Rationale, design, and baseline characteristics of the Canagliflozin Cardiovascular Assessment Study (CANVAS) A randomized placebo-controlled trial

Transcription

1 Rationale, design, and baseline characteristics of the Canagliflozin Cardiovascular Assessment Study (CANVAS) A randomized placebo-controlled trial Bruce Neal, PhD, a,g Vlado Perkovic, PhD, a,g Dick de Zeeuw, PhD, b,g Kenneth W. Mahaffey, MD, c,g Greg Fulcher, MD, d,g Peter Stein, MD, e,g Mehul Desai, MD, e,g Wayne Shaw, DSL, e,g Joel Jiang, PhD, e,g Frank Vercruysse, MD, e,g Gary Meininger, MD, e,g and David Matthews, DPhil f,g New South Wales, Australia; Groningen, The Netherlands; Durham, NC; Raritan, NJ; and Oxford, United Kingdom Sodium glucose co-transporter 2 inhibition is a novel mode of treatment for type 2 diabetes mellitus (T2DM). The sodium glucose co-transporter 2 inhibitor canagliflozin lowered blood glucose, blood pressure, and body weight, with increased risk of urogenital infections in Phase 2 studies. Effects on macrovascular complications of diabetes remain to be determined. CANVAS is a double-blind, placebo-controlled trial designed to evaluate the effects of canagliflozin on the risk of cardiovascular disease and to assess safety and tolerability in patients with inadequately controlled T2DM and increased cardiovascular risk. The first of 2 planned phases randomized 4,330 individuals to placebo, canagliflozin 100 or 300 mg (1:1:1) with planned follow-up of about 2 years to substantiate potential cardiovascular protection by assessing key biomarkers and to achieve initial safety objectives. By the end of mid-september 2012, a total of 7174 patient-years of follow-up were accrued. Mean baseline age was 62 years, duration of diabetes 13 years; hemoglobin A 1c 8.2%, fasting plasma glucose 9.3 mmol/l, and body mass index 32 kg/m 2.Of the participants, 34% are female and 57% had a history of atherosclerotic vascular disease. Participants will be followed up to achieve primary safety and tolerability objectives and to investigate secondary outcomes. The planned second phase will not be undertaken. CANVAS will define the effects of canagliflozin on biomarkers and provide data on cardiovascular safety against established regulatory parameters. (Am Heart J 2013;166: e11.) Background Type 2 diabetes mellitus (T2DM) and its complications cause a large and growing burden of disease around the world. 1 Effective control of blood glucose levels is central to the management of symptoms and the prevention of complications. 2 There are a broad range of different pharmacological treatments that can be used to lower blood glucose levels but many patients still fail to achieve targets or suffer significant side effects from therapy. 3 As From a The George Institute for Global Health and University of Sydney, Sydney, New South Wales, Australia, b Department of Clinical Pharmacology, University of Groningen, University Medical Centre Groningen, Groningen, The Netherlands, c Duke Clinical Research Institute, Durham, NC, d The Royal North Shore Hospital and University of Sydney, Sydney, New South Wales, Australia, e Janssen Research & Development, LLC, Raritan, NJ, and f University of Oxford, Oxford, United Kingdom. g On behalf of the CANVAS Trial Collaborative Group. See the online Appendix E for a complete listing. RCT reg #NCT Deepak L. Bhatt, MD, MPH served as guest editor for this article. Submitted December 20, 2012; accepted May 8, Reprint requests: Bruce Neal, PhD, The George Institute for Global Health, University of Sydney and the Royal Prince Alfred Hospital, Level 10, King George V Building, Royal Prince Alfred Hospital, Missenden Rd, Sydney, NSW 2050, Australia. bneal@georgeinstitute.org.au , Mosby, Inc. Open access under CC BY-NC-ND license. such, there remains a need for novel agents that can lower blood glucose levels and reduce vascular risk with good safety and tolerability. Inhibition of the sodium glucose co-transporter 2 (SGLT2) is a novel strategy for glucose management in T2DM. 4,5 Glucose is filtered by the kidney and after glomerular passage is reabsorbed by the proximal tubule and returned to the blood stream. 6-9 SGLT2 is a highcapacity and low-affinity glucose transporter expressed exclusively in the luminal membranes of the S1 and S2 segments of the proximal renal tubules and is responsible for the majority of glucose reabsorption from the tubular lumen. SGLT1, expressed in the S3 segment, is a lowcapacity and high-affinity transporter that reabsorbs the remaining glucose in the glomerular ultrafiltrate. SGLT1 is also expressed in the intestine where it absorbs intestinal glucose and galactose. An inherited deficiency of SGLT2 can produce renal glycosuria, with some affected individuals excreting as much as 100 g of urinary glucose per day but otherwise suffering no significant short- or long-term adverse effects. 5 Pharmacological interventions with SGLT2 inhibitors result in less proximal tubular glucose reabsorption and greater urinary excretion. In this way the body loses glucose and plasma levels are decreased. An advantage of this approach is that the risk

2 218 Neal et al American Heart Journal August 2013 of hypoglycemia is reduced because there is a built-in physiological control of the system. A disadvantage is that the kidney must be filtering effectively for the intervention to be fully effective. Canagliflozin is an orally active inhibitor of SGLT2. 10 Completed Phase 2 studies of canagliflozin in patients with diabetes have demonstrated beneficial effects on glycemic control and body weight with good tolerability and safety. 10,11 Following recent concerns about the cardiovascular safety of rosiglitazone, 12 the US Food and Drug Administration provided new guidance regarding the conditions under which new therapies for the management of diabetes will be considered for registration and marketing. 13 Based on this background, the canagliflozin cardiovascular assessment study (CANVAS) was designed to define the effects of canagliflozin on safety, tolerability, and efficacy outcomes, and thereby provide the evidence on which the clinical utility of the agent could be determined. Objectives The intended primary objectives of the study were 2- fold the first being to determine the effects of canagliflozin compared to placebo (against a background of standard of care) on the risk of cardiovascular disease, and the second being to provide data on safety and tolerability. The primary null hypothesis to be tested was that there would be no difference in the risk of cardiovascular disease between patients treated with canagliflozin compared to patients treated with placebo. The secondary objectives of the study were to determine the effects of canagliflozin compared to placebo on intermediate efficacy measures, specifically, beta-cell function (using the homeostasis model assessment [HOMA] B and the proinsulin:insulin ratio), 14 progression of albuminuria, the albumin:creatinine ratio and the estimated glomerular filtration rate. The study was also designed to define the short-term, medium, and longterm effects of canagliflozin compared to placebo on hemoglobin A 1c [HbA 1c ], fasting plasma glucose [FPG], body weight, blood pressure, fasting plasma lipids, and side effects. Overall design CANVAS is a randomized, double-blind, placebo-controlled, parallel group, multicenter trial. The study was intended to be implemented in 2 consecutive phases to maximize statistical and operational efficiency. The first phase comprises 4,330 individuals randomized on a 1:1:1 basis to placebo, canagliflozin 100 mg, or canagliflozin 300 mg who were scheduled to be followed initially for an average of about 2 years. The intent was that this first phase of the study (1) substantiate the potential for demonstrating cardiovascular protection by examining effects on biomarkers, (2) confirm the dose(s) of canagliflozin to be taken forward into the second phase, and (3) achieve initial safety objectives for the compound. If the potential for cardiovascular protection was confirmed and initial safety criteria were met, the trial was planned to then enter a second phase that would recruit and follow an additional 14,000 individuals with the goal of demonstrating cardiovascular protection. Methods Recruitment The recruitment of the initial cohort was done at 386 centers in 24 countries, commencing in December 2009 and completing in March 2011 (ClinicalTrials.gov Identifier: NCT ). Regulatory approval for the conduct of the trial was obtained in each country and ethics approval was received at every site prior to initiation. The proposed recruitment of the second cohort of 14,000 was predicated on the effects of the agent on intermediate outcomes being likely to translate into a 15% or greater reduction in vascular risk, the compound proceeding to registration, a futility analysis being passed (based on the hazard ratio being b0.95), and there being no un-blinding of the observed effects of canagliflozin on the primary efficacy outcome during the regulatory submission and review process. This latter point was important to the integrity of the trial because interim un-blinding of the effects on the primary efficacy outcome would potentially influence the actions of the investigators and the care of the patients involved in the study, and therefore represent a threat to the validity of the final result. Participant inclusion and exclusion criteria Participants in the CANVAS trial are men and women with T2DM who have inadequate glycemic control (HbA 1c 7.0 and 10.5%) and are at an elevated risk of cardiovascular disease. All participants were required to provide informed consent, be willing and able to adhere to the study protocol requirements, and successfully complete the trial screening and run-in period. To ensure the recruitment of a broad population group there were minimal restrictions on the use of background therapies for glucose control or vascular risk factor management. To be eligible to participate, subjects had to be at an elevated risk of cardiovascular disease to ensure that sufficient events are accrued within the scheduled study time frame. The inclusion criteria were designed to achieve a primary outcome event rate of at least 2.25% per year. Accordingly, participants were either 30 years or older with a history of symptomatic atherosclerotic vascular disease (coronary, cerebrovascular, or peripheral) or they were 50 years or older with 2 or more of the following risk factors for vascular disease duration of diabetes 10 years or more, systolic blood pressure above 140 mmhg while on 1 more or antihypertensive agents, current smoking, micro- or macroalbuminuria, or high-density lipoprotein cholesterol [HDL-C] b1 mmol/l). The initial intent was to include 70% of individuals with a vascular disease history and 30% with 2 or more disease risk factors alone through capping of recruitment, although this was modified (to 60% with a vascular disease history) to facilitate a more rapid enrollment. Other inclusion and exclusion criteria were typical of those usually invoked for a

3 American Heart Journal Volume 166, Number 2 Neal et al 219 trial of this stage in this patient group (online Appendix A). Background antihyperglycemic agents needed to have been administered at a stable dose for the 8 weeks prior to screening and a vascular event in the preceding 3 months, a history of New York Heart Association heart failure class IV or a scheduled revascularization procedure all led to exclusion. Screening and run-in All potential participants completed a screening visit to establish initial eligibility and progressed to a 2-week, singleblind, placebo run-in period if screening criteria were met. The screening visit was to be scheduled 3 weeks prior to randomization, and individuals returned a week later for the run-in visit during which eligibility was reviewed in light of the screening visit laboratory results. If eligibility was confirmed, the participant was counseled about diet and exercise, trained in the recognition of hypoglycemia, provided with materials for making self-monitored blood glucose measurements, and then prescribed single-blind placebo therapy. The primary purpose of the run-in period was to exclude prior to randomization those individuals unlikely to adhere to the long-term treatment and follow-up regimen required by the trial. Randomization Randomization was done centrally through an interactive voice response system or an interactive web response system. Participants were randomly assigned on the basis of a computergenerated randomization schedule prepared by the study sponsor. Randomization was in a 1:1:1 ratio to 1 of 3 treatment groups: canagliflozin 100 mg, canagliflozin 300 mg, or matching placebo using randomly permuted blocks with stratification by baseline use of background glucose-lowering therapy. Post-randomization study treatment and control Study treatment is provided in identical blister cards of canagliflozin 300 mg, canagliflozin 100 mg, or placebo. Participants take their study treatment once daily before the first meal of the day until completion of the study or premature treatment discontinuation. Participants and all study staff will remain masked to individual treatment allocation until the completion of the study. Background drug treatments Participants were required to have stable background glucose-lowering therapy for 8 weeks before screening and wherever possible to persist with this treatment regimen unchanged for the first 18 weeks after randomization. The rationale for this strategy was the desire to evaluate the shortterm (18-week) effects of canagliflozin on biomarkers while on constant background therapy. Criteria for the institution of glycemic rescue therapy were provided for patients with significant derangements of blood glucose levels during this initial 18-week period (online Appendix B). Glycemic management is otherwise at the discretion of the responsible investigator in line with applicable local guidelines. Likewise, the use of all other therapies is according to best practice throughout the course of the study and instituted according to local guidelines and policies. Follow-up schedule Post-randomization follow-up in the first 6 months was scheduled at 2, 4, 6, 9, 12, 18, and 26 weeks after randomization and thereafter follow-up is at approximately 3-month intervals. The follow-up done at 2, 4, and 9 weeks after randomization was by telephone, unless a face-to-face consultation was deemed necessary. Every follow-up contact includes inquiry about primary outcomes, adverse events and concomitant therapies, and a review of the patient diary entries and self-monitored blood glucose results with additional biomarker collections for selected outcomes. All face-toface visits also include prescription of study medication, recording of physical measurements, and documentation of medical resource utilization. After the Year 1 visit, clinical laboratory assessments of biomarkers are variously done every 6 or 12 months and an ECG is recorded annually. Individuals who prematurely discontinue study treatment are encouraged to return for regular assessments to ensure full ascertainment of study outcomes and to support an intention-to-treat analysis for all outcomes. Outcomes and endpoint adjudication The primary outcome specified for the evaluation of the effects of canagliflozin on the risk for cardiovascular disease was the composite of cardiovascular death, nonfatal myocardial infarction (MI), and nonfatal stroke (ie, MACE). For the program-wide cardiovascular safety evaluation to which CANVAS contributed, the composite vascular outcome was expanded to include hospitalization for unstable angina (ie, MACE-plus). Secondary efficacy outcomes are fasting measures of beta-cell function using the HOMA-B and the proinsulin: insulin ratio. These HOMA-B assessments will be made in a subset of sites and will include about 1,200 individuals not using insulin at baseline. A first morning void urine will be collected to determine the effects on progression/regression between normal albuminuria, micro-albuminuria, and macroalbuminuria as well as to quantify the exact effects on the albumin:creatinine ratio. The estimated glomerular filtration rate will be calculated using the Modification of Diet in Renal disease formula. 15 The short-term, medium, and long-term effects will also be determined for HbA 1c, FPG, body weight, blood pressure, fasting plasma lipids (triglycerides, highdensity lipoprotein cholesterol [HDL-C], low-density lipoprotein cholesterol [LDL-C], total cholesterol, and the ratio of LDL-C to HDL-C). Safety will be evaluated on the basis of the incidence of all adverse events, all serious adverse events, and all deaths. Effects on laboratory parameters, clinical signs, and ECG findings will also be reviewed. Tolerability will be determined based on the proportion of participants permanently discontinuing randomized treatment. Adverse events pre-specified as being of specific interest included vulvovaginal infections in women, genital mycotic infections in men, urinary tract infections, hypoglycemia, fractures, cardiovascular events, and skin adverse events. In addition, cancer events were carefully assessed following a reported imbalance in breast and bladder cancer incidence in association with dapagliflozin, another SGLT2 inhibitor in clinical development. 16 Independent Endpoint Adjudication Committees assessed all events that could potentially be a primary outcome in addition to all other deaths and suspected venous thromboembolic events,

4 220 Neal et al American Heart Journal August 2013 congestive heart failure events, fractures, renal events, and hepatic events. Statistical power The trial was designed to provide 90% power (α =.05) to detect a relative risk reduction of 15% or greater for the effect of canagliflozin compared to placebo on the primary cardiovascular endpoint across the 2 trial phases among a total study population of about 18,000. This estimate assumed an event rate of 2.25% per year and a 5% per year rate of premature discontinuation of active treatment. The 15% relative risk reduction was based on projected favorable effects of the agent on HbA 1c, blood pressure, and weight. To achieve this power the study was planned to continue until 1,600 MACE events had accrued. The recruitment of 4,330 individuals to phase one of the study was based on this being a sufficient number of participants to meet initial cardiovascular safety requirements when considered in concert with the data from other controlled trials in the canagliflozin development program. It was projected that within about 2 years there would be more than 90% power with 160 MACE-plus events across the program to exclude a non-inferiority margin of 1.8, assuming a point estimate of effect of 1.0. This cohort of 4,330 individuals will continue to be followed and will contribute to the ongoing assessment of cardiovascular safety and the assessments of other outcomes as planned. In light of the decision to enroll only the initial cohort of patients there will, however, be power to detect only large effects of canagliflozin on the primary efficacy outcome 80% power (2-sided α = 0.05) to detect a 25% or greater risk reduction assuming about 420 events are accrued at study completion. Analysis Analysis will be by intention-to-treat and a P value of 0.05 will be taken to indicate a significant effect. The effect of canagliflozin compared to placebo on intermediate outcome measures will utilize methods based on analysis of covariance with the models including terms for treatment and the stratification factors used at randomization. A Cox proportional hazards model will be used to estimate effects on cardiovascular safety outcomes. A series of prespecified 18-week substudy analyses will be done to define the short-term effects of the compound on biomarkers. An Independent Data Monitoring Committee has been established to provide interim monitoring of safety and efficacy data throughout the course of the trial. The primary role of the Independent Data Monitoring Committee is to implement a prespecified interim monitoring program that will see the trial stopped prematurely if there is early evidence of significant benefit or significant harm. For all comparisons, the significance level will be preserved at 0.05 through the application of gate-keeping. Scientific responsibility for the design, analysis, and reporting of the trial lies with the Steering Committee, which comprises 6 independent academics and a representative of the trial sponsor. After trial completion, the Steering Committee will have full access to the trial database and will conduct independent analyses of the main trial outcomes. Day-to-day operation of the trial is done jointly by the sponsor and an academic research organization broadly, the academic research organization is responsible for the management of the various trial committees and operations in the Asia-Pacific region, with the sponsor having primary responsibility for other geographic regions and functions. Current status During a recruitment period of 15 months there were 7,691 individuals who were screened and 4,330 that were randomized. Of the 3,361 subjects who were not randomized, 51% had an HbA 1c out of range, 14% were using metformin and had a concurrent low estimated glomerular filtration rate, 9% met another laboratory exclusion criterion, 6% did not meet the cardiovascular disease entry criteria, 6% had a consent issue, 4% were not on stable therapy for their diabetes, and 10% were excluded for other reasons. By Mid- September 2012 the total follow-up accrued was 7,174 patient-years with a mean duration of 86.5 weeks, a median duration of 93.3 weeks, and a maximum followup of weeks. The mean age of participants at baseline was 62 years and 34% were female (Table and online Appendix C). Almost 3 quarters are white, with most of the remainder being of Asian descent. The average time since diagnosis of diabetes was a little over 13 years, with about 3 quarters of participants using metformin, one half a sulfonylurea, and one half insulin. As would be anticipated for a group with long-standing diabetes, many were on multiple glucose-lowering therapies, with the majority also using a statin and large numbers prescribed antithrombotic therapy. About one-fifth and one-sixth of those randomized had diagnoses of retinopathy and nephropathy, respectively, while 31% of participants had neuropathy. Fifty-seven percent reported a history of macrovascular atherosclerotic vascular disease at entry into the study. Mean baseline body mass index was 32 kg/m 2, mean HbA 1c was 8.2%, mean fasting plasma glucose was 9.3 mmol/l and mean estimated glomerular filtration rate was 77 ml/min per 1.73 m 2. With measurements made with patients using background therapy, baseline blood pressure was in the pre-hypertensive range (136/78 mmhg) and mean total cholesterol was 4.36 mmol/l, mean HDL-C was 1.19 mmol/l, mean LDL-C was 2.30 mmol/l, and mean triglycerides were 1.96 mmol/l. Discussion The CANVAS trial has successfully recruited 4,330 individuals with T2DM to address the initial objectives of the study. The characteristics of the participants enrolled are mostly representative of a high-risk diabetic population, both in terms of the nature of the complications reported and the types of treatments they are receiving. The proportion of participants with a history of atherosclerotic vascular disease is higher than is typically

5 American Heart Journal Volume 166, Number 2 Neal et al 221 Table. Baseline characteristics of randomized participants Age (mean, SD) 62.4 (8.0) Female (%) 34 Race (%) White 73 Asian 18 Black 2 Other 6 Current smoker (%) 18 History of hypertension (%) 87 History of heart failure (%) 11 Duration of diabetes, years (mean, SD) 13.4 (7.5) Drug therapy (n, %) Insulin 2,171 (50.1) Sulphonylurea 2,032 (46.9) Metformin 3,158 (72.9) Statin 3,119 (72.0) Anti-thrombotic 3,073 (71.0) Micro-vascular disease history (n, %) Retinopathy 842 (19.4) Nephropathy 656 (15.2) Neuropathy 1,329 (30.7) Atherosclerotic vascular disease history (n, %) Coronary 2,201 (50.8) Cerebrovascular 669 (15.5) Peripheral 658 (15.2) Any 2,714 (62.7) Body mass index, kg/m 2 (mean, SD) 32.1 (6.2) Blood pressure, mmhg (SBP/diastolic 136 (15.7)/78 (9.7) blood pressure) (mean, SD) Hemoglobin A 1c, % (mean, SD) 8.2 (0.9) Fasting plasma glucose, mmol/l (mean, SD) 9.3 (2.6) Total cholesterol, mmol/l (mean, SD) Triglycerides, mmol/l (mean, SD) 4.36 (1.15) 1.96 (1.36) HDL-C, mmol/l (mean, SD) 1.19 (0.32) LDL-C, mmol/l (mean, SD) 2.30 (0.93) LDL-C:HDL-C ratio (mean, SD) 2.02 (0.90) egfr, ml/min per 1.73 m 2 (mean, SD) 77.2 (18.8) egfr b 60 (n, %) 711 (16.4) Microalbuminuria (n, %) 966 (22.4) Macroalbuminuria (n, %) 268 (6.2) New York Heart Association class I-III. Some participants had more than one type of atherosclerotic disease. Figure Screening (n = 7,692) Randomization (n = 4,330) (n = 4,327 received at least 1 dose of study drug) Canagliflozin 300 mg (n = 1,441) Canagliflozin 100 mg (n = 1,445) Excluded (n = 3,362) 38% - HbA 1c low 14% - HbA 1c high 13% - GFR low 6% - CV criteria not met 3% - Not on stable AHA 6% - Consent issue 10% - Lab/ECG issue 10% - Other 18- and 52-week analyses for sub-studies Placebo (n = 1,441) 2-year safety analyses in support of New Drug Application 5-year safety and efficacy analyses in support of continued drug registration Trial flow chart. GFR, glomerular filtration rate; CV, cardiovascular; AHA, antihyperglycemic agent. observed in this patient group but this reflects the inclusion criteria that targeted the enrolment of individuals with ischemic macrovascular complications. The study is well placed to provide a robust assessment of the effects of canagliflozin in a diverse population group to which the agent might ultimately be applied in clinical practice. The study was designed to be done in 2 phases using a novel, scientifically robust approach. The rationale for the phased design was that it would enable key safety information to be collected for regulatory purposes, concurrent with the accrual of the data required to address the primary hypothesis of cardiovascular protection. The data from the initial follow-up of the 4,330 participants included in phase one of the study was also intended to provide insight into the effects of the agent on biomarkers, with second phase recruitment to progress only if there was evidence of effects that would be expected to translate into cardiovascular protection. Fundamental to this study design was the capacity to prevent the release of interim data defining the effects of the compound on cardiovascular outcomes. Unblinding of the primary outcome effect estimate at any point before completion of the study would represent a significant breach in the integrity of the trial and undermine the design principles. Accordingly, a series of procedures were put in place to ensure the effect estimate remained blinded and plans were developed such that the review of the data by the regulators would not require disclosure of the effects on the cardiovascular outcome. This was based upon identifying a specific and

6 222 Neal et al American Heart Journal August 2013 very limited group of individuals responsible for data analysis and report preparation who would have access to these data. Ultimately, however, a strategy that would ensure concealment of the hazard ratio for the primary outcome was not implemented, with the sponsor electing to un-blind the data to obtain better insight into the effects of the compound while preparing materials for submission to the regulators. Accordingly, recruitment to the second phase of the study will not proceed but a separate large outcome trial may be done. Follow-up of the 4,330 participants already enrolled in the first phase of the trial is scheduled to continue with the goal of addressing the primary safety and tolerability objectives and the effects on the prespecified secondary outcomes (Figure). The last few years have seen new warnings about cardiovascular harm for rosiglitazone followed by withdrawal of the agent from the market in some countries, 12 new warnings about the potential for bladder cancer for pioglitazone 17 and the completion of a series of large trials, at least one of which has suggested possible adverse effects from intensive drug regimens targeting low blood glucose levels. 2 The uncertainty that these studies have introduced into the management of T2DM serves to highlight the need for definitive large-scale outcome trials that objectively define the effects of treatments on key efficacy and safety outcomes. For canagliflozin to have a significant role in the management of diabetes will require the demonstration of meaningful benefits and a good safety and tolerability profile. In addition, many countries will require evidence of costeffectiveness. The provision of this evidence will require data from adequately powered studies of the compound able to define the effects on cardiovascular and other important outcomes. The CANVAS trial will make an important contribution by precisely and reliably defining the effects of the compound on key biomarkers as well as documenting the safety profile in this high-risk patient group. Evidence for protection against major cardiovascular events will, however, need to be obtained from a future study. The exploration of SGLT2 inhibitors as another strategy for the management of diabetes reflects the suboptimal control of hyperglycemia and vascular risks for many patients with diabetes. 18 Data from multiple studies of canagliflozin and related compounds show considerable promise for inhibition of SGLT2 as a novel therapeutic strategy in this large group of patients. Not only has this approach demonstrated significant effects on HbA 1c control, but it has also been shown that these drugs reduce body weight as well as lower blood pressure. The combined beneficial effects on glycemia, blood pressure, and weight could ultimately translate into reductions in the risks of macrovascular and microvascular complications. Acknowledgements Steering Committee: D Matthews (Co-chair), B Neal (Co-chair), G Fulcher, K Mahaffey, V Perkovic, P Stein, D de Zeeuw. Independent Data Monitoring Committee: P Home (Chair), J Anderson, I Campbell, J Lachin, D Scharfstein, S Solomon, R Uzzo. Endpoint Adjudication Committee: G Fulcher, J Amerena, G Figtree, J French, B Jenkins, R Lindley, B McGrath, A Street, J Watson. Collaborating sites (online Appendix E). Disclosures Sponsor: Janssen Research & Development, LLC. Authorship: The Steering Committee designed the study in conjunction with the Sponsor. BN wrote the first draft of the paper. The Sponsor provided the data. All authors provided extensive input into subsequent drafts. Technical editorial assistance was provided by Kimberly Dittmar of MedErgy and was funded by Janssen Global Services, LLC. References 1. Unwin N, Gan D, Mbanya J, et al. Diabetes atlas. Brussels, Belgium: International Diabetes Federation; Turnbull FM, Abraira C, Anderson RJ, et al. Intensive glucose control and macrovascular outcomes in type 2 diabetes. Diabetologia 2009; 52(11): American Diabetes Association. Executive summary: standards of medical care in diabetes Diabetes Care 2010;33(Suppl 1): S4-S List JF, Woo V, Morales E, et al. Sodium-glucose cotransport inhibition with dapagliflozin in type 2 diabetes. Diabetes Care 2009; 32(4): Wright EM. Renal Na(+)-glucose cotransporters. Am J Physiol Renal Physiol 2001;280(1):F Bailey CJ. Renal glucose reabsorption inhibitors to treat diabetes. Trends Pharmacol Sci 2011;32(2): Mogensen CE. Maximum tubular reabsorption capacity for glucose and renal hemodynamcis during rapid hypertonic glucose infusion in normal and diabetic subjects. Scand J Clin Lab Invest 1971;28(1): Mather A, Pollock C. Glucose handling by the kidney. Kidney Int Suppl 2011;120:S Rahmoune H, Thompson PW, Ward JM, et al. Glucose transporters in human renal proximal tubular cells isolated from the urine of patients with non-insulin-dependent diabetes. Diabetes 2005;54(12): Rosenstock J, Aggarwal N, Polidori D, et al. Dose-ranging effects of canagliflozin, a sodium-glucose cotransporter 2 inhibitor, as add-on to metformin in subjects with type 2 diabetes. Diabetes Care 2012; 35(6): Inagaki N, Kondo K, Iwasaki T, et al. Canagliflozin, a novel inhibitor of sodium glucose co-transporter 2 (SGLT2) improves glycemic control and reduces body weight in Japanese type 2 diabetes mellitus (T2DM). Diabetes 2011;60(Suppl 1):A274 Abstract 0999-P.

7 American Heart Journal Volume 166, Number 2 Neal et al Cohen D. Rosiglitazone: what went wrong? BMJ 2010;341:c Food and Drug Administration. Guidance for industry diabetes mellitus evaluating cardiovascular risk in new antidiabetic Therapies to treat type 2 diabetes; Vallon V, Richter K, Blantz RC, et al. Glomerular hyperfiltration in experimental diabetes mellitus: potential role of tubular reabsorption. J Am Soc Nephrol 1999;10(12): Levey A, Green T, Kusek J, et al. A simplified equation to predict glomerular filtration rate from serum creatinine. J Am Soc Nephrol 2000;11:155A. 16. Bristol-Myers Squibb, AstraZeneca.. U.S. Food and Drug Administration endocrinologic & metabolic advisory committee background document: dapagliflozin, BMS , NDA Princeton, NJ: Bristol-Myers Squibb; [cited 2013 Apr. 4]; URL: AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/ EndocrinologicandMetabolic DrugsAdvisoryCommittee/ UCM pdf. 17. Lewis JD, Ferrara A, Peng T, et al. Risk of bladder cancer among diabetic patients treated with pioglitazone: interim report of a longitudinal cohort study. Diabetes Care 2011;34(4): Gaede P, Lund-Andersen H, Parving HH, et al. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008;358(6):

8 American Heart Journal Volume 166, Number 2 Neal et al 223.e1 Appendix A. Full inclusion and exclusion criteria The study included subjects with a diagnosis of T2DM and a history or high risk of cardiovascular disease. Inclusion criteria Potential subjects were required to satisfy all of the following criteria to be enrolled in the study: Inclusion Criteria at Screening Visit Man or woman with a diagnosis of T2DM with HbA 1c level 7.0% to 10.5% at screening and be either (1) not currently on antihyperglycemic agent (AHA) therapy or (2) on AHA monotherapy or combination therapy with any approved agent: eg, sulfonylurea, metformin, pioglitazone, alpha-glucosidase inhibitor, glucagon-like peptide 1 analogue, dipeptidyl peptidase 4 inhibitor, or insulin. History or high risk of cardiovascular disease defined on the basis of either: Age 30 years with documented symptomatic atherosclerotic cardiovascular disease: including stroke; MI; hospital admission for unstable angina; coronary artery bypass graft; percutaneous coronary intervention (with or without stenting); peripheral revascularization (angioplasty or surgery); symptomatic with documented hemodynamically-significant carotid or peripheral vascular disease; or amputation secondary to vascular disease Age 50 years with 2 or more of the following risk factors determined at the screening visit: duration of T2DM of 10 years or more, systolic blood pressure N140 mmhg while the subject is on at least one blood pressure-lowering treatment, current daily cigarette smoker, documented micro- or macro-albuminuria or documented HDL-C of b1 mmol/l (b39 mg/dl) Women were required to be: Postmenopausal, defined as N45 years of age with amenorrhea for at least 18 months, or N45 years of age with amenorrhea for at least 6 months and less than 18 months and a serum follicle stimulating hormone level N40 IU/L, or Surgically sterile (have had a hysterectomy or bilateral oophorectomy, tubal ligation), or otherwise incapable of pregnancy, or Heterosexually active and practicing a highly effective method of birth control, including hormonal prescription oral contraceptives, contraceptive injections, contraceptive patch, intrauterine device, double-barrier method (eg, condoms, diaphragm, or cervical cap with spermicidal foam, cream, or gel), or male partner sterilization, consistent with local regulations regarding use of birth control methods for subjects participating in clinical trials, for the duration of their participation in the study, or Not heterosexually active Women of childbearing potential were required to have a negative urine β-human chorionic gonadotropin pregnancy test at screening and baseline (predose, day 1) Note: subjects who were not heterosexually active at screening were required to agree to utilize a highly effective method of birth control if they became heterosexually active during their participation in the study All were required to be willing and able to adhere to the prohibitions and restrictions specified in the protocol All were required to have signed an informed consent document indicating that they understood the purpose of and procedures required for the study and were willing to participate in the study To participate in the optional pharmacogenomic component of the study, subjects were required to have signed the informed consent form for pharmacogenomic research indicating willingness to participate in the pharmacogenomic component of the study (where local regulations permit). Refusal to give consent for this component did not exclude a subject from participation in the clinical study Inclusion criterion at randomization visit Subject must have taken 80% of their single-blind placebo capsules during the 2-week run-in period at Day 1 to have been eligible for randomization Exclusion criteria Potential subjects who met any of the following criteria were excluded from participating in the study: Diabetes-related/metabolic History of diabetic ketoacidosis, T1DM, pancreas or beta-cell transplantation, or diabetes secondary to pancreatitis or pancreatectomy On an AHA and not on a stable regimen (ie, agents and doses) for at least 8 weeks before the screening visit and through the screening/run-in period Note: a stable dose of insulin was defined as no change in the insulin regimen (ie, type[s] of insulin) and 15% change in the total daily dose of insulin (averaged over 1 week to account for day to day variability) Fasting fingerstick glucose at home or at investigational site N270 mg/dl (N15 mmol/l) at baseline/day 1 For patients on a sulfonylurea agent or on insulin: fasting fingerstick glucose at home or at investigational site b110 mg/dl (b6 mmol/l) at Baseline/Day 1

9 223.e2 Neal et al American Heart Journal August 2013 Note: at the investigator s discretion, based upon an assessment of recent self-monitored blood glucose values, subjects meeting either of these fingerstick glucose exclusion criteria could continue the singleblind placebo and return to the investigational site within 14 days and were eligible to be randomized if the repeat fasting fingerstick value no longer met the exclusion criterion. Subjects with fingerstick glucose N270 mg/dl (N15 mmol/l) were able to have their AHA regimen adjusted, and be rescreened once on a stable regimen for at least 8 weeks History of one or more severe hypoglycemic episode within 6 months before screening Note: a severe hypoglycemic episode was defined as an event that required the help of another person. History of hereditary glucose-galactose malabsorption or primary renal glucosuria Ongoing, inadequately controlled thyroid disorder Note: subjects on thyroid hormone replacement therapy were required to be on a stable dose for at least 6 weeks before Day 1 Renal/cardiovascular Renal disease that required treatment with immunosuppressive therapy or a history of chronic dialysis or renal transplant Note: subjects with a history of treated childhood renal disease, without sequelae, were eligible to participate Myocardial infarction, unstable angina, revascularization procedure, or cerebrovascular accident within 3 months before screening, or a planned revascularization procedure, or history of New York Heart Association Class IV cardiac disease Findings on a 12-lead electrocardiogram (ECG) that would require urgent diagnostic evaluation or intervention (eg, new clinically important arrhythmia or conduction disturbance) Gastrointestinal History of hepatitis B surface antigen or hepatitis C antibody positive (unless associated with documented persistently stable/normal range aspartate aminotransferase and alanine aminotransferase levels), or other clinically active liver disease Any history of or planned bariatric surgery Laboratory Estimated glomerular filtration rate (egfr) b30 ml/ min per 1.73m 2 at screening (provided by the central laboratory) For subjects taking metformin: at screening, serum creatinine 1.4 mg/dl (124 μmol/l) for men or 1.3 mg/dl (115 μmol/l) for women; no contraindication to the use of metformin (including egfr) based on the label of the country of investigational site Alanine aminotransferase levels N2.0 times the upper limit of normal (ULN) or total bilirubin N1.5 times the ULN at screening, unless in the opinion of the investigator and as agreed upon by the sponsor s medical officer, the findings are consistent with Gilbert s disease Other conditions History of malignancy within 5 years before screening (exceptions: squamous and basal cell carcinomas of the skin and carcinoma of the cervix in situ, or a malignancy that in the opinion of the investigator, with concurrence with the sponsor s medical monitor, is considered cured with minimal risk of recurrence) History of HIV antibody positive A current clinically important hematological disorder (eg, symptomatic anemia, proliferative bone marrow disorder, thrombocytopenia) Investigator s assessment that the subject s life expectancyislessthan1year,oranyconditionthatinthe opinion of the investigator would make participation not in the best interest of the subject, or could prevent, limit, or confound the protocol-specified safety or efficacy assessments Major surgery (ie, requiring general anesthesia) within 3 months of the screening visit or any surgery planned during the subject s expected participation in the study (except minor surgery, ie, outpatient surgery under local anesthesia) Any condition that, in the opinion of the investigator, would compromise the wellbeing of the subject or prevent the subject from meeting or performing study requirements Medications/therapies Current use of other SGLT2 inhibitor; use of rosiglitazone within 8 weeks of screening Note: subjects identified as taking rosiglitazone who were already in screening were not eligible for randomization Known allergies, hypersensitivity, or intolerance to canagliflozin or its excipients Current use of a corticosteroid medication or immunosuppressive agent, or likely to require treatment with a corticosteroid medication (for longer than 2 weeks in duration) or an immunosuppressive agent Note: subjects using inhaled, intranasal, intra-articular, or topical corticosteroids, or corticosteroids in therapeutic replacement doses may participate Received an active investigational drug (including vaccines) or used an investigational medical device within 3 months before day 1/baseline or received at least one dose of canagliflozin in a prior study General History of drug or alcohol abuse within 3 years before screening

10 American Heart Journal Volume 166, Number 2 Neal et al 223.e3 Pregnant or breast-feeding or planning to become pregnant or breast-feed during the study Employees of the investigator or study center, with direct involvement in the proposed study or other studies under the direction of that investigator or study center, as well as family members of the employees or the investigator Note: Investigators were required at randomization to assure that all study enrollment criteria had been met and determined that the subject has not had any interval change in clinical status since screening. Before randomization, subjects whose status changed after screening, such that they now met an exclusion criterion, were excluded from participation. Appendix B. Criteria for the institution of glycemic rescue therapy during initial 18-week period During the first 18-weeks of follow-up glycemic rescue criteria were applied. After Week 18, investigators are free to determine participants glycemic goals and the need for adjustments in the glucose lowering drug regimen. From Day 1 to Week 18, the criteria for starting glycemic rescue therapy were based on an FPG value exceeding the glucose cut-points shown in the table below. Subjects were counselled to contact the site if their self-monitored blood glucose levels consistently exceeded these values. A FPG measurement (ie, venous blood collection) was then done to determine eligibility for initiation of glycemic rescue therapy. Rescue criteria through week 18 Time point After day 1 through week 6 After week 6 through week 12 After week 12 through week 18 Glucose N270 mg/dl (15 mmol/l) N240 mg/dl (13.3 mmol/l) N200 mg/dl (11.1 mmol/l) Glycemic rescue therapy was determined by the investigator: either up-titration of current medications or the stepwise addition of non-insulin AHAs and then insulin therapies. After initiation of rescue therapy, the glycemic goals were based upon standard diabetes guidance, individualized for the subject, as considered appropriate by the investigator. Appendix C Baseline characteristics of randomized participants according to whether included on basis of disease history or disease risk factors Vascular disease history Basis for inclusion Two or more disease risk factors Age (mean, SD) 62.5 (8.6) 62.4 (7.2) Female (%) Race (%) White Asian Black 2 3 Other 6 5 Current smoker (%) History of hypertension (%) History of heart failure (%) 14 6 Duration of diabetes, years 12.8 (8.2) 14.2 (6.5) (mean, SD) Drug therapy (%) Insulin Sulphonylurea Metformin Statin Anti-thrombotic Micro-vascular disease history (%) Retinopathy Nephropathy Neuropathy Atherosclerotic vascular disease history (%) Coronary Cerebrovascular 23 6 Peripheral 21 8 Any Body mass index, kg/m 2 (mean, SD) 32.0 (6.2) 32.3 (6.3) Blood pressure, mmhg (SBP/diastolic blood pressure) (mean, SD) 134 (16.5)/ 77 (9.9) 139 (14.4)/79 (9.3) Hemoglobin A 1c, % (mean, SD) 8.2 (0.9) 8.2 (0.9) Fasting plasma glucose, 9.2 (2.6) 9.3 (2.6) mmol/l (mean, SD) Total cholesterol, mmol/l (mean, SD) 4.25 (1.2) 4.49 (1.1) Triglycerides, mmol/l (mean, SD) 1.96 (1.43) 1.96 (1.26) HDL-C, mmol/l (mean, SD) 1.16 (0.30) 1.23 (0.34) LDL-C, mmol/l (mean, SD) 2.23 (0.94) 2.39 (0.91) LDL-C:HDL-C ratio (mean, SD) 2.01 (0.93) 2.05 (0.86) egfr, ml/min per 1.73 m (18.8) 79.0 (18.8) (mean, SD) egfr b60 (%) Microalbuminuria (%) Macroalbuminuria (%) HDL-C b1 mmol/l (%) New York Heart Association class I-III. Some participants had more than one type of atherosclerotic disease.

11 223.e4 Neal et al American Heart Journal August 2013 Appendix D. Endpoint definitions for major clinical endpoints Death All deaths will be reviewed by the adjudicators and a cause determined at least to the level of the ICD chapter headings. Because the main role is to identify cardiovascular deaths, the approach used will be to present all deaths as potential cardiovascular deaths and ask the committee to confirm or refute that the cause was cardiovascular. Because there is often confusion in reporting cause of death the study will seek a proximate and underlying cause of death in every case (although it is understood that it may not be possible to assign both for all deaths). The question about cardiovascular cause will be applied to both the proximate and underlying causes. The reason for assigning a death as cardiovascular or noncardiovascular, and the reasoning behind the adjudicator s assignment of the cause of death, will be documented. The determination of the specific cause of cardiovascular death is complicated by the fact that the interest is particularly in one underlying cause of death (acute MI and several modes of death (arrhythmia and heart failure/ low output). It is noted that heart attack-related deaths are manifested as sudden death or heart failure, so these events need to be carefully defined. Definition of cardiovascular death. Cardiovascular death includes death resulting from an acute MI, sudden cardiac death, death due to heart failure, death due to stroke, and death due to other cardiovascular causes, as follows: 1. Death due to Acute Myocardial Infarction refers to a death by any mechanism (arrhythmia, heart failure, low output) within 30 days after a MI related to the immediate consequences of the myocardial infarction, such as progressive congestive heart failure (CHF), inadequate cardiac output, or recalcitrant arrhythmia. If these events occur after a break (eg, a CHF and arrhythmia free period of at least a week), they should be designated by the immediate cause, even though the MI may have increased the risk of that event (eg, late arrhythmic death becomes more likely after an acute MI [AMI]). The acute MI should be verified to the extent possible by the diagnostic criteria outlined for acute MI or by autopsy findings showing recent MI or recent coronary thrombus. Sudden cardiac death, if accompanied by symptoms suggestive of myocardial ischemia, new ST elevation, new left bundle-branch block (LBBB), or evidence of fresh thrombus by coronary angiography and/or at autopsy should be considered death resulting from an acute myocardial infarction, even if death occurs before blood samples or 12-lead ECG could be obtained, or at a time before the appearance of cardiac biomarkers in the blood. Death resulting from a procedure to treat a MI (percutaneous coronary intervention [PCI], coronary artery bypass graft surgery (CABG), or to treat a complication resulting from MI, should also be considered death due to acute MI. Death resulting from a procedure to treat myocardial ischemia (angina) or death due to a myocardial infarction that occurs as a direct consequence of a cardiovascular investigation/procedure/operation should be considered as a death due to other cardiovascular causes. 2. Sudden Cardiac Death refers to a death that occurs unexpectedly, not following an acute MI, and includes the following deaths: a. Death witnessed and instantaneous without new or worsening symptoms b. Death witnessed within 60 minutes of the onset of new or worsening cardiac symptoms, unless the symptoms suggest AMI c. Death witnessed and attributed to an identified arrhythmia (eg, captured on an ECG recording, witnessed on a monitor, or unwitnessed but found on implantable cardioverter-defibrillator review) d. Death after unsuccessful resuscitation from cardiac arrest e. Death after successful resuscitation from cardiac arrest and without identification of a non-cardiac etiology (post-cardiac arrest syndrome) f. Unwitnessed death without other cause of death (information regarding the patient s clinical status preceding death should be provided, if available) General considerations A subject seen alive and clinically stable hours prior to being found dead without any evidence or information of a specific cause of death should be classified as sudden cardiac death. Typical scenarios include: Subject well the previous day but found dead in bed the next day Subject found dead at home on the couch with the television on Deaths for which there is no information beyond Patient found dead at home may be classified as death due to other cardiovascular causes or in some trials, undetermined cause of death. Please see Definition of Undetermined Cause of Death, for full details. 3. Death due to Heart Failure or Cardiogenic Shock refers to a death occurring in the context of clinically worsening symptoms and/or signs of heart failure (see Chapter 7) without evidence of another cause of death and not following an AMI. Note that deaths due to heart failure can have various etiologies, including one or more AMIs (late effect), ischemic or non-ischemic cardiomyopathy, or valve disease.