Running Title: Effect of Patisiran on Cardiac hattr Amyloidosis

Transcription

1 1.1161/CIRCULATIONAHA Effects of Patisiran, an RNA Interference Therapeutic, on Cardiac Parameters in Patients with Hereditary Transthyretin-Mediated Amyloidosis: An Analysis of the APOLLO Study Running Title: Effect of Patisiran on Cardiac hattr Amyloidosis Scott D. Solomon, MD 1 ; David Adams, MD, PhD 2 ; Arnt Kristen, MD 3 ; Martha Grogan, MD 4 ; Alejandra González-Duarte, MD 5 ; Mathew S. Maurer, MD 6 ; Giampaolo Merlini, MD 7 ; Thibaud Damy, MD, PhD 8 ; Michel S. Slama, MD 9 ; Thomas H. Brannagan III, MD 1 ; Angela Dispenzieri, MD 11 ; John L. Berk, MD 12 ; Amil M. Shah, MD 1 ; Pushkal Garg, MD 13 ; Akshay Vaishnaw, MD, PhD, FRCP 13 ; Verena Karsten, PhD 13 ; Jihong Chen, PhD 13 ; Jared Gollob, MD 13 ; John Vest, MD 13 ; Ole Suhr, MD, PhD 14 Downloaded from by on September 15, Brigham and Women s Hospital, Harvard Medical School, Boston, MA; 2 APHP, National Reference Center for FAP (NNERF), CHU Bicêtre, INSERM U1195, Université Paris Sud, Le Kremlin-Bicêtre, France; 3 Department of Cardiology, University of Heidelberg, Heidelberg, Germany; 4 Mayo Clinic, Rochester, MN; 5 Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, México D.F., CP, Mexico; 6 Department of Medicine/Cardiology, Columbia University Medical Center, New York, NY; 7 Amyloidosis Research and Treatment Center, Foundation IRCCS Policlinico San Matteo and University of Pavia, Pavia, Italy; 8 French Referral Center for Cardiac Amyloidosis, Amyloidosis Mondor Network, GRC Amyloid Research Institute and Department of Cardiology, all at APHP, CHU Henri Mondor, and INSERM U955, Clinical Investigation Center 6, and DHU ATVB all at Creteil, France; 9 Hôpital Bichat, Cardiology Department, Université Paris Sud, Paris, France; 1 Neurology Department, Columbia University, College of Physicians and Surgeons, New York, NY; 11 Division of Hematology, Mayo Clinic, Rochester, MN; 12 Amyloidosis Center, Boston Medical Center, Boston, MA; 13 Alnylam Pharmaceuticals, Cambridge, MA; 14 Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden Address for Correspondence: Scott Solomon, MD Cardiovascular Division Brigham and Women s Hospital Harvard Medical School 75 Francis Street, Boston, MA 2115, USA Tel. 1 (857) Fax 1 (857) ssolomon@rics.bwh.harvard.edu 1

2 1.1161/CIRCULATIONAHA Abstract Downloaded from by on September 15, 218 Background: Hereditary transthyretin-mediated (hattr) amyloidosis is a rapidly progressive, multisystem disease that presents with cardiomyopathy and/or polyneuropathy. The APOLLO study assessed the efficacy and tolerability of patisiran in patients with hattr amyloidosis. The effects of patisiran on cardiac structure and function in a pre-specified subpopulation of patients with evidence of cardiac amyloid involvement at baseline, were assessed. Methods: APOLLO was an international, randomized, double-blind, placebo-controlled Phase 3 trial in patients with hattr amyloidosis. Patients were randomized 2:1 to receive.3 mg/kg patisiran or placebo, via intravenous infusion, once every 3 weeks for 18 months. The prespecified cardiac subpopulation comprised patients with a baseline left ventricular (LV) wall thickness 13 mm and no history of hypertension or aortic valve disease. Pre-specified exploratory cardiac endpoints included mean LV wall thickness, global longitudinal strain, and N-terminal prohormone of brain natriuretic peptide (NT-proBNP). Cardiac parameters in the overall APOLLO patient population were also evaluated. A composite endpoint of cardiac hospitalizations and all-cause mortality was assessed in a post hoc analysis. Results: In the cardiac subpopulation (n=126; 56% of total population), patisiran reduced mean LV wall thickness (least-squares mean difference±sem:.9±.4 mm, P=.17), interventricular septal wall thickness, posterior wall thickness, and relative wall thickness at Month 18, compared with placebo. Patisiran also led to increased end-diastolic volume (8.3±3.9 ml, P=.36), decreased global longitudinal strain ( 1.4±.6%, P=.15), and increased cardiac output (.38±.19 L/min, P=.44) compared with placebo at Month 18. Patisiran lowered NTproBNP at 9 and 18 months (at 18 months, ratio of fold-change patisiran/placebo.45, P<.1). A consistent effect on NT-proBNP at 18 months was observed in the overall APOLLO patient population (n=225). Median follow-up duration was 18.7 months. The exposure-adjusted rates of cardiac hospitalizations and/or all-cause death were 18.7 and 1.1 per 1 patient-years in the placebo and patisiran groups, respectively (Andersen Gill hazard ratio.54, 95% confidence interval: ). Conclusions: Patisiran decreased mean LV wall thickness, global longitudinal strain, NTproBNP, and adverse cardiac outcomes compared with placebo at Month 18, suggesting that patisiran may halt or reverse the progression of the cardiac manifestations of hattr amyloidosis. Clinical Trial Registration: URL: Unique Identifier: NCT Key Words: APOLLO; cardiac amyloidosis; hattr amyloidosis; cardiomyopathy; patisiran; RNA interference 2

3 1.1161/CIRCULATIONAHA Clinical Perspective What is new? The Phase 3 APOLLO study assessed the effect of patisiran, a novel RNA interference therapeutic that inhibits transthyretin synthesis, in patients with hereditary transthyretinmediated (hattr) amyloidosis. The current analysis presents results for exploratory endpoints in a pre-specified subpopulation of patients from the APOLLO study with evidence of cardiac amyloid involvement. In this subpopulation, patisiran resulted in decreased mean left ventricular wall thickness, global longitudinal strain, and N-terminal prohormone of brain natriuretic peptide compared with placebo at Month 18. In a post hoc analysis, patisiran treatment lowered combined all-cause hospitalization and mortality compared with placebo at Month 18. Downloaded from by on September 15, 218 What are the clinical implications? hattr amyloidosis, an inherited, rapidly progressive, life-threatening disease with limited treatment options has a clinical presentation that includes both cardiomyopathy and polyneuropathy. Results from this pre-specified subgroup analysis of the Phase 3 study suggest that patisiran may provide benefit to patients with the cardiac manifestations of hattr amyloidosis. 3

4 1.1161/CIRCULATIONAHA Hereditary transthyretin-mediated (hattr) amyloidosis is an inherited, rapidly progressive, lifethreatening disease 1 3 caused by mutation of the transthyretin (TTR) gene. Pathogenic mutations cause TTR protein to misfold and accumulate as amyloid fibrils, typically consisting of both mutant and wild-type (wt) TTR protein. The amyloid fibrils deposit in multiple tissues including the heart, nerves, gastrointestinal tract, and kidneys, 1,4 6 resulting in a multisystem disease with a heterogeneous clinical presentation that includes cardiomyopathy and polyneuropathy. 1,7,8 Patients with hattr amyloidosis and cardiomyopathy typically experience progressive symptoms of heart failure (HF) and cardiac arrhythmias, with death typically occurring 2.5 to 5 years after diagnosis. 2,3,9,1 Cardiac infiltration of the extracellular matrix by TTR amyloid fibrils leads to a progressive increase of ventricular wall thickness and a marked increase in chamber stiffness, resulting in impaired diastolic function. Systolic function is also impaired, typically reflected by abnormal longitudinal strain despite a normal ejection fraction, which is preserved Downloaded from by on September 15, 218 until late stages of the disease. 4,9 11 In patients with ATTR amyloidosis and light-chain (AL) cardiac amyloidosis, both longitudinal strain and N-terminal prohormone of brain natriuretic peptide (NT-proBNP) have been shown to be independent predictors of survival Patisiran is an RNAi therapeutic composed of a small interfering RNA (sirna) formulated as a lipid nanoparticle that enables delivery to hepatocytes, the main site of TTR production. After intracellular release, the sirna blocks production of mutant and wt TTR protein, by inducing cleavage of TTR messenger RNA. 15 Patisiran has demonstrated dosedependent reduction of TTR protein in healthy volunteers and patients with hattr amyloidosis. 15,16 The recently completed Phase 3 APOLLO study assessed the efficacy and safety of patisiran in patients with hattr amyloidosis. As described separately, 17 the APOLLO study met its primary endpoint demonstrating improvements in neuropathy as well as 4

5 1.1161/CIRCULATIONAHA polyneuropathy and all secondary endpoints including quality of life, ambulatory function, and autonomic symptoms, with patisiran compared with placebo. Furthermore, the APOLLO study demonstrated an acceptable safety profile for patisiran. Here, we present the effect of patisiran on cardiac structure and function in a pre-specified subpopulation of APOLLO patients with evidence of cardiac involvement at study entry, as well as cardiac safety in the overall APOLLO patient population. Methods Due to the sensitive nature of the data collected for this study, the dataset will not be made available to other researchers. However, requests from qualified researchers for additional analyses may be sent to Alnylam Pharmaceuticals (medinfo@alnylam.com). Study Oversight Downloaded from by on September 15, 218 APOLLO (NCT196348) was a multicenter, international, randomized, double-blind, placebocontrolled, Phase 3 study of patisiran in patients with hattr amyloidosis. The study was approved by the central and local institutional review boards or ethics committees and was conducted according to the International Conference on Harmonization for Good Clinical Practice, the World Health Organization Declaration of Helsinki, and the 1996 Health Insurance Portability and Accountability Act. All participants provided written informed consent. The full methodology of APOLLO is described in detail elsewhere. 18 Study Participants Eligible patients were aged years, had a diagnosis of hattr amyloidosis with a documented TTR mutation and symptomatic neuropathy, were ambulatory (with or without walking aids), had adequate liver function (aspartate transaminase and alanine transaminase 5

6 1.1161/CIRCULATIONAHA levels 2.5 times the upper limit of normal), and adequate renal function (creatinine levels 2 times the upper limit of normal). Patients with previous liver transplant, Type I diabetes, New York Heart Association (NYHA) classification >2, acute coronary syndrome within the past 3 months, uncontrolled cardiac arrhythmia, or unstable angina were excluded from the study. The use of tafamidis, diflunisal, doxycycline, tauroursodeoxycholic acid, or any investigational agent other than patisiran was prohibited during treatment with study drug and a wash-out period was mandated if these agents were used prior to screening. A cardiac subpopulation was pre-specified in the statistical analysis plan and comprised patients with evidence of cardiac amyloid involvement, defined as a baseline LV wall thickness 13 mm and no history of aortic valve disease or hypertension. Study Design and Treatment Patients were enrolled between December 213 and January 216 at 44 sites in 19 countries. Downloaded from by on September 15, 218 Patients were randomly assigned 2:1 (as previously described 18 ) to receive.3 mg/kg patisiran or placebo, via intravenous infusion, once every 3 weeks for 18 months. To mitigate against infusion-related reactions, patients received premedication (dexamethasone [1 mg], paracetamol/acetaminophen [5 mg], and H1 [diphenhydramine 5 mg] and H2 blockers [ranitidine 5 mg], or equivalent) prior to infusion. As TTR is a transporter of retinol-binding protein, 19 patients received oral supplements of vitamin A at the recommended daily allowance to prevent deficiency. Both patisiran and placebo groups received premedication and vitamin A supplements. Cardiac Measures and Safety Assessments Assessment of cardiac structure and function via 2-dimensional echocardiography, as well as biomarkers of cardiac stress and injury including NT-proBNP and troponin I, were exploratory 6

7 1.1161/CIRCULATIONAHA endpoints in the APOLLO study. Measurements of cardiac parameters were conducted at baseline, Month 9, and Month 18. Echocardiography was used to assess cardiac structure and function; parameters pre-specified in the statistical analysis plan include mean LV wall thickness, LV mass, longitudinal strain, and ejection fraction. Cardiac output, left atrial size, LV end-diastolic volume (LVEDV), and LV end-systolic volume (LVESV) were also analyzed. Echocardiograms were obtained at the study sites according to a pre-specified protocol and underwent blinded assessment in a cardiac imaging core laboratory. Myocardial strain was assessed with speckle tracking using vendor-independent software (TOMTEC, Munich, Germany). Reproducibility of echocardiographic measures and myocardial strain has been previously reported. 2,21 Analysis of NT-proBNP and troponin I was also pre-specified in the statistical analysis plan and biomarkers were analyzed at a central laboratory using chemiluminescence assays (Roche Diagnostic Cobas, Indianapolis, IN, USA for NT-proBNP; Downloaded from by on September 15, 218 Siemens Centaur XP, Camberley, Surrey, UK for troponin I). Creatinine levels were measured at baseline in a central laboratory, and estimated glomerular filtration rate (egfr) was calculated from creatinine levels using the Modification of Diet in Renal Disease study formula. 22 Adverse events (AEs) were coded according to the Medical Dictionary for Regulatory Activities (MedDRA). Cardiac AEs were AEs that mapped within the cardiac disorders system organ class (SOC). Cardiac arrhythmia AEs were AEs that mapped within the cardiac arrhythmias MedDRA high-level group term. Cardiac failure AEs were AEs that mapped within the cardiac failure standardized MedDRA query: narrow search. A blinded, independent clinical endpoint adjudication committee determined whether deaths were of cardiovascular (CV) or non-cv origin according to a pre-specified charter; hospitalizations were not adjudicated. 7

8 1.1161/CIRCULATIONAHA Statistical Analysis The primary population for efficacy and safety analyses was the modified intention-to-treat (mitt) population (all randomized patients who received 1 dose of study drug). The cardiac parameters were analyzed in the pre-specified cardiac subpopulation, using the mixed-effects model repeated measures (MMRM) method. The model included change from baseline as the outcome variable, baseline value as a covariate, and fixed-effect terms including treatment arm, visit (Month 9 or 18), and treatment-by-visit interaction. The primary comparison was the difference in the least-squares (LS) means between the patisiran and placebo groups at 18 months. The normally distributed assumption for the response variable in the MMRM model was examined using normal Q-Q plots. For echocardiographic parameters, there was no indication of a violation of the distributional assumption. For NT-proBNP, normality assumption was apparently violated, therefore a logarithmic transformation was applied to normalize the data Downloaded from by on September 15, 218 before fitting the MMRM model. The adjusted geometric mean fold-change and the ratio of the fold-change (patisiran/placebo) from baseline were calculated along with the corresponding 95% confidence intervals (CIs). The proportions of patients achieving the following clinically meaningful threshold values were also estimated for evaluable patients who had both baseline and Month 18 assessments: mean LV wall thickness change (increase or decrease) of >2 mm, 12 longitudinal strain absolute change (increase or decrease) of >2%, 12 and NT-proBNP values >3 ng/l or change from baseline (increase or decrease) 3% and 3 ng/l in patients with baseline NT-proBNP 65 ng/l. 23,24 Post hoc analyses of recurrent hospitalization and death events were conducted to assess the composite endpoints of 1) any hospitalization and/or all-cause death and 2) cardiac hospitalization and/or all-cause death. Any hospitalization or death event associated with a 8

9 1.1161/CIRCULATIONAHA serious adverse event (SAE) that occurred within 28 days of last dose of study drug was included in the analysis. Cardiac hospitalizations were SAEs that coded to the cardiac disorders SOC that led to hospitalization or prolongation of existing hospitalization. For both composite endpoints, the exposure-adjusted event rates were estimated by treatment arm. A non-parametric mean cumulative function method was used to estimate the average number of events expected per patient at a certain time point. In addition, negative binomial regression and Andersen Gill models were applied to analyze the recurrent-events data. Overall type I error was not controlled for the exploratory endpoints. Differences in baseline characteristics were tested using t tests for log-transformed continuous variables including years since hattr amyloidosis diagnosis, estimated glomerular filtration rate, NT-proBNP, Wilcoxon rank sum test for age which deviated from normal distribution, and Fisher s exact tests for all categorical variables. Further details of the statistical analyses used in APOLLO are described elsewhere. 18 Downloaded from by on September 15, 218 Results Patient Population Among the 225 total patients enrolled in APOLLO, 126 (56%) fulfilled criteria to be included in the pre-specified cardiac subpopulation. Among the cardiac subpopulation, 36 of 126 (28.6%) patients received placebo and 9 of 126 (71.4%) patients received patisiran. Ninety-nine patients did not meet the pre-specified criteria for the cardiac subpopulation and are referred to as all other patients. Among these other patients, 55.6% also had a mean LV wall thickness 13 mm but were excluded from the cardiac subpopulation primarily due to history of hypertension. Baseline demographic and disease characteristics and baseline echocardiographic parameters, for both the cardiac subpopulation and all other patients, are shown in Tables 1 and 2, respectively. 9

10 1.1161/CIRCULATIONAHA Among the cardiac subpopulation, the median age of patients was 61 years (interquartile range [IQR] 54 67), and 78% were male (Table 1). The median time from diagnosis was 1.4 years (IQR. 21.) and 73% of patients were non-val3met genotype. A higher proportion of patients receiving placebo were Asian (5.% versus 25.6% in the patisiran arm), otherwise there were no differences in demographic characteristics between treatment groups. At baseline, the majority (6%) of patients in the cardiac subpopulation were classified as NYHA Class II, with the remainder (4%) classified as NYHA Class I. Echocardiographic features in the cardiac subpopulation were notable for increased LV wall thickness (median [IQR] 16.4 [ ] mm), reduction in global longitudinal strain (median [IQR] 15.1% [ 17.5% to 12.6%]), and preserved LV ejection fraction (mean±standard deviation 61±1.%) (Table 2). The median NT-proBNP level was 837 pg/ml (IQR ). Within the cardiac subpopulation, treatment groups were generally balanced with respect to cardiac measures. Downloaded from by on September 15, 218 Compared with all other patients, a higher proportion of patients in the cardiac subpopulation were NYHA Class II (6.3% versus 38.1%) and were non-val3met genotype (73.% versus 37.4%). Genotypes more commonly found in the cardiac subpopulation, compared with all other patients, included Ala97Ser, Thr6Ala, and Ser5Arg (Supplemental Table 1). Compared with all other patients, in the cardiac subpopulation there were signs of greater cardiac dysfunction at baseline, including greater abnormalities in cardiac structure and function and higher NT-proBNP levels (Tables 1 and 2). Effects of Patisiran on Cardiac Parameters in the Cardiac Subpopulation In the cardiac subpopulation, a reduction in mean LV wall thickness (LS mean difference±standard error of the mean;.9±.4 mm, P=.17) was observed with patisiran compared with placebo, and corresponding reductions were also seen in interventricular septal 1

11 1.1161/CIRCULATIONAHA wall thickness, posterior wall thickness, and relative wall thickness. In patisiran-treated patients compared with placebo, global longitudinal strain was decreased ( 1.4%±.6%, P=.15), cardiac output was increased (.38±.19 L/min, P=.44), and LVEDV was increased (8.31±3.91 ml, P=.36) (Figure 1A). The treatment effect of patisiran on these echocardiographic parameters was observed across all change thresholds at 18 months (Figure 1B). For LV mass, a reduction compared with baseline was observed in the patisiran arm (mean change 15.1 g, 95% CI: 25.8 to 4.4), and a trend toward reduction relative to placebo was also observed. LVESV decreased less in patients receiving patisiran than those receiving placebo, although the difference was not statistically significant (2.66±2.11 ml, P=.211). There were no differences in LV ejection fraction or left atrial volume between the treatment groups (Supplemental Table 2). A greater proportion of patients in the patisiran group had a reduction from baseline in mean LV wall thickness >2 mm compared with placebo (29.1% versus 4.%) Downloaded from by on September 15, 218 and an absolute decrease (indicating improvement in function) from baseline in global longitudinal strain of >2% compared with placebo (21.3% versus 8.%) whereas a smaller proportion of patients in the patisiran group had an absolute increase (indicating worsening in function) from baseline in global longitudinal strain of >2% at 18 months (25.3% versus 44.% in the placebo arm) (Figure 2). Patisiran reduced NT-proBNP compared with placebo at 9 months (ratio of fold-change patisiran/placebo.63, 95% CI:.5.8) and 18 months (ratio of fold-change patisiran/placebo.45, 95% CI:.34.59, P= ), corresponding to a 55% reduction relative to placebo (Figure 3A). The effect of patisiran (decrease in NT-proBNP compared with placebo) was observed across all change thresholds at both 9 and 18 months (Figure 3B). In the patisiran group 31.6% of evaluable patients had a decrease (change from baseline) of NT- 11

12 1.1161/CIRCULATIONAHA probnp 3% and 3 pg/ml at Month 18, whereas no placebo patients had a decrease in NTproBNP of this magnitude. Conversely, at Month 18, a lower proportion of evaluable patients in the patisiran group compared with placebo had an increase (change from baseline) in NTproBNP 3% and 3 pg/ml (21.1% versus 58.3%) (Figure 3C). The majority of troponin I values (9.2%) were reported as <.1 μg/l, which is also the limit of detection for the troponin I assay used in the study. Accordingly, the troponin I data prohibited an accurate assessment of patisiran treatment effect on troponin I. Patisiran treatment also impacted 1-meter walk test (1- MWT) gait speed in the cardiac subpopulation, with an increase compared to placebo of.161 m/s (95% CI: ) at 9 months and.354 m/s (95% CI: ) at 18 months. Effects of Patisiran on Cardiac Parameters in All Other Patients and the mitt Population In the 44% of patients who did not fulfil the pre-specified cardiac criteria (all other patients), no significant impact on echocardiographic parameters was observed with patisiran compared with Downloaded from by on September 15, 218 placebo (data not shown). However, NT-proBNP was reduced relative to placebo in all other patients (51% reduction with patisiran treatment relative to placebo), which was similar to the effect seen in the cardiac subpopulation (Supplemental Table 3). Furthermore, in all other patients, 1-MWT gait speed increased by.283 m/s (95% CI: ) in the patisiran group compared with the placebo group at 18 months. Analyses were also performed in the mitt population (all patients enrolled in APOLLO irrespective of cardiac status). Results for assessments of echocardiographic parameters were directionally similar to those observed in the cardiac subpopulation (Supplemental Table 4). NTproBNP was reduced relative to placebo in the mitt population and in all pre-specified subgroups (Supplemental Figure 1). The effect on NT-proBNP was also seen regardless of baseline NT-proBNP level. In all NT-proBNP baseline categories, NT-proBNP levels decreased 12

13 1.1161/CIRCULATIONAHA or were stable in the patisiran group and nearly doubled in the placebo group (Supplemental Figure 2). In the mitt population, regardless of treatment group, survival was worse in patients with baseline NT-proBNP levels >3 ng/l compared with patients with levels 3 ng/l (hazard ratio 19.3, 95% CI: , Supplemental Figure 3). Cardiac Safety The overall safety profile of patisiran has been described previously. 17 In addition, a detailed evaluation of cardiac events was performed in the mitt population (Table 3). The proportion of patients with cardiac AEs, cardiac SAEs, and cardiac failure AEs were similar in the patisiran and placebo groups. The rate of cardiac arrhythmia AEs was lower in the patisiran group compared with placebo (18.9% versus 28.6%). Median survival follow-up duration on the study was 18.7 months. In the mitt population, there were 7 (4.7%) deaths in the patisiran group all adjudicated as CV in nature) and Downloaded from by on September 15, (7.8%) deaths in the placebo group (3 adjudicated as CV in nature, 2 non-cv, and 1 unknown origin). All deaths were considered unlikely or not related to study drug by the investigators (Supplemental Table 5). The exposure-adjusted death rate per 1-patient years was 6.2 (95% CI: ) in the placebo group and 3.2 (95% CI: ) in the patisiran group. In a post hoc analysis of the safety data, the rates of any hospitalization and/or all-cause death were 71.8 and 34.7 per 1 patient-years in the placebo and patisiran groups respectively, while the rates of cardiac hospitalizations and/or all-cause death were 18.7 and 1.1 per 1 patient-years in the placebo and patisiran groups respectively (Table 3). This approximates a reduction in event rate of approximately 5% for all-cause hospitalization and mortality and approximately 45% for cardiac hospitalization and all-cause mortality (Figure 4). 13

14 1.1161/CIRCULATIONAHA In the cardiac subpopulation, the proportion of patients with cardiac AEs (32.2% versus 36.1%) and cardiac SAEs (14.4% versus 11.1%) was comparable across the patisiran and placebo groups respectively. The proportion of patients with cardiac arrhythmia AEs was higher in the placebo group (3.6%) than the patisiran group (18.9%). The proportion of patients with cardiac failure AEs was higher in the patisiran group (11.1%) as compared with the placebo group (5.6%). This difference may be due to baseline imbalances in the cardiac subpopulation with respect to cardiac history and NYHA status. There were 5 (5.6%) deaths in the patisiran group (all CV in nature) and 4 (11.1%) deaths in the placebo group (1 CV, 2 non-cv, and 1 unknown origin). Discussion In this pre-specified analysis of patients with evidence of cardiac involvement due to hattr Downloaded from by on September 15, 218 amyloidosis, treatment with patisiran for up to 18 months resulted in improvement relative to placebo in important measures of cardiac structure and function. These included reduction in ventricular wall thickness and decrease (improvement) in global longitudinal strain. In addition, patisiran treatment led to a reduction in NT-proBNP as early as 9 months after initiating treatment. These findings collectively suggest that patisiran may provide benefit to patients with cardiac manifestations of hattr amyloidosis. The pre-specified cardiac subpopulation was intended to allow for more sensitive and specific detection of patisiran treatment effects on cardiac parameters, which might be difficult to discern in the overall population. In this subpopulation, treatment with patisiran compared with placebo improved important measures of cardiac structure and function. LV wall thickening caused by myocardial infiltration of TTR amyloid is associated with impaired myocardial 14

15 1.1161/CIRCULATIONAHA function, with increases of 1 mm (the approximate difference observed between patisiran and placebo groups) being associated with worsening of global longitudinal strain. 12 The change in wall thickness observed in the present study was also associated with an increase in LVEDV. Given that ventricular capacitance, reflected by LVEDV, is central to the pathophysiology of diastolic dysfunction in hattr amyloidosis, this change in cavity volume may reflect decreased myocardial infiltration and potentially a reduction in pre-existing amyloid deposits. Global longitudinal strain, a sensitive measure of systolic function, is known to increase (indicating worse function) in amyloid heart disease, 12,25 often in the setting of a normal ejection fraction. The magnitude of difference in global longitudinal strain observed between patisiran and placebo groups has been shown to independently predict survival in patients with ATTR and AL amyloidosis, 12 providing further support for the potential clinical significance of the observed treatment effect. Downloaded from by on September 15, 218 In addition, a treatment effect of patisiran on NT-proBNP was observed. NT-proBNP is a potent marker of the severity of HF, and is a prognostic of outcome in patients with HF of various etiologies, 26,27 including amyloidosis. 28,29 It was modestly elevated in the cardiac subpopulation at baseline, with a median value in the range that would be considered diagnostic for HF. Reduction in NT-proBNP with patisiran was apparent based on both mean values and the proportion of patients achieving established treatment response thresholds. Of note, the relative reduction with patisiran was apparent as early as 9 months, despite there being no discernible improvement in echocardiographic measures at this time point. This suggests that NT-proBNP may be an earlier indicator of the effect of patisiran on cardiac function than echocardiography and that circulating forms of TTR may have a cardiotoxic effect beyond amyloid deposition in the heart. 15

16 1.1161/CIRCULATIONAHA Substantial disease progression was observed in the placebo group, with increasing NTproBNP levels and worsening global longitudinal strain along with decreasing cardiac output and LVEDV, although wall thickness was relatively stable. There are limited natural history data available for patients with hattr amyloidosis with cardiomyopathy; however, increasing NTproBNP and decreasing LVEDV have been observed in an independent cohort of patients with hattr amyloidosis over 18 months. 11 Our findings, including improvements in LV wall thickness and NT-proBNP relative to baseline, suggest that patisiran may impede cardiac disease progression, and possibly reflect clearance of amyloid deposits from the myocardium. Patisiran treatment also resulted in an improvement in functional status relative to the placebo group as measured by 1-MWT gait speed. While an improvement in 1-MWT gait speed was also seen in all other patients, the effect seen in the cardiac subpopulation was greater. Both peripheral neuropathy and cardiomyopathy can compromise ambulatory ability; as such, it Downloaded from by on September 15, 218 is possible that the effect of patisiran on gait speed was the result of its impact on both aspects of the disease. It is probable that cardiac amyloid involvement was present outside of the predefined cardiac subpopulation. Indeed, evidence of potential cardiac amyloid involvement was seen in most patients in the study; 8% had LV wall thickness >13 mm and 79% had abnormal NTproBNP levels. Accordingly, the treatment effect was also assessed in all other patients, who did not meet criteria for the pre-specified cardiac subpopulation, to further understand the impact of patisiran on cardiac parameters in patients with less overt (wall thickness <13 mm) or less specific (wall thickness 13 mm with confounding medical history) cardiac manifestations. Among these patients there were no discernible differences between treatment arms in the effect on echocardiographic parameters over 18 months. As expected, abnormalities in baseline 16

17 1.1161/CIRCULATIONAHA echocardiographic parameters were less pronounced in these patients compared with the cardiac subpopulation, and confounding effects of hypertension on parameters including wall thickness were present. Consequently, echocardiographic assessment may have lacked the sensitivity to discern a treatment effect in these patients. By contrast, the treatment effect of patisiran on NTproBNP observed in these patients likely reflects the higher sensitivity of the biomarker assay compared with echocardiography. These findings suggest that most patients with hattr amyloidosis may have cardiac involvement, and that NT-proBNP could be used as a biomarker to assess response to treatment. Given that patisiran s treatment effect relative to placebo increased over time on both echocardiographic parameters and NT-proBNP, it is possible that further improvement could be achieved with a longer treatment duration. In the mitt population, a patisiran treatment effect on NT-proBNP was observed regardless of baseline disease severity, suggesting patisiran may benefit patients with hattr amyloidosis regardless of Downloaded from by on September 15, 218 disease stage, and that early intervention may provide benefit in this rapidly progressive disease. Patisiran had an acceptable safety profile, based on an in-depth analysis of cardiac events in both the mitt population and cardiac subpopulation. Importantly, a post hoc analysis of safety data showed that rates of all-cause death and cardiac hospitalization were decreased in the patisiran group compared with the placebo group. Some limitations of this analysis should be noted. The cardiac subpopulation was defined by LV wall thickness, a measure that increases the likelihood of, but does not definitively establish, cardiac amyloid involvement. Similarly, the lack of this degree of ventricular wall thickness does not rule out cardiac involvement. A further limitation is the exploratory nature of the reported cardiac assessments, where overall type I error was not controlled. Nevertheless, the consistency of the biomarker findings, in concert with the observed effect on multiple 17

18 1.1161/CIRCULATIONAHA echocardiographic parameters and the clinical endpoint of 1-MWT gait speed suggest a patisiran treatment effect on cardiac disease manifestations. It should also be noted that the study did not include patients with NYHA Class III or IV at baseline and that the majority of patients had preserved ejection fraction. While this study was not designed or powered to investigate the effect of patisiran on clinical outcomes such as death or CV hospitalization, the safety data suggest promising trends in the reduction in cardiac hospitalizations and all-cause mortality. In conclusion, treatment with the RNAi therapeutic patisiran resulted in improvement in several important measures of cardiac structure and function over 18 months, including reduction in LV wall thickness and NT-proBNP compared with both baseline and placebo, and relative improvement in global longitudinal strain compared with placebo. These findings suggest that patisiran may halt, or possibly reverse, the progression of hattr amyloid heart disease. Downloaded from by on September 15, 218 Acknowledgments Alnylam Pharmaceuticals would like to thank all patients and their families as well as all investigators involved in this APOLLO study subgroup analysis. Sources of Funding This study was funded by Alnylam Pharmaceuticals. Editorial support was provided by Adelphi Communications Ltd, funded by Alnylam Pharmaceuticals. Disclosures Dr. Solomon reports receiving personal and consulting fees from Alnylam Pharmaceuticals, as well as grants (through Brigham and Women s Hospital) from Alnylam Pharmaceuticals, Ionis 18

19 1.1161/CIRCULATIONAHA Pharmaceuticals, and GSK. Dr. Adams reports receiving a grant and consulting fees from Alnylam Pharmaceuticals. Prof. Kristen reports receiving compensation for study participation from Alnylam Pharmaceuticals. Dr. Grogan reports receiving consulting fees from Alnylam Pharmaceuticals and Pfizer. Dr. González-Duarte reports receiving personal fees from Alnylam Pharmaceuticals. Prof. Maurer reports receiving grants from Alnylam Pharmaceuticals, Eidos Therapeutics, GSK, and Pfizer, consulting fees from Prothena Inc and GSK, and being a member of a steering committee for Pfizer. Dr. Merlini reports none. Prof. Damy reports research grants, scientific meeting expenses, or honoraria from Alnylam Pharmaceuticals, Pfizer, Prothena, and GSK. Dr. Slama reports receiving grants from Alnylam Pharmaceuticals, Ionis Pharmaceuticals, and Pfizer, as well as personal fees from Alnylam Pharmaceuticals and Pfizer, and non-financial support from Ionis Pharmaceuticals. Dr. Brannagan reports receiving a grant (through Columbia University) for participation in this trial from Alnylam Pharmaceuticals, speaker fees from Downloaded from by on September 15, 218 Alnylam Pharmaceuticals, consulting fees from Alnylam Pharmaceuticals and Ionis, and an additional grant (through Columbia University) from Ionis Pharmaceuticals. Dr. Dispenzieri reports receiving grants for study trials from Alnylam Pharmaceuticals, Celgene, Jannsen, Pfizer, Prothena, and Takeda. Dr. Berk reports receiving honoraria from Alnylam Pharmaceuticals and Ionis Pharmaceuticals, as well as consulting fees from Intellia Therapeutics. Dr. Shah reports receiving consulting fees from Philips Ultrasound and Bellerophon and research support (through Brigham and Women s Hospital) from Novartis and Bellerophon. Drs. Garg, Vaishnaw, Karsten, Chen, and Gollob are employees of Alnylam Pharmaceuticals. Dr. Vest is an employee of Alnylam Pharmaceuticals and a former employee of Novartis. Dr. Suhr reports fees (through Umeå University) for advisory board services, lectures, and expert testimony from Alnylam Pharmaceuticals, travel support and speaker service fees (through Umeå University) from Pfizer, 19

20 1.1161/CIRCULATIONAHA and advisory board service fees (through Umeå University) from Prothena and Eidos Therapeutics. References Downloaded from by on September 15, Hanna M. Novel drugs targeting transthyretin amyloidosis. Curr Heart Fail Rep. 214;11: Damy T, Judge DP, Kristen AV, Berthet K, Li H, Aarts J. Cardiac findings and events observed in an open-label clinical trial of tafamidis in patients with non-val3met and non- Val122Ile hereditary transthyretin amyloidosis. J Cardiovasc Transl Res. 215;8: Hawkins PN, Ando Y, Dispenzeri A, Gonzalez-Duarte A, Adams D, Suhr OB. Evolving landscape in the management of transthyretin amyloidosis. Ann Med. 215;47: Mohty D, Damy T, Cosnay P, Echahidi N, Casset-Senon D, Virot P, Jaccard A. Cardiac amyloidosis: updates in diagnosis and management. Arch Cardiovasc Dis. 213;16: Adams D, Coelho T, Obici L, Merlini G, Mincheva Z, Suanprasert N, Bettencourt BR, Gollob JA, Gandhi PJ, Litchy WJ, Dyck PJ. Rapid progression of familial amyloidotic polyneuropathy: a multinational natural history study. Neurology. 215;85: Rowczenio DM, Noor I, Gillmore JD, Lachmann HJ, Whelan C, Hawkins PN, Obici L, Westermark P, Grateau G, Wechalekar AD. Online registry for mutations in hereditary amyloidosis including nomenclature recommendations. Hum Mutat. 214;35:E243 E Conceicao I, Gonzalez-Duarte A, Obici L, Schmidt HH, Simoneau D, Ong ML, Amass L. "Red-flag" symptom clusters in transthyretin familial amyloid polyneuropathy. J Peripher Nerv Syst. 216;21: Shin SC, Robinson-Papp J. Amyloid neuropathies. Mt Sinai J Med. 212;79: Castano A, Drachman BM, Judge D, Maurer MS. Natural history and therapy of TTRcardiac amyloidosis: emerging disease-modifying therapies from organ transplantation to stabilizer and silencer drugs. Heart Fail Rev. 215;2: Dungu JN, Anderson LJ, Whelan CJ, Hawkins PN. Cardiac transthyretin amyloidosis. Heart. 212;98: Ruberg FL, Maurer MS, Judge DP, Zeldenrust S, Skinner M, Kim AY, Falk RH, Cheung KN, Patel AR, Pano A, Packman J, Grogan DR. Prospective evaluation of the morbidity and mortality of wild-type and V122I mutant transthyretin amyloid cardiomyopathy: the Transthyretin Amyloidosis Cardiac Study (TRACS). Am Heart J. 212;164: e Quarta CC, Solomon SD, Uraizee I, Kruger J, Longhi S, Ferlito M, Gagliardi C, Milandri A, Rapezzi C, Falk RH. Left ventricular structure and function in transthyretin-related versus light-chain cardiac amyloidosis. Circulation. 214;129: Gillmore JD, Damy T, Fontana M, Hutchinson M, Lachmann HJ, Martinez-Naharro A, Quarta CC, Rezk T, Whelan CJ, Gonzalez-Lopez E, Lane T, Gilbertson JA, Rowczenio D, Petrie A, Hawkins PN. A new staging system for cardiac transthyretin amyloidosis. Eur Heart J. 218;39: Grogan M, Scott CG, Kyle RA, Zeldenrust SR, Gertz MA, Lin G, Klarich KW, Miller WL, Maleszewski JJ, Dispenzieri A. Natural history of wild-type transthyretin cardiac 2

21 1.1161/CIRCULATIONAHA Downloaded from by on September 15, 218 amyloidosis and risk stratification using a novel staging system. J Am Coll Cardiol. 216;68: Coelho T, Adams D, Silva A, Lozeron P, Hawkins PN, Mant T, Perez J, Chiesa J, Warrington S, Tranter E, Munisamy M, Falzone R, Harrop J, Cehelsky J, Bettencourt BR, Geissler M, Butler JS, Sehgal A, Meyers RE, Chen Q, Borland T, Hutabarat RM, Clausen VA, Alvarez R, Fitzgerald K, Gamba-Vitalo C, Nochur SV, Vaishnaw AK, Sah DW, Gollob JA, Suhr OB. Safety and efficacy of RNAi therapy for transthyretin amyloidosis. N Engl J Med. 213;369: Suhr OB, Coelho T, Buades J, Pouget J, Conceicao I, Berk J, Schmidt H, Waddington- Cruz M, Campistol JM, Bettencourt BR, Vaishnaw A, Gollob J, Adams D. Efficacy and safety of patisiran for familial amyloidotic polyneuropathy: a phase II multi-dose study. Orphanet J Rare Dis. 215;1: Adams D, Gonzalez-Duarte A, O'Riordan WD, Yang CC, Ueda M, Kristen AV, Tournev I, Schmidt HH, Coelho T, Berk JL, Lin KP, Vita G, Attarian S, Plante-Bordeneuve V, Mezei MM, Campistol JM, Buades J, Brannagan TH, 3rd, Kim BJ, Oh J, Parman Y, Sekijima Y, Hawkins PN, Solomon SD, Polydefkis M, Dyck PJ, Gandhi PJ, Goyal S, Chen J, Strahs AL, Nochur SV, Sweetser MT, Garg PP, Vaishnaw AK, Gollob JA and Suhr OB. Patisiran, an RNAi therapeutic, for hereditary transthyretin amyloidosis. N Engl J Med. 218;379: Adams D, Suhr OB, Dyck PJ, Litchy WJ, Leahy RG, Chen J, Gollob J, Coelho T. Trial design and rationale for APOLLO, a Phase 3, placebo-controlled study of patisiran in patients with hereditary ATTR amyloidosis with polyneuropathy. BMC Neurol. 217;17: Monaco HL, Rizzi M, Coda A. Structure of a complex of two plasma proteins: transthyretin and retinol-binding protein. Science (New York, NY). 1995;268: Shah AM, Cheng S, Skali H, Wu J, Mangion JR, Kitzman D, Matsushita K, Konety S, Butler KR, Fox ER, Cook N, Ni H, Coresh J, Mosley TH, Heiss G, Folsom AR, Solomon SD. Rationale and design of a multicenter echocardiographic study to assess the relationship between cardiac structure and function and heart failure risk in a biracial cohort of community-dwelling elderly persons: the Atherosclerosis Risk in Communities study. Circ Cardiovasc Imaging. 214;7: Kraigher-Krainer E, Shah AM, Gupta DK, Santos A, Claggett B, Pieske B, Zile MR, Voors AA, Lefkowitz MP, Packer M, McMurray JJ, Solomon SD, Investigators P. Impaired systolic function by strain imaging in heart failure with preserved ejection fraction. J Am Coll Cardiol. 214;63: Klahr S. The modification of diet in renal disease study. N Engl J Med. 1989;32: Merlini G, Lousada I, Ando Y, Dispenzieri A, Gertz MA, Grogan M, Maurer MS, Sanchorawala V, Wechalekar A, Palladini G, Comenzo RL. Rationale, application and clinical qualification for NT-proBNP as a surrogate end point in pivotal clinical trials in patients with AL amyloidosis. Leukemia. 216;3: Amyloidosis Research Consortium ( Guidance for industry AL amyloidosis developing drugs for treatment Available from: content/uploads/218/4/guidance-for-industry-al-amyloidosis-developing-drugs-for- Treatment-12_16.pdf (accessed August 8, 218) 25. Di Bella G, Pizzino F, Minutoli F, Zito C, Donato R, Dattilo G, Oreto G, Baldari S, Vita G, Khandheria BK, Carerj S. The mosaic of the cardiac amyloidosis diagnosis: role of imaging in subtypes and stages of the disease. Eur Heart J Cardiovasc Imaging. 214;15:

22 1.1161/CIRCULATIONAHA Seino Y, Ogawa A, Yamashita T, Fukushima M, Ogata K, Fukumoto H, Takano T. Application of NT-proBNP and BNP measurements in cardiac care: a more discerning marker for the detection and evaluation of heart failure. Eur J Heart Fail. 24;6: Januzzi JL, van Kimmenade R, Lainchbury J, Bayes-Genis A, Ordonez-Llanos J, Santalo-Bel M, Pinto YM, Richards M. NT-proBNP testing for diagnosis and short-term prognosis in acute destabilized heart failure: an international pooled analysis of 1256 patients: the International Collaborative of NT-proBNP Study. Eur Heart J. 26;27: Kristen AV, Maurer MS, Rapezzi C, Mundayat R, Suhr OB, Damy T, THAOS Investigators. Impact of genotype and phenotype on cardiac biomarkers in patients with transthyretin amyloidosis - Report from the Transthyretin Amyloidosis Outcome Survey (THAOS). PLoS One. 217;12:e Klaassen SHC, Tromp J, Nienhuis HLA, van der Meer P, van den Berg MP, Blokzijl H, van Veldhuisen DJ, Hazenberg BPC. Frequency of and prognostic significance of cardiac involvement at presentation in hereditary transthyretin-derived amyloidosis and the value of N- terminal pro-b-type natriuretic peptide. Am J Cardiol. 218;121: Downloaded from by on September 15,

23 Table 1. Baseline Demographics and Disease Characteristics Cardiac Subpopulation Characteristic Placebo (n=36) Patisiran (n=9) Overall (n=126) All Other Patients (n=99)* Age, years Median (IQR) Age 75 years, n (%) 62 ( ) 5 (13.9) 6 ( ) 5 (5.6) 61 ( ) 1 (7.9) 65 ( ) 8 (8.1) Male sex, n (%) 3 (83.3) 68 (75.6) 98 (77.8) 69 (69.7) Race, n (%) Asian Black White 18 (5.) 1 (2.8) 16 (44.4) 23 (25.6) 2 (2.2) 63 (7.) 41 (32.5) 3 (2.4) 79 (62.7) 11 (11.1) 2 (2.) 84 (84.9) Downloaded from by on September 15, 218 Region, n (%) North America Western Europe Rest of the world Asia Central and South America Eastern Europe 7 (19.4) 12 (33.3) 17 (47.2) 15 (41.7) 2 (5.6) 21 (23.3) 31 (34.4) 38 (42.2) 2 (22.2) 11 (12.2) 7 (7.8) 28 (22.2) 43 (34.1) 55 (43.7) 35 (27.8) 13 (1.3) 7 (5.6) 19 (19.2) 55 (55.6) 25 (25.3) 9 (9.1) 6 (6.1) 1 (1.1) Median (IQR) time since hattr diagnosis, years 1.2 (.1 8.8) 1.7 (. 21.) 1.4 (. 21.) 1.4 (. 17.2) TTR genotype, n (%) non-val3met Val3Met FAP stage, n (%) NIS score, n (%) < (66.7) 12 (33.3) 13 (36.1) 23 (63.9) 12 (33.3) 24 (66.7) 68 (75.6) 22 (24.4) 42 (46.7) 48 (53.3) 4 (44.4) 5 (55.6) 92 (73.) 34 (27.) 55 (43.7) 71 (56.3) 52 (41.3) 74 (58.7) 37 (37.4) 62 (62.6) 49 (49.5) 49 (49.5) 1 (1.) 45 (45.5) 54 (54.5) Previous TTR stabilizer use, n (%) 17 (47.2) 46 (51.1) 63 (5.) 56 (56.6) 23

24 1.1161/CIRCULATIONAHA egfr, ml/min per 1.73 m 2 Median (IQR) egfr <6, n (%) 97.8 ( ) 1 (2.8) ( ) 1 (11.1) 19.3 ( ) 11 (8.7) 97.2 ( ) 1 (1.1) Cardiac implanted devices, n (%) Pacemaker Defibrillator 9 (25.) 9 (25.5) 13 (14.4) 11 (12.2) 2 (2.2) 22 (17.5) 2 (15.9) 2 (1.6) 18 (18.2) 18 (18.2) Medical history, n (%) Hypertension Supraventricular arrhythmias Atrial fibrillation Diabetes Aortic valve disease 6 (16.7) 3 (8.3) 17 (18.9) 11 (12.2) 3 (3.3) 23 (18.3) 14 (11.1) 3 (2.4) 62 (62.6) 15 (15.2) 7 (7.1) 3 (3.) 3 (3.) Downloaded from by on September 15, 218 Baseline treatment, n (%) blockers ACE inhibitors ARBs Furosemide Spironolactone Hydrochlorothiazide Torasemide NYHA Class, n (%) Class I Class II NT-proBNP, pg/ml Median (IQR) Geometric mean (CV %) 2 (5.6) 7 (19.4) 2 (5.6) 1 (2.8) 16 (44.4) 2 (55.6) ( ) (19.8) 3 (3.3) 1 (1.1) 1 (1.1) 12 (13.3) 6 (6.7) 1 (1.1) 3 (3.3) 34 (37.8) 56 (62.2) ( ) (22.3) 5 (4.) 1 (.8) 1 (.8) 19 (15.1) 8 (6.3) 2 (1.6) 3 (2.4) 5 (39.7) 76 (6.3) ( ) (21.1) 27 (27.3) 18 (18.2) 13 (13.1) 9 (9.1) 3 (3.) 8 (8.1) 3 (3.) 6 (61.9) 37 (38.1) ( ) 36.7 (23.2) ACE indicates angiotensin converting enzyme; ARB, angiotensin receptor blocker; CV, coefficient of variation; egfr, estimated glomerular filtration rate; FAP, familial amyloid neuropathy; hattr, hereditary transthyretin-mediated amyloidosis; IQR, interquartile range; NIS, neuropathy impairment score; NT-proBNP, N-terminal prohormone of brain natriuretic peptide; NYHA, New York Heart Association; and TTR, transthyretin. *Includes patients with <13 mm mean left ventricular wall thickness or 13 mm mean left ventricular wall thickness and history of hypertension or aortic valve disease. Statistically significant difference between patisiran and placebo in cardiac subpopulation (P<.5). North America: USA, Canada; Europe: Germany, Spain, France, UK, Italy, Netherlands, Portugal, Sweden; Rest of the world: Asia: Japan, Korea, Taiwan; Central and South America: Mexico, Argentina, Brazil; Eastern Europe: Bulgaria, Cyprus, Turkey. Represents 31 different TTR mutations; listed in Supplemental Table 1. Statistically significant difference between cardiac subpopulation and all other patients (P<.1). 24