is not permitted. Subjects in whom the investigator has requested either who either had no uptitration in study medication or had a down titration in study medication based on tolerability reasons prior to initiation of blinded combination therapy may be re-uptitrated at any point as part of the post-clinical failure blinded combination therapy at the request of the investigator. This uptitration will follow the format in the first Medical Intervention table above below. The Investigator will be asked prior to the initiation of blinded combination treatment if the subject should receive full dose of blinded combination therapy or if the subject should stay on any previously requested lower dose of IP. Should the investigator choose the subject receive combination therapy following an event it will be provided in the following example formats to ensure that the blind is maintained (investigator will know that the subject is on combination therapy, but not what group the subject was randomised to): 122

236 GM2008/00365/04 CONFIDENTIAL Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg (event ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg prior to week 4) (TAD optimised) ABS 5mg + TAD 40mg (event ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg prior to week 8) (ABS dose optimised) ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD ABS 10mg + TAD 40mg 40mg ABS 10mg + TAD 40mg (event ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg post week 8) Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo ABS 10mg + TAD PBO ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg (event prior to week 8) (ABS optimised) ABS 10mg + TAD Placebo ABS 10mg + TAD 20mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg (event post week 8) Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS Placebo + TAD 40mg ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg (event prior to week 4) (TAD optimised) ABS Placebo + TAD 40mg (event post week 4) ABS 5mg + TAD 40mg ABS 10mg + TAD 40mg ABS 10mg + TAD 40mg For those Subjects in whom the investigator has requested either no uptitration in Investigational Product at weeks 4 or 8 or a down titration in Investigational Product based on tolerability reasons, the following format will apply: The following example formats will apply in: a) subjects in whom the investigator has requested no uptitration in tadalafil due to renal impairment b) subjects in whom the investigator has requested no uptitration in ambrisentan at week 8 due to tolerability reasons, or a down titration in either Investigational Product based on tolerability reasons. Exception: Subjects receiving placebo will receive full dose of the corresponding IP (see table above). E.g. Prior to clinical failure the investigator requests down titration of ambrisentan in a subject receiving tadalafil monotherapy (i.e. subject continues to receive ambrisentan placebo). Following clinical failure the investigator requests blinded combination treatment and the subject is up-titrated to 5mg then 10mg of ambrisentan). 123

237 GM2008/00365/04 CONFIDENTIAL Treatment at Clinical Failure Stage 1 Stage 2 Stage 3 Combination Therapy Arm ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg [prior to week 8 and renal impairment] ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg Monotherapy arm: ambrisentan group ABS 5mg + TAD Placebo ABS 5mg + TAD 20mg ABS 5mg + TAD 40mg ABS 5mg + TAD 40mg ABS 5mg + TAD Placebo [renal impairment] ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD 20mg ABS 5mg + TAD Placebo ABS 10mg + TAD PBO ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg [prior to week 8 and renal impairment] ABS 10mg + TAD Placebo ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg [renal impairment] Monotherapy arm: tadalafil group ABS Placebo + TAD 20mg ABS 5mg + TAD 20mg ABS 10mg + TAD 20mg ABS 10mg + TAD 20mg RATIONALE The process for administering blinded combination therapy was considered as requiring further clarity to ensure that the process could be more easily understood by investigators. Therefore the following changes have been made to help improve clarity: The dose combination of ABS 10mg and TAD 20mg is permitted in the protocol for those subjects with mild to moderate renal impairment and for subjects who are down titrated from TAD 40mg to TAD 20mg between the week 4 and 8 visits. This option was omitted in the tables that describe stage 1 of blinded combination treatment. This amendment therefore clarifies the dosing for blinded combination therapy in these subjects and addresses the 10mg ABS + 20mg TAD dose omission, allowing subjects with mild to moderate renal impairment or those unable to tolerate 20mg TAD between the week 4 and 8 visits to receive the maximum dose of ABS. The description of when these events may occur in parenthesise in the first table have been removed as they did not apply to all events and may cause confusion. Additional text to clarify the process including which patients follow which dose stages. The investigator may request that the subject re-uptitrate their IP directly following a TtCF event, but not at any point following a TtCF event. This change simplifies the dosing and clarifies when the investigator may request this. 124

238 GM2008/00365/04 CONFIDENTIAL Section 5.6.1, Permitted Medications and Non-Drug Therapies DELETED TEXT Cyclosporine A: Steady state co administration of ambrisentan and cyclosporine A resulted in a 2 fold increase in ambrisentan exposure in healthy volunteers. This may be due to the inhibition by cyclosporine A of transporters and metabolic enzymes involved in the pharmacokinetics of ambrisentan. Therefore the dose of ambrisentan should be limited to 5 mg once daily when co administered with cyclosporine A. Multiple doses of ambrisentan had no effect on cyclosporine A exposure, and no dose adjustment of cyclosporine A is warranted. RATIONALE In a controlled clinical trial the Sponsor does not want to be allowing concomitant use of medications that are known to increase the mean exposure of AMB by two fold. This aligns with the Sponsor s other ambrisentan PAH protocols, where subjects receiving cyclosporine A are excluded. Section 5.6.2, Prohibited Medications and Non-Drug Therapies PREVIOUS TEXT Drugs prohibited while receiving investigational product: Other ERAs, such as Bosentan (Tracleer) and Sitaxentan (Thelin) Commercial ambrisentan (Volibris or Letairis) Inhaled nitric oxide Intravenous inotropes (e.g. dopamine, dobutamine) PDE-5i such as sildenafil (Revatio or Viagra) and vardenafil (Levitra) Commercial tadalafil (Adcirca or Cialis) Any other investigational therapy All forms of prostanoids are prohibited unless a subject has met the definition of clinical failure (Section 3.1) Nitrates Potent inhibitors of CYP3A4 (e.g. protease inhibitors, systemic ketoconazole, or systemic itraconazole) Potent inducers of CYP3A4 (e.g. rifampicin) 125

239 GM2008/00365/04 REVISED TEXT CONFIDENTIAL Drugs prohibited while receiving investigational product: Other ERAs, such as Bosentan (Tracleer) and Sitaxentan (Thelin) Commercial ambrisentan (Volibris or Letairis) Inhaled nitric oxide Intravenous inotropes (e.g. dopamine, dobutamine) PDE-5i such as sildenafil (Revatio or Viagra) and vardenafil (Levitra) Commercial tadalafil (Adcirca or Cialis) Any other investigational therapy All forms of prostanoids are prohibited unless a subject has met the definition of clinical failure (Section 3.1) Nitrates Potent inhibitors of CYP3A4 (e.g. protease inhibitors, systemic ketoconazole, or systemic itraconazole) Potent inducers of CYP3A4 (e.g. rifampicin) Cyclosporine A RATIONALE In a controlled clinical trial the Sponsor does not want to be allowing concomitant use of medications that are known to increase the mean exposure of AMB by two fold. This aligns with the Sponsor s other ambrisentan PAH protocols, where subjects receiving cyclosporine A are excluded. Section 6 Study Assessments and Procedures; Table footnote 2 PREVIOUS TEXT 2. Unscheduled / Change of randomised treatment visit. Also refer to foot note 12. If a subject permanently withdraws from investigational product and from the study at the same time, the End of Study Visit should be completed. REVISED TEXT 2 Unscheduled / Change of randomised treatment visit. Also refer to foot note 12. If a subject permanently withdraws from investigational product and from the study at the same time, the End of Study Visit should be completed. RATIONALE The cross-reference was included in the original protocol in error. It is redundant and was left after changing the footnotes in a previous draft of the protocol. 126

240 GM2008/00365/04 CONFIDENTIAL Section 6 Study Assessments and Procedures ADDITIONAL TEXT Subjects will have the option to receive study related SMS/text messages including visit reminders and a reminder not to take investigational product less than 16 hours prior to the week 16 visit. RATIONALE This is a long-term study with multiple clinic visits as well as monthly safety visits. In addition the week 16 visit requires that subjects hold that days dose of IP until they attend the clinic. Therefore an optional text message service is offered to participating centres and subjects to remind them when their visits are scheduled and any visit specific instructions. Section Peak-trough assessment of 6MWD; first paragraph PREVIOUS TEXT At the Randomisation visit (Week 0), all subjects will be randomly assigned to perform the Week 16 6MWD assessment at either peak or trough ambrisentan concentrations. Prior to arriving at the investigative site for the Week 16 visit, subjects will not administer their daily dose of either Investigational Products (ABS or TAD). The site will dispense blinded Investigational Product (ambrisentan/ambrisentan matching placebo) as directed by the IVRS for the peak-trough assessment. Subjects in the ambrisentan monotherapy arm and the combination arm will receive either ambrisentan followed by placebo (peak assessment) or placebo followed by ambrisentan (trough assessment). Subjects randomised to the tadalafil monotherapy arm will receive placebo followed by placebo to maintain the blind. REVISED TEXT At the Randomisation visit (Week 0), All subjects will be randomly assigned to perform the Week 16 6MWD assessment at either peak or trough ambrisentan concentrations. Prior to arriving at the investigative site for the Week 16 visit, subjects will not administer their daily dose of either Investigational Products (ABS or TAD). The site will dispense blinded Investigational Product (ambrisentan/ambrisentan matching placebo) as directed by the IVRS for the peak-trough assessment. Subjects in the ambrisentan monotherapy arm and the combination arm will receive either ambrisentan followed by placebo (peak assessment) or placebo followed by ambrisentan (trough assessment). Subjects randomised to the tadalafil monotherapy arm will receive placebo followed by placebo to maintain the blind. RATIONALE The randomisation will be performed at week 16, not week 0, as stated in Section 5.2. This was an error in the original protocol. This change will make the protocol consistent. 127

241 GM2008/00365/04 CONFIDENTIAL Section Management, Monitoring and Follow-up of Serum Aminotransferase Abnormalities; Table 2, first row PREVIOUS TEXT Serum Aminotransferase Elevations ALT>3xULN and bilirubin 2XULN or if ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. Actions to be Taken Stop investigational product treatment immediately. Report the liver event to the Medical Monitor within 24 hours of learning of its occurrence and report as an SAE. Make every reasonable attempt to have subjects return to clinic within 24 hours for repeat liver chemistries, liver event follow up assessments, and close monitoring Complete liver event CRF within 24 hours. After discontinuation of investigational product, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. A specialist or hepatology consultation is recommended investigational product may not be reintroduced. AMENDED TEXT Serum Aminotransferase Elevations ALT>3xULN and bilirubin >2XULN or if ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. Actions to be Taken Stop investigational product treatment immediately. Report the liver event to the Medical Monitor within 24 hours of learning of its occurrence and report as an SAE. Make every reasonable attempt to have subjects return to clinic within 24 hours for repeat liver chemistries, liver event follow up assessments, and close monitoring Complete liver event CRF within 24 hours. After discontinuation of investigational product, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. A specialist or hepatology consultation is recommended Investigational product may not be reintroduced. RATIONALE The definition of liver events that potentially meet the criteria for Hy s law cases was amended from bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. 128

242 GM2008/00365/04 CONFIDENTIAL Section Liver event follow-up assessment; first paragraph 4 th bullet point and 2 nd paragraph PREVIOUS TEXT Fractionate bilirubin, if total bilirubin 2xULN.. Serum bilirubin fractionation should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick. The following are required for subjects with ALT 3xULN and bilirubin 2xULN (>35% direct) but are optional for other abnormal liver chemistries: REVISED TEXT Fractionate bilirubin, if total bilirubin >2xULN.. Serum bilirubin fractionation should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick. The following are required for subjects with ALT >3xULN and bilirubin >2xULN (>35% direct) but are optional for other abnormal liver chemistries: RATIONALE The definition of liver events that potentially meet the criteria for Hy s law cases was amended from ALT value from 3xULN to >3xULN and bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. Further, text related to fractionation of bilirubin as part of Hy s law cases was removed to simplify the definition of these events. Section Definition of an SAE, parts g & h PREVIOUS TEXT g. All events of possible drug-induced liver injury with hyperbilirubinaemia defined as ALT 3xULN and bilirubin 2xULN (>35% direct) (or ALT 3xULN and INR>1.5, if INR measured) termed Hy s Law events (INR measurement is not required and the threshold value stated will not apply to patients receiving anticoagulants). h. NOTE: bilirubin fractionation is performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If testing is unavailable and a subject meets the criterion of total bilirubin 2xULN, then the event is still reported as an SAE. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury. 129

243 GM2008/00365/04 AMENDED TEXT CONFIDENTIAL g. All events of possible drug-induced liver injury with hyperbilirubinaemia defined as ALT >3xULN and bilirubin >2xULN (>35% direct) (or ALT 3xULN and INR>1.5, if INR measured) termed Hy s Law events (INR measurement is not required and the threshold value stated will not apply to patients receiving anticoagulants). For the purposes of this protocol, ALT >3xULN associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia will be considered a protocol defined SAE. h. NOTE: bilirubin fractionation is should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If testing is unavailable and a subject meets the criterion of total bilirubin 2xULN, then the event is still reported as an SAE. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury. RATIONALE The definition of liver events that potentially meet the criteria for Hy s law cases was amended from ALT value from 3xULN to >3xULN and bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. Further, text related to fractionation of bilirubin as part of Hy s law cases was removed to simplify the definition of these events. Additionally, text was added to section to clarify that symptomatic elevations >3xULN are considered protocol defined SAEs. Section Laboratory and Other Safety Assessment Abnormalities Reported as AEs and SAEs; Paragraphs 3,4 & 5 PREVIOUS TEXT All events of possible drug-induced liver injury with hyperbilirubinaemia (defined as ALT 3xULN plus bilirubin 2xULN and/or INR>1.5) or Hy s Law events, require immediate Investigational Product cessation and reporting as an SAE (see Section g above). NOTE: bilirubin fractionation is performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If testing is unavailable and a subject meets the criterion of total bilirubin 2.0 xuln, then the event is still reported as an SAE. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury. 130

244 GM2008/00365/04 REVISED TEXT CONFIDENTIAL All events of possible drug-induced liver injury with hyperbilirubinaemia (defined as ALT >3xULN plus bilirubin >2xULN and/or INR>1.5) or Hy s Law events, require immediate Investigational Product cessation and reporting as an SAE (see Section g above). NOTE: bilirubin fractionation is should be performed if testing is available. If testing is unavailable, record presence of detectable urinary bilirubin on dipstick indicating direct bilirubin elevations and suggesting liver injury. If testing is unavailable and a subject meets the criterion of total bilirubin 2.0 xuln, then the event is still reported as an SAE. If INR is obtained, include values on the SAE form. INR elevations >1.5 suggest severe liver injury. RATIONALE The definition of liver events that potentially meet the criteria for Hy s law cases was amended from ALT value from 3xULN to >3xULN and bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. Further, text related to fractionation of bilirubin as part of Hy s law cases was removed to simplify the definition of these events. 131

245 GM2008/00365/04 CONFIDENTIAL Section Prompt Reporting of Serious Adverse Events and Other Events to Sponsor; table PREVIOUS TEXT Initial Reports Follow-up Information on a Previous Report Type of Event Time Frame Documents Time Frame Documents All SAEs 24 hours SAE data 24 hours Updated SAE collection tool Pregnancy 24 hours Pregnancy Notification Form Liver chemistry abnormalities ALT 3xULN and 24 hours* SAE data Bilirubin 2xULN collection tool. (>35% direct) (or **Liver Event ALT 3xULN and Case Report Form INR>1.5, if INR (CRF) and liver measured)*** imaging and/or biopsy CRFs if ALT 5xULN; ALT 3xULN with hepatitis or rash or 3xULN 4 weeks ALT 3xULN and <5xULN and bilirubin <2xULN data collection tool 24 hours Pregnancy Follow up Form 24 hours Updated SAE data collection tool. **Updated Liver Event CRF applicable 24 hours* **Liver Event CRF 24 hours **Updated Liver Event CRF 24 hours* **Liver Event CRF does not need completing unless elevations persist for 4 weeks or subject cannot be monitored weekly for 4 weeks *Sponsor to be notified at onset of liver chemistry elevations to discuss subject safety. ** Liver event documents should be completed as soon as possible *** INR measurement is not required; if measured, the threshold value stated will not apply to patients receiving anticoagulants. 132

246 GM2008/00365/04 CONFIDENTIAL REVISED TEXT Initial Reports Follow-up Information on a Previous Report Type of Event Time Frame Documents Time Frame Documents All SAEs 24 hours SAE data 24 hours Updated SAE collection tool Pregnancy 24 hours Pregnancy Notification Form Liver chemistry abnormalities ALT >3xULN and Bilirubin >2xULN (>35% direct) (or ALT >3xULN and INR>1.5, if INR measured)*** or ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. ALT 5xULN; ALT 3xULN with hepatitis or rash or 3xULN 4 weeks ALT/AST >3xULN and <5xULN and bilirubin < 2xULN 24 hours* SAE data collection tool. **Liver Event Case Report Form (CRF) and liver imaging and/or biopsy CRFs if applicable data collection tool 24 hours Pregnancy Follow up Form 24 hours Updated SAE data collection tool. **Updated Liver Event CRF 24 hours* **Liver Event CRF 24 hours **Updated Liver Event CRF 24 hours* **Liver Event CRF does not need completing unless elevations persist for 4 weeks or subject cannot be monitored weekly for 4 weeks 24 Hours **Update Liver Event CRF *Sponsor to be notified at onset of liver chemistry elevations to discuss subject safety. ** Liver event documents should be completed as soon as possible *** INR measurement is not required; if measured, the threshold value stated will not apply to patients receiving anticoagulants. 133

247 GM2008/00365/04 RATIONALE CONFIDENTIAL The definition of liver events that potentially meet the criteria for Hy s law cases was amended from ALT value from 3xULN to >3xULN and bilirubin value from 2xULN to >2xULN for the sake of consistency with other liver event definitions in the protocol. Further, text related to fractionation of bilirubin as part of Hy s law cases was removed to simplify the definition of these events. Additionally, text was added to clarify that symptomatic elevations >3xULN are considered protocol defined SAEs. As described in detail in protocol Section [Management, Monitoring and Follow-up of Serum Aminotransferase Abnormalities], Table 2, all ALT/AST >3xULN and bilirubin 2xULN will be reported using the liver event CRF as well as follow-up information collected within 24 hours. The revised wording reflects these inconsistencies in the protocol. Appendix 2, 6 Minute Walk Distance Test PREVIOUS TEXT The Day Before the Test Tell the subject the time their test will take place. Give the subject the following instructions: - Take a light meal 3 to 4 hours before the test. - Take nothing by mouth (po) except clear liquids for 1 hour before the test. - Do not take heart medication within 2 hours of the test, including short-acting nitrates. - Do not smoke for at least 2 hours before the test. - Prior to all clinic visits, withhold the daily dose of Investigational Product such that the last dose of Investigational Product is administered more than 16 hours prior to the 6MWD and BDI assessments REVISED TEXT The Day Before the Test Tell the subject the time their test will take place. Give the subject the following instructions: - Take a light meal 3 to 4 hours before the test. - Take nothing by mouth (po) except clear liquids for 1 hour before the test. - Do not take heart medication within 2 hours of the test, including short-acting nitrates. - Do not smoke for at least 2 hours before the test. 134

248 GM2008/00365/04 CONFIDENTIAL Prior to all clinic visits, withhold the daily dose of Investigational Product such that the last dose of Investigational Product is administered more than 16 hours prior to the 6MWD and BDI assessments RATIONALE This is a clarification to reflect the main body of the protocol. Subjects will not be asked to hold their dose of either Investigational Product on the day of their clinic visit, with the exception of the week 16 visit to enable the peak-trough assessment. This is dealt with separately in the protocol Section In addition, nitrates are excluded from the study therefore the reference to nitrates has been removed to avoid any confusion. 135

249 GM2008/00365/04 CONFIDENTIAL Appendix 9: Protocol Amendment 02 Changes AMENDMENT 2 Where the Amendment Applies This amendment is applicable to all countries and sites Summary of Amendment Changes with Rationale Medical Monitor: Change to Sponsor GSK Medical Monitor and associated contact information due to change in study personnel. Affects Protocol Summary Sponsor Information Page GSK Primary Endpoint: Clarification of primary endpoint wording second component to allow any hospitalisation due to worsening PAH, not just non-elective hospitalisation. Change to fourth component- removal wording that could potentially be interpreted as contradictory to aid clarity. Affects Protocol Summary, Section 3.1 Study Design and Section Primary End Point. Change to allow use of local lab tests at screening to assess eligibility for randomisation. Amendments to Section 4.3 Exclusion Criteria and Section Safety Tests. Changes to eligibility criteria: Review of blinded demographic data for subjects enrolled during the first six months of the study showed a relatively high prevalence of risk factors for left ventricular dysfunction (e.g. CAD, diabetes, hypertension) compared to previous PAH trials. The current amendment includes revisions to the eligibility criteria in order to reduce the likelihood of patients with PH due to covert diastolic dysfunction being enrolled (i.e WHO Group 2). Excluding subjects with three or more risk factors for left heart disease, along with more stringent hemodynamic requirements, will increase the likelihood that subjects with WHO Group 2 pulmonary hypertension are not enrolled. In addition, investigators will be required to document a negative V/Q scan (plus pulmonary arteriogram, if needed) for all subjects to confirm that elevated pulmonary pressures are not due to CTEPH (WHO Group 4). Amendments to sections Protocol Summary [study design], Section 4.2 [Inclusion Criteria 3, 6 & 7]. United States Specific Amendment: Allow RHC as a study procedure during the Screening Period in pre-defined circumstances. Additional text added to Section 4.2 Inclusion Criteria. Reason for this change is that in the United States, many insurance companies will not approve payment for a second RHC procedure even where the first investigation is incomplete or inaccessible. Diagnostic test results (e.g. RHC) received by specialist centres are occasionally incomplete or inaccessible for screening, and a second RHC is required to confirm diagnosis. 136

250 GM2008/00365/04 CONFIDENTIAL Clarification that ophthalmic formulation of cyclosporine is not excluded or prohibited medication. Additional text added to Section 4.3 Exclusion Criteria and Section Prohibited Medications and Non-Drug Therapies. Basis of this exclusion is potential drug-drug interaction with ambrisentan. Since there is no systemic absorption of the ophthalmic formulation of cyclosporine and blood concentrations are undetectable after topical administration there is no risk of a drug-drug interaction with cyclosporine eye drops. Study Assessments and Procedures table: Relocation of footnote 5 requesting subjects withhold IP for 16 hours prior to clinic visits from all visits to just the week 16 visit. This was incorrect and the contradictory text was unintentionally left over following amendment 01. Clarification that reporting of deaths should take place from the signing of informed consent. Affects Section Increase sample size, recruitment and study duration and removal of interim analyses, to increase power for comparison of combination therapy to individual monotherapy arms. Changes to Protocol Summary, Section 3.1, Section 4.1, Section 8.2.1, Section 8.2.2, Section 8.3.1, Section 8.3.2, Section and Section Change to testing of secondary endpoints and addition of a hierarchical ordering for testing in response to regulatory agency feedback. Changes to Protocol Summary (Study Endpoints and assessments) and Section Adds an as treated population for the reporting of safety data. Appendix 2: Removal of the requirement for the 6MWD track to be exactly 30m. The ATS guidelines contain apparently contradictory information that the track must be 30m but track lengths from 15 to 50m approx., don t affect measurements. Interpretation of ATS guidelines by clinicians varies widely, such that some use 30m tracks but many don t and never have (some clinicians working environment does not permit a 30m track, some use tracks that are longer). Thus the current protocol statement, whilst consistent with ATS guidelines, is inconsistent with clinical interpretation of the guidelines and standard clinical practice for many PAH specialists. Therefore, to make the protocol more consistent with the clinical use of the guidelines, the 30m requirement is removed from the protocol and additional clarification, based on ATS guidelines, will be provided in the Study Procedures Manual. List of Specific Changes Sponsor Information Page: GSK PREVIOUS TEXT Sponsor Medical Monitor Contact Information: Dr Telephone: 137

251 GM2008/00365/04 REVISED TEXT CONFIDENTIAL Sponsor Medical Monitor Contact Information: Dr Telephone: Protocol Summary: Study Design Current Text: Subjects must have a confirmed diagnosis of PAH with documented mean pulmonary arterial pressure (mpap) 25 mmhg, pulmonary vascular resistance (PVR) 240 dyne sec/cm 5, and pulmonary capillary wedge pressure (PCWP) or left ventricle end diastolic pressure (LVEDP) of 15 mmhg based on a right heart catheterization (RHC) prior to Screening. Revised Text: Subjects must have a confirmed diagnosis of PAH with documented mean pulmonary arterial pressure (mpap) 25 mmhg, pulmonary vascular resistance (PVR) dyne sec/cm 5, and pulmonary capillary wedge pressure (PCWP) or left ventricle end diastolic pressure (LVEDP) of 12 mmhg if PVR 300 to <500 dyne sec/cm 5, or PCWP/LVEDP 15 mmhg if PVR 500 dyne sec/cm 5, based on a right heart catheterization (RHC) prior to Screening. Subjects must not have 3 or more of the following left ventricular disease/dysfunction risk factors: i. Body Mass Index (BMI) 30 ii. iii. iv. History of Essential Hypertension Diabetes Mellitus any type Historical evidence of significant coronary disease established by any one of: - history of myocardial infarction - history of percutaneous intervention - angiographic evidence of CAD (>50% stenosis in at least one vessel), either by invasive angiography or by CT Angiography - positive stress test with imaging (either pharmacologic or with exercise) - previous coronary artery surgery - chronic stable angina 138

252 GM2008/00365/04 CONFIDENTIAL Subjects are required to have a negative V/Q scan. If results of the V/Q scan are not definitive, evidence of a pulmonary arteriogram (negative for CTEPH) is required prior to screening. Protocol Summary: Study Endpoints/Assessments; primary endpoint, second and fourth components PREVIOUS TEXT Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non-elective hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) REVISED TEXT Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non elective Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months 139

253 GM2008/00365/04 CONFIDENTIAL A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) Section 3.1: Study Design: description of primary endpoint, second and fourth components PREVIOUS TEXT Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non-elective hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) REVISED TEXT Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non elective Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) 140

254 GM2008/00365/04 CONFIDENTIAL Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) Section 4.2: Inclusion Criteria 3, 6 & 7 PREVIOUS TEXT 4. Subjects must have a diagnosis of PAH due to the following: a. idiopathic or heritable PAH b. PAH associated with: iv. connective tissue disease (e.g., limited scleroderma, diffuse scleroderma, mixed CTD, systemic lupus erythematosus, or overlap syndrome) v. drugs or toxins vi. HIV infection vii. congenital heart defects repaired greater than 1 year prior to screening (i.e., atrial septal defects, ventricular septal defects, and patent ductus arteriosus) NB: subjects with portopulmonary hypertension and PVOD are NOT eligible for the study 6. Subject must meet all of the following haemodynamic criteria by means of a RHC prior to screening: i. mpap of 25 mmhg ii. PVR 240 dyne sec/cm5 iii. PCWP or LVEDP of 15 mmhg 7. Subject must meet all of the following pulmonary function tests completed no more than 24 weeks before the Screening visit: i. Total lung capacity (TLC) 60% of predicted normal and ii. Forced expiratory volume in one second (FEV1) 55% of predicted normal REVISED TEXT 3. Subjects must have a diagnosis of PAH due to the following: a. idiopathic or heritable PAH b. PAH associated with: i. connective tissue disease (e.g., limited scleroderma, diffuse scleroderma, mixed CTD, systemic lupus erythematosus, or overlap syndrome) 141

255 GM2008/00365/04 CONFIDENTIAL ii. drugs or toxins iii. HIV infection iv. congenital heart defects repaired greater than 1 year prior to screening (i.e., atrial septal defects, ventricular septal defects, and patent ductus arteriosus) NB: subjects with portopulmonary hypertension and PVOD are NOT eligible for the study NB: Subjects must not have 3 or more of the following left ventricular disease/dysfunction risk factors: v. Body Mass Index (BMI) 30 vi. History of Essential Hypertension vii. Diabetes Mellitus any type viii. Historical evidence of significant coronary disease established by any one of: history of myocardial infarction history of percutaneous intervention angiographic evidence of CAD (>50% stenosis in at least one vessel), either by invasive angiography or by CT Angiography positive stress test with imaging (either pharmacologic or with exercise treadmill) previous coronary artery surgery chronic stable angina 6. Subject must meet all of the following haemodynamic criteria by means of a RHC prior to screening: i. mpap of 25 mmhg ii. PVR dyne sec/cm5 iii. PCWP or LVEDP of 12 mmhg if PVR 300 to <500 dyne sec/cm 5, or PCWP/LVEDP 15 mmhg if PVR 500 dyne sec/cm 5 U.S. Specific Text: Subjects in the United States who have had previous RHC and have evidence to suggest the presence of PH but require an RHC during the Screening Period to confirm the diagnosis of PH (as defined above) may be permitted to receive a RHC as a study procedure, provided approval from the medical monitor is obtained in advance. Approval will be based upon the following criteria: The existing RHC data in the subject's medical record is deemed to be inadequate or non-interpretable per the investigator's judgment An adequate RHC record could not be obtained or was not accessible from the subject's medical record 142

256 GM2008/00365/04 CONFIDENTIAL Subject has met all other screening requirements prior to undergoing RHC The site/subject are able to perform the RHC in advance and within the screening period Documentation of approval will be maintained in the source documents and at the Sponsor 5. Subject must meet all of the following pulmonary function tests completed no more than 24 weeks before the Screening visit: i. Total lung capacity (TLC) 60% of predicted normal and ii. Forced expiratory volume in one second (FEV1) 55% of predicted normal Subjects are required to have a documented negative V/Q scan or pulmonary arteriogram confirming the absence of CTEPH prior to screening. Section 4.3: Exclusion Criteria PREVIOUS TEXT Subjects meeting any of the following criteria must not be enrolled in the study: REVISED TEXT Subjects meeting any of the following criteria must not be enrolled in the study (please note that screening visit laboratory tests may be performed by local or central laboratory): PREVIOUS TEXT 4. Subject is receiving treatment with cyclosporine A REVISED TEXT 4. Subject is receiving treatment with cyclosporine A (except ophthalmic formulation) Section 5.6.2: Prohibited Medications and Non-Drug Therapies PREVIOUS TEXT Cyclosporine A REVISED TEXT Cyclosporine A (except ophthalmic formulation) 143

257 GM2008/00365/04 CONFIDENTIAL Section 6: Study Assessments and Procedures PREVIOUS TEXT Wk Wk Wk 12-week Procedures ScreenBase line Week (Wk)/Visit Wk Wk Wk intervals Treatment FAV 3 EOS Visit 4 Safety Visits / telephone contact 4 Weekly 1 Unscheduled/ Follow Up Change of Call 13 Tx 2 Visit Windows All visit windows are ± 3 days with the exception of screen baseline. If eligible, subjects should be randomised within 4 weeks of the screening visit. Written Informed X Consent Subject X Demography Medical History X Disease History X Therapy History X Inclusion/Exclusio X Criteria Efficacy Assessments 6 MWD 5 X X X X X X X X X X BDI Score 5 X X X X X X X X X X WHO FC assessment X X X X X X X X X X Clinical Failure X X X X X X X X X Health Outcome Assessments SF-36 X X X X X X X 14 CAMPHOR X X X X X X X 14 Safety Assessments Concomitant Medication X X X X X X X X X X X Physical X X X Examination Vital Signs 6 X X X X X X X X X X 12-lead ECG X X X Pulse Oximetry X Pulmonary Function Test 7 X Adverse Events X X X X X X X X X X X Serious Adverse X X X X X X X X X X X X Events Laboratory Assessments Haematology X X X X X X X X X X Chemistry X X X X X X X X X X X (inc. LFT s) 8 Testicular X X X X X 12 X X X Function 9 Urinalysis X X X Pregnancy Test 10 S U S S S S S S S S S 144

258 GM2008/00365/04 CONFIDENTIAL Wk Wk Wk 12-week Procedures ScreenBase line Week (Wk)/Visit Wk Wk Wk intervals Treatment FAV 3 EOS Visit 4 Safety Visits / telephone contact 4 Weekly 1 NT-pro-BNP X X X X X X X X X Coagulation X Pharmacogenetics X (optional) Investigational product (IP) Dispense IP X X X X X X X X 11 Unscheduled/ Follow Up Change of Call 13 Tx 2 Assess IP X X X X X X X X compliance IVRS X X X X X X X X X X 11 Peak-Trough X 1. Safety Visits will be required every 4 weeks after week 4 unless a clinic visit is scheduled. A phone call is required to assess concomitant medications and adverse events (including serious). See the Clinical Laboratory Test section for details of the laboratory tests required. If any additional procedures are performed at the safety visit, an unscheduled visit should be completed. 2. Unscheduled / Change of randomised treatment visit. If a subject permanently withdraws from investigational product and from the study at the same time, the End of Study Visit should be completed. 3. Final Assessment Visit 4. End of Study. (This visit should also be completed by subjects who prematurely withdraw from the study) 5. Subjects will take their investigational products dose at least 16 hours prior to the 6MWD and BDI assessments 6. Includes weight, blood pressure and heart rate plus height (height at screening only). 7. Pulmonary function test may also be carried out at screening if not done within 24 weeks prior. 8. Liver Function Test (LFT) should be measured and Investigational Product dose adjusted / stopped as specified in Section and Section Testicular function (Total testosterone, sex hormone binding globulin (SHBG needed to calculate free testosterone), FSH, LH, and inhibin B. For males only. Every effort should be made to collect blood samples for testicular function between 8am and 1pm 10. Pregnancy test: urine = U, serum = S. For women of child bearing potential only 11. If applicable 12. Alternate visits only (i.e. every 24 weeks) 13. Follow up telephone contact 30 days after last dose of Investigational Product 14. Only if change of randomised treatment See Appendix 2 for instructions on how to conduct the 6 minute walk test. (2 tests i.e. 1 at screening and 1 at randomisation should be performed during the screening period such that 2 consecutive tests do not differ by more than 10%) 145

259 GM2008/00365/04 REVISED TEXT CONFIDENTIAL Procedures Week (Wk)/Visit Screen Base line Wk Wk -4 0 Treatment Wk Wk Wk Wk 12-week FAV 3 EOS intervals Visit 4 Safety Visits / telephone contact 4 Weekly 1 Unscheduled/ Change of Tx 2 Follow U Call 13 Visit Windows All visit windows are ± 3 days with the exception of screen baseline. If eligible, subjects should be randomised within 4 weeks of the screening visit. Written Informed X Consent Subject X Demography Medical History X Disease History X Therapy History X Inclusion/Exclusion X Criteria Efficacy Assessments 6 MWD 5 X X X X X 5 X X X X X BDI Score 5 X X X X X 5 X X X X X WHO FC assessment X X X X X X X X X X Clinical Failure X X X X X X X X X Health Outcome Assessments SF-36 X X X X X X X 14 CAMPHOR X X X X X X X 14 Safety Assessments Concomitant Medication X X X X X X X X X X X Physical X X X Examination Vital Signs 6 X X X X X X X X X X 12-lead ECG X X X Pulse Oximetry X Pulmonary Function Test 7 X Adverse Events X X X X X X X X X X X Serious Adverse X X X X X X X X X X X X Events Laboratory Assessments Haematology X X X X X X X X X X Chemistry X X X X X X X X X X X (inc. LFT s) 8 Testicular X X X X X 12 X X X Function 9 Urinalysis X X X Pregnancy Test 10 S U S S S S S S S S S NT-pro-BNP X X X X X X X X X Coagulation X 146

260 GM2008/00365/04 CONFIDENTIAL Procedures Week (Wk)/Visit Screen Base line Wk Wk -4 0 Treatment Wk Wk Wk Wk 12-week FAV 3 EOS intervals Visit 4 Safety Visits / telephone contact 4 Weekly 1 Unscheduled/ Change of Tx 2 Follow U Call 13 Pharmacogenetics X (optional) Investigational product (IP) Dispense IP X X X X X X X X 11 Assess IP X X X X X X X X compliance IVRS X X X X X X X X X X 11 Peak-Trough X 1. Safety Visits will be required every 4 weeks after week 4 unless a clinic visit is scheduled. A phone call is required to assess concomitant medications and adverse events (including serious). See the Clinical Laboratory Test section for details of the laboratory tests required. If any additional procedures are performed at the safety visit, an unscheduled visit should be completed. 2. Unscheduled / Change of randomised treatment visit. If a subject permanently withdraws from investigational product and from the study at the same time, the End of Study Visit should be completed. 3. Final Assessment Visit 4. End of Study. (This visit should also be completed by subjects who prematurely withdraw from the study) 5. Subjects will take their investigational products dose at least 16 hours prior to the 6MWD and BDI assessments 6. Includes weight, blood pressure and heart rate plus height (height at screening only). 7. Pulmonary function test may also be carried out at screening if not done within 24 weeks prior. 8. Liver Function Test (LFT) should be measured and Investigational Product dose adjusted / stopped as specified in Section and Section Testicular function (Total testosterone, sex hormone binding globulin (SHBG needed to calculate free testosterone), FSH, LH, and inhibin B. For males only. Every effort should be made to collect blood samples for testicular function between 8am and 1pm 10. Pregnancy test: urine = U, serum = S. For women of child bearing potential only 11. If applicable 12. Alternate visits only (i.e. every 24 weeks) 13. Follow up telephone contact 30 days after last dose of Investigational Product 14. Only if change of randomised treatment See Appendix 2 for instructions on how to conduct the 6 minute walk test. (2 tests i.e. 1 at screening and 1 at randomisation should be performed during the screening period such that 2 consecutive tests do not differ by more than 10%) Section : Time to Clinical Failure, second and fourth components PREVIOUS TEXT Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non-elective hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy 147

261 GM2008/00365/04 CONFIDENTIAL Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) REVISED TEXT Death (all-cause) Hospitalization for worsening PAH (adjudicated) Non elective Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving randomised treatment for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days Sustained WHO class III symptoms for 6 months (WHO class III symptoms assessed at two clinic visits separated by 6 months) Section : Safety Tests, second paragraph PREVIOUS TEXT The following tests are required at all clinic visits (including FAV, EOS visit, Unscheduled/Change of Randomised Treatment Visit (if applicable)) and will be performed by a central laboratory: 148

262 GM2008/00365/04 REVISED TEXT CONFIDENTIAL The following tests are required at all clinic visits (including FAV, EOS visit, Unscheduled/Change of Randomised Treatment Visit (if applicable)) and will be performed by a central laboratory, except for the screening visit, where a local or central laboratory may be used: Section 6.3.8: Time Period and Frequency of Detecting AEs and SAEs PREVIOUS TEXT All deaths, regardless of cause or relationship, must be reported for subjects on study and for all deaths occurring within 30 days of last investigational product dose or within 30 days of last study evaluation, whichever is longer. REVISED TEXT All deaths, regardless of cause or relationship, must be reported for subjects on study (defined as the point from which the subject is consented into the study) and for all deaths occurring within 30 days of last investigational product dose or within 30 days of last study evaluation, whichever is longer. Protocol Summary: Study Design; Paragraphs 2, 12 and 13 PREVIOUS TEXT Enrolment of 352 subjects is planned with 176 subjects in the combination arm and 176 in the monotherapy arm. All subjects are eligible to receive a minimum of 24 weeks of therapy. At the time that 82 adjudicated primary endpoint events are projected to have occurred, i.e. 82 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety assessment will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed. An Independent Data Monitoring Committee (IDMC) will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will also review formal interim analysis and recommend continuation or early termination of the study based on stopping criteria defined in the IDMC charter. REVISED TEXT Enrolment of subjects is planned with subjects in the combination arm and in the monotherapy arm (100 subjects receiving ambrisentan and 100 subjects receiving tadalafil). 149

263 GM2008/00365/04 CONFIDENTIAL All subjects are eligible to receive a minimum of 24 weeks of therapy. At the time that adjudicated primary endpoint events are projected to have occurred, i.e subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety assessment will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed. An Independent Data Monitoring Committee (IDMC) will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will also review formal interim analysis and recommend continuation or early termination of the study based on stopping the criteria defined in the IDMC charter. Protocol Summary: Study Endpoints/assessments PREVIOUS TEXT This event-driven study requires 82 subjects with a clinical failure event in order to have 90% power with an overall type I error rate (alpha level) of 5%, accounting for two interim analyses as detailed in Section Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of 10% per year, i.e. an absolute difference of 10%. This leads to an approximate hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial. Alpha spending associated with the stopping criteria are described in Section The study will require 352 subjects (176 per treatment group) assuming a recruitment period of 65 weeks (based on an average recruitment rate of 5.4 subjects per week) and a total study duration of 129 weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 10% per treatment group at one year has been considered for this calculation. The primary comparison will be between the combination therapy arm and the pooled monotherapy arms. For the primary analysis, time to the first clinical failure event of PAH, the log rank test will be used to compare the two treatment arms. Time to clinical failure will be displayed by treatment group as a Kaplan-Meier event-free curve from randomisation up to the Final Assessment Visit (FAV). Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using a Cox proportional hazards regression model. 150

264 GM2008/00365/04 CONFIDENTIAL The two interim analyses are planned after 50% and 75% of the primary endpoints events have been reached. Statistical stopping rules are in place to stop the study early for either superiority or futility. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints. Inferences on the superiority of combination versus monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint is statistically significant will inferences on the Week 24 change from baseline 6MWD endpoint be evaluated. Only if the Week 24 change from baseline 6MWD endpoint is found to be significant will inferences on the percentage of subjects with satisfactory response at Week 24 endpoint be evaluated. This gate keeping approach, described above, will be implemented for all secondary endpoints in the order listed above. With this strategy, all tests will be based on a two-side p-value at a 5% significance level. REVISED TEXT This event-driven study requires subjects with a clinical failure event in order to have approximately 9097% power for the comparison of combination therapy with pooled monotherapy, and 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a an overall type I error rate (alpha level) of 5% for each of the three comparisons., accounting for two interim analyses as detailed in Section Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to a an approximate hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial. Alpha spending associated with the stopping criteria are described in Section The study will require subjects ( subjects in the combination arm and 200 in the monotherapy arm 100 subjects receiving ambrisentan and 100 subjects receiving tadalafil per treatment group) assuming a recruitment period of 6582 weeks (based on an average recruitment rate of subjects per week) and a total study duration of weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 105% per treatment group at one year has been considered for this calculation. The primary comparison will be between the combination therapy arm and the pooled monotherapy arms. For the primary analysis, time to the first clinical failure event of PAH, the log rank test will be used to compare the two treatment arms. Time to clinical failure will be displayed by treatment group as a Kaplan-Meier event-free curve from randomisation up to the Final Assessment Visit (FAV). Events which occur after the FAV will not be used in the 151

265 GM2008/00365/04 CONFIDENTIAL primary analysis, but may be considered for various sensitivity analyses. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using a Cox proportional hazards regression model. The primary comparison will be between the combination therapy arm and the pooled monotherapy arms. Comparison of the combination therapy to the individual monotherapy arms is a secondary comparison. These comparisons will only be performed if the comparison of the combination vs. pooled monotherapy is significant. The two interim analyses are planned after 50% and 75% of the primary endpoints events have been reached. Statistical stopping rules are in place to stop the study early for either superiority or futility. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints and treatment comparisons. Inferences on the superiority of combination versus pooled monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance for combination therapy vs pooled monotherapy therapy. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint is statistically significant will inferences on the Week 24 change from baseline 6MWD endpoint be evaluated. Only if the Week 24 change from baseline 6MWD endpoint is found to be significant will inferences on the percentage of subjects with satisfactory response at Week 24 endpoint be evaluated. This gate keeping approach, described above, will be implemented for all secondary endpoints in the order listed above. With this strategy, all tests will be based on a two side p value at a 5% significance level. Only if the primary endpoint combination vs pooled comparison is statistically significant will inferences on the first secondary endpoint be evaluated. Only if the first secondary endpoint combination vs pooled comparison is found to be significant will inferences on the subsequent secondary endpoint be evaluated. This gate keeping approach will be implemented for all secondary endpoints in pre-define order, and all tests will be performed at a 5% significance level. Additionally, for each endpoint, if the combination treatment arm is demonstrated as statistically significantly superior to the pooled monotherapy arm using the gate keeping approach described above, comparisons of combination treatment versus each individual monotherapy treatment will be performed at a 5% significance level. 152

266 GM2008/00365/04 CONFIDENTIAL Section 3.1: Study Design; Figure 1, Study Schematic, paragraph 2, penultimate and final paragraph and Figure 2 PREVIOUS TEXT Figure 3 Study Schematic This is an event driven trial. The duration of the study will be determined by the rate of first occurrence of time to clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. The study will be terminated when it is projected that approximately 82 events have occurred. Based on current assumptions, the median treatment duration is anticipated to be 1.6 years with the total study duration estimated at 2.5 years. All subjects are eligible to receive a minimum of 24 weeks of therapy. When 82 adjudicated primary endpoint events are projected to have occurred, i.e. 82 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety follow-up by phone will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed (see Figure 2). An IDMC will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will also review formal interim analysis and recommend continuation or early termination of the study based on stopping criteria defined in the IDMC charter. 153

267 GM2008/00365/04 CONFIDENTIAL Figure 4 End of Study Strategy Including Unblinding Following Database Lock REVISED TEXT Figure 5 Study Schematic This is an event driven trial. The duration of the study will be determined by the rate of first occurrence of time to clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. The study will be terminated when it is projected that approximately events have occurred. Based on current assumptions, the median treatment duration is anticipated to be years with the total study duration estimated at 2.85 years. 154

268 GM2008/00365/04 CONFIDENTIAL All subjects are eligible to receive a minimum of 24 weeks of therapy. When adjudicated primary endpoint events are projected to have occurred, i.e subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety follow-up by phone will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed (see Figure 2). An IDMC will monitor the safety and welfare of the study subjects and will meet at designated intervals to review accumulated data. The IDMC will also review formal interim analysis and recommend continuation or early termination of the study based on stopping the criteria defined in the IDMC charter. Figure 6 End of Study Strategy Including Unblinding Following Database Lock Figure 2: Study Visit Strategy Primary analysis Sensitivity analysis Projected events Database Lock 1 (triggers unblinding of subjects) Database Lock 2 *Subjects with events who do not return within 28 days will not have their events counted in the primary analysis, but will in the sensitivity analysis. Subjects to return within 28* days to clinic Last subject Final Assessment Visit [all events to FAS visit included in primary analysis ] 4 weeks (standard timelines for ecrf between last subject last visit and database lock) Blinded Study Drug Subjects to return to clinic within 4* weeks for unblinding following database lock Last subject End of Study visit. Subject returns to clinic and is unblinded. Investigator prescribes standard care 4 weeks (30 day safety follow up after last dose of study medication) Last subject follow up phone call for safety 4 weeks (standard timelines for ecrf between last subject last visit and database lock) Standard of Care 155

269 GM2008/00365/04 CONFIDENTIAL Section 4.1: Number of Subjects PREVIOUS TEXT Planned estimated enrolment is 352 subjects (176 subjects per arm) with PAH. Approximately 125 investigational sites world-wide will be selected for participation in the study. REVISED TEXT Planned estimated enrolment is subjects (176 subjects per arm) with PAH (200 subjects in the combination arm and 200 in the monotherapy arm 100 subjects receiving ambrisentan and 100 subjects receiving tadalafil). Approximately investigational sites world-wide will be selected for participation in the study. Section 8.2.1: Sample size assumptions PREVIOUS TEXT This event-driven study requires 82 subjects with a clinical failure event in order to have 90% power with an overall type I error rate (alpha level) of 5%, accounting for two interim analyses as detailed in Section Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of 10% per year, i.e. an absolute difference of 10%. This leads to an approximate hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial. The alpha spending associated with the stopping criteria are described in Section Assuming a recruitment period of 65 weeks (based on a recruitment rate of 5.4 subjects per week) and a total study duration of 129 weeks the study requires 352 subjects (176 per treatment group). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 10% per treatment group at one year has been considered for this calculation. The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have 75% power for each comparison. For this calculation, it has been assumed that there are 55 clinical failure events for 264 subjects (176 on combination and 88 on each monotherapy), no interim analyses and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been assumed. Calculations were performed in EAST version 5.2, Cytel Inc, using simulated power. 156

270 GM2008/00365/04 CONFIDENTIAL Central randomisation is planned stratified by aetiology of PAH (IPAH/HPAH and Non- IPAH) and WHO Functional Class (II and III). Additionally, subjects will be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 176 ABS/TAD subjects and 88 PBO/TAD subjects yields approximately 86% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. Treatment effect and standard deviation were based on data from Phase 2 and Phase 3 clinical studies of ambrisentan. Sample size and final alpha level were calculated using nquery Advisor Version 6.0. Statistical Solutions, Cork, Ireland. Although the sample size was calculated using a 2-sided t-test, the specified analysis for this endpoint uses a Wilcoxon rank sum test; therefore, the actual power may vary slightly. REVISED TEXT This event-driven study requires subjects with a an adjudicated clinical failure event in order to have approximate 9097% power for the comparison of combination therapy with pooled monotherapy, and approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with an overall a type I error rate (alpha level) of 5% for each of the three comparisons. accounting for two interim analyses as detailed in Section Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to an approximate a hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial as defined in the IDMC Charter. The alpha spending associated with the stopping criteria are described in Section Assuming a recruitment period of 6582 weeks (based on a recruitment rate of subjects per week) and a total study duration of weeks the study requires eligible subjects (200 subjects in the combination arm and 200 in the monotherapy arm 100 subjects receiving ambrisentan and 100 subjects receiving tadalafil 172 per treatment group). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 10 5% per treatment group at one year has been considered for this calculation. The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have approximately 8575% power for each comparison. For this calculation, it has been assumed that there are 5570 clinical failure events for subjects ( on combination and on each monotherapy), no interim analysis and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been assumed. 157

271 GM2008/00365/04 CONFIDENTIAL Calculations were performed in EAST version 5.2, Cytel Inc. using simulated power Central randomisation is planned stratified by aetiology of PAH (IPAH/HPAH and Non- IPAH) and WHO Functional Class (II and III). Additionally, subjects will be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with ABS/TAD subjects and PBO/TAD subjects yields approximately 8696% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. Treatment effect and standard deviation were based on data from Phase 2 and Phase 3 clinical studies of ambrisentan. Sample size and final alpha level were calculated using nquery Advisor Version 6.0. Statistical Solutions, Cork, Ireland. Although the sample size was calculated using a 2-sided t-test, the specified analysis for this endpoint uses a Wilcoxon rank sum test; therefore, the actual power may vary slightly. Section 8.2.2: Sample size assumptions PREVIOUS TEXT The sample size required to achieve a given number of necessary primary endpoint events depends on the basic assumptions as outlined in Section The power (based on simulation) resulting from an observed 82 events for various reductions in risk different from the assumed 53% are shown in the table below. Table 3 Assumptions for Power Calculation Reduction in Risk Hazard Ratio Approximate Power 40% % 45% % 50% % 53% % 60% % 65% 0.35 >99% Assuming 352 subjects, 82 events, alpha of 0.05, 10% yearly dropout rate per group, and varying hazard ratios. If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at 352, the study duration may differ from the current planned 129 weeks (see Table 4.). 158

272 GM2008/00365/04 CONFIDENTIAL Table 4 Sample Size Sensitivity for 90% Power and 52.8% Reduction in Risk and Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n=352) 25% 106 weeks 20% 129 weeks 15% 171 weeks 10% 268 weeks Assuming 352 subjects, 82 events, alpha of 0.05, power of approximately 90%, 10% yearly combination therapy event rate, 10% yearly dropout rate per group and varying Monotherapy yearly event rates If the recruitment rate differs from the assumed recruitment rate of 5.4 subjects per week, the study duration may differ from the current planned 129 weeks (see Table 6). Table 5 Sample Size Sensitivity for 90% Power, 52.8% Reduction in Risk, 352 Subjects, and Varying Recruitment Rates Recruitment Period (Weeks) Recruitment Rate (pts/week) Follow-up Period after Last Subject Enrolled (Weeks) Total Study Duration (Weeks) Assuming 352 subjects total, 82 events, alpha of 0.05; power of approximately 90%, 10% yearly combination therapy event rate, 20% yearly monotherapy event rate,10% yearly dropout rate per group, and varying recruitment rates All calculations were performed using EAST versions 5.2, Cytel Inc, using simulated power. 159

273 GM2008/00365/04 REVISED TEXT CONFIDENTIAL The sample size required to achieve a given number of necessary primary endpoint events depends on the basic assumptions as outlined in Section The power (based on simulation) resulting from an observed 82 events for various reductions in risk different from the assumed 53% are shown in the table below. Power calculations for combination therapy vs the pooled monotherapy resulting from a range of treatment effects and monotherapy event rates are shown in the table below. Table 7 Assumptions for Power Calculation Reduction in Risk Hazard Ratio Approximate Power 40% % 45% % 50% % 53% % 60% % 65% 0.35 >99% Assuming 352 subjects, 82 events, alpha of 0.05, 10% yearly dropout rate per group, and varying hazard ratios. Table 3 Power Calculations for Varying Event Rates (Combination Therapy vs Pooled Monotherapy) Monotherapy Event Rate 15% 20% 25% 0.70 (30% ) 40% 50% 58% Hazard Ratio (relative reduction) 0.60 (40% ) 66% 78% 85% 0.50 (50% ) 88% 95% 98% 0.47 (53% ) 92% 97% 99% Assuming: 400 subjects (200 combination therapy; 200 pooled monotherapy [100 ambrisentan and 100 tadalafil]), 82 wks recruitment, 146 wks study duration (~2yrs 10 months) alpha of 0.05, 5% annual dropout rate per group and varying Monotherapy annual event rates and varying treatment effects 160

274 GM2008/00365/04 CONFIDENTIAL Simulations were run to determine the empirical probability of different study outcomes, based on the assumptions outlined in Section The results are presented in the table below. Table 4 Sample Size Sensitivity for 90% Power and 52.8% Reduction in Risk and Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n=352) 25% 106 weeks 20% 129 weeks 15% 171 weeks 10% 268 weeks Assuming 352 subjects, 82 events, alpha of 0.05, power of approximately 90%, 10% yearly combination therapy event rate, 10% yearly dropout rate per group and varying Monotherapy yearly event rates If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at 352, the study duration may differ from the current planned 129 weeks (see Table 4.). Table 4 Empirical Probabilities for Study Outcomes Based on Simulations Outcome Combination therapy better than pooled monotherapy Combination therapy better than an individual monotherapy Combination therapy better than pooled monotherapy and better than each individual monotherapy Empirical Probability 96% 88% 79% Assuming 400 subjects, 105 events, alpha of 0.05, 5% yearly dropout rate per group, 10% combination therapy yearly event rate, and 20% monotherapy yearly event rate (HR=0.047). Simulations were performed using R software. Note: Simulation results differ slightly from power calculations generated using EAST (large sample theory based method). If the recruitment rate differs from the assumed recruitment rate of 5.4 subjects per week, the study duration may differ from the current planned 129 weeks (see Table 6). If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at 400, the study duration may differ from the current planned 146 weeks (see Table 5). 161

275 GM2008/00365/04 CONFIDENTIAL Table 5 Study Duration Sensitivity for Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n=400) 25% 122 weeks 20% 146 weeks 15% 188 weeks Assuming 400 subjects, 105 events, alpha of 0.05, power of approximately 97% for combination vs pooled monotherapy, 53% Risk Reduction between groups, 5% yearly dropout rate per group and varying Monotherapy yearly event rates If the recruitment rate differs from the assumed recruitment rate of 4.9 subjects per week, the study duration may differ from the current planned 146 weeks (see Table 6). Table 11 Study Duration Sensitivity for Varying Recruitment Rates Recruitment Period (Weeks) Recruitment Rate (pts/week) Follow-up Period after Last Subject Enrolled (Weeks) Total Study Duration (Weeks) Assuming 400 subjects total, 105 events, alpha of 0.05; power of approximately 97%, 20% yearly monotherapy event rate, hazard ratio of 0.47, 5% yearly dropout rate per group, and varying recruitment rates Sample size calculations were performed using EAST versions 5.2, Cytel Inc. using simulated power. Simulations in Table 4 were performed using R software. Section 8.3.1: Analysis Populations PREVIOUS TEXT The populations to be used in the analysis of this study are: 162

276 GM2008/00365/04 CONFIDENTIAL ITT- All Randomised Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. PP-The per-protocol population will consist of the subset of subjects in the ITT without any major protocol violation. If the PP population is greater than 85% of the ITT population or less than 50% of the ITT population, a PP analysis will not be performed. REVISED TEXT The populations to be used in the analysis of this study are: ITT- All Randomised Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. PP-The per-protocol population will consist of the subset of subjects in the ITT without any major protocol violation. If the PP population is greater than 85% of the ITT population or less than 50% of the ITT population, a PP analysis will not be performed. Safety- As Treated Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment they received. Section : Primary comparisons of Interest PREVIOUS TEXT The primary comparison is of the hazard rates of clinical failure between the combination therapy arm (ambrisentan and tadalafil) and the pooled monotherapy arm (ambrisentan and placebo plus tadalafil and placebo). The ITT population will be utilized for this comparison. The comparison will be made at a value close to 5% significance level, adjusting for interim analyses using the alpha-spending approach detailed in Section REVISED TEXT The primary comparison is of the hazard rates of clinical failure between the combination therapy arm (ambrisentan and tadalafil) and the pooled monotherapy arm (ambrisentan and placebo plus tadalafil and placebo). The ITT population will be utilized for this comparison. The comparison will be made at a value close to 5% significance level., adjusting for interim analyses using the alpha spending approach detailed in Section

277 GM2008/00365/04 CONFIDENTIAL Section : Other Comparisons of Interest PREVIOUS TEXT The primary comparison will be between the combination and pooled monotherapy treatment arms. If the combination treatment arm is demonstrated as statistically significantly superior to the pooled monotherapy arm, comparisons of combination treatment versus each individual monotherapy may be performed. Alpha adjustments to accommodate the two monotherapy treatment comparisons will be implemented with a Hochberg approach. It is acknowledged that these comparisons are underpowered. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary endpoints. Inferences on the superiority of combination versus pooled monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint is statistically significant will inferences on the Week 24 change from baseline 6MWD endpoint be evaluated. Only if the Week 24 change from baseline 6MWD endpoint is found to be significant will inferences on the percentage of subjects with satisfactory response at Week 24 endpoint be evaluated. This gate keeping approach, described above, will be implemented for all secondary endpoints in the order listed. With this approach, no adjustment to the significance level at each step is needed. The primary, secondary and exploratory analyses may be repeated for the PP population. REVISED TEXT For each endpoint, Tthe primary comparison will be between the combination and pooled monotherapy treatment arms. Comparison of the combination therapy to the individual monotherapy arms is a secondary comparison. These comparisons will only be performed if the comparison of the combination vs. pooled monotherapy is significant. If the combination treatment arm is demonstrated as statistically significantly superior to the pooled monotherapy arm, comparisons of combination treatment versus each individual monotherapy may will be performed at a 5% significance level. Alpha adjustments to accommodate the two monotherapy treatment comparisons will be implemented with a Hochberg approach. It is acknowledged that these comparisons are underpowered. To minimise the likelihood of Type 1 errors, the following strategy will be implemented to address the multiple inferential tests among the primary and secondary comparisons and endpoints. Inferences on the superiority of combination versus pooled monotherapy treatment with respect to secondary endpoints will only be tested if the primary endpoint has been shown to demonstrate statistical significance for combination therapy vs. pooled monotherapy therapy. A step down procedure will then be adopted among the secondary endpoints. Only if the primary endpoint combination vs. pooled monotherapy comparison is statistically significant will inferences on the first secondary endpoint be evaluated. Only if the first secondary endpoint combination vs. pooled monotherapy comparison is found to be significant 164

278 GM2008/00365/04 CONFIDENTIAL will inferences on the subsequent secondary endpoint be evaluated. This gate keeping approach will be implemented for all secondary endpoints in pre-define order (defined below), and all tests will be performed at a 5% significance level. 1. Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP) 2. Percentage of subjects with satisfactory clinical response measured at week Change from baseline in 6MWD measured at week Change from baseline measured at week 24 in WHO Functional Class 5. Change from baseline measured at week 24 in Borg Dyspnea Index Additionally, for each endpoint, if the combination treatment arm is demonstrated as statistically significantly superior to the pooled monotherapy arm using the gate keeping approach described above, comparisons of combination treatment versus each individual monotherapy treatment will be performed at a 5% significance level. Week 24 change from baseline 6MWD endpoint be evaluated. Only if the Week 24 change from baseline 6MWD endpoint is found to be significant will inferences on the percentage of subjects with satisfactory response at Week 24 endpoint be evaluated. This gate keeping approach, described above, will be implemented for all secondary endpoints in the order listed. With this approach, no adjustment to the significance level at each step is needed. The primary, secondary and exploratory analyses may be repeated for the PP population. Section 8.3.4: Interim Analysis PREVIOUS TEXT The IDMC will monitor progress of the study, review safety data to protect subject safety and assess efficacy data against the pre-specified stopping rules identified in the protocol. An external statistician will produce the summaries and analyses required by the IDMC. Further details will be specified in the IDMC charter. Only the IDMC will be authorised to review unblinded interim efficacy and safety analyses during the study. Based on the interim analyses, the IDMC will give recommendations to the Sponsors and Chair of the Steering Committee regarding whether the trial should be stopped. Safety data will be monitored throughout the study on an ongoing basis. Formal interim analyses of efficacy data will be performed when approximately 50% and 75% of the expected number of primary endpoint events have been reached. Statistical stopping rules for superiority and futility based on O Brien Fleming will then be applied. 165

279 GM2008/00365/04 CONFIDENTIAL Table 12 p-values at each Interim Analysis Interim 1 Interim 2 Approximate Percentage of Events 50% 75% Average Time of Analysis (weeks) Probability of stopping for Superiority / Futility 25% / 1% 46% / 4% Efficacy p-value for superiority Futility p-value (non-binding) As the alpha spend associated with each interim is relatively small the final analysis will be performed at alpha= If the study is to be terminated early as a result of an interim analysis, a further additional sensitivity analysis will be performed including all subjects and blinded data collected, i.e. including additional blinded data that did not contribute to the interim analysis. Calculations were performed in EAST version 5.2, Cytel Inc, using simulated power. Further details of the IDMC will be provided in the IDMC Charter. REVISED TEXT There are no planned interim analysis of efficacy data The IDMC will monitor progress of the study, review safety data to protect subject safety and assess efficacy data against the pre specified stopping rules identified in the protocol. An external statistician will produce the summaries and analyses required by the IDMC. Further details will be specified in the IDMC charter. Only the IDMC will be authorised to review unblinded interim efficacy and safety analyses during the study. Based on the interim analyses, the IDMC will give recommendations to the Sponsors and Chair of the Steering Committee regarding whether the trial should be stopped. Safety data will be monitored throughout the study on an ongoing basis. Formal interim analyses of efficacy data will be performed when approximately 50% and 75% of the expected number of primary endpoint events have been reached. Statistical stopping rules for superiority and futility based on O Brien Fleming will then be applied. 166

280 GM2008/00365/04 CONFIDENTIAL Table 13 p values at each Interim Analysis Interim 1 Interim 2 Approximate Percentage of Events 50% 75% Average Time of Analysis (weeks) Probability of stopping for Superiority / Futility 25% / 1% 46% / 4% Efficacy p value for superiority Futility p value (non binding) As the alpha spend associated with each interim is relatively small the final analysis will be performed at alpha= If the study is to be terminated early as a result of an interim analysis, a further additional sensitivity analysis will be performed including all subjects and blinded data collected, i.e. including additional blinded data that did not contribute to the interim analysis. Calculations were performed in EAST version 5.2, Cytel Inc, using simulated power. Further details of the IDMC will be provided in the IDMC Charter. Section : Primary Analysis, second paragraph PREVIOUS TEXT The primary analysis will include all events (target of 82 events) up to the end of the Final Assessment Visit of the main study phase. Any events occurring after this timepoint will also be adjudicated and be reported as a sensitivity analysis or as a separate summary. REVISED TEXT The primary analysis will include all adjudicated events (target of events) up to the end of the Final Assessment Visit of the main study phase. Any events occurring after this timepoint will be reported as a sensitivity analysis or as a separate summary for that particular endpoint. Section : Secondary Analysis PREVIOUS TEXT A multiplicity adjustment will be applied for the secondary endpoints as detailed in Section

281 GM2008/00365/04 REVISED TEXT CONFIDENTIAL The gatekeeper approach described in Section will be applied for the secondary endpoints to protect against type I error. A multiplicity adjustment will be applied for the secondary endpoints as detailed in Section A complete list of sensitivity analyses will be defined in the RAP. Appendix 2: 6 Minute Walk Distance Test; Test Location PREVIOUS TEXT The Test Location Select a quiet, enclosed corridor, free of distractions and drafts. Depending on your site s standard operating procedures, please mark out a 30 m or 100 ft course. The start and end of the course must be visibly marked by placing tape on the floor, using chairs etc. The course should be sub-divided into 3m or 10ft sections using a method unnoticeable to the subject. REVISED TEXT The Test Location Select a quiet, enclosed corridor, free of distractions and drafts. Depending on your site s standard operating procedures, please mark out a 30 m or 100 ft 6MWT course. See Study Procedures Manual for further guidance. The start and end of the course must be visibly marked by placing tape on the floor, using chairs etc. The course should be sub-divided into 3m or 10ft sections using a method unnoticeable to the subject. 168

282 GM2008/00365/04 CONFIDENTIAL Appendix 10: Protocol Amendment 03 Changes AMENDMENT 3 Where the Amendment Applies This amendment is applicable to all countries and sites Summary of Amendment Changes with Rationale Change to Sponsor (GSK) Medical Monitor contact information. Change to a modified Intent to Treat and additional subject recruitment: A blinded review of subject baseline characteristics revealed a high proportion of subjects with borderline haemodynamics and/or significant Coronary Artery Disease risk factors. Changes to inclusion criteria were put in place in Amendment 02. However, a number of subjects have already been recruited that do not meet the Amendment 02 eligibility criteria. Expert advice is that the response to treatment is not fully characterised in subjects with multiple CAD risk-factors, and therefore a modified ITT is proposed on subjects meeting the amendment 02 eligibility criteria. To maintain adequate power and minimise study duration it is proposed that an additional 145 subjects be recruited and study duration increased by 13 weeks in order to have 105 events from 456 mitt subjects. The subjects who do not meet amendment 02 eligibility criteria will be assessed as part of an overall sensitivity analysis using the whole population as well as being assessed as a separate sub-group. Changes to sections: Study Summary (Study Design and Study Endpoints and Assessments), Section 3.1, Section 4.1, Section 8.2.1, Section 8.2.2, Section 8.3.1, Section , Section , Section Changes to the exclusion criteria regarding previous PAH therapies to clarify that the intent of the study is to recruit subjects who have not previously received chronic PAH specific medications to treat their PAH. Changes have been made to exclude subjects who have previously received greater than 14 days of continuous PAH treatment at any time prior to screening. Changes to Study Summary (Study Design and Study Endpoints and Assessments), Section 4.3 Exclusion Criteria 1. Change to Exclusion Criteria 10 regarding previous conditions requiring long or short acting nitrate use. The use of nitrates for conditions other than Angina Pectoris has been reduced from 12 weeks to 48 hours prior to screening. Coronary Artery Disease risk factors have been addressed in Amendment 02. For other conditions, use of nitrates greater than 48 hours prior to initiation of tadalafil is acceptable. Changes to the liver stopping criteria to allow the stopping of AMB/PBO only, not both IPs, if chosen by the investigator, in the event of a liver enzyme elevation. The change in implemented based on steering committee recommendation and agreed by the Sponsors that the risk:benefit of withdrawing both active treatments is greater than the risk:benefit of a subject with elevated liver enzymes continuing on tadalafil. However, should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. The investigator is reminded that only withdrawing AMB/PBO does 169

283 GM2008/00365/04 CONFIDENTIAL not guarantee the subjects will be receiving active treatment in the form of tadalafil as the subjects could be receiving placebo and subjects should be carefully monitored. Affects Section 4.4, and Section Changes to Visit Schedule and assessments for subjects discontinuing IP, but agreeing to remain in the study, to aid subject retention and gather outcome data for all patients. The visit schedule may be extended to 24 weeks, from 12 weeks and only assessments relevant to the primary endpoint or safety need to be collected. The 30 day after discontinuation of IP safety follow-up will be retained. It is still preferable, and subjects may continue to follow the full visit schedule if they agree. This change is made to encourage as many subjects to agree to be monitored in the study as possible. Change to Section 4.4: Withdrawal Criteria. List of Specific Changes Sponsor Information Page (GSK): Medical Monitor Contact Information PREVIOUS TEXT Sponsor Medical Monitor Contact Information: Dr Telephone: REVISED TEXT Sponsor Medical Monitor Contact Information: Dr : + Study Summary: Study Design (paragraphs 2 & 6) PREVIOUS TEXT Enrolment of 400 subjects is planned with 200 subjects in the combination arm and 200 in the monotherapy arm (100 subjects receiving ambrisentan and 100 subjects receiving tadalafil). Subjects must not have previously received chronic PAH therapy (i.e., prostanoids, ERAs, or PDE-5 inhibitors) within 4 weeks prior to Screening. REVISED TEXT Enrolment of total subjects is planned to provide 456 modified ITT subjects who meet the PAH Diagnosis and classification eligibility criteria defined in protocol amendment 02 with subjects in the combination arm and in the monotherapy arm ( subjects receiving ambrisentan and subjects receiving tadalafil). 170

284 GM2008/00365/04 CONFIDENTIAL Subjects must not have previously received PAH therapy continuously for 14 days or more chronic PAH therapy (i.e., prostanoids, ERAs, or PDE-5 inhibitors) within 4 weeks prior to Screening. Subjects who previously received PAH therapy for less than 14 days must not have received any PAH therapy within 7 days prior to Screening Visit. Study Summary: Study Endpoints and Assessments PREVIOUS TEXT The intent to treat (ITT) population will include all subjects randomised to treatment who receive at least one dose of Investigational Product, analyzed according to randomised treatment group. This will be the primary analysis population for assessing efficacy and safety. A per protocol (PP) population will also be defined and used to alternatively describe the primary and the most important secondary and tertiary efficacy endpoints. The PP population will consist of the subset of subjects contained in the ITT without any major protocol violation. If the PP population is greater than 85% or less than 50% of the ITT population, a PP analysis will not be performed. The CEC will provide adjudication of primary and secondary clinical events reported during the study. The members of this committee will be blinded to treatment assignment and investigator. This event-driven study requires 105 subjects with a clinical failure event in order to have approximately 97% power for the comparison of combination therapy with pooled monotherapy, and 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. The study will require 400 subjects (200 subjects in the combination arm and 200 in the monotherapy arm [100 subjects receiving ambrisentan and 100 subjects receiving tadalafil]) assuming a recruitment period of 82 weeks (based on an average recruitment rate of 4.9 subjects per week) and a total study duration of 146 weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 5% per treatment group at one year has been considered for this calculation. REVISED TEXT The modified intent to treat (mitt) population will include all subjects randomised to treatment who receive at least one dose of Investigational Product and who meet the PAH diagnosis and classification eligibility criteria defined in Protocol Amendment No. 2, analyzed according to randomised treatment group. This will be the primary analysis population for assessing efficacy and safety. A per protocol (PP) population will 171

285 GM2008/00365/04 CONFIDENTIAL also be defined and used to alternatively describe the primary and the most important secondary and tertiary efficacy endpoints. The PP population will consist of the subset of subjects contained in the mitt without any major protocol violation. If the PP population is greater than 85% or less than 50% of the mitt population, a PP analysis will not be performed. Additional sensitivity analyses of all randomised and dosed patients (ITT) and non-mitt patients will also be performed for the primary endpoint. Safety data will be summarised for the mitt, full ITT and non-mitt patients. The CEC will provide adjudication of primary and secondary clinical events reported during the study. The members of this committee will be blinded to treatment assignment and investigator. This event-driven study requires 105 subjects from the mitt population with a clinical failure event in order to have approximately 97% power for the comparison of combination therapy with pooled monotherapy, and 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. The study will require 400approximately 545 subjects to obtain 456 subjects in the mitt population ( subjects in the combination arm and in the monotherapy arm [ subjects receiving ambrisentan and subjects receiving tadalafil]) assuming a recruitment period of weeks (based on an average recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 5% per treatment group at one year has been considered for this calculation. 172

286 GM2008/00365/04 Section 3.1: Study Design PREVIOUS TEXT Figure 1 Study Schematic CONFIDENTIAL This Phase III/IV, randomised, double-blind study will compare the safety and efficacy of initiating combination therapy (with ambrisentan and tadalafil) or monotherapy (with ambrisentan or tadalafil) as first-line therapy in subjects with WHO FC II and III PAH. This is an event driven trial. The duration of the study will be determined by the rate of first occurrence of time to clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. The study will be terminated when it is projected that approximately 105 events have occurred. Based on current assumptions, the median treatment duration is anticipated to be 2 years with the total study duration estimated at 2.8 years. All subjects are eligible to receive a minimum of 24 weeks of therapy. When 105 adjudicated primary endpoint events are projected to have occurred, i.e. 105 subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety follow-up by phone will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed (see Figure 2). 173

287 GM2008/00365/04 REVISED TEXT Figure 1 Study Schematic CONFIDENTIAL This Phase III/IV, randomised, double-blind study will compare the safety and efficacy of initiating combination therapy (with ambrisentan and tadalafil) or monotherapy (with ambrisentan or tadalafil) as first-line therapy in subjects with WHO FC II and III PAH. This is an event driven trial. The duration of the study will be determined by the rate of first occurrence of time to clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. The study will be terminated when it is projected that approximately 105 events in the mitt population have occurred. Based on current assumptions, the median treatment duration is anticipated to be years with the total study duration estimated at years. All subjects are eligible to receive a minimum of 24 weeks of therapy. When 105 adjudicated primary endpoint events are projected to have occurred, i.e. 105 mitt subjects with an event, all subjects will be notified to return for a Final Assessment Visit (FAV) within 28 days or their week 24 visit, whichever is later. Subjects will continue to receive blinded treatment until the first database lock, after which subjects will return for an End of Study (EOS) visit within 4 weeks of sponsor notification. Events which occur after the FAV will not be used in the primary analysis, but may be considered for various sensitivity analyses. At this EOS visit, subjects will be unblinded and treated per the discretion of the Investigator. A safety follow-up by phone will be performed 30 days after the subjects last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed (see Figure 2). Section 4.1: Number of Subjects PREVIOUS TEXT Planned estimated enrolment is 400 subjects with PAH (200 subjects in the combination arm and 200 in the monotherapy arm [100 subjects receiving ambrisentan and 100 subjects receiving tadalafil]). 174

288 GM2008/00365/04 REVISED TEXT CONFIDENTIAL Planned enrolment is 400 approximately 545 subjects to provide 456 mitt subjects with PAH i.e. meeting the inclusion/exclusion criteria defined in Protocol Amendment No.2 ( subjects in the combination arm and in the monotherapy arm [ subjects receiving ambrisentan and subjects receiving tadalafil]). Section 4.3: Exclusion Criteria; 1 & 10 PREVIOUS TEXT - Subject received previous PAH therapy (PDE5i, ERA, chronic prostanoid*) within 4 weeks prior to the screening visit *Chronic prostanoid use is considered >7 days of treatment Other Therapies 10. Subject has a history of angina pectoris or other condition that was treated with long or short-acting nitrates <12 weeks of screening REVISED TEXT PAH Treatments 1. Subject has received previous PAH therapy continuously for 14 days or more (PDE5i, ERA, chronic prostanoid). Subjects who previously received PAH therapy for less than 14 days must not have received any PAH therapy within 7 days prior to Screening Visit. within 4 weeks prior to the screening visit has received previous PAH therapy continuously for 14 days or more *Chronic prostanoid use is considered >7 days of treatment Other Therapies 10. Subject has a history of angina pectoris or other condition that was treated with long or short-acting nitrates <12 weeks of screening or nitrate use for any other condition within 48 hours of screening Section 4.4: Withdrawal Criteria; paragraphs 2 & 5 PREVIOUS TEXT If a subject discontinues Investigational Product prematurely, the subject should be encouraged to continue all study Clinic Visits according to the planned schedule of events, after completing the Change of Randomised Treatment Visit. Possible reasons for subject discontinuation from IP include, but are not limited to, the following: 175

289 GM2008/00365/04 CONFIDENTIAL Development of an AE where continuation of the subject s participation in the study is thought by the Investigator to be inappropriate Subject meets liver stopping Subject begins treatment with a prohibited concomitant therapy Lost to follow-up REVISED TEXT If a subject discontinues Investigational Product prematurely, the subject should be encouraged to continue all study Clinic Visits according to the planned schedule of events, after completing the Change of Randomised Treatment Visit. If, following discontinuation of IP, a subject does not wish to continue with the burden of all Clinic Visits and assessments, it is acceptable to continue in the study with a reduced clinic visit schedule and set of assessments. The subject may perform 24 weekly clinic visits and only the key information required to assess the TTCF primary EP (TTCF, 6MWD, WHO FC), safety (AEs. SAE s) and concomitant medications need to be collected. The safety follow-up call 30 days after discontinuation of IP should still be completed. Possible reasons for subject discontinuation from IP include, but are not limited to, the following: Development of an AE where continuation of the subject s participation in the study is thought by the Investigator to be inappropriate Subject meets liver stopping criteria and discontinues both Investigational products Subject begins treatment with a prohibited concomitant therapy Lost to follow-up 176

290 GM2008/00365/04 CONFIDENTIAL Section 6.3.4: Management, monitoring and follow-up of serum aminotransferase abnormalities; Table 2 PREVIOUS TEXT Table 2 Management, Monitoring and Follow-up of Subjects Experiencing Serum Aminotransferase Elevations Serum Aminotransferase Elevations ALT>3xULN and bilirubin >2xULN or if ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. ALT/AST >3x and 5xULN ALT/AST >5x and 8xULN Actions to be Taken Stop investigational product treatment immediately. Report the liver event to the Medical Monitor within 24 hours of learning of its occurrence and report as an SAE. Make every reasonable attempt to have subjects return to clinic within 24 hours for repeat liver chemistries, liver event follow up assessments, and close monitoring Complete liver event CRF within 24 hours. After discontinuation of investigational product, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. A specialist or hepatology consultation is recommended Investigational product may not be reintroduced. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. If confirmed, interrupt therapy. Continue to monitor serum aminotransferase concentrations every 2 weeks or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of investigational product may be considered by the Investigator after consultation with the Medical Monitor. Stop investigational product treatment immediately. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. Continue to monitor serum aminotransferase concentrations every 2 weeks, or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of investigational product may be considered by the Investigator after consultation with the Medical Monitor. 177

291 GM2008/00365/04 CONFIDENTIAL Serum Aminotransferase Actions to be Taken Elevations ALT/AST >8xULN Stop investigational product treatment immediately. Report the liver event to the Medical Monitor and within 24 hours of learning of its occurrence. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. After discontinuation of investigational product, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. Investigational product may not be reintroduced. *If investigational product is reintroduced, serum aminotransferase concentrations should be checked within 3 days, and then every 2 weeks or more frequently if clinically indicated. Once investigator, in consultation with Medical Monitor, feels that subject is clinically stable, liver safety testing may be resumed at standard protocol defined intervals. If at any time these subjects experience serum aminotransferase elevations, proceed as described above. REVISED TEXT Table 2 Management, Monitoring and Follow-up of Subjects Experiencing Serum Aminotransferase Elevations Serum Aminotransferase Elevations ALT>3xULN and bilirubin >2xULN or if ALT >3xULN and associated with the appearance or worsening of symptoms of hepatitis or hypersensitivity such as fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash or eosinophilia. Actions to be Taken Stop investigational product AMB/PBO treatment immediately. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Report the liver event to the Medical Monitor within 24 hours of learning of its occurrence and report as an SAE. Make every reasonable attempt to have subjects return to clinic within 24 hours for repeat liver chemistries, liver event follow up assessments, and close monitoring Complete liver event CRF within 24 hours. After discontinuation of investigational product AMB/PBO, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. The subject should be followed until the serum aminotransferase concentrations decrease below 3xULN. A specialist or hepatology consultation is recommended Investigational product AMB/PBO may not be reintroduced. 178

292 GM2008/00365/04 CONFIDENTIAL Serum Aminotransferase Elevations ALT/AST >3x and 5xULN ALT/AST >5x and 8xULN Actions to be Taken Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. If confirmed, interrupt Investigational product AMB/PBO therapy. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Continue to monitor serum aminotransferase concentrations every 2 weeks or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of Investigational product AMB/PBO therapy may be considered by the Investigator after consultation with the Medical Monitor. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. Stop Investigational product AMB/PBO treatment immediately. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. Continue to monitor serum aminotransferase concentrations every 2 weeks, or more frequently if clinically indicated. If the serum aminotransferase concentrations decrease below 3xULN, reintroduction* of investigational product AMB/PBO may be considered by the Investigator after consultation with the Medical Monitor. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. ALT/AST >8xULN Stop Investigational product AMB/PBO treatment immediately. Subjects may remain on TAD/PBO if clinically indicated. It should be noted that if TAD/PBO is continued the subject may be only receiving placebo, subjects should therefore be carefully monitored and consideration of additional therapy should be undertaken at the discretion of the investigator. Report the liver event to the Medical Monitor and within 24 hours of learning of its occurrence. Complete liver event CRF within 24 hours. Confirm by another serum aminotransferase test within 7 days. After discontinuation of Investigational product AMB/PBO, continue to monitor serum aminotransferase concentrations weekly, or more frequently if clinically indicated. The subject should be followed until the serum aminotransferase 179

293 GM2008/00365/04 CONFIDENTIAL Serum Aminotransferase Actions to be Taken Elevations concentrations decrease below 3xULN. Investigational product AMB/PBO therapy may not be reintroduced. Should acute liver failure develop or adverse events consistent with tadalafil toxicity become clinically apparent, consideration of cessation of tadalafil should be undertaken and discussed with the medical monitor. *If Investigational product AMB/PBO therapy is reintroduced, serum aminotransferase concentrations should be checked within 3 days, and then every 2 weeks or more frequently if clinically indicated. Once investigator, in consultation with Medical Monitor, feels that subject is clinically stable, liver safety testing may be resumed at standard protocol defined intervals. If at any time these subjects experience serum aminotransferase elevations, proceed as described above. Section 6.3.6: Laboratory and Other Safety Assessment Abnormalities Reported as AEs and SAEs; Paragraph 3 PREVIOUS TEXT All events of possible drug-induced liver injury with hyperbilirubinaemia (defined as ALT >3xULN plus bilirubin >2xULN and/or INR>1.5) or Hy s Law events, require immediate Investigational Product cessation and reporting as an SAE (see Section above). REVISED TEXT All events of possible drug-induced liver injury with hyperbilirubinaemia (defined as ALT >3xULN plus bilirubin >2xULN and/or INR>1.5) or Hy s Law events, require immediate Investigational Product AMB/PBO cessation and reporting as an SAE (see Section above). Section 8.2.1: Sample Size Assumptions PREVIOUS TEXT This event-driven study requires 105 subjects with an adjudicated clinical failure event in order to have approximate 97% power for the comparison of combination therapy with pooled monotherapy, and approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial as defined in the IDMC Charter. 180

294 GM2008/00365/04 CONFIDENTIAL Assuming a recruitment period of 82 weeks (based on a recruitment rate of 4.9 subjects per week) and a total study duration of 146 weeks the study requires 400 eligible subjects (200 subjects in the combination arm and 200 in the monotherapy arm [100 subjects receiving ambrisentan and 100 subjects receiving tadalafil]). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 5% per treatment group at one year has been considered for this calculation. The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have approximately 85% power for each comparison. For this calculation, it has been assumed that there are 70 clinical failure events for 300 subjects (200 on combination and 100 on each monotherapy) and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been assumed. Calculations were performed in EAST version 5.2, Cytel Inc. Central randomisation is planned stratified by aetiology of PAH (IPAH/HPAH and Non- IPAH) and WHO Functional Class (II and III). Additionally, subjects will be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 200 ABS/TAD subjects and 100 PBO/TAD subjects yields approximately 96% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. REVISED TEXT This event-driven study requires 105 mitt subjects with an adjudicated clinical failure event in order to have approximate 97% power for the comparison of combination therapy with pooled monotherapy, and approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. The IDMC will review data periodically throughout the trial as defined in the IDMC Charter. 181

295 GM2008/00365/04 CONFIDENTIAL Assuming a recruitment period of weeks (based on a recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 159 weeks the study requires approximately 545 subjects enrolled to obtain 400 eligible 456 mitt subjects ( subjects in the combination arm and in the monotherapy arm [ subjects receiving ambrisentan and subjects receiving tadalafil]). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 5% per treatment group at one year has been considered for this calculation. The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have approximately 85% power for each comparison. For this calculation, it has been assumed that there are approximately 70 clinical failure events for mitt subjects ( on combination and on each monotherapy) and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been assumed. Calculations were performed in EAST version 5.2, Cytel Inc. Central randomisation is planned stratified by aetiology of PAH (IPAH/HPAH and Non- IPAH) and WHO Functional Class (II and III). Additionally, subjects will be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with ABS/TAD subjects and PBO/TAD subjects yields approximately 96 97% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. The primary analysis for the peak/trough assessment will be in the mitt population; but the analysis may be repeated for the ITT population and non-mitt population. These analyses will be defined further in the RAP prior to unblinding. Section 8.2.2: Sample Size Sensitivity PREVIOUS TEXT The sample size required to achieve a given number of necessary primary endpoint events depends on the basic assumptions as outlined in Section Power calculations for combination therapy vs the pooled monotherapy resulting from a range of treatment effects and monotherapy event rates are shown in the table below. 182

296 GM2008/00365/04 CONFIDENTIAL Table 4 Power Calculations for Varying Event Rates (Combination Therapy vs Pooled Monotherapy) Monotherapy Event Rate 15% 20% 25% 0.70 (30% ) 40% 50% 58% Hazard Ratio (relative reduction) 0.60 (40% ) 66% 78% 85% 0.50 (50% ) 88% 95% 98% 0.47 (53% ) 92% 97% 99% Assuming: 400 subjects (200 combination therapy; 200 pooled monotherapy [100 ambrisentan and 100 tadalafil]), 82 wks recruitment, 146 wks study duration (~2yrs 10 months) alpha of 0.05, 5% annual dropout rate per group and varying Monotherapy annual event rates and varying treatment effects Simulations were run to determine the empirical probability of different study outcomes, based on the assumptions outlined in Section The results are presented in the table below. Table 5 Empirical Probabilities for Study Outcomes Based on Simulations Outcome Combination therapy better than pooled monotherapy Combination therapy better than an individual monotherapy Combination therapy better than pooled monotherapy and better than each individual monotherapy Empirical Probability 96% 88% 79% Assuming 400 subjects, 105 events, alpha of 0.05, 5% yearly dropout rate per group, 10% combination therapy yearly event rate, and 20% monotherapy yearly event rate (HR=0.047). Simulations were performed using R software. Note: Simulation results differ slightly from power calculations generated using EAST (large sample theory based method). If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at 400, the study duration may differ from the current planned 146 weeks (see Table 5). 183

297 GM2008/00365/04 CONFIDENTIAL Table 6 Study Duration Sensitivity for Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n=400) 25% 122 weeks 20% 146 weeks 15% 188 weeks Assuming 400 subjects, 105 events, alpha of 0.05, power of approximately 97% for combination vs pooled monotherapy, 53% Risk Reduction between groups, 5% yearly dropout rate per group and varying Monotherapy yearly event rates If the recruitment rate differs from the assumed recruitment rate of 4.9 subjects per week, the study duration may differ from the current planned 146 weeks (see Table 6). Table 7 Study Duration Sensitivity for Varying Recruitment Rates Recruitment Period (Weeks) Recruitment Rate (pts/week) Follow-up Period after Last Subject Enrolled (Weeks) Total Study Duration (Weeks) Assuming 400 subjects total, 105 events, alpha of 0.05; power of approximately 97%, 20% yearly monotherapy event rate, hazard ratio of 0.47, 5% yearly dropout rate per group, and varying recruitment rates Sample size calculations were performed using EAST versions 5.2, Cytel Inc. Simulations in Table 4 were performed using R software. REVISED TEXT The sample size required to achieve a given number of necessary primary endpoint events depends on the basic assumptions as outlined in Section Power calculations for combination therapy vs the pooled monotherapy resulting from a range of treatment effects and monotherapy event rates are shown in the table below. 184

298 GM2008/00365/04 CONFIDENTIAL Table 3 Power Calculations for Varying Event Rates (Combination Therapy vs Pooled Monotherapy) Monotherapy Event Rate 15% 20% 25% 0.70 (30% ) 40% 50% 58% Hazard Ratio (relative reduction) 0.60 (40% ) 66% 78% 85% 0.50 (50% ) 88% 95% 98% 0.47 (53% ) 92% 97% 99% Assuming: subjects ( combination therapy; pooled monotherapy [ ambrisentan and tadalafil]), wks recruitment, wks study duration (~2 3yrs 10 2 months) alpha of 0.05, 5% annual dropout rate per group and varying Monotherapy annual event rates and varying treatment effects Simulations were run to determine the empirical probability of different study outcomes, based on the assumptions outlined in Section The results are presented in the table below. Table 4 Empirical Probabilities for Study Outcomes Based on Simulations Outcome Combination therapy better than pooled monotherapy Combination therapy better than an individual monotherapy Combination therapy better than pooled monotherapy and better than each individual monotherapy Empirical Probability 96% 88% 79% Assuming mitt subjects, with105 events, alpha of 0.05, 5% yearly dropout rate per group, 10% combination therapy yearly event rate, and 20% monotherapy yearly event rate (HR=0.047). Simulations were performed using R software. Note: Simulation results differ slightly from power calculations generated using EAST (large sample theory based method). If the annual monotherapy event rate differs from the assumed 20% but the sample size remains at , the study duration may differ from the current planned weeks (see Table 4) 185

299 GM2008/00365/04 CONFIDENTIAL Table 5 Study Duration Sensitivity for Varying Monotherapy Annual Event Rates Monotherapy Annual Event Rate Total Study Duration (n= mitt) 25% weeks 20% weeks 15% weeks Assuming mitt subjects, with 105 events, alpha of 0.05, power of approximately 97% for combination vs pooled monotherapy, 53% Risk Reduction between groups, 5% yearly dropout rate per group and varying Monotherapy yearly event rates If the recruitment rate differs from the assumed recruitment rate of 4.9 subjects per week, the study duration may differ from the current planned 146 weeks (see Table 5) Table 5 Study Duration Sensitivity for Varying Recruitment Rates Recruitment Period (Weeks) Recruitment Rate (pts/week) Follow up Period after Last Subject Enrolled (Weeks) Total Study Duration (Weeks) Assuming 400 subjects total, 105 events, alpha of 0.05; power of approximately 97%, 20% yearly monotherapy event rate, hazard ratio of 0.47, 5% yearly dropout rate per group, and varying recruitment rates Sample size calculations were performed using EAST versions 5.2, Cytel Inc. Simulations in Table 4 were performed using R software. Section 8.3.1: Analysis Population PREVIOUS TEXT The populations to be used in the analysis of this study are: 186

300 GM2008/00365/04 CONFIDENTIAL ITT- All Randomised Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. PP-The per-protocol population will consist of the subset of subjects in the ITT without any major protocol violation. If the PP population is greater than 85% of the ITT population or less than 50% of the ITT population, a PP analysis will not be performed. Safety- As Treated Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment they received. REVISED TEXT The populations to be used in the analysis of this study are: mitt Population: All randomised subjects who meet the inclusion/exclusion criteria defined in Protocol Amendment No. 2 and who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. ITT- All Randomised Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. PP-The per-protocol population will consist of the subset of subjects in the mitt without any major protocol violation. If the PP population is greater than 85% of the mitt population or less than 50% of the mitt population, a PP analysis will not be performed. Safety- As Treated Population: All randomised subjects who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment they received. An As Treated population will be defined to correspond to each of the mitt, ITT and non-mitt populations. Non-mITT Population: All randomised subjects who fail to meet the inclusion/exclusion criteria defined in Protocol Amendment No. 2 and who also received at least one dose of Investigational Product. Subjects will be analyzed according to the treatment group to which they were randomised. Section : Primary Comparisons of Interest PREVIOUS TEXT The primary comparison is of the hazard rates of clinical failure between the combination therapy arm (ambrisentan and tadalafil) and the pooled monotherapy arm (ambrisentan and placebo plus tadalafil and placebo). The ITT population will be utilized for this comparison. The comparison will be made at a 5% significance level. 187

301 GM2008/00365/04 REVISED TEXT CONFIDENTIAL Since the primary objective of this study is to compare the difference between two treatment strategies (first-line combination therapy (ambrisentan and tadalafil) versus first-line monotherapy (ambrisentan or tadalafil) in subjects with PAH), the primary comparison is of the hazard rates of clinical failure between the combination therapy arm (ambrisentan and tadalafil) and the pooled monotherapy arm (ambrisentan and placebo plus tadalafil and placebo). The mitt population will be utilized for this comparison. The comparison will be made at a 5% significance level. PREVIOUS TEXT For each endpoint, the primary comparison will be between the combination therapy arm and the pooled monotherapy arms. The primary, secondary and exploratory analyses may be repeated for the PP population. REVISED TEXT For each endpoint, the primary comparison will be between the combination therapy arm and the pooled monotherapy arms, and the primary analysis will use the mitt population. The primary, secondary and exploratory analyses may be repeated for the ITT population, PP population and non-mitt population. These analyses will be defined further in the RAP prior to unblinding. PREVIOUS TEXT The primary analysis will include all adjudicated events (target of 105 events) up to the end of the Final Assessment Visit of the main study phase. Any events occurring after this timepoint will also be adjudicated and be reported as a sensitivity analysis or as a separate summary for that particular endpoint. REVISED TEXT The primary analysis will include all adjudicated events in the mitt population (target of 105 events) up to the end of the Final Assessment Visit of the main study phase. Any events occurring after this timepoint will also be adjudicated and be reported as a sensitivity analysis or as a separate summary for that particular endpoint. 188

302 GM2008/00365/04 CONFIDENTIAL Appendix 11: Protocol Amendment 04 Changes AMENDMENT 4 Where the Amendment Applies This amendment is applicable to all countries and sites Summary of Amendment Changes with Rationale Changes to anticipated sample size: The anticipated sample size for the study is increased from 545 subjects, to 614 subjects. This is due to a lower than anticipated event rate. The wording is also changed to make it clear that study aims to recruit subjects up to 24 weeks prior to the predicated 105 event. Therefore the study duration and sample size are variable based on the event rate. Based on current event rate 614 subjects are required to produce 105 events and continue to recruit to 24 weeks prior to predicted 105 th event. This may change as the event rate changes, to require fewer or more than 614 subjects (up to a maximum of 680 subjects). This flexibility if proposed to be added to the protocol to avoid multiple further protocol amendments. Affects sections Protocol Summary, Section 3.1 [Study Design], Section 4.1 [Sample Size], [Sample Size Estimations], and [Sample Size Restimation]. Time to Clinical Failure, fourth component: The fourth component of the primary EP, [unsatisfactory long-term clinical response] wording has been amendment slightly to make sure it is fully in line with the ITT principle, and therefore change to wording to clarify that subjects are eligible if received at least one dose of study medication, and in the study for 6 months or more (previously wording was that the subject needed to be on randomized treatment for 6 months). Affects sections Protocol Summary, Section 3.1 [Study Design] and Section [Primary Endpoint]. List of Specific Changes Protocol Summary; Study Design, Paragraph 2 The study is event driven and therefore enrolment and study duration are dependent on study event rate. Enrolment of subjects will continue up until 24 weeks prior to the anticipated 105 th first event in the mitt population (up to a maximum of 680 subjects). Currently of total subjects is planned to provide modified ITT subjects who meet the PAH Diagnosis and classification eligibility criteria defined in protocol amendment 02 with subjects in the combination arm and in the monotherapy arm ( subjects receiving ambrisentan and subjects receiving tadalafil). Protocol Summary; Study Endpoints/Assessments, Primary EP The primary efficacy endpoint is the time (days) to the first clinical failure event of PAH. Time to clinical failure (TtCF) is defined as the time from randomisation to the first occurrence of: 189

303 GM2008/00365/04 CONFIDENTIAL Death (all-cause) Hospitalization for worsening PAH (adjudicated) Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO class III symptoms assessed at two clinic visits separated by 6 months Time to clinical worsening (death, hospitalisation for worsening PAH, or disease progression) and long-term survival (time to death) will be examined as supportive analyses of the primary endpoint. Protocol Summary; Study Endpoints/Assessments, Sample Size This event-driven study requires 105 subjects from the mitt population with a clinical failure event in order to have approximately 97% power for the comparison of combination therapy with pooled monotherapy, and 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Assumptions of the event rates are based on Steering Committee recommendations and are were originally projected as a monotherapy event rate of 20% per year and a combination event rate of approximately10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. Following a blinded review of event rate after approximately 2 years of recruitment, the number of estimated adjudicated events (adjudicated events plus events pending adjudication adjusted for concordance) was approximately 77% of the predicted overall event rate of 15%. The overall event rate was re-estimated to be approximately 12%, and a combination event rate of 8% and a monotherapy event rate of 16% were assumed to maintain the reduction in the hazard of 53% (see Section 8.2.3). Enrolment of subjects will continue up until 24 prior to the anticipated 105 th first event in the mitt population (up to a maximum of 680 subjects). Based on current blinded overall event rate estimate the study will require approximately

304 GM2008/00365/04 CONFIDENTIAL subjects to obtain subjects in the mitt population ( subjects in the combination arm and in the monotherapy arm [ subjects receiving ambrisentan and subjects receiving tadalafil]) assuming a recruitment period of weeks (based on an average recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of weeks. Every possible effort will be made to follow all subjects until the required number of events has been observed, unless the subject withdraws consent to be followed. To account for the loss of subjects, a drop-out rate of 5% per treatment group at one year has been considered for this calculation. Section 3.1: Study Design; Study Schematic, Sample size and primary EP This is an event driven trial. The duration of the study and sample size will be determined by the rate of first occurrence of time to clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. Recruitment will continue up to 24 weeks prior to the predicated 105 th event in the mitt population, and is anticipated to be subjects based on the current blinded event rate estimate. Should the estimated event rate increase over projections, recruitment will be stopped prior to 614 subjects, or should it reduce, may continue past 614 subjects, up to a maximum of 680. The study will be terminated when it is projected that approximately 105 events in the mitt population have occurred. Based on current assumptions, the median treatment duration is anticipated to be years with the total study duration estimated at 3.14 years. The primary efficacy endpoint is the time to the first clinical failure event of PAH. Time to clinical failure is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy 191

305 GM2008/00365/04 CONFIDENTIAL Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO class III symptoms assessed at two clinic visits separated by 6 months Section 4.1: Number of Subjects Planned enrolment is approximately subjects to provide mitt subjects with PAH i.e. meeting the inclusion/exclusion criteria defined in Protocol Amendment No.2 ( subjects in the combination arm and in the monotherapy arm [ subjects receiving ambrisentan and subjects receiving tadalafil]). Since this is an event-driven study, the final number of subjects will depend on the rate of clinical failure events and may be increased to a maximum of 680 subjects. Approximately 150 investigational sites world-wide will be selected for participation in the study. Section : Primary Endpoint; Time to Clinical Failure The primary efficacy endpoint is the time to the first clinical failure event of PAH. Time to clinical failure is defined as the time from randomisation to the first occurrence of: Death (all-cause) Hospitalization for worsening PAH (adjudicated) Any hospitalization for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) >15% decrease from baseline in 6MWD combined with WHO class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months 192

306 GM2008/00365/04 CONFIDENTIAL A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO class III symptoms assessed at two clinic visits separated by 6 months Section 8.2.1: Sample Size Assumptions, paragraphs 2, 3, 4 and 6 This section presents sample size calculations based on an overall event rate of 15% (a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year). Based on the current blinded event rate estimate, the sample size has been adjusted. Section presents additional sample size calculations based on an overall event rate of 12% (a monotherapy event rate of 16% per year and a combination event rate of approximately 8% per year). Section 8.2.3: Sample Size Restimation Following a blinded review of the overall event rate after approximately 2 years of recruitment, the number of estimated adjudicated events (adjudicated events plus events pending adjudication adjusted for concordance) was approximately 77% of the predicted overall event rate of 15%. The overall event rate was re-estimated to be approximately 12%. Sample size calculations were performed assuming a combination event rate of 8% and a monotherapy event rate of 16% to maintain the original assumed reduction in the hazard of 53%. Assuming a recruitment period of 148 weeks (based on a recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 175 weeks, the study requires approximately 614 subjects enrolled to obtain 520 mitt subjects (260 subjects in the combination arm and 260 in the monotherapy arm [130 subjects receiving ambrisentan and 130 subjects receiving tadalafil]). Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a dropout rate of 5% per treatment group per year has been considered for this calculation. Based on the revised sample size of 520 mitt patients, 105 mitt subjects with an adjudicated clinical failure event are required in order to maintain the same approximate 97% power for the comparison of combination therapy with pooled monotherapy, and the same approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Calculations were performed in EAST version 5.2, Cytel Inc. 193

307 GM2008/00365/04 CONFIDENTIAL To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 260 ABS/TAD subjects and 130 PBO/TAD subjects yields approximately 98% power to detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. As this is an event driven trial, the duration of the study and sample size will be determined by the rate of clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. Recruitment will continue up to 24 weeks prior to the predicted 105 th event in the mitt population, and up to a maximum of 680 subjects. 194

308 CONFIDENTIAL The GlaxoSmithKline group of companies Division: World Wide Development Retention Category: GRS019 Information Type: Reporting and Analysis Plan Title: Reporting and Analysis Plan for AMBITION (AMB112565/ GS-US ): A Randomised, Double-Blind, Multicenter Study of First-Line Ambrisentan and Tadalafil Combination Therapy in Subjects with Pulmonary Arterial Hypertension Compound Number: Effective Date: GSK MAY-2014 Description: A reporting and analysis plan to describe the planned analyses for a randomised multicenter study comparing the effect on time to clinical failure of initial combination therapy (ambrisentan and tadalafil) and initial monotherapy (ambrisentan or tadalafil) in subjects with pulmonary arterial hypertension Subject: AMB112565, Analysis Plan, Pulmonary Arterial Hypertension Author s Name, Title and Functional Area: : Biostatistician, Hartington Statistics and Data Management Ltd : Managing Director, Hartington Statistics and Data Management Ltd Statistical Programmer, Hartington Statistics and Data Management Ltd Approved by: Director, Biostatistics and Programming GlaxoSmithKline Manager, Statistics GlaxoSmithKline Senior Manager, Biostatistics Gilead Sciences Date: 09-MAY-2014 Date: 09-MAY-2014 Date: 09-MAY-2014 Copyright 2014 the GlaxoSmithKline group of companies. All rights reserved. Unauthorised copying or use of this information is prohibited. 1

309 CONFIDENTIAL TABLE OF CONTENTS PAGE 1. INTRODUCTION STUDY OBJECTIVE(S) AND OUTCOME(S) STUDY OBJECTIVE(S) PRIMARY OBJECTIVE SECONDARY OBJECTIVES STUDY OUTCOME(S) PRIMARY OUTCOME SECONDARY OUTCOMES EXPLORATORY OUTCOMES HEALTH OUTCOMES SAFETY OUTCOMES STATISTICAL HYPOTHESES STUDY DESIGN PLANNED ANALYSES CHANGES FROM PROTOCOL SPECIFIED ANALYSES INTERIM ANALYSES FINAL ANALYSIS SAMPLE SIZE CONSIDERATIONS SAMPLE SIZE ASSUMPTIONS SAMPLE SIZE SENSITIVITY SAMPLE SIZE RE-ESTIMATION ANALYSIS POPULATIONS TREATMENT COMPARISONS GENERAL CONSIDERATIONS FOR DATA ANALYSES MULTICENTRE STUDIES

310 CONFIDENTIAL 8.2. RANDOMISATION OTHER STRATA AND COVARIATES STRATA COVARIATES SUBGROUPS MULTIPLE COMPARISONS AND MULTIPLICITY STATISTICAL METHODOLOGIES ANCOVA COCHRAN MANTEL HAENSZEL COX REGRESSION LOG TRANSFORMATIONS LOGISTIC REGRESSION MIXED MODELS REPEATED MEASURES (MMRM) WILCOXON RANK SUM TEST DATA HANDLING CONVENTIONS PREMATURE WITHDRAWAL AND MISSING DATA DISCONTINUATION OF IP STUDY COMPLETION MISSING PRIMARY OUTCOME DATA MISSING SECONDARY OUTCOME DATA MISSING ITEMS IN HEALTH OUTCOME DATA MISSING AE DATA MISSING DATES (OTHER THAN FOR AES) DERIVED AND TRANSFORMED DATA ADJUDICATION AGE AGE GROUP

311 CONFIDENTIAL BASELINE BCT INITIATION BMI CAMPHOR CHANGE FROM BASELINE AND % CHANGE FROM BASELINE CLINICAL FAILURE EVENT CLINICAL WORSENING EVENT CUMULATIVE DOSE EXPOSURE MITT PRIOR AND CONCOMITANT MEDICATIONS SATISFACTORY CLINICAL RESPONSE SHORT FORM 36 HEALTH QUESTIONNAIRE (SF-36) TRANSFORMATIONS FOR EFFICACY OUTCOMES TREATMENT COMPLIANCE RATES VALUES OF CLINICAL CONCERN WHO FC CHANGE FROM BASELINE CATEGORISATION WHO FC CHANGE FROM BASELINE CATEGORISATION ASSESSMENT WINDOWS VISIT WINDOWING STUDY POPULATION EFFICACY ANALYSES PRIMARY EFFICACY ANALYSIS PRIMARY ANALYSIS SENSITIVITY ANALYSES OF PRIMARY ANALYSIS SECONDARY EFFICACY ANALYSES PRO-B-TYPE NATRIURETIC PEPTIDE

312 CONFIDENTIAL SATISFACTORY CLINICAL RESPONSE MWD WHO FC SCORES BDI SCORES EXPLORATORY OUTCOMES HEALTH OUTCOME OUTCOMES SF CAMPHOR SAFETY ANALYSES EXTENT OF EXPOSURE ADVERSE EVENTS LABORATORY DATA VITAL SIGNS LIVER EVENTS ECG VIRAL RESULTS REFERENCES ATTACHMENTS Table of Contents for Data Displays Population Tables Primary Outcome Tables Secondary Outcome Tables Exploratory Outcome Tables Health Outcome Tables SAFETY TABLES Population Listings Efficacy Listings Safety Listings Efficacy Figures Safety Figures...95 APPENDIX 1 STUDY ASSESSMENTS AND PROCEDURES...96 APPENDIX 2 SCORING ALGORITHMS FOR CAMPHOR...98 APPENDIX 3 CALCULATION OF COHEN KAPPA STATISTIC IN SAS

313 CONFIDENTIAL ABBREVIATIONS 6MWD AE ALT AMB ANCOVA AST ATC BCT BDI BMI BP CAD CAMPHOR CCB CEAC CMH CSR ECG ecrf EOS FAV FC GH HIV HPAH HR HT IDMC IDSL IP IPAH ITT IVRS LOCF LVEDP MCS MedDRA MH mitt MMRM mpap Six Minute Walk Distance Adverse Event Alanine Transaminase Ambrisentan Analysis of Covariance Aspartate Transaminase Anatomical Therapeutic Chemical Blinded Combination Therapy Borg Dyspnea Index Body Mass Index Bodily Pain Coronary Artery Disease Cambridge Pulmonary Hypertension Outcome Review Calcium Channel Blockers Clinical Endpoint Adjudication Committee Cochran Mantel Haenszel Clinical Study Report Electrocardiogram Electronic Case Report Form End of Study Final Assessment Visit (First Database Lock) Functional Class General Perception of Health Human Immunodeficiency Virus Hereditary Pulmonary Arterial Hypertension Hazard Ratio Reported Health Transition Independent Data Monitoring Committee Integrated Data Standards Library Investigational Product Idiopathic Pulmonary Arterial Hypertension Intent-to-Treat Interactive Voice Response System Last Observation Carried Forward Left Ventricle End Diastolic Pressure Mental Component Summary Medical Dictionary for Drug Regulatory Activities Mental Health Modified Intent-to-Treat Mixed Models for Repeated Measures Mean Pulmonary Arterial Pressure 6

314 CONFIDENTIAL NT-proBNP PAH PCS PF PP PVR RAP RE RHC RP SAE SF SF-36 TAD TEAE TFL TTCF VT WHO N-terminal pro-b-type natriuretic peptide Pulmonary Arterial Hypertension Physical Component Summary Physical Functioning Per Protocol Pulmonary Vascular Resistance Reporting and Analysis Pan Role Emotional Right heart catheterization Role-Physical Serious Adverse Event Social Functioning Short Form 36 Health Survey Tadalafil Treatment emergent adverse event Table, Figure, Listing Time To Clinical Failure Vitality World Health Organisation Trademark Information Trademarks of the GlaxoSmithKline group of companies None Trademarks not owned by the GlaxoSmithKline group of companies Borg Dyspnea Index East CAMPHOR nquery Advisor Quality Metric Health Outcomes Scoring Software 3.0 SAS SF-36 7

315 CONFIDENTIAL 1. INTRODUCTION The purpose of this reporting and analysis plan (RAP) is to provide details of the statistical analyses that have been outlined within the AMBITION study (AMB GS-US ) protocol [document number: GM2008/00365/02, finalisation date 15 April 2010 and subsequent amendments]. This RAP will be used for production of the clinical study report (CSR). 2. STUDY OBJECTIVE(S) AND OUTCOME(S) 2.1. STUDY OBJECTIVE(S) The study objectives (also given in protocol section 2), are as follows Primary Objective The primary objective of this study is to compare the difference between two treatment strategies; first-line combination therapy (ambrisentan (AMB) 10mg once daily and tadalafil (TAD) 40mg once daily) versus pooled monotherapy (with AMB 10mg once daily or TAD 40mg once daily) in subjects with pulmonary arterial hypertension (PAH). This will be assessed by time to first clinical failure (TTCF) adjudicated event (see RAP section 9.2.9) Secondary Objectives Comparison of the combination therapy to the individual monotherapy arms with respect to TTCF adjudicated event is a secondary objective and will only be performed if the combination therapy versus the pooled monotherapy arms is statistically significant (pvalue <0.05). This also applies to all additional secondary efficacy objectives. Additional secondary objectives of this study are as follows. To compare the change in other clinical measures of PAH (see RAP section 2.2.2) after initiating combination therapy or monotherapy in subjects with PAH. 8

316 CONFIDENTIAL The safety and tolerability of first-line combination therapy (ambrisentan and tadalafil) will be compared to first-line monotherapy (ambrisentan or tadalafil). In addition, the effect of AMB on exercise capacity at both peak and trough plasma concentrations will be assessed as an exploratory analysis STUDY OUTCOME(S) The study outcomes (also given in protocol sections to 6.2.3), are as follows Primary Outcome Time (days) from randomisation to the first clinical failure adjudicated event of PAH (see RAP section 9.2.9). Additionally, time to first clinical failure adjudicated event component (see RAP section 9.2.9) (irrespective of whether it contributes to TTCF) and time to clinical worsening adjudicated (see RAP section ) will be evaluated as supportive analyses of the primary outcome. An independent Clinical Endpoint Adjudication Committee (CEAC) will review and adjudicate all clinical failure events. Additionally, the CEAC will adjudicate all serious adverse event (SAE) hospitalisations (different from the event hospitalisations) to ensure that all hospitalisations due to worsening PAH are captured. Further details of the adjudication process are contained in the separate CEAC charter Secondary Outcomes Change from baseline measured at week 24 in N-terminal pro-b-type natriuretic peptide (NT-proBNP). Percentage of subjects with satisfactory clinical response measured at week 24 (see RAP section ). Change from baseline in six minute walk distance (6MWD) measured at week 24. Change from baseline measured at week 24 in World Health Organisation (WHO) Functional Class (FC). 9

317 CONFIDENTIAL Change from baseline measured at week 24 in Borg Dyspnea Index (BDI) immediately following exercise Exploratory Outcomes An assessment of the effects of AMB on 6MWD at peak and trough plasma concentrations at Week 16, which includes: The AMB placebo-corrected mean change from baseline 6MWD following 16 weeks of treatment (i.e., TAD+AMB combination therapy versus TAD +placebo monotherapy). The ratio of the placebo-corrected mean change from baseline trough 6MWD at week 16 to the placebo-corrected mean change from baseline peak 6MWD at week 16 (see RAP section 11.3) Health Outcomes Change from baseline to week 24 in Quality of Life as measured by the Short Form 36 Health Survey (SF-36) will be presented for each individual domain (physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, mental-health, reported-health transition), the mental component summary and the physical component summary. The SF-36 and individual scoring of each domain can be found in the Quality Metric Health Outcomes Scoring Software 4.5, User s Guide. The Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) will be summarized for each of the subscales (Symptoms, Symptom-Energy, Symptom Breathlessness, Symptom Mood, QoL and Activity subscales). The CAMPHOR review and descriptions of how each of the individual subscales are scored are found in Appendix 2 (selected countries where language translation is validated) Safety Outcomes Incidence, relationship to treatment and severity of adverse events. Vital signs (i.e. blood pressure and heart rate). Laboratory tests: clinical chemistry, haematology, testicular function (males only). 10

318 CONFIDENTIAL Liver events (as assessed by alkaline phosphatase, alanine transaminase (ALT), aspartate transaminase (AST) and total bilirubin). Electrocardiogram (ECG) STATISTICAL HYPOTHESES The statistical hypothesis (also given in protocol section 8.1) is as follows. The null hypothesis tested for the primary outcome of efficacy is that there is no difference in TTCF of PAH when treated with monotherapy compared to combination therapy (i.e., H 0 : Hazard Ratio (HR) = 1). The two-sided alternative hypothesis is that there is a difference (i.e., H 1 : HR 1, a two-sided test). The study is designed to show superiority of combination therapy relative to pooled monotherapy (i.e. a hazard ratio less than one would represent a reduction in the risk of clinical failure on combination therapy vs monotherapy). 3. STUDY DESIGN The study design (also given in protocol section 3.1) is as follows. Study design: The study is a phase III/IV, randomised, parallel group, double-blind study to compare the safety and efficacy of initiating combination therapy or monotherapy as first-line therapy in subjects with WHO FC II and III PAH. Blinding will be accomplished by providing ambrisentan or placebo in tablets that are indistinguishable from each other and by providing tadalafil or placebo in tablets that are indistinguishable from each other. Sample size: The original sample size calculations were based on an overall event rate of 15% (a monotherapy event rate of 20% per year and a combination therapy event rate of approximately 10% per year). Based on the blinded event rate estimate of 12% per year observed in December 2012 (a combination therapy event rate of 8% and a monotherapy event rate of 16% to maintain the original assumed reduction in the hazard of 53%), the sample size was adjusted. Details on sample size adjustment are described in protocol section Sample size Re-estimation and in section of the RAP. 11

319 CONFIDENTIAL Planned sample size was 545 and increased to 614 subjects based on the December 2012 blinded review of the event rate to provide an increase from 456 to 520 modified intention-to-treat (mitt) (see protocol section 8.2.3) subjects who meet the PAH diagnosis and classification eligibility criteria defined in Protocol Amendment No. 2, with 260 subjects in the combination arm and 260 subjects in the monotherapy arm [130 subjects receiving AMB and 130 subjects receiving TAD]. Recruitment will continue until 24 weeks prior to the estimated 105 th event in the mitt population. At the time of Protocol Amendment No. 4, this was estimated to be approximately 614 subjects. Study duration: As this is an event driven trial, the duration of the study and sample size will be determined by the rate of clinical failure events. Subjects will be encouraged to remain in the study after their events for the duration of the trial and continue on treatment until the target number of events is achieved. Recruitment will continue up to 24 weeks prior to the predicted 105 th event in the mitt population, and up to a maximum of 680 subjects. Based on current assumptions, the median treatment duration is anticipated to be 2 years with the total study duration estimated at 2.8 years. Randomisation: See RAP section 8.2. Medication dosing: The target doses of the study medications will be 10mg AMB once daily and 40mg TAD once daily. However, subjects will start on lower doses as follows and will be up-titrated to the target doses as appropriate. Up titration to tadalafil 40mg is mandatory (except for renal impairment) but up titration to 10mg AMB is a target based on tolerability and subjects may remain on 5mg. Combination therapy arm: Subjects randomised to combination treatment will receive both: One tablet of 5mg AMB and one tablet of AMB-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg AMB (10mg once daily), thereafter One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and two tablets of 20mg TAD (40mg once daily), thereafter Subjects with mild to moderate renal impairment (Creatinine Clearance >30mL/min and <80mL/min) should be carefully assessed for tolerability at the week 4 visit and a risk to 12

320 CONFIDENTIAL benefit decision made by the investigator whether the subject remain on TAD 20mg once daily or up-titrate to TAD 40mg once daily. AMB monotherapy arm: Subjects randomised to AMB monotherapy will receive both: One tablet of 5mg AMB and one tablet of AMB-matching placebo for the first 8 weeks and, if well tolerated, two tablets of 5mg AMB (10mg once daily), thereafter Two tablets of TAD-matching placebo. TAD monotherapy arm: Subjects randomised to TAD monotherapy will receive both: Two tablets of AMB-matching placebo One tablet of 20mg TAD and one tablet of TAD-matching placebo for the first 4 weeks and, two tablets of 20mg TAD (40mg once daily), thereafter. Subjects with mild to moderate renal impairment should be treated as described for the combination therapy arm. Medication tolerability issues: Should tolerability issues be experienced the investigator may separate dosing of Investigational Products into morning and evening dosing (i.e., AMB in the morning and TAD in the evening). If tolerability issues are still experienced the investigator may request a down titration of AMB from 10mg to 5mg once daily. If tolerability issues are still experienced the investigator may then also request a down titration of TAD from 40mg once daily to 20mg once daily (Sponsor Medical Monitor approval required). Study visits: Study visits are summarised below. A summary of the study assessments and procedures as taken from section 6 of the protocol is given in RAP Appendix 1. Clinic visits: Subjects will be assessed for efficacy and safety at Screening, Randomisation, Weeks 4, 8, 16, 24, and every 12 weeks thereafter until their final assessment visit (FAV). Laboratory safety assessments: In between clinic visits subjects will have monthly laboratory safety assessments. 13

321 CONFIDENTIAL Final assessment visit (FAV): All subjects are eligible to receive a minimum of 24 weeks of therapy. When 105 adjudicated primary outcome events are projected to have occurred (i.e. 105 mitt subjects with an event), all subjects will be notified to return for a FAV within 28 days or their week 24 visit, whichever is later. First database lock will occur after all subjects have completed the FAV. All adjudicated events occurring prior to or at the FAV will be included in the primary analysis. End of study (EOS): Subjects will continue to receive blinded treatment until the first database lock (approximately 4 weeks following last subject last FAV), after which subjects will return for an EOS visit within 4 weeks of sponsor notification. At this EOS visit, subjects will be un-blinded and treated per the discretion of the Investigator. All adjudicated events occurring prior to or at the EOS will be included in the sensitivity analysis. Safety follow-up: A safety follow-up by phone will be performed 30 days after the subject s last dose of Investigational Product, after which the data will be cleaned and a second database lock will be performed. Change of randomised treatment visit: The investigator is required to complete a Change of Randomised Treatment Visit to assess safety and efficacy prior to any treatment changes this includes change in dose, return to investigational product (IP) following an interruption or temporary discontinuation, permanent discontinuation of IP, separation of IP dosing to the morning and evening or if dosing was recombined after a previous separation of dose, request blinded combination therapy (BCT), change in any PAH treatment, use of prostanoid therapy. Medical Interventions for PAH: Parenteral prostanoid therapy may be given at any point during the study. Addition of parenteral prostanoid therapy is considered a primary outcome event and subjects may continue in the study following initiation. Medical Interventions for clinical failure event: If a subject experiences an event of clinical failure the following medical interventions may also be utilised. The clinical failure event necessary for medical intervention is based on the opinion of the investigator and does not require prior adjudication (see RAP section 9.2.1) from the CEAC. 14

322 CONFIDENTIAL The investigator may choose to request BCT to ensure that the subject is receiving AMB/TAD combination therapy. The investigator may choose to add prostanoid therapy (parenteral or nonparenteral). For either of these two options, the investigator must assess response to intervention within 8 weeks following initiation. If not satisfied after completion of the first intervention the investigator may do the following (if not done so already): Initiate non-parenteral prostanoid therapy. Initiate parenteral prostanoid therapy. Request blinded AMB/TAD combination therapy. 4. PLANNED ANALYSES 4.1. CHANGES FROM PROTOCOL SPECIFIED ANALYSES There are no changes to the analyses specified in the protocol INTERIM ANALYSES An Independent Data Monitoring Committee (IDMC) will meet periodically to monitor the safety within the study. The primary objective of the IDMC is to protect the ethical and safety interests of subjects recruited into. The IDMC will review unblinded data, however there will be no formal interim analyses, and no stopping rules based on assessment of the efficacy data FINAL ANALYSIS The efficacy analysis, initial safety analysis and associated tables, figures and listings (TFLs) will be produced following first database lock which will occur after the last subject has completed their FAV (see RAP section 3). Additional safety, as well as sensitivity / exploratory analyses and TFLs relating specifically to the EOS visit will be produced following second database lock after the last subject has completed their end of study and follow-up visits. Specific details of the populations and endpoints are given 15

323 CONFIDENTIAL in RAP sections 10, 11, and 12. Specific details of the statistical methodologies that will be used to analyse the study outcomes are given in RAP section SAMPLE SIZE CONSIDERATIONS The following sample size data have been taken from protocol sections to SAMPLE SIZE ASSUMPTIONS This section presents sample size calculations based on an overall event rate of 15% (a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year). Based on a review of blinded event rate estimate in December 2012, the sample size has been adjusted. Details on sample size adjustment are described in protocol section Sample size Re-estimation and RAP section Original sample size calculation: This event-driven study requires 105 mitt subjects with an adjudicated clinical failure event in order to have approximate 97% power for the comparison of combination therapy with pooled monotherapy, and approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either AMB or TAD alone); with a two-sided type I error rate (alpha level) of 5% for each of the three comparisons, assuming an exponential distribution for time to clinical failure. Initial assumptions of the event rates are based on Steering Committee recommendations and are projected as a monotherapy event rate of 20% per year and a combination event rate of approximately 10% per year, i.e. an absolute difference of approximately 10%. This leads to a hazard ratio of 0.47, i.e. a 53% reduction in risk. Assuming a recruitment period of 134 weeks (based on a recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 159 weeks the study requires approximately 545 subjects enrolled to obtain 456 mitt subjects (228 subjects in the combination arm and 228 in the monotherapy arm [114 subjects receiving ambrisentan and 114 subjects receiving tadalafil]). Every possible effort will be made to follow all subjects until the required number of events has been observed. To account 16

324 CONFIDENTIAL for the loss of subjects a drop-out rate of 5% per treatment group at one year has been considered for this calculation. The secondary comparison of combination therapy versus each monotherapy treatment is estimated to have approximately 85% power for each comparison. For this calculation, it has been assumed that there are approximately 70 clinical failure events for 342 mitt subjects (228 on combination and 114 on each monotherapy) and an alpha level of 5%. Similar assumptions for the main sample size with respect to event rates and hazard ratio have been made. Calculations of event driven sample sizes and power were performed using East version 5.2, Cytel Inc SAMPLE SIZE SENSITIVITY The sample size sensitivity is described in protocol section Sample Size Re-estimation The event rate will be monitored using data blinded with respect to treatment assignment. The protocol states that should the observed, blinded event rate be lower than anticipated or discontinuation rates (i.e., withdrawal from study or discontinuation of IP) be higher than expected, then the sponsors and the Steering Committee will use blinded data to assess the value of increasing study size and/or length of duration. The table below provides details on the approximate expected number of mitt subjects to be enrolled and the approximate expected number of events at last patient in (LPI), 4 months prior to LPI, and 6 months prior to LPI. The expected number of events is based on the assumption of an exponential distribution for time to clinical failure. In order to have time to increase enrollment if necessary, the blinded event rate will be examined between 4 and 6 months prior to LPI. Table 1. Expected Enrollment and Event Rate in Protocol Version 5 Time to LPI Estimated Date Weeks Expected mitt Subjects Enrolled Expected mitt Events 80% of Expected mitt Events 6 months 07 December (50%) 42 17

325 CONFIDENTIAL 4 months 07 February (60%) 50 0 months 07 June (76%) 64 Following a blinded review of the overall event rate after approximately 2 years of recruitment, the number of estimated adjudicated events (adjudicated events plus events pending adjudication adjusted for concordance) was approximately 77% of the predicted overall event rate of 15%. The overall event rate was re-estimated to be approximately 12% per year. Sample size calculations were performed assuming a combination event rate of 8% and a monotherapy event rate of 16% to maintain the original assumed reduction in the hazard of 53%. Assuming a recruitment period of 148 weeks (based on a recruitment rate of 4.9 subjects per week prior to implementation of Protocol Amendment No.2 and 3.4 subjects per week thereafter) and a total study duration of 175 weeks, the study requires approximately 614 subjects enrolled to obtain 520 mitt subjects (260 subjects in the combination arm and 260 in the monotherapy arm [130 subjects receiving ambrisentan and 130 subjects receiving tadalafil]. Every feasible effort will be made to follow all subjects until the required number of subjects with events has been observed. To account for the loss of subjects a drop-out rate of 5% per treatment group per year has been considered for this calculation. Based on the revised sample size of 520 mitt subjects, 105 mitt subjects with an adjudicated clinical failure event are required in order to maintain the same approximate 97% power for the comparison of combination therapy with pooled monotherapy, and the same approximate 85% power for the comparison of combination therapy to individual monotherapy (i.e. either ambrisentan or tadalafil alone); with a type I error rate (alpha level) of 5% for each of the three comparisons. Calculations were performed in EAST version 5.2, Cytel Inc. To evaluate the effects of ambrisentan on 6MWD at both maximum (i.e. peak) and minimum (i.e. trough) ambrisentan plasma concentrations, a test of the null hypothesis of no treatment group difference in change from baseline to week 16 in the 6MWD with 260 ABS/TAD subjects and 130 PBO/TAD subjects yields approximately 98% power to 18

326 CONFIDENTIAL detect an average placebo-adjusted treatment effect of 30m based on a 2-sample t-test and a standard deviation of 65m. Treatment effect and standard deviation were based on data from Phase 2 and Phase 3 clinical studies of AMB. Sample size and final alpha level were calculated using nquery Advisor Version 6.0. Statistical Solutions, Cork, Ireland. Although the sample size was calculated using a 2-sided t-test, the specified analysis for this outcome uses a Wilcoxon rank sum test; therefore, the actual power may vary slightly. 6. ANALYSIS POPULATIONS The following populations (with the exception of the peak / trough population) are also described in protocol section Tables will be produced in triplicate for the mitt (those subjects meeting amendment 2 criteria), Intent-to-Treat (ITT) (all randomised subjects receiving at least one dose of double-blind treatment), and non-mitt (those subjects NOT meeting amendment 2 criteria) population. Inferential statistical analyses will not be performed for the non-mitt population. Primary outcome tables will be repeated for the Per Protocol (PP) population only if the criteria for use of the PP population are met. Note that for the Safety tables, the mitt, ITT, and non-mitt populations will be displayed before and after initiation of BCT. Treatment headers will be randomized treatment. The split before and after initiation of BCT will enable interpretation of safety based on treatment that was actually received. Modified Intent-to-Treat (mitt) Population - All subjects randomised to treatment who receive at least one dose of Investigational Product and who meet the PAH diagnosis and classification eligibility criteria defined in Protocol Amendment No. 2 (see protocol section 4.2 and RAP section ). Subjects will be analysed according to randomised treatment group. This will be the primary analysis population for assessing efficacy. Data will be displayed in terms of the following treatment categories: Combination therapy, Pooled Monotherapy, Monotherapy: TAD, Monotherapy: AMB. Intent-to-Treat (ITT) - All subjects randomised to treatment who receive at least one dose of Investigational Product. Subjects will be analysed according to the treatment group to which they were randomised. Data will be displayed in terms 19

327 CONFIDENTIAL of the following treatment categories: Combination therapy, Pooled Monotherapy, Monotherapy: TAD, Monotherapy: AMB. Non-mITT Population: All subjects randomised to treatment who receive at least one dose of Investigational Product and who did not meet the PAH diagnosis and classification eligibility criteria defined in Protocol Amendment No. 2. Subjects will be analysed according to randomised treatment group. Data will be displayed in terms of the following treatment categories: Combination therapy, Pooled Monotherapy, Monotherapy: TAD, Monotherapy: AMB. Per-Protocol (PP) Population -The per-protocol population will consist of the subset of subjects contained in the mitt without any major protocol violation. If the PP population is greater than 85% or less than 50% of the mitt population, a PP efficacy analysis will not be performed. Data will be displayed in terms of the following treatment categories: Combination therapy, Pooled Monotherapy, Monotherapy: TAD, Monotherapy: AMB. Per protocol analyses will be limited to the primary endpoint (TTCF), the individual components of the primary endpoint and TTCW. The following constitute major protocol violations and subjects who meet these criteria should be excluded from the PP population. These violations are based on a subset of the inclusion and exclusion criteria listed in the protocol. For definitions of specific criteria such as permitted diagnoses of PAH and stable human immunodeficiency virus (HIV) refer to the relevant criteria in the protocol; text in brackets at end of each violation refers to the related inclusion (I) or exclusion (E) criterion in the Protocol Amendment No Subject less than 18 years old (I1) - Subject did not have a permitted diagnosis of PAH (I3) - Subject did not have WHO FC of II or III at screening (I4) - Subject had unstable HIV (I5) - Subject did not meet haemodynamic criteria for right heart catheterization (RHC) (I6) - Subject did not meet pulmonary function test criteria (I7) - Subject did not walk a distance of 125m and 500m at screening. [Subject must walk a distance of 125m and 500m at the screening visit. In addition 20

328 CONFIDENTIAL the screening and baseline 6MWD tests must not vary by greater than 10%.] (I8) - Subject enrolled in non-permitted exercise program (I10) - Subject enrolled in another interventional study (I12) - -Subject received previous PAH therapy continuously for 14 days or more (PDE5i, ERA, prostanoid) prior to screening or within 7 days prior to screening (E1) - Subject received intravenous inotropes within 2 weeks prior to screening (E5) - Subject treated with potent inhibitor of CYP3A4 at time of screening (E6) - Subject treated with potent inducer of CYP3A4 at time of screening (E7) - Subject treated with calcium channel blockers (CCBs) or HMG-CoA within 4 weeks prior to screening (E9) - Subject treated with short-acting nitrates within 12 weeks prior to screening (E10) - Subject has severe renal impairment at screening visit (E13) - Subject has severe hepatic impairment at screening visit (E14) - Subject had clinically significant anaemia at screening visit (E15) - Subject had an acute myocardial infarction within 90 days prior to screening visit (E19) - Subject has non-permitted cardiovascular disease (E20) - Subject non-compliant with previous medical regimens (E27). Furthermore, subjects will be excluded from the PP population if they fail to meet the criteria for the baseline 6MWD other than that listed for I8 above. Subjects will be excluded if the two tests that constitute baseline are not separated by at least two hours. For the per protocol efficacy analysis TTCF, TTCW and the individual components of each measure, subjects will be included if they fulfil the PP criteria. However, if study medication compliance is <80% or > 120% as measured at any visit up to the occurrence of TTCF, TTCW or the individual components of each measure, the subject would be censored at that time-point in the Cox regression analysis. 21

329 CONFIDENTIAL Safety - Safety data will be displayed for the mitt, non-mitt, and ITT populations and will be presented prior to BCT (i.e. on randomised treatment) and after the initiation of BCT (for subjects who received BCT). Displays will be in terms of the randomized treatment. Treatment categories prior to BCT are Combination therapy, Monotherapy: TAD, Monotherapy: AMB. Treatment categories during BCT are Combination therapy - BCT, Monotherapy: TAD - BCT, Monotherapy: AMB BCT. Pooled monotherapy as a category will not be displayed unless otherwise specified below. Peak / Trough - Subjects in the mitt population who were randomised to combination therapy or TAD monotherapy at baseline and who were included in the peak/trough randomisation at Week 16. This population only includes subjects who have not, by week 16, had any medical intervention or a down-titration of AMB or TAD. This population will only be used in the Peak / Trough assessment of the 6MWD. The analysis will be repeated based on those subjects in the ITT and non-mitt populations. 7. TREATMENT COMPARISONS The primary comparison of interest is between the combination therapy arm and the pooled monotherapy arm. The secondary comparisons of interest are between the combination therapy arm and each individual monotherapy arm. The analysis methods used for the secondary comparisons will be the same as those used for the primary comparison. 8. GENERAL CONSIDERATIONS FOR DATA ANALYSES 8.1. MULTICENTRE STUDIES Sites will be grouped into two regions for the purpose of covariate analyses (see RAP section 8.3.2): North America (i.e. sites in USA and Canada), and Rest of the World (non-usa/non-canada sites). 22

330 CONFIDENTIAL 8.2. RANDOMISATION Subjects will be assigned to study treatment in accordance with the randomisation schedule. The randomisation code will be generated using the Sponsor s randomisation system RandAll. Randomisation will be performed by an interactive voice response system (IVRS) which will direct the investigator to assign the appropriate treatment to each subject. Randomisation of eligible subjects will be stratified based on the underlying aetiology of PAH and WHO FC (see protocol section 5.2). Subjects will be randomised 2:1:1 to either the combination therapy arm or to one of the monotherapy arms. Subjects will also be randomly assigned in a 1:1 allocation to peak or trough measurements at week 16 within each treatment group OTHER STRATA AND COVARIATES The following sections define the list of strata, covariates and subgroups of specific clinical interest that will be used in the analyses. Note: if the percentage of subjects within a particular subgroup is small (<25% of their randomized treatment), then the subgroup categories may be refined once enrollment is completed and in any case prior to unblinding the trial Strata As randomisation is stratified (at baseline) by PAH aetiology (Idiopathic (IPAH)/ Hereditary (HPAH) and non-ipah) and WHO FC (II and III) analyses will be stratified/adjusted for these variables where specified (see RAP section 8.5). In case of differences between randomised and actual data, the randomised stratification factors will be used in the analysis Covariates As part of a sensitivity analysis the following covariates will be individually investigated for inclusion in Cox and Logistic regression models where specified (see RAP section 8.5). Region (North America, Rest of World) Age (continuous) Sex (male, female) 23

331 CONFIDENTIAL The strata variables (PAH Aetiology and WHO FC) will also be included as covariates (see RAP section 8.5) to investigate the significance of (two-way) interactions between treatment and the strata variables Subgroups Where specified (see RAP section 8.5), summary and analysis tables relating to the primary and secondary outcomes will be subset by the following subgroups. Subgroup analyses will not be stratified. Aetiology (IPAH/HPAH and non-ipah) Baseline WHO functional class (II, III) Region (North America, Rest of World) Baseline age group (<65, 65 years) Baseline age group (above and below the study median) Sex (Male, Female) Baseline 6MWD (above and below the study median) 8.4. MULTIPLE COMPARISONS AND MULTIPLICITY The following relates to protocol section A step-down procedure will be adopted among the outcomes. Only if the primary outcome combination versus pooled monotherapy comparison is statistically significant will inferences on the first secondary outcome be evaluated. Only if the first secondary outcome combination versus pooled monotherapy comparison is found to be significant will inferences on the second secondary outcome be evaluated. This gate keeping approach will be implemented for all secondary outcomes in pre-defined order (defined below). 1. Change from baseline measured at week 24 in NT-proBNP. 24

332 CONFIDENTIAL 2. Percentage of subjects with satisfactory clinical response measured at week Change from baseline in 6MWD measured at week Change from baseline measured at week 24 in WHO FC. 5. Change from baseline measured at week 24 in BDI. Additionally, for each outcome where the combination treatment arm is demonstrated as significant to the pooled monotherapy, comparisons of combination treatment versus each individual monotherapy treatment will be performed. The significance level for all comparisons will be 5% (two-sided). Note that to aid with the planning, programming, and production of outcome tables, all tables will be produced irrespective of whether they meet the above criteria. However, the above criteria will be used for interpretation of the study results. The gatekeeping approach described above will be used for the primary statistical analysis for the primary and each secondary endpoint. The gatekeeping approach will not be used for sensitivity analyses as assessed by alternative statistical methods STATISTICAL METHODOLOGIES In this section, specific details of the statistical methodologies that will be used to analyse the study outcomes (see RAP section 2.2), are given. Methodologies are presented in alphabetical order for ease of reference ANCOVA Analysis of covariance (ANCOVA) (SAS Institute Ltd., 2008a) will be used as a sensitivity analysis for many of the secondary and exploratory endpoints relating to change from baseline at week 24 (NT-proBNP [note that data will be log-transformed] (see RAP section 8.5.4), 6MWD and BDI). Covariates will be the randomised stratification factors (see RAP section 8.3.1), treatment, and baseline data for each outcome of interest. Treatment will have three levels and contrast statements will be used to test each of the pair-wise treatment comparisons of interest (see RAP section 7). Treatment difference will be estimated using the difference in least squares means and 25

333 CONFIDENTIAL presented together with the 2-sided 95% CI and p-value. Model checking will be performed Cochran Mantel Haenszel Analysis on change from baseline in WHO FC categorization [(+2, +1, 0, -1, -2) and (Improved, No Change, and Deteriorated)] will be performed using Cochran Mantel Haenszel (CMH) tests. Covariates will be the stratification factors (see RAP section 8.3.1) and treatment. The 3 comparisons of interest (randomized combination therapy vs. pooled monotherapy, randomized combination therapy vs. ambrisentan monotherapy, and randomized combination therapy vs. tadalafil monotherapy) will be looked at separately Cox regression Cox proportional hazards regression (SAS Institute Ltd., 2008b; SAS Institute Ltd., 2008c), adjusting for the randomisation stratification factors (PAH aetiology and WHO FC), will be used to analyse the primary endpoint. Time to clinical failure will be displayed as Kaplan Meier event-free curves from randomisation to the final assessment visit for the primary analysis. Differences between the curves will be tested for significance by the stratified log rank test. The hazard ratio will be used to characterize the treatment effect and 95% confidence intervals will be calculated using the Cox proportional hazards model. Analyses of time to first clinical worsening event and time to each of the components of the primary endpoint (time to death, time to hospitalisation, time to disease progression and time to unsatisfactory long-term clinical response) through FAV (First Data Base Lock) will also be performed. FAV will be determined based on each individual subject s FAV. (NB: some subjects will have their week 24 visit coincide with the FAV visit. For those subjects only the week 24 CRF page will be completed. All such subjects will be able to be identified because their week 24 visit will be on or after 03 March For subjects without a FAV (i.e. subjects who terminated early from both study drug and study), FAV will be defined as the last subject s FAV. That is, subjects without a FAV will be censored at the time they discontinue the study; however, if an event is reported for such a subject after study 26

334 CONFIDENTIAL discontinuation but prior to the last subject s FAV, the event will be included in the primary analysis. Supportive analyses for TTCF, TTCW and the individual components will be performed up to EOS utilizing a similar strategy. Additionally, time to death will be analysed up to the last contact (defined as the last clinic visit or the 30 day follow-up phone call, whichever occurs last). Subjects who have an event (as opposed to no event) will be modelled. Additional sensitivity analyses of TTCW through FAV and through EOS will be performed that will censor subjects (with or without a clinical worsening event) at the time of initiation of BCT. Possible scenarios for analyses of the primary event are presented below:- 1. Subjects who have an event which is overturned by adjudication, and who have withdrawn from the study without having an adjudicated event, will be censored at their time of study discontinuation. If an adjudicated event is reported after study discontinuation, it will be included in the analyses. If the adjudicated event occurred prior to the last subject s FAV, it will be included in the primary analysis of events through FAV. If the adjudicated event occurred after the last subject s FAV, it will be included in the sensitivity analysis of events through EOS. 2. Subjects who withdraw from Investigational Product before a TTCF event but remain in the study and subsequently have an event that is confirmed by adjudication will have their adjudicated event counted at the time it occurred. 3. Subjects for whom investigators initiate BCT based on an event that is overturned by adjudication and subsequently have a second event confirmed by adjudication will have the second event counted in the analysis and will be analysed as randomised. 4. Subjects who have a first adjudicated event after their FAV will have that event counted in the EOS analysis. With respect to adjudicated events which occurred on the same day:- 27

335 CONFIDENTIAL Example 1: A subject had a 6MWD which reflected a greater than 15% decrease from baseline. They were admitted to the hospital for worsening of PAH on the same day. More than 14 days later they had a second walk test that was greater than 15% decrease from baseline (meeting the criteria for disease progression). The hospitalization for worsening of PAH is the first event and counted in the primary endpoint. The disease progression is the second event and will not be counted in the primary endpoint but will be counted in the analysis of the disease progression component. Example 2; A subject presented to the clinic and had a greater than 15% decrease from baseline. Prior to the second repeat 6MWD the subject died. The death would be counted in the primary endpoint. There would be no disease progression. Example 3. A subject was admitted to hospital for worsening PAH and died on the same day (within 24 hours). The death would be counted in the primary analysis of TTCF and in the secondary analyses of (a) TTCW and (b) death. The hospitalisation would not be reported as an event by the CEAC. Example 4. A subject was admitted to hospital for worsening PAH and died 2 days later. The hospitalization for worsening PAH would be counted in the primary analysis of TTCF and in the secondary analyses of TTCW. The death would be counted in the analysis of death. In general, for events that occur on the same day, the most severe event will be counted as the first event in the primary endpoint TTCF and in the secondary endpoint TTCW. Treatment will be included as a covariate. Separate models will be produced for each of the treatment comparisons of interest (i.e. 3 separate models, see RAP section 7). The stratification variables (see RAP section 8.3.1) will be modelled by means of the STRATA statement to allow for more robust estimates of their effect. Ties in survival times will be handled using Breslow s approximation. The assumption of proportionality will be investigated with a time-dependent explanatory variable which is defined as treatment*[log(time to event)] for the primary endpoint model only (i.e. TTCF at FAV). If the p-value from the Wald chi squared statistic for this variable is less than 10% (indicating a departure from proportional hazards) further investigation of the primary outcome would then be based on nonparametric analyses 28

336 CONFIDENTIAL including the Wilcoxon-Gehan censored data rank sum test. The primary analysis will be the Cox regression analysis, even if the assumption of proportionality is violated. Irrespective of the statistical significance for the test for proportional hazards, hazard ratios and 95% confidence intervals using the Cox Proportional Hazards Regression model will still be calculated and presented to characterize the treatment effect. Covariate Analyses Covariate analyses will be performed as a sensitivity analysis to investigate the significance of (two-way) interactions between treatment and the strata variables (PAH Aetiology and WHO FC, see RAP section 8.3.1), and the significance of other covariates of interest (region, age and sex, see RAP section 8.3.2) in the primary outcome of time to first clinical failure event models (at FAV) that relate to each of the 3 treatment comparisons of interest (see RAP section 7). For all investigations of interaction a significance level of 10% will be used. Strata variables will first be analysed as individual (i.e. one at a time - one model for PAH Aetiology by treatment interaction and a separate model for WHO FC by treatment interaction) covariates (as opposed to in the STRATA statement) to allow for the significance of interactions with treatment to be investigated. The model will include treatment, covariate (e.g. PAH Aetiology) and treatment by covariate interaction (e.g. treatment by PAH aetiology). If no treatment by covariate interactions are found to be significant the strata variables will be modelled in the STRATA statement (as for the primary model) before the significance of other covariates for inclusion are investigated. If interactions between treatment and stratification variables (eg PAH aetiology) are found to be significant then the main effect (eg PAH aetiology) and interaction (eg treatment by PAH Aetiology) effect will be kept in the model as covariates before the significance of other covariates for inclusion are investigated. Inclusion of any other covariate in the model will be based on using a backward elimination procedure as follows:- All main effects of other covariates and pairwise interactions with treatment will be included. Backward selection will be used to eliminate non-significant (p 0.10) pairwise interactions (forcing the main effects into the model at this stage). 29

337 CONFIDENTIAL Once non-significant pairwise interactions have been eliminated, backwards selection will be used to eliminate any non-significant (p 0.10) main effects that are not involved in remaining interaction terms. Subgroup Analyses Time to first clinical failure event (to FAV) will also be analysed for each subgroup of interest: PAH Aetiology (IPAH/HPAH and non-ipah), baseline WHO functional class (II, III), Region (North America, Rest of World), baseline age group (<65, 65), baseline age group (above and below the study median), sex (male, female) and baseline 6MWD (above and below the study median). For each subgroup of interest there will be 3 class terms in the model - treatment, subgroup and treatment by subgroup interaction, with the interaction p-value reported. If the interaction is significant (p<0.10) then the hazard ratio for treatment, 95% confidence interval and p-value will be reported from the Cox regression analysis run within each level of the subgroup. If the interaction is not significant (p 0.10) the interaction p-value will still be reported and the hazard ratio for treatment and 95% confidence interval will be reported from the Cox regression analysis run with treatment, subgroup and treatment by subgroup interaction in the model. Per Protocol Analyses (if performed) If study medication compliance is <80% or >120% as measured at any visit up to the occurrence of TTCF, TTCW or the individual components of each measure, the subject will be censored at that time-point point at which non-compliance was first recorded in the Cox regression analysis for the per protocol population Log transformations For log transformed data (NT-proBNP) the geometric mean and coefficient of variation will be calculated as below (based on the log-transformed data) and the geometric mean ratio will be calculated as the ratio between the value of the endpoint at the time point of interest and the baseline value (based on the log-transformed data). Geometric mean = exp(µ) Percent change = 100 x (Geometric mean 1) Coefficient of variation = 100 x exp(σx2)

338 CONFIDENTIAL Logistic regression For the logistic regression model (Hosmer and Lemeshow, 2000; SAS Institute Ltd., 2008d; SAS Institute Ltd., 2008e), the outcome variable will be the outcome of interest - satisfactory clinical response - with response (as opposed to non-response) being modelled. The covariates will be the stratification factors (see RAP section 8.3.1), and the treatment variables; these covariates will remain in the model irrespective of their significance. Separate models will be produced for each of the 3 treatment comparisons of interest (see RAP section 7). Covariate Analyses Covariate analyses will be conducted in a similar way as described for the Cox regression model (although stratification variables will be modelled as covariates), using a backward elimination procedure. Subgroup Analyses Subgroup analysis will be conducted in a similar way as described for the Cox regression model Mixed Models Repeated Measures (MMRM) MMRM will be implemented using PROC Mixed (SAS Institute Inc., 2008a) as a secondary analysis of the secondary and exploratory outcomes relating to change from baseline. Each of the time points of the parameter of interest up to and including week 24 will be included in the model. The parameters of interest are NT-proBNP (note that data will be log-transformed [see RAP section 8.5.4]), 6MWD, BDI, SF-36, and CAMPHOR. Covariates will be the stratification factors (see RAP section 8.3.1), treatment, and baseline data for each outcome of interest. In the model statement an interaction between treatment and visit will be included to enable treatment comparisons at each visit. A repeat effect will be included for visit with subjects as the blocking variable; an unstructured covariance matrix will be specified. Treatment will have three levels and contrast statements will be used to test each of the pair-wise treatment comparisons of interest (see RAP section 7). Treatment difference will be estimated using the difference in least squares means and differences will be summarized for each time point up to week 24 enabling an evaluation of the time to onset and maintenance of treatment effect. The p-value for treatment by visit interaction will be reported. 31

339 CONFIDENTIAL NB: Prior to the implementation of MMRM, missing data will be left missing and not be imputed as described in RAP section (Missing Secondary Outcome Data) or RAP section (Visit Windowing) below Wilcoxon Rank Sum Test 6MWD, WHO FC and BDI will be analysed using the Wilcoxon rank sum test stratified by PAH aetiology and WHO Functional Class at baseline. The test will be based on Last Observation Carried Forward (LOCF) data (including cases of an adjudicated clinical failure event of disease progression or un-satisfactory clinical response preceding the missing observation). However, in the cases of an adjudicated clinical failure event of death or hospitalisation preceding the missing data observation, the missing observations will be assigned the worst-rank scores relative to those actually observed and will be assigned ranks reflecting the relative order of the actual event times [Lachin JM, et al. Worst Rank Score Analysis with Informatively Missing Observations in Clinical Trials. Controlled Clinical Trials : ]. Subgroup Analyses 6MWD and BDI will also be analysed for each subgroup of interest: PAH Aetiology (IPAH/HPAH and non-ipah), baseline WHO functional class (II, III), Region (North America, Rest of World), baseline age group (<65, 65), baseline age group (above and below the study median), sex (male, female) and baseline 6MWD (above and below the study median). The unstratified Wilcoxon rank sum test will be performed within each level of the subgroup, using the method described above. Treatment by subgroup interaction will not be investigated or reported. 9. DATA HANDLING CONVENTIONS All data displays will be presented according to GSK s Integrated Data Standards Library (IDSL) statistical display principles. All reporting will be performed using the SAS System, Version 8.2 or higher (SAS is a registered trademark of the SAS Institute Inc., Cary, NC, USA). The file extension used for landscape tables and listings will be L10, with point size of 10, linesize of 108 and 43 lines per page. Where data are sparse, 32

340 CONFIDENTIAL empty tables may be produced with the Data too sparse for table to be produced, No Data to Report or similar PREMATURE WITHDRAWAL AND MISSING DATA Discontinuation of IP Note that, if a subject discontinues Investigational Product prematurely they will be encouraged to continue all study clinic visits according to the planned schedule of events, after completing the change of randomised treatment visit. If, following discontinuation of IP, a subject does not wish to continue with the burden of all clinic visits and assessments, subjects will be permitted to continue in the study with a reduced clinic visit schedule and set of assessments. The subject may perform 24 weekly clinic visits and only the key information required to assess the primary endpoint (TTCF), safety (AEs, SAEs) and concomitant medications need to be collected. The safety follow-up call 30 days after discontinuation of IP should still be completed. The primary analyses will include all post-randomisation events (through the FAV visit, if applicable) and include all subjects who took at least one dose of randomised treatment irrespective of whether the subject discontinued IP prior to the first database lock. If data for some subjects are not available following discontinuation of IP, their event times will be treated as censored for statistical analyses at their last assessment time Study Completion All subjects will be considered to have completed the study upon completion of assessments and procedures up to and including the Follow-up visit. A subject is considered to have completed the blinded phase of the study if the answer to the question Did patient fail to complete the blinded phase of the study? on the Study Conclusion electronic case report form (ecrf) page is answered No ; otherwise the subject is considered to be an early withdrawal from the blinded phase of the study. A subject is considered to have completed the study if the answer to the question Was the subject withdrawn from the study? on the Study Conclusion electronic case report form (ecrf) page is answered No ; otherwise the subject is considered to be an early withdrawal. 33

341 CONFIDENTIAL Completion of the follow-up visit will mark the subject s formal termination from the study Missing Primary Outcome Data For the primary outcome and each component of the primary outcome, any subject who is withdrawn from the study or lost to follow up and does not have a subsequent adjudicated event prior to the last subject s FAV, will be considered to be censored on the date of their last clinical failure event assessment prior to them being lost from the study. Any subject who remains in the study and has not experienced a clinical failure event at the time-point of interest will be censored at that time point. For the analysis of time to death (a supportive analysis of the primary outcome), any subject who is withdrawn or lost to follow up and does not have a subsequent adjudicated event prior to the last subject s FAV, will be considered to be censored at the date last known alive. Any subject who remains in the study and has not died by the time-point of interest will be censored at that time point. For the analysis of time to clinical worsening (a supportive analysis of the primary outcome), any subject who is withdrawn or lost to follow up and does not have a subsequent adjudicated event prior to the last subject s FAV, will be considered to be censored on the date of their last clinical worsening event assessment prior to them being lost from the study. Any subject who remains in the study and has not experienced an event at the time-point of interest will be censored at that time point Missing Secondary Outcome Data For the secondary outcomes relating to 6MWD, WHO FC and BDI, missing data will be handled as follows:- LOCF/Worst Case imputation In descriptive summaries and analyses of covariance models for 6MWD and BDI, as well as the CMH test for WHO FC, subjects who died or had an adjudicated hospitalisation (and were unable to perform the test) on or before the date of week 24 visit, worst case imputation will be used for missing data following death or adjudicated hospitalisation:- for 6MWD, missing data will be imputed with zero, 34

342 CONFIDENTIAL for WHO FC, missing data will be imputed with IV, for BDI, missing data will be imputed with the overall maximum BDI score observed up to the week 24 visit for all subjects in the study. This is the observed score across all subjects irrespective of treatment. A single value that will be used for all subjects. For subjects whose missing data occurred prior to death or adjudicated hospitalisation (or if these events were not experienced by the subject), missing data will be imputed with LOCF. Baseline data will not be carried forward. LOCF/Worst Rank imputation In the Wilcoxon Rank Sum analyses for 6MWD, WHO FC and BDI, in those cases of subjects who experience an adjudicated clinical failure event of death or hospitalisation, the missing observations following the event will be assigned the worst-rank scores relative to those actually observed and will be assigned ranks reflecting the relative order of the actual event times [Lachin JM, et al. Worst Rank Score Analysis with Informatively Missing Observations in Clinical Trials. Controlled Clinical Trials : ]. For subjects whose missing data occurred prior to an adjudicated clinical failure event of death or hospitalisation (or if an adjudicated clinical failure event was not experienced by the subject or if an adjudicated clinical failure event of disease progression or unsatisfactory clinical response was experienced), missing data will be imputed by means of LOCF methodology. Baseline data will not be carried forward. No imputation For the analysis of satisfactory clinical response, no imputation will be performed. If a subject has an event of clinical worsening prior to or at the week 24 visit, the subject will be counted as a non-satisfactory clinical response. If a subject does not have an event of clinical worsening prior to or at the week 24 visit, and the subject does not have a 6MWD value or a WHO functional class value at Week 24, the subject will be excluded from the analysis. For the analysis of NT-proBNP, no imputation will be performed. If a subject does not have a value at the time point of interest, the subject will be excluded from the analysis at that time point. 35

343 CONFIDENTIAL 6MWD Conduct In tables of 6MWD conduct summarising supplemental oxygen use and 6MWD, for If no death or adjudicated hospitalisation occurs, LOCF will be used to impute missing data. If death or hospitalisation occurs, then the following worst case imputation will be those walking less than 6 minutes the following rules for imputed data will be used:- performed:- Missing supplemental oxygen, assume used. If the subject has used supplemental oxygen previously then impute maximum amount observed up to the visit. If the subject has not used, then keep as missing. Missing 6 minute walk time, assume 6 minutes Missing Items in Health Outcome Data SF-36 SF-36 scores will be determined by means of a given programme Quality Metric Health Outcomes Scoring Software 3.0. Only missing item scores will be imputed by means of the software. The scoring algorithm is given in the Quality Metric Health Outcomes Scoring Software 4.5, User s Guide. CAMPHOR Only missing item scores will be imputed as described below. The scoring algorithm is given in RAP Appendix 2. For respondents with between one and five missing responses in the symptom or QoL sections (i.e. cases with no more than 20% missing data), the prorated total score will be 36

344 CONFIDENTIAL calculated as given below. Subjects with more than five missing responses will not be allocated a total score. Two missing responses are permitted in the Energy subscale and 1 each in the Breathlessness and Mood subscales. For respondents with no more than three missing responses in the Activity section (that is, cases with no more than 20% missing data), the prorated total score is calculated as given below. Subjects with more than the permitted missing responses will not be allocated a total score Missing AE Data Where a start date for an adverse event (AE) is partial or missing, the following imputation rules will be applied: If day portion is missing, set day to 01. If month portion is missing, set month to January. If the date is completely missing, or if the date imputed using the above rules is prior to the first dose date, set the date to first dose date. Where an end date for an AE is partial or missing, the following imputation rules will be applied. These will only be applied to AEs that are resolved; if they are not resolved then nothing will be imputed. If day portion is missing, set day to the last day of the month. If month portion is missing, set month to December. If the date is completely missing, or if the date imputed using the above rules is after the treatment end date + 30 days, set the date to the treatment end date + 30 days or if treatment end date is missing then set to date of discontinuation + 30 days. 37

345 CONFIDENTIAL Missing Dates (other than for AEs) For partial date of birth, date last smoked, date of diagnosis, medical history start date, date of surgery or procedure, impute missing day as 01 and missing month as January. If date is completely missing then date should remain missing. Where IP date is partial or missing the last known date of dosing will be used as the last date for determining length of exposure DERIVED AND TRANSFORMED DATA Details of how any data variables are defined, derived, or transformed to enable analyses of the study outcomes are given in the following sections (in alphabetical order) Adjudication The CEAC will provide adjudication of primary and supportive events reported during the study where appropriate (clinical failure events, clinical worsening events). The members of this committee will be blinded to treatment assignment and investigator. Unless otherwise specified all tabulations will be based on adjudicated events. The level of concordance between investigator-determined events and adjudicated events will be reported overall by each TTCF component Age Age (years) at the screening visit will be derived as a whole number according to the IDSL standard algorithm Age group Two age categories will be derived: <65, 65. Additionally, age will be categorised as above or below the median age across all subjects in each population being analysed Baseline With the exception of baseline values for 6MWD (see Baseline 6MWD ) and BDI (see Baseline BDI ), the baseline value for a variable will be the last value prior to administration of study drug; this may be prior to or on the day of study drug initiation. 38

346 CONFIDENTIAL Baseline 6MWD Baseline 6MWD will comprise the average of the last two consecutive measurements prior to randomisation that varied by no greater than 10%. If only one measurement is available, that measurement will be used. If no two consecutive measures vary by no greater than 10% then baseline will be based on the last two consecutive measures for a subject. Baseline BDI Baseline BDI score will comprise the average of the two BDI values obtained following the two 6MWD tests used in determining the baseline 6MWD. If only one measurement is available, that measurement will be used. Baseline WHO functional class The baseline WHO functional class for a subjects will be the latest assessment prior to dosing (i.e., at Randomisation or Screening). By definition, this will be either II or III to be eligible for entry into the trial BCT INITIATION BCT initiation is taken as the BCT start date + 1 day BMI Baseline Body Mass Index (BMI) will be derived as: BMI= weight (kg) / height (m) CAMPHOR The CAMPHOR questionnaire comprises three sections that are scored individually as follows. Symptoms section: The first section is the symptoms section which has 25 negatively weighted items (statements) consisting of three sub-scales relating to energy, breathlessness, and mood. Each item is in the form of a simple statement to which subjects indicate whether or not it is true for them at that moment. A yes response is scored 1 and a no response 0. Item scores are added to give a total symptoms score which can range from 0 (best) to 25 (worst). For the subscale scores, the energy subscale consists of items 1 to 10, the breathlessness subscale of items 11 to 18, and the mood subscale of items 19 to

347 CONFIDENTIAL Activities section: The second section is the activity section which has 15 items, each of which relate to activities described by subjects as being affected by PAH. The subjects rate themselves as being Able to perform each activity on their own without difficulty (scored 0), Able to do on own with difficulty (scored 1), and Unable to do on own (scored 2). Item scores are then summed to give the total activity score ranging from 0 (best) to 30 (worst). Quality of Life: The third section is the quality of life section which contains 25 negatively weighted items, and is scored using the same method as for the symptom score Change from baseline and % change from baseline For untransformed data change from baseline at Week X will be calculated as Week X value minus baseline value. Similarly, percentage change from baseline at Week X will be calculated as follows: ((Week X value - baseline value) / baseline value) * 100. For log-transformed data (see RAP section 8.5.4), percent change from baseline will be calculated by taking the mean change on the log scale, exponentiating, subtracting 1 and multiplying by 100. (Exp((Log x - Log baseline)) -1)* Clinical failure event A clinical failure event is defined as the occurrence of: Death (all-cause), or Hospitalisation for worsening PAH (adjudicated) and as defined below, or Any hospitalisation for worsening PAH Lung or heart/lung transplant Atrial septostomy Initiation of parenteral prostanoid therapy Disease progression (adjudicated) and as defined below, or 40

348 CONFIDENTIAL >15% decrease from baseline in 6MWD combined with WHO functional class III or IV symptoms (at two consecutive post-baseline clinic visits separated by 14 days) Unsatisfactory long-term clinical response as defined below (adjudicated, all criteria required) Receiving at least one dose of randomised treatment and being in the study for at least 6 months A decrease from baseline in 6MWD at two consecutive post-baseline clinic visits separated by 14 days WHO functional class III symptoms assessed at two clinic visits separated by > 6 months (NB: Symptoms can improve from class III over the 6 month period but there needs to be two assessments of class III symptoms 6 months apart). NB: Deaths will be classified by the CEAC as 1. Sudden cardiac death 2. Death from progressive heart failure 3. Pulmonary embolism 4. Death due to other cardiac causes 5. Death due to vascular causes (stroke) 6. Death due to non-cardiovascular causes 7. Cannot be determined Time to clinical failure event is calculated as:- Date of clinical failure event date of randomisation Clinical worsening event A clinical worsening event is defined as the occurrence of any of the following components of the clinical failure event definition: death, hospitalisation for worsening PAH, or disease progression. 41

349 CONFIDENTIAL Time to clinical worsening event is calculated as:- Date of clinical worsening event date of randomisation Cumulative dose Cumulative dose is calculated as the daily dose multiplied by the number of days the dose was taken, summed over time Exposure Exposure (days) will be calculated as (stop date of drug - start date of drug) + 1. For subjects randomised to combination therapy, exposure will be calculated separately for each drug. For subjects randomised to monotherapy who subsequently change to combination therapy, exposure will be calculated for whilst on monotherapy and then separately for the combination therapy drugs (e.g. for a subject that started on AMB and subsequently went to BCT, exposure will be calculated separately for AMB monotherapy, AMB BCT, and TAD BCT). Exposure will be summarized prior to and after the initiation of BCT. This will be summarized by randomised treatment mitt Subjects in the mitt population are those subjects randomised to treatment who receive at least one dose of Investigational Product and who meet the PAH diagnosis and classification eligibility criteria defined in Protocol Amendment No. 2. The PAH diagnosis and classification eligibility criteria are as follows:- Subjects must have a confirmed diagnosis of PAH with documented mean pulmonary arterial pressure (mpap) 25 mmhg, pulmonary vascular resistance (PVR) 300 dynesec/cm 5, 42

350 CONFIDENTIAL and pulmonary capillary wedge pressure (PCWP) or left ventricle end diastolic pressure (LVEDP) of 12 mmhg if PVR 300 to <500 dyne sec/cm 5, or PCWP/LVEDP 15 mmhg if PVR 500 dynesec/cm 5 based on a right heart catheterization (RHC) prior to Screening. Subjects must not have 3 or more of the following left ventricular disease/dysfunction risk factors: i. Body Mass Index (BMI) 30 ii. iii. iv. History of Essential Hypertension Diabetes Mellitus any type Historical evidence of significant coronary disease established by any one of: - history of myocardial infarction - history of percutaneous intervention - angiographic evidence of coronary artery disease (CAD) (>50% stenosis in at least one vessel), either by invasive angiography or by computerized tomographic (CT) Angiography - positive stress test with imaging (either pharmacologic or with exercise) - previous coronary artery surgery - chronic stable angina The mitt population will be identified programmatically by GSK using hemodynamic data and medical history data. The medical history terms that identify mitt subjects and the classification of mitt or non-mitt for each subject will be finalized and documented prior to unblinding Prior and concomitant medications Prior and concomitant medications will be defined using the start date, the stop date, and on-going fields recorded in the ecrf. A prior medication will be defined as any medication taken up to, but not including the randomisation date. Post-treatment medications will be defined as any medication taken after the date of last dose of IP. For subjects who did not take IP, all medications taken on or after the randomisation date will be considered post-treatment medications. For subjects who received IP, medications that started or stopped on the date of randomisation or on the date of the 43

351 CONFIDENTIAL last dose of IP will be considered concomitant medications. Note that a medication can be counted in more than one phase. The definitions for prior, concomitant, pre-treatment, on-treatment, and post-treatment are shown schematically in Table 2. Table 2. Concomitant Medication Classification Pre-treatment On-treatment Posttreatment Prior Concomitant x x x x x x------x x x x x x Y Y Y N N N N Y Y Y Y N? x? ? x x Y Y 2 Y 2 N Y Y 2 x Randomisation x ? ? x ? End of Treatment Y N N Y 1 Y N x = medication start or stop date? = missing date 1 A medication that is started pre-treatment or on-treatment and has no stop date will be assumed to be on-going for the remainder of the study. 2 A medication that is stopped on-treatment or post-treatment and has no start date recorded will be assumed to have been on-going from the pre-treatment phase Satisfactory clinical response A satisfactory clinical response at week 24 is defined as a subject who meets all of the following criteria: 10% improvement in 6MWD at Week 24 compared to baseline. 6MWD must be observed at this time without LOCF. No imputation is performed. WHO functional class I or II symptoms at Week 24. WHO FC must be assessed at week 24 with no LOCF/imputation. No events of clinical worsening prior to or at the week 24 visit. 44

352 CONFIDENTIAL If a subject has an event of clinical worsening prior to or at the week 24 visit, he will be counted as a non-satisfactory clinical response. If a subject does not have an event of clinical worsening prior to or at the week 24 visit, and the subject does not have a 6MWD value or a WHO functional class value at Week 24, the subject will be excluded from the analysis Short Form 36 Health questionnaire (SF-36) The SF-36 is a 36-item quality of life instrument designed to measure the subject s level of performance in nine health domains: Physical Functioning (PF) - 10 items, Role-Physical (RP) 4 items, Social Functioning (SF) 2 items, Bodily Pain (BP) - 2 items, Mental Health (MH) 5 items, Role-Emotional (RE) 3 items, Vitality (VT) - 4 items, and General Perception of Health (GH) 5 items, Reported Health Transition (HT) 1 item. Two component scores can also be calculated: Physical Component Summary (PCS) score (PF, RP, BP, GH) and Mental Component Summary (MCS) score (VT, SF, RE and MH). Note that HT is not used in the computation of the component scores. Further information and details of the scoring process are given in the Quality Metric Health Outcomes Scoring Software 4.5, User s Guide and Ware, Kosinski, Bjorner et al. (2008) Transformations for Efficacy Outcomes A log transformation will be applied to NT-proBNP data Treatment Compliance Rates Compliance is recorded at each visit in one of the following categorical groups:- 0% compliant (subject did not take any doses) 0% to < 80% compliant (subject missed a number of doses) 80% and 120%, (number of doses taken was within compliance range) > 120% compliant (number of doses taken exceeds compliance limits) Compliance is not captured separately for each therapy component. At the subject level compliance rate is calculated as 100*(the number of visits at which the subject was compliant (i.e. 80% and 120%) / (the sum of all study visits for the subject). At a treatment group level compliance rate is calculated as 45

353 CONFIDENTIAL 100*(the total number of visits at which all subjects in that group were compliant (i.e. 80% and 120%)) / (the sum of all study visits for all subjects in that group) Values of Clinical Concern Values of clinical concern for haematology, clinical chemistry and testicular function are as follows:- Testicular function: Free Testosterone, Inhibin and Follicle Stimulating Hormone less than 80% of the lower and more than 120% of upper limit of normal Creatinine: above/below normal range BUN: above/below normal range Anaemia: Haemoglobin <10g/dL Values of clinical concern for liver function parameters are defined in section There are no values of clinical concern for vital signs WHO FC change from baseline categorisation 1 Note that based on the study inclusion criteria subjects must have a WHO FC of II or III at baseline. Change from baseline at Week X will be calculated as Week X value minus baseline value, thus categories may be -2, -1, 0, +1, +2. Post Baseline WHO FC I II III IV Baseline WHO FC II III WHO FC change from baseline categorisation 2 Improved = -1 or -2 in Change from Baseline Categorisation No change = 0 in Change from Baseline Categorisation Deteriorated = +1 or +2 in Change from Baseline Categorisation 46

354 CONFIDENTIAL 9.3. ASSESSMENT WINDOWS Visit windowing All data will be assessed against the nominal visit for which it is collected. If multiple measurements are made at a scheduled visit the last measurement will be used in summaries, unless otherwise specified. No visit windows will be assigned, however data recorded as screening or baseline data that are post initiation of study drug will be reclassified as unscheduled. Any data relating to unscheduled visits will not be slotted to a particular time point. However for efficacy analyses, data from unscheduled visits will be included in the derivations of LOCF data (6MWD, BDI and WHO FC) if the unscheduled data are between a time point of interest and a missing observation. Additionally, for subjects with only baseline and EOS visits, the EOS measures will be included in LOCF summaries at all time points of interest. However, for the MMRM analyses unscheduled visits will NOT be carried forward for analyses. 10. STUDY POPULATION Population data will be summarised for the combination therapy, pooled monotherapy, for each individual monotherapy arm and for those who initiated BCT (this is a subset of all randomised columns). Unless otherwise specified, unscheduled data will only be listed. Table numbers referenced are for the mitt population, and will also be produced for the ITT (Tables 2.1 to 2.28) and non-mitt (Tables 3.1 to ) populations. Disposition of subjects: The number and percentage of subjects who complete or withdraw from the study will be tabulated, as will the number and percentage of subjects by primary reason for withdrawal for those subjects who withdraw from the study, overall and by whether prior to or after adjudicated clinical failure event (Table 1.1). Disposition data will also be listed (Listing 1.1). Investigational Product Discontinuation: The number and percentage of subjects who discontinued IP will be tabulated, as will the reason for discontinuation, overall and by whether prior to or after adjudicated clinical failure event (Table 1.2) and by visit (Table 1.3). These two tables relate to permanent discontinuation of IP, not interruptions. Investigational Product Discontinuation data will also be listed (Listing 1.2). Subjects 47

355 CONFIDENTIAL randomised but not receiving investigational product will be listed (Listing 1.3). The first patient first visit and last patient last visit will be listed (Listing 1.4) Study populations: The number and percentage of subjects in the mitt, ITT, non-mitt and PP populations will be summarised as will the number and percentage of subjects excluded from each population by reason for exclusion (Table 1.4). The number and percentage of subjects in the mitt, ITT, non-mitt and PP populations will also be summarised for those who initiated BCT therapy (Table 1.4). The number and percentage of subjects in the peak/trough, non-mitt that have 3, 4 or >4 criteria of the left ventricular disease/dysfunction risk factors and mitt with 0, 1 or 2 risk factors populations will also be summarised (Table 1.4). Unblinded subjects will be listed (Listing 1.5). Study population data will also be listed (Listings 1.6, 1.7, 1.8, 1.9). Inclusion / exclusion criteria deviations: The number and percentage of subjects who have inclusion and / or exclusion criteria deviations will be tabulated overall and by deviation (Table 1.5). This table will also be used to summarise inclusion and / or exclusion deviations relating to the Per-Protocol population (Table 1.6). Important protocol deviations will be tabulated (Table 1.7). Inclusion / exclusion criteria deviations data will also be listed (Listing 1.10). All important protocol deviations will be listed (Listing 1.11). Demographic and baseline characteristics: The number and percentage of subjects by region, country and investigator will be tabulated (Table 1.8). The number and percentage of subjects by sex, child bearing potential, ethnicity and region, and descriptive summary statistics for age will be tabulated. The number and percentage of subjects who had concomitant calcium channel blockers, creatinine clearance categories, aetiology of PAH, and WHO FC will also be tabulated in the same table (Table 1.9). Smoking history will also be tabulated (Table 1.10). Subject recruitment by region, country and site number will be listed (Listing 1.12). Demographic characteristics data will also be listed, as will data for smoking history (Listings 1.13, 1.14 & 1.15). Listing 1.8 which presents the randomisation strata (aetiology of PAH and WHO FC), will also present PAH diagnosis and WHO FC at screening, flagging subjects where there are differences compared with randomisation strata. 48

356 CONFIDENTIAL Race and racial combinations: The number and percentage of subjects will be summarised by race and racial combinations (Table 1.11). Race data will also be listed (Listing 1.16). Medical history: GSK will provide groupings for tabulation of unique verbatim terms for conditions of interest, Table 1.12 will summarise these groupings by treatment. Specific mappings of verbatim terms to those conditions will be summarised (Table 1.13). Medical History data will also be listed (Listing 1.17). Pulmonary Arterial Hypertension: The number and percentage of subjects who have a PAH history and idiopathic or heritable PAH will be summarised as will data on association of the disease. Descriptive statistics will also be presented to summarise time from diagnosis to randomisation (Table 1.14). Pulmonary Arterial Hypertension data will also be listed (Listing 1.18). PAH therapy history: PAH therapy history will be summarised in terms of the number and percentage of subjects who had at least one therapy and the number and percentage of subjects who had each GSK coded therapy (Table 1.15). PAH therapy history data will also be listed (Listing 1.19). Historical / baseline right heart catheterisation: Right heart catheterisation data will be summarised by means of descriptive statistics (Table 1.16). Right heart catheterisation data will also be listed (Listing 1.20). Medical / surgical procedure history: The number and percentage of subjects who had at least one prior medical / surgical procedure will be summarised, as will the number and percentage of subjects by verbatim term (Table 1.17). GSK will provide groupings for tabulation of unique verbatim terms that may affect exercise tolerance or haemodynamics, Table 1.18 will summarise these groupings by treatment group. Specific mappings of verbatim terms to those groupings will be summarized (Table 1.19). Procedure history data will also be listed (Listing 1.21). Oxygen saturation: Oxygen saturation at baseline will be summarised by means of descriptive statistics (Table 1.20). Oxygen saturation data will also be listed (Listing 1.22). 49

357 CONFIDENTIAL Pulmonary function test: Pulmonary function test results will be summarised by means of descriptive statistics (Table 1.21). Pulmonary function test data will also be listed (Listing 1.23). Treatment changes: The number and percentage of subjects with protocol driven up titrations, separation of dosing, down-titrations, and initiation of BCT will be tabulated by visit (Table 1.22), before and after initiation of BCT. Treatment titration data will also be listed including blinded combination therapy stage where appropriate (Listing 1.24). Treatment compliance: Treatment compliance will be summarised overall, by whether prior to/after clinical failure event and by visit, overall, by whether prior/to after clinical failure event and by whether prior to/after BCT, by compliance categories (Tables 1.23 and 1.24). Treatment compliance data will also be listed (Listing 1.25). See RAP section for algorithm for calculating compliance. Concomitant medications: The number and percentage of subjects who had at least one concomitant medication will be tabulated, as will the number and percentage of subjects who had each concomitant medication by Anatomical Therapeutic Chemical (ATC) level 1 and ingredients (Table 1.25). Concomitant medications data will also be listed (Listing 1.26). Prostanoid therapy use: The number and percentage of subjects who had and did not have prostanoid therapy will be tabulated, as will the number and percentage of subjects who had parenteral and non-parenteral prostanoid therapy pre and post first adjudicated clinical failure event (Table 1.26). This information will also be summarised by visit (Table 1.27). Initiation of parenteral prostanoid therapy is a clinical failure event. By convention, if it is the first adjudicated clinical failure event, it will be counted under pre event. Prostanoid therapy use data will also be listed (Listing 1.27). Pregnancy: The number and percentage of female subjects who became pregnant during the study will be tabulated (Table 1.28). Pregnancy data will also be listed (Listing 1.28). 11. EFFICACY ANALYSES Efficacy data will be summarised for the combination therapy, pooled monotherapy, and for each individual monotherapy arm. Efficacy analysis tables will be presented for the 50

358 CONFIDENTIAL treatment arm comparisons of interest as detailed in RAP section 7. Unless otherwise specified, unscheduled data will only be listed. Table and Figure numbers referenced are for the mitt population, and will also be produced for the ITT (Tables 5.1 to 5.63, 9.1 to 9.61, 12.1 to 12.2, 15.1 to and Figures 2.1 to 2.39), non-mitt populations (Tables 6.1 to 6.59, 10.1 to 10.61, 13.1 to 13.2 and 16.1 to note that summary tables only will be populated, analysis tables will be shells with Analysis not performed for Non-mITT population ) PRIMARY EFFICACY ANALYSIS The primary analysis, which will be based on TTCF, will use all adjudicated (positive) events determined by the CEAC. Calculation of the time to event or censoring is described previously in RAP section Primary analysis Summary Tables: The number and percentage of subjects who had one or more adjudicated clinical failure event between randomisation and FAV will be summarised by first event, both overall and for each event type (Table 4.1). Analysis Tables: The number and percentage of subjects who had an adjudicated clinical failure event between randomisation and FAV will be summarised, along with the number and percentage of subjects who had their information censored before, or at, their FAV. Kaplan-Meier estimates of the probability of failure at 1 year, 2 years and 3 years will be included. The adjusted hazard ratio and 95% confidence intervals, calculated by means of a Cox regression model on time to first adjudicated clinical failure event will also be tabulated (see RAP section 8.5.3). Statistical significance will be assessed by the stratified log rank test (Table 4.3). Covariate Analysis Table: A further supportive analysis will be presented for time from randomisation to FAV for subjects with clinical failure events, testing for inclusion of interactions between treatment and strata variables and the inclusion of the covariates mentioned in RAP section in the model (see section covariate analyses): (Table 4.4). 51

359 CONFIDENTIAL Primary outcome data will also be listed (Listing 1.29). A Kaplan-Meier plot will be produced to graphically display time to first adjudicated clinical failure from randomisation to FAV (Figure 1.1) Sensitivity analyses of primary analysis The primary summary table will also be repeated for All adjudicated clinical failure events between randomisation and FAV by event type (Table 4.2) First adjudicated clinical failure event between randomisation and EOS (Table 4.5) All adjudicated clinical failure events between randomisation and EOS by event type (Table 4.6) First adjudicated clinical worsening event between randomisation and FAV (Table 4.8) All adjudicated clinical worsening events between randomisation and FAV by event type (Table 4.9) First adjudicated clinical worsening event between randomisation and EOS (Table 4.12) All adjudicated clinical worsening events between randomisation and EOS by event type (Table 4.13) First adjudicated clinical failure event between randomisation and FAV by subgroups of interest: aetiology of PAH (Table 4.28), baseline WHO functional class (Table 4.30), region (Table 4.32), baseline age group (<65, 65) (Table 4.34), baseline age above and below median (Table 4.36), sex (Table 4.38), baseline 6MWD above or below median (Table 4.40). First investigator assessed clinical failure event between randomisation and FAV (Table 4.42) 52

360 CONFIDENTIAL All investigator assessed clinical failure events between randomisation and FAV by event type (Table 4.43) First investigator assessed clinical failure event between randomisation and EOS (Table 4.45) All investigator assessed clinical failure events between randomisation and EOS by event type (Table 4.46) First investigator assessed clinical worsening event between randomisation and FAV (Table 4.48) All investigator assessed clinical worsening events between randomisation and FAV by event type (Table 4.49) First investigator assessed clinical worsening event between randomisation and EOS (Table 4.51) All investigator assessed clinical worsening events between randomisation and EOS by event type (Table 4.52) The following additional summary tables will also be presented:- Death classifications by CEAC between randomisation and FAV (Table 4.16) Death classifications by CEAC between randomisation and EOS (Table 4.18) Death classifications by CEAC between randomisation and last contact (Table 4.20) As supportive analyses, the primary analysis table will be additionally presented for First adjudicated clinical failure event between randomisation and EOS (Table 4.7) First adjudicated clinical worsening event between randomisation and FAV (Table 4.10) 53

361 CONFIDENTIAL First adjudicated clinical worsening event between randomisation and FAV, censoring at initiation of BCT (Table 4.11) First adjudicated clinical worsening event between randomisation and EOS (Table 4.14) First adjudicated clinical worsening event between randomisation and EOS, censoring at initiation of BCT (Table 4.15) Death from randomisation to FAV (Table 4.17) Death from randomisation to EOS (Table 4.19) Death from randomisation to last contact. This analysis includes all deaths reported during the study. (Table 4.21) First hospitalisation (adjudicated) from randomisation to FAV (Table 4.22) First hospitalisation (adjudicated) from randomisation to EOS (Table 4.23) First disease progression (adjudicated) from randomisation to FAV (Table 4.24) First disease progression (adjudicated) from randomisation to EOS (Table 4.25) First unsatisfactory long term clinical response (adjudicated) from randomisation to FAV (Table 4.26) First unsatisfactory long term clinical response (adjudicated) from randomisation to EOS (Table 4.27) First adjudicated clinical failure event between randomisation and FAV by subgroups of interest (see section subgroup analyses): aetiology of PAH (Table 4.29), baseline WHO functional class (Table 4.31), region (Table 4.33), baseline age group (<65, 65) (Table 4.35), baseline age above and below median (Table 4.37), sex (Table 4.39), baseline 6MWD above or below median (Table 4.41). Analysis by subgroups will not be stratified. First investigator assessed clinical failure event between randomisation and FAV (Table 4.44) 54

362 CONFIDENTIAL First investigator assessed clinical failure event between randomisation and EOS (Table 4.47) First investigator assessed clinical worsening event between randomisation and FAV (Table 4.50) First investigator assessed clinical worsening event between randomisation and EOS (Table 4.53) First hospitalisation (investigator assessed) from randomisation to FAV (Table 4.54) First hospitalisation (investigator assessed) from randomisation to last contact (Table 4.55) First disease progression (investigator assessed) from randomisation to FAV (Table 4.56) First disease progression (investigator assessed) from randomisation to EOS (Table 4.57) First unsatisfactory long term clinical response (investigator assessed) from randomisation to FAV (Table 4.58) First unsatisfactory long term clinical response (investigator assessed) from randomisation to EOS (Table 4.59) As supportive figures, the primary Kaplan-Meier plot will be additionally presented for First adjudicated clinical failure event between randomisation and EOS (Figure 1.2) First adjudicated clinical worsening event between randomisation and FAV (Figure 1.3) First adjudicated clinical worsening event between randomisation and EOS (Figure 1.4) Death from randomisation to FAV (Figure 1.5) 55

363 CONFIDENTIAL Death from randomisation to EOS (Figure 1.6) Death from randomisation to last contact (Figure 1.7) First hospitalisation (adjudicated) from randomisation to FAV (Figure 1.8) First hospitalisation (adjudicated) from randomisation to EOS (Figure 1.9) First disease progression (adjudicated) from randomisation to FAV (Figure 1.10) First disease progression (adjudicated) from randomisation to EOS (Figure 1.11) First unsatisfactory long term clinical response (adjudicated) from randomisation to FAV (Figure 1.12). First unsatisfactory long term clinical response (adjudicated) from randomisation to EOS (Figure 1.13). The above plots will also be repeated for investigator assessed events (Figures 1.14 to 1.23). As further supportive figures, a Forest plot will be produced to graphically display the hazard ratio point estimate and 95% confidence intervals for the treatment comparisons of interest from the analysis of time to first clinical failure event, clinical worsening, death, hospitalisation, disease progression, unsatisfactory long term clinical response, from randomisation to FAV for adjudicated events (Figure 1.24) and randomisation to EOS for adjudicated events (Figure 1.25), from randomisation to FAV for investigator assessed events (Figure 1.33) and randomisation to EOS for investigator assessed events (Figure 1.34). Forest plots will also be produced to graphically display the hazard ratio point estimate and 95% confidence intervals for the treatment comparisons of interest from the analysis of time to first clinical failure event from randomisation to FAV for adjudicated events by subgroups of interest: aetiology of PAH (Figure 1.26), baseline WHO functional class (Figure 1.27), region (Figure 1.28), baseline age group (<65, 65) (Figure 1.29), baseline age above and below median (Figure 1.30), sex (Figure 1.31), baseline 6MWD above or below median (Figure 1.32). 56

364 CONFIDENTIAL As an informative table, the concordance of adjudicated and investigator assessed first clinical failure event will be summarised from randomisation to FAV (Table 4.60) and from randomisation to EOS (Table 4.62). The Cohen Kappa Statistic and 95% confidence interval for overall agreement between adjudicated and investigator assessed events will be presented (see Appendix 3 for details). In addition, the concordance of adjudicated and investigator assessed clinical failure events by event type will be summarised from randomisation to FAV (Table 4.61) and from randomisation to EOS (Table 4.63) SECONDARY EFFICACY ANALYSES Secondary outcomes are defined in RAP section Testing of these outcomes will follow the hierarchical strategy outlined in RAP section Pro-B-type natriuretic peptide Summary and analysis tables in relation to NT-proBNP will be presented as follows: Descriptive statistics will be used to summarise log transformed NT-proBNP data by visit for each of the study arms, including % change from baseline summaries (Table 8.1). % change from baseline in NT-proBNP at week 24 will be analysed by means of a ANCOVA model (including terms for treatment, strata variables, and baseline) based on the log-transformed data (Table 8.2). No imputation will be performed for missing data. MMRM analyses will also be performed on log-transformed data to evaluate the time course of treatment effect (Table 8.2). Test for treatment by visit interaction will be conducted at p<0.10. Irrespective of the statistical significance for interaction between treatment comparisons will be made at each time point where NT-proBNP is measured to evaluate the onset of treatment effect. A plot will be included to illustrate the treatment effect over time. There will be 4 lines on each plot: one for each monotherapy, the pooled monotherapy and combination therapy (Figure 1.35). 57

365 CONFIDENTIAL NT-proBNP data will also be listed (Listing 1.30) Satisfactory clinical response Summary and analysis tables in relation to satisfactory clinical response will be presented as follows: The number and percentage of subjects who had, and did not have, a satisfactory clinical response at week 24 will be summarised, as will the number and percentage of subjects who failed each of the satisfactory clinical response criteria (a subject may have failed more than one criterion by week 24) (Table 8.3). Satisfactory clinical response at week 24 will be analysed by means of logistic regression with baseline stratification variables as covariates (see RAP section 8.5.5) (Table 8.4). Covariate Analysis Table: A further supportive logistic regression analysis will be presented testing for inclusion of interactions between the treatment and strata variables and the inclusion of covariates (see RAP section 8.5.5) (Table 8.5). Satisfactory clinical response at week 24 will be summarised for subgroups of interest (see RAP section 8.5.5): aetiology of PAH (Table 8.6), baseline WHO functional class (Table 8.8), region (Table 8.10), baseline age group (<65, 65) (Table 8.12), baseline age above and below median (Table 8.14), sex (Table 8.16), baseline 6MWD above or below median (Table 8.18). Analysis by subgroups will not be stratified. Satisfactory clinical response at week 24 will be analysed by logistic regression for subgroups of interest (see RAP section 8.5.5): aetiology of PAH (Table 8.7), baseline WHO functional class (Table 8.9), region (Table 8.11), baseline age group (<65, 65) (Table 8.13), baseline age above and below median (Table 8.15), sex (Table 8.17), baseline 6MWD above or below median (Table 8.19). Analysis by subgroups will not be stratified. Satisfactory clinical response data will also be listed (Listing 1.31). 58

366 CONFIDENTIAL MWD Summary and analysis tables in relation to 6MWD will be presented as follows: Descriptive statistics will be used to summarise 6MWD (meters) by visit for each of the study arms. This table will include change from baseline and percentage change from baseline summaries. The table will be produced for raw data and imputed data (LOCF/worst case see RAP section 9.1.4) (Table 8.20). A further table will be produced with and without imputed data (LOCF/worst case) to summarise the change and percentage change between baseline and week 24 only (Table 8.21). The table will also be repeated by subgroups of interest: aetiology of PAH (Table 8.25), baseline WHO functional class (Table 8.27), region (Table 8.29), baseline age group (<65, 65) (Table 8.31), baseline age above and below median (Table 8.33), sex (Table 8.35), baseline 6MWD above or below median (Table 8.37). Change from baseline in 6MWD will be analysed using the Wilcoxon Rank Sum test stratified by PAH aetiology and WHO functional class (Table 8.22). Results will be presented for each time point where 6MWD is measured. This includes the measurements taken at 12 weekly intervals following the week 24 clinic visit. The test will be based on LOCF data, except in the cases of an adjudicated clinical failure event preceding the missing data observation. In those cases, the missing observations will be assigned the worst-rank scores relative to those actually observed and will be assigned ranks reflecting the relative order of the actual event times [Lachin JM, et al. Worst Rank Score Analysis with Informatively Missing Observations in Clinical Trials. Controlled Clinical Trials : ] Sensitivity analyses will include analysis of the change from baseline in 6MWD by means of o MMRM enabling an evaluation of the onset and maintenance of treatment effect (Table 8.22). Test for treatment by visit interaction will be conducted at p<0.10. Irrespective of the statistical significance for interaction, between treatment comparisons will be made at each time point up to and including week 24 where 6MWD is measured to evaluate 59

367 CONFIDENTIAL the onset of treatment effect. Time points beyond week 24 will not be included in the MMRM model. No imputation of missing values will be performed. o A plot will be included to illustrate the treatment effect over time (displaying LSmeans +/1 SE). There will be 4 lines on each plot: one for each monotherapy, the combined monotherapy and combination therapy (Figure 1.36). o ANCOVA (see RAP section 8.5) at week 24. The analysis will be based on imputed (LOCF/worst case) data at week 24 only (Table 8.22). The above Wilcoxon Rank Sum Test analyses will also be repeated for subgroups of interest at week 24: aetiology of PAH (Table 8.26), baseline WHO functional class (Table 8.28), region (Table 8.30), baseline age group (<65, 65) (Table 8.32), baseline age above and below median (Table 8.34), sex (Table 8.36), baseline 6MWD above or below median (Table 8.38). A summary table will be produced to summarise the number and percentage of subjects who used supplemental oxygen and the number and percentage of subjects who walk less than 6 minutes by visit (Table 8.23), for both observed and imputed data. In this table summary statistics will also be presented to summarise oxygen use for those subjects who used supplemental oxygen and time walked for those subjects who walked for less than 6 minutes. The number and percentage of subjects by change in 6MWD from baseline at week 24 in 30m increments (i.e. -90, >-90 to -60, >-60 to -30, >-30 to <0, 0 to <30, 30 to <60, 60 to <90, 90) will be tabulated (Table 8.24), for both observed and imputed data, and graphically displayed (Figure 1.37). Similar summaries using 20m increments (i.e. -80, >-80 to -60, >-60 to -40, >-40 to - 20, >-20 to <0, 0 to <20, 20 to <40, 40 to <60, 60 to <80, 80) will be tabulated (Table 8.24) and graphically displayed (Figure 1.38). 6MWD data will also be listed (Listings 1.32 and 1.33). 60

368 CONFIDENTIAL WHO FC scores Summary and analysis tables in relation to WHO FC scores will be presented as follows: Descriptive statistics will be used to summarise WHO FC scores by visit for each of the study arms. This table will include change from baseline summaries. The table will be produced for raw data and imputed data (LOCF/worst case see RAP section 9.1.4) (Table 8.39). A shift table will be produced to summarise shifts from baseline by visit for each of the study arms. The table will be produced for raw data and imputed data (LOCF/worst case see RAP section 9.1.4) (Table 8.40). A further table will be produced with and without imputed data (LOCF/worst case) to summarise the change between baseline and week 24 only (Table 8.41). Change from baseline WHO FC scores will be analysed as for 6MWD using the Wilcoxon Rank Sum Test stratified by PAH aetiology and WHO FC. Results will be presented for each time point where WHO FC is measured (Table 8.42). Further tables will be produced summarising change in WHO FC scores categorised in terms of +2, +1, 0, -1, -2 and in terms of Improved, No Change, and Deteriorated (Table 8.43). Change from baseline categorization (+2,+1,0.-1,-2) and change from baseline categorization using the following three categories: Improved, No change, and Deteriorated will be analysed at week 24 by means of Cochran Mantel Haenszel tests with baseline stratification variables as covariates (see RAP section 8.5.2). The analysis will be based on imputed data (LOCF/worst case) (Table 8.44). WHO FC data will also be listed (Listing 1.34) BDI Scores Summary and analysis tables in relation to BDI scores will be presented as follows: Descriptive statistics will be used to summarise BDI by visit for each of the study arms. This table will include change from baseline summaries. The table will be 61

369 CONFIDENTIAL produced for raw data and imputed data (LOCF/worst case see RAP section 9.1.4) (Table 8.45). A further table will be produced with and without imputed data (LOCF/worst case) to summarise the change between baseline and week 24 only (Table 8.46). The table will also be repeated by subgroups of interest: aetiology of PAH (Table 8.48), baseline WHO functional class (Table 8.50), region (Table 8.52), baseline age group (<65, 65) (Table 8.54), baseline age above and below median (Table 8.56), sex (Table 8.58), baseline 6MWD above or below median (Table 8.60). Change from baseline BDI will be analysed as for 6MWD using the Wilcoxon Rank Sum Test (as per 6MWD) stratified by PAH aetiology and WHO functional class. Results will be presented for each time point where BDI is measured (Table 8.47). Sensitivity analyses will include analysis of the change from baseline in BDI at week 24 by means of o MMRM enabling an evaluation of the onset and maintenance of treatment effect (Table 8.47). Test for treatment by visit interaction will be conducted at p<0.10. Irrespective of the statistical significance for interaction, between treatment comparisons will be made at each time point where BDI is measured to evaluate the onset of treatment effect. No imputation of missing values will be performed. o A plot will be included to illustrate the treatment effect over time (displaying LSmeans +/1 SE). There will be 4 lines on each plot: one for each monotherapy, the combined monotherapy and combination therapy (Figure 1.39). o ANCOVA (see RAP section 8.5.1) at week 24. The analysis will be based on imputed (LOCF/worst case) data at week 24 (Table 8.47). The above Wilcoxon Rank Sum Test analyses will also be repeated by subgroups of interest at week 24: aetiology of PAH (Table 8.49), baseline WHO functional class (Table 8.51), region (Table 8.53), baseline age group (<65, 65) 62

370 CONFIDENTIAL (Table 8.55), baseline age above and below median (Table 8.57), sex (Table 8.59), baseline 6MWD above or below median (Table 8.61). BDI data will also be listed (Listing 1.35) EXPLORATORY OUTCOMES The change from baseline in 6MWD at week 16 will be analysed by means of an ANCOVA model (see RAP section 8.5.1) comparing the combination therapy group versus the TAD monotherapy group (Table 11.1). Stratification factors will be baseline aetiology of PAH (IPAH/HPAH vs Non-IPAH) and WHO functional class (II vs III). The analysis will be based on observed data as well as imputed data (LOCF/worst case). The ratio of the tadalafil-corrected mean change from baseline trough 6MWD at week 16 to the tadalafil-corrected mean change from baseline peak 6MWD at week 16 will be calculated. The ratio will be based on observed data with no imputation. The ratio will be determined as follows: 6MWD trough - 6MWD tad 6MWD peak - 6MWD tad where, 6MWD trough = mean change from baseline 6MWD for the subjects in the ambrisentan/tadalafil combination group whose 6MWD was assessed at trough ambrisentan plasma concentrations 6MWD peak = mean change from baseline 6MWD for the subjects in the ambrisentan/tadalafil combination group whose 6MWD was assessed at peak ambrisentan plasma concentrations 6MWD tad = mean change from baseline 6MWD for the tadalafil monotherapy group. Example:- 63

371 CONFIDENTIAL If 6MWD trough = 35.99m, 6MWD peak = 36.57m and 6MWD tad = 30.29m the Ratio = ( ) / ( ) = 5.70 / 6.28 = In addition to reporting the above ratio, an interval estimate for the ratio using Hinkley s Method (Hinkley, 1969) will be provided (Table 11.2) HEALTH OUTCOME OUTCOMES Health outcome data will be summarised for the combination therapy, pooled monotherapy, and for each individual monotherapy arm. Health outcome analysis tables will be presented for the treatment arm comparisons of interest as detailed in RAP section SF-36 Summary and analysis tables in relation to SF-36 scores will be presented as follows: Descriptive statistics will be used to summarise SF-36 scores by visit for each of the study arms for each domain and component score. The tables will include change from baseline summaries (Tables 14.1 to 14.11). Further tables will be produced to summarise the change between baseline and week 24 only for each domain and component scores (Tables to 14.22). The change from baseline in domain and component scores will be analysed by means of MMRM (Tables to 14.33). Test for treatment by visit interaction will be conducted at p<0.10. Irrespective of the statistical significance for interaction, between treatment comparisons will be made at each time point where SF-36 is measured to evaluate the onset of treatment effect. No imputation of missing values will be performed. SF-36 data will also be listed for each domain and component scores (Listing 1.36 to 1.46) CAMPHOR Summary and analysis tables in relation to CAMPHOR scores will be presented as follows for subjects from sites in countries where language translation is validated 64

372 CONFIDENTIAL (Canada (Canadian French and Canadian English language), Germany, Sweden, UK, US, and Australia). Descriptive statistics will be used to summarise CAMPHOR scores by visit for each of the study arms for symptoms scores, symptoms subscale scores, QoL scores and activities scores. The tables will include change from baseline summaries (Tables to 14.39). Further tables will be produced to summarise the change between baseline and week 24 only for symptoms scores, symptoms subscale scores, QoL scores and activities scores (Tables to 14.45). The change from baseline in symptoms scores, symptoms subscale scores, QoL scores and activities scores will be analysed by means of MMRM (Tables to 14.51). Test for treatment by visit interaction will be conducted at p<0.10. Irrespective of the statistical significance for interaction, between treatment comparisons will be made at each time point where CAMPHOR is measured to evaluate the onset of treatment effect. No imputation of missing values will be performed. CAMPHOR data will also be listed for symptoms scores, symptoms subscale scores, QoL scores and activities scores (Listing 1.47 to 1.52). 12. SAFETY ANALYSES Safety analyses will be performed for the mitt, ITT, and non-mitt populations. All Safety displays will be presented by randomised treatment group. Two sets will be presented (a) before initiation of BCT (i.e. on randomised treatment) and (b) after initiation of BCT. Safety data will be displayed in terms of the following treatment categories:- Before initiation of BCT: Combination therapy, Ambrisentan Monotherapy, Tadalafil Monotherapy. After initiation of BCT: BCT, BCT Ambrisentan Monotherapy, BCT Tadalafil Monotherapy. The monotherapy groups here are for subjects on low dose or prior to 65

373 CONFIDENTIAL protocol up-titration who will be first maximized at their randomised therapy before a second therapy is added. Thereafter subjects will be tabulated under blinded combination therapy. NB: Some investigator-assessed clinical failure events (events need not be adjudicated as events but BCT is initiated by the investigator) may not lead to initiation of BCT so this is an important qualifier as the intent of the display is to lend some interpretation to assessment of safety profiles knowing that all subjects are receiving active tadalafil and active ambrisentan. There will be a small number of subjects randomised to ambrisentan monotherapy or tadalafil monotherapy who remain on monotherapy after an investigator initiated event which leads to initiation of BCT. Hence there will be some confounding in these summaries of monotherapy with monotherapy subjects who transition to active BCT. These will be subjects who are (1) ambrisentan monotherapy subjects who have an investigator assessed event which leads to BCT before the forced week 8 titration (2) tadalafil monotherapy subjects who have an investigator assessed event which leads to BCT before the forced week 4 titration (3) ambrisentan or tadalafil monotherapy subjects who are on low dose of their respective therapies who will be first maximized at their randomised therapy before a second therapy is added. NB: Recognising the protocol defines the safety population as summarising subjects according to the treatment received and any medication errors that occur should be of short term and not sustained duration it is reasonable to summarize safety by randomised treatment group. However, some assessment of agreement with randomised treatment should be summarised. To enable this evaluation, the following is proposed: 1. The duration of time a subject is on randomised treatment will be calculated. This is date of last randomised dose minus date of first randomised dose plus 1 (eg. a subject was randomised to ambrisentan on 31 December 2010 and their last dose of randomised treatment was on 01 January This is 367 days). 2. The concordance between actual treatment received and randomised treatment will be summarised (Table 17.2). 66

374 CONFIDENTIAL 3. If the concordance is < 90%, further assessment as to lack of concordance will be performed. Table numbers referenced are for the mitt population, and will also be produced for the ITT (Tables 18.1 to 18.72) and non-mitt (Tables 19.1 to 19.72) populations Extent of Exposure Exposure (in days) will be calculated as described in RAP section Descriptive statistics will be used to summarise overall exposure to the study drugs in terms of daily dose (mg), cumulative dose (mg) (see RAP section ), and days on study drug (Table 17.1). Exposure will be split out into exposure to randomised therapy and exposure to BCT. Exposure data will also be listed (Listing 1.53) Adverse Events Adverse events (AEs) will be coded using the Medical Dictionary for Drug Regulatory Activities (MedDRA) coding dictionary. A mapping of the MedDRA primary system organ class and preferred term to which each verbatim term has been coded will be provided in a listing. Only treatment emergent adverse events (TEAEs) will be tabulated. TEAEs are defined as those events that start on the day of or after the day of IP initiation up to 30 days post last dose of IP. Each TEAE will be categorised into periods according to its onset date. The periods are as follows: Baseline to Final Assessment Visit, Final Assessment Visit to End of Study Visit and End of Study Visit to Last Contact. Within these periods events are categorized further into those with onset prior to BCT initiation and those with onset from BCT initiation. An AE can only be categorised into one group. All TEAEs will be summarised by system organ class and preferred term (Table 17.4). This table will be repeated for adverse events occurring in the following study phases: from baseline to FAV (Table 17.6), from FAV to EOS (Table 17.8) and from EOS to Last Contact (Table 17.9 produced after second database lock). This table will also be repeated for most common ( 5% in any group) adverse events overall (Table 17.5) and from baseline to FAV (Table 17.7) 67

375 CONFIDENTIAL The above tables will be repeated for those adverse events occurring prior to the first adjudicated clinical failure event overall (Table 17.10) and additionally for those adverse events occurring after the first adjudicated clinical failure event overall (Table 17.11). Treatment related TEAEs will be tabulated overall, by study phase and for those occurring before and after the first adjudicated clinical failure event overall (Tables 17.12, 17.14, to 17.19). This table will also be repeated for most common ( 5% in any group) treatment-related TEAEs overall (Table 17.13) and from baseline to FAV (Table 17.15) If relationship is missing, it should be counted as related in tables summarising TEAE relationship and left missing in the listings. Tables will be footnoted describing the handling of missing AE relationship. Treatment related TEAEs of special interest (anemia, fluid retention, hypersensitivity, hypotension, liver events) will be tabulated overall, by study phase and for those occurring before and after the first adjudicated clinical failure event overall (Tables to 17.25). These events will be identified by blinded review by GSK. TEAEs leading to withdrawal from the study will be tabulated overall, by study phase (Tables to 17.29). TEAEs leading to permanent discontinuation of investigational product will be tabulated overall, by study phase (Tables to 17.33). Serious TEAEs will be tabulated overall, by study phase and for those occurring before and after the first adjudicated clinical failure event overall (Tables to 17.39). Fatal serious TEAEs will be tabulated overall, by study phase (Tables to 17.43) The most common ( 5% in any group) non-serious TEAEs will be tabulated overall (Table 17.44). TEAEs will be tabulated by maximum severity from baseline to FAV and FAV to EOS (Tables to 17.46). If the severity is missing for a given TEAE it would 68

376 CONFIDENTIAL be imputed and counted as severe. In the individual subject listings it would remain missing. Tables will be footnoted describing the handling of missing TEAE severity. An overview of TEAEs, serious non-fatal, fatal and leading to study withdrawal relative to relationship to investigational product will be tabulated pooling monotherapy (Table 17.3). Adverse event data will also be listed (Listing 1.54 to 1.58). It is possible there may be subjects with adverse events whose onset date is after their FAV visit but before the last subject s FAV. These adverse events could either resolve prior to the last subject s FAV or be ongoing. Consequently, they could have been in the first database freeze. However, these adverse events would not be considered to be part of the primary analyses because their onset date was after the last subject s FAV. Following the second database freeze a PROC COMPARE of the two adverse event data sets will be performed to identify for each subject any AE s that occurred after their FAV date but before the last subject s FAV date. This includes an individual subject s adverse events that were continuing at their FAV and changed afterwards (i.e. either resolved, severity, relationship). These adverse events will be listed, and if appropriate summarized. The cumulative incidence plots for adverse events of special interest and most common AE relative risk plots will be produced for the FAV SAC. Following review of the data and in particular the incidence of these specific events, similar plots may be produced following the FAV SAC as ad hoc plots, from BCT initiation Laboratory Data Laboratory data will be summarised (up to 30 days post last dose of IP) as follows:- Haematology data will be summarised by means of descriptive statistics (Table 17.47). The table will be repeated for change from baseline data (Table 17.48). Haematology data of potential clinical concern will also be summarised at each visit through Week 24 and cumulatively from the start of IP 69

377 CONFIDENTIAL to Week 24, start to FAV and start to EOS (Table 17.49). Haematology, Coagulation, Immunology and Virology data will also be listed (Listing 1.59). Chemistry data will be summarised by means of descriptive statistics (Table 17.50). The table will be repeated for change from baseline data (Table 17.51). Chemistry data of potential clinical concern will also be summarised at each visit through Week 24 and cumulatively from the start of IP to Week 24, start to FAV and start to EOS (Table 17.52). Chemistry data will also be listed (Listing 1.60). The number and percentage of subjects with ALT (Table 17.53) or AST (Table 17.54) within the following categories at each visit through Week 24 and cumulatively from the start of IP to Week 24, start to FAV and start to EOS will be presented. 1xULN >1xULN and 3xULN > 3xULN >3xULN and 5xULN >5xULN and 8xULN >8xULN The number and percentage of subjects with alkaline phosphatase (Table 17.55) or total bilirubin (Table 17.56) within the following categories at each visit through Week 24 and cumulatively from the start of IP to Week 24, start to FAV and start to EOS will be presented: <2xULN 2xULN The number and percentage of subjects with ALT or AST >3xULN and total bilirubin >2xULN at each visit through Week 24 and cumulatively from the start of IP to Week 24, start to FAV and start to EOS will be presented (Table 17.57). The number and percentage of subjects with shifts from baseline ALT (Table 17.58), AST (Table 17.59), alkaline phosphatase (Table 17.60) and total 70

378 CONFIDENTIAL bilirubin (Table 17.61), at each visit through Week 24 and cumulatively from the start of IP to week 24, start to FAV and start to EOS will be presented. Testicular function data (males only) will be summarised by means of descriptive statistics (Table 17.62). The table will be repeated for change from baseline data (Table 17.63). Testicular function data of potential clinical concern will also be summarised at each visit through Week 24 and cumulatively from the start of IP to Week 24, start to FAV and start to EOS (Table 17.64). Testicular function data will also be listed (Listing 1.64). Subjects meeting Hy s Law criteria will be listed (Listing 1.61) ALT, AST, alkaline phosphatase, GGT, total bilirubin for subjects who have experienced a Liver Event will be listed (Listing 1.62). Local laboratory data will be listed (Listing 1.63) Urinalysis Dipstick Results will be tabulated (Table 17.65). All Urinalysis data will also be listed (Listing 1.65) Vital signs Vital signs data will be summarised (up to 30 days post last dose of IP) by means of descriptive statistics (Table 17.66). The table will be repeated for change from baseline data (Table 17.67). Vital signs data will also be listed (Listing 1.66) Liver events Liver event data will be summarised (up to 30 days post last dose of IP) (Table 17.68). Time on treatment before liver event will be summarised (Table 17.69). Liver biopsy details (Table 17.70), liver imaging details (Table 17.71). Liver event data will also be listed (Listing 1.67 to 1.71) ECG Shifts from baseline in ECG findings will be summarised (up to 30 days post last dose of IP) (Table 17.72). ECG data will also be listed (Listing 1.72). 71

379 CONFIDENTIAL 12.7 VIRAL RESULTS Viral results will be listed (Listing 1.73). 72

380 CONFIDENTIAL 13. REFERENCES Cohen, J., A Coefficient of Agreement for Nominal Scales. Educational and Psychological Measurement, 20, Fleiss, J. L., Levin, B., and Paik, M. C., 2003, Statistical Methods for Rates and Proportions, Third Edition, New York: John Wiley & Sons. Fleiss, J. L. and Cohen, J., The Equivalence of Weighted Kappa and the Intraclass Correlation Coefficient as Measures of Reliability. Educational and Psychological Measurement, 33, Fleiss, J. L., Cohen, J., and Everitt, B. S., Large-Sample Standard Errors of Kappa and Weighted Kappa. Psychological Bulletin, 72, Hinkley, D.V., On the ratio of two correlated normal random variables. Biometrika, 56 (3): 635. Hosmer, D.W. and Lemeshow, S., Applied Logistic Regression. 2nd Ed. John Wiley & Sons, USA. Lachin JM, et al Worst Rank Score Analysis with Informatively Missing Observations in Clinical Trials. Controlled Clinical Trials, 20: Quality Metric Health Outcomes Scoring Software 4.5, User s Guide, Appendix D, SAS Institute Inc., 2008a. SAS/STAT 9.2 User s Guide. The MIXED Procedure. Cary, NC: SAS Institute Inc. SAS Institute Inc., 2008b. SAS/STAT 9.2 User s Guide. The PHREG Procedure. Cary, NC: SAS Institute Inc. SAS Institute Inc., 2008c. SAS/STAT 9.2 User s Guide. The LIFETEST Procedure. Cary, NC: SAS Institute Inc. SAS Institute Inc., 2008d. SAS/STAT 9.2 User s Guide. The GENMOD Procedure. Cary, NC: SAS Institute Inc. SAS Institute Inc., 2008e. SAS/STAT 9.2 User s Guide. The LOGISTIC Procedure. Cary, NC: SAS Institute Inc. SAS Institute Inc., 2008f. SAS/STAT 9.2 User s Guide. Introduction to Nonparametric Analysis. Cary, NC: SAS Institute Inc. Ware, J.E., Kosinski, M., Bjorner, J.B. et al. (2008). SF-36v2 Health Survey: Administration guide for clinical trials investigators. Lincoln, RI: QualityMetric Incorporated. 73

381

382 CONFIDENTIAL Summary of Medical/Surgery Procedure History by Grouped Term and Verbatim - Procedures Affecting Y Table 1.19 Exercise Capacity or Haemodynamics Table 1.20 Summary of Oxygen Saturation at Baseline Y Table 1.21 Summary of Pulmonary Function Test at Baseline Y Table 1.22 Summary of Treatment Changes by Visit Y Table 1.23 Summary of Investigational Product Compliance Overall Y Table 1.24 Summary of Investigational Product Compliance by Visit Y Table 1.25 Summary of Concomitant Medications Y Table 1.26 Summary of Prostanoid Therapy Use Y Table 1.27 Summary of Prostanoid Therapy Use by Visit Y Table 1.28 Summary of Pregnancy During Study (Female subjects only) Y The above tables will be repeated for the ITT population (Tables 2.1 to 2.28), and for the Non-mITT population (Table 3.1 to 3.28). 75

383

384 CONFIDENTIAL Table 4.21 Analysis of Time to Death (Baseline to Last Contact) Y Table 4.22 Analysis of Time to First Hospitalisation (Adjudicated) (Baseline to Final Assessment Visit) Y Table 4.23 Analysis of Time to First Hospitalisation (Adjudicated) (Baseline to End of Study Visit) Y Table 4.24 Analysis of Time to First Disease Progression (Adjudicated) (Baseline to Final Assessment Visit) Y Table 4.25 Analysis of Time to First Disease Progression (Adjudicated) (Baseline to End of Study Visit) Y Table 4.26 Analysis of Time to First Unsatisfactory Long Term Clinical Response (Adjudicated) (Baseline to Final Y Assessment Visit) Table 4.27 Analysis of Time to First Unsatisfactory Long Term Clinical Response (Adjudicated) (Baseline to End of Y Study Visit) Table 4.28 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Aetiology of Y PAH Table 4.29 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Aetiology of Y PAH Table 4.30 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Baseline Y WHO Functional Class Table 4.31 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Baseline Y WHO Functional Class Table 4.32 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Region Y Table 4.33 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Region Y Table 4.34 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Baseline Y Age Group Table 4.35 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Baseline Age Y Group Table 4.36 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Baseline Y Age Above or Below Median Baseline Age Table 4.37 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Baseline Age Y Above or Below Median Baseline Age Table 4.38 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Sex Y Table 4.39 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Sex Y Table 4.40 Summary of First Adjudicated Clinical Failure Criteria (Baseline to Final Assessment Visit) by Baseline Six Y Minute Walk Distance (6MWD) Above or Below Median Baseline Six Minute Walk Distance (6MWD) Table 4.41 Analysis of Time to First Adjudicated Clinical Failure (Baseline to Final Assessment Visit) by Baseline Six Y Minute Walk Distance (6MWD) Above or Below Median Baseline Six Minute Walk Distance (6MWD) Table 4.42 Summary of First Investigator Assessed Clinical Failure Criteria (Baseline to Final Assessment Visit) Y Table 4.43 Summary of Investigator Assessed Clinical Failure Events by Event Type (Baseline to Final Assessment Y Visit) 77

385 CONFIDENTIAL Table 4.44 Analysis of Time to First Investigator Assessed Clinical Failure (Baseline to Final Assessment Visit) Y Table 4.45 Summary of First Investigator Assessed Clinical Failure Criteria (Baseline to End of Study Visit) Y Table 4.46 Summary of Investigator Assessed Clinical Failure Events by Event Type (Baseline to End of Study Visit) Y Table 4.47 Analysis of Time to First Investigator Assessed Clinical Failure (Baseline to End of Study Visit) Y Table 4.48 Summary of First Investigator Assessed Clinical Worsening Criteria (Baseline to Final Assessment Visit) Y Summary of Investigator Assessed Clinical Worsening Events by Event Type (Baseline to Final Y Table 4.49 Assessment Visit) Table 4.50 Analysis of Time to First Investigator Assessed Clinical Worsening (Baseline to Final Assessment Visit) Y Table 4.51 Summary of First Investigator Assessed Clinical Worsening Criteria (Baseline to End of Study Visit) Y Summary of Investigator Assessed Clinical Worsening Events by Event Type (Baseline to End of Study Y Table 4.52 Visit) Table 4.53 Analysis of Time to First Investigator Assessed Clinical Worsening (Baseline to End of Study Visit) Y Table 4.54 Analysis of Time to First Hospitalisation (Investigator Assessed) (Baseline to Final Assessment Visit) Y Table 4.55 Analysis of Time to First Hospitalisation (Investigator Assessed) (Baseline to Last Contact) Y Table 4.56 Analysis of Time to First Disease Progression (Investigator Assessed) (Baseline to Final Assessment Y Visit) Table 4.57 Analysis of Time to First Disease Progression (Investigator Assessed) (Baseline to End of Study Visit) Y Table 4.58 Analysis of Time to First Unsatisfactory Long Term Clinical Response (Investigator Assessed) (Baseline Y to Final Assessment Visit) Table 4.59 Analysis of Time to First Unsatisfactory Long Term Clinical Response (Investigator Assessed) (Baseline Y to End of Study Visit) Table 4.60 Summary of Concordance of Adjudicated and Investigator Assessed First Clinical Failure Event Y (Baseline to Final Assessment Visit) Table 4.61 Summary of Concordance of Adjudicated and Investigator Assessed Clinical Failure Events by Event Y Type (Baseline to Final Assessment Visit) Table 4.62 Summary of Concordance of Adjudicated and Investigator Assessed First Clinical Failure Event (Baseline Y to End of Study Visit) Table 4.63 Summary of Concordance of Adjudicated and Investigator Assessed Clinical Failure Events by Event Y Type (Baseline to End of Study Visit) The above tables will be repeated for the ITT population (Tables 5.1 to 5.63), for the Non-mITT population (Table 6.1 to 6.63 note that summary tables only will be populated, analysis tables will be shells with Analysis not performed for Non-mITT population ), and for the PP population (if analysed) (Tables 7.1 to note summary and analysis tables will only be populated for endpoints up to FAV visit, summary and analysis tables up to EOS and Last Contact will be shells with Table only produced up to Final Assessment Visit for Per-Protocol population ) 78

386

387 CONFIDENTIAL Above or Below Median Baseline Six Minute Walk Distance (6MWD) Table 8.20 Summary of 6 Minute Walk Distance Results (meters) by Visit Y Table 8.21 Summary of 6 Minute Walk Distance Results (meters) at Week 24 Y Table 8.22 Analysis of 6 Minute Walk Distance Results (meters) Change from Baseline Y Table 8.23 Summary of 6 Minute Walk Distance Conduct by Visit Y Table 8.24 Summary of 6 Minute Walk Distance at Week 24 by Change from Baseline Categories (meters) Y Table 8.25 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Aetiology of PAH Y Table 8.26 Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Aetiology of PAH Y Change from Baseline Table 8.27 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline WHO Functional Class Y Table 8.28 Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline WHO Functional Class Y Change from Baseline Table 8.29 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Region Y Table 8.30 Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Region Y Change from Baseline Table 8.31 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline Age Group Y Table 8.32 Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline Age Group Y Change from Baseline Table 8.33 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline Age Above or Below Y Median Baseline Age Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline Age Above or Below Y Table 8.34 Median Baseline Age Change from Baseline Table 8.35 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Sex Y Table 8.36 Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Sex Y Change from Baseline Table 8.37 Summary of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline Six Minute Walk Distance Y (6MWD) Above or Below Median Baseline Six Minute Walk Distance (6MWD) Analysis of 6 Minute Walk Distance Results (meters) at Week 24 by Baseline Six Minute Walk Distance Y Table 8.38 (6MWD) Above or Below Median Baseline Six Minute Walk Distance (6MWD) Change from Baseline Table 8.39 Summary of WHO Functional Class by Visit Y Table 8.40 Summary of WHO Functional Class Shifts from Baseline by Visit Y Table 8.41 Summary of WHO Functional Class at Week 24 Y 80

388 CONFIDENTIAL Analysis of WHO Functional Class Y Table 8.42 Change from Baseline Table 8.43 Summary of WHO Functional Class Change from Baseline Categorisation by Visit Y Table 8.44 Analysis of WHO Functional Class Change from Baseline Categorisation at Week 24 Y Table 8.45 Summary of BDI Scores by Visit Y Table 8.46 Summary of BDI Scores at Week 24 Y Table 8.47 Analysis of BDI Scores Change from Baseline Table 8.48 Summary of BDI Scores at Week 24 by Aetiology of PAH Y Table 8.49 Analysis of BDI Scores at Week 24 by Aetiology of PAH Y Change from Baseline Table 8.50 Summary of BDI Scores at Week 24 by Baseline WHO Functional Class Y Table 8.51 Analysis of BDI Scores at Week 24 by Baseline WHO Functional Class Y Change from Baseline Table 8.52 Summary of BDI Scores at Week 24 by Region Y Table 8.53 Analysis of BDI Scores at Week 24 by Region Y Change from Baseline Table 8.54 Summary of BDI Scores at Week 24 by Baseline Age Group Y Table 8.55 Analysis of BDI Scores at Week 24 by Baseline Age Group Y Change from Baseline Table 8.56 Summary of BDI Scores at Week 24 by Baseline Age Above or Below Median Baseline Age Y Table 8.57 Analysis of BDI Scores at Week 24 by Baseline Age Above or Below Median Baseline Age Y Change from Baseline Table 8.58 Summary of BDI Scores at Week 24 by Sex Y Table 8.59 Analysis of BDI Scores at Week 24 by Sex Y Change from Baseline Table 8.60 Summary of BDI Scores at Week 24 by Baseline Six Minute Walk Distance (6MWD) Above or Below Y Median Baseline Six Minute Walk Distance (6MWD) Analysis of BDI Scores at Week 24 by Baseline Six Minute Walk Distance (6MWD) Above or Below Y Table 8.61 Median Baseline Six Minute Walk Distance (6MWD) Change from Baseline The above tables will be repeated for the ITT population (Tables 9.1 to 9.61), and for the Non-mITT population (Table 10.1 to note that summary tables only will be populated, analysis tables will be shells with Analysis not performed for Non-mITT population ). Y 81

389

390

391 CONFIDENTIAL Table Summary of SF-36 Scores for Physical Component Summary at Week 24 Y Table Analysis of SF-36 Scores for Physical Functioning Domain Y Change from Baseline Table Analysis of SF-36 Scores for Role-Physical Domain Y Change from Baseline Table Analysis of SF-36 Scores for Bodily Pain Domain Y Change from Baseline Table Analysis of SF-36 Scores for General Health Domain Y Change from Baseline Table Analysis of SF-36 Scores for Vitality Domain Y Change from Baseline Table Analysis of SF-36 Scores for Social Functioning Domain Y Change from Baseline Table Analysis of SF-36 Scores for Role-Emotional Domain Y Change from Baseline Table Analysis of SF-36 Scores for Mental Health Domain Y Change from Baseline Table Analysis of SF-36 Scores for Reported Health Transition Domain Y Change from Baseline Table Analysis of SF-36 Scores for Mental Component Summary Y Change from Baseline Table Analysis of SF-36 Scores for Physical Component Summary Change from Baseline Y Table Summary of CAMPHOR Symptom Scores by Visit Y Table Summary of CAMPHOR Symptom Energy Subscale Scores by Visit Y Table Summary of CAMPHOR Symptom Breathlessness Subscale Scores by Visit Y Table Summary of CAMPHOR Symptom Mood Subscale Scores by Visit Y Table Summary of CAMPHOR QoL Scores by Visit Y Table Summary of CAMPHOR Activity Scores by Visit Y Table Summary of CAMPHOR Symptom Scores at Week 24 Y Table Summary of CAMPHOR Symptom Energy Subscale Scores at Week 24 Y Table Summary of CAMPHOR Symptom Breathlessness Subscale Scores at Week 24 Y Table Summary of CAMPHOR Symptom Mood Subscale Scores at Week 24 Y Table Summary of CAMPHOR QoL Scores at Week 24 Y Table Summary of CAMPHOR Activity Scores at Week 24 Y 84

392 CONFIDENTIAL Table Table Table Table Table Table Analysis of CAMPHOR Symptom Scores Change from Baseline Analysis of CAMPHOR Symptom Energy Subscale Scores Change from Baseline Analysis of CAMPHOR Symptom Breathlessness Subscale Scores Change from Baseline Analysis of CAMPHOR Symptom Mood Subscale Scores Change from Baseline Analysis of CAMPHOR QoL Scores Change from Baseline Analysis of CAMPHOR Activities Scores Change from Baseline The above tables will be repeated for the ITT population (Tables 15.1 to 15.51), and for the Non-mITT population (Table 16.1 to 16.51) Y Y Y Y Y Y 85

393