Rationale, design, and organization of a randomized, controlled Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS) in patients with type 2 diabetes and established cardiovascular disease Jennifer B. Green, MD, a M. Angelyn Bethel, MD, b Sanjoy K. Paul, PhD, c Arne Ring, PhD, b Keith D. Kaufman, MD, d Deborah R. Shapiro, DrPH, d Robert M. Califf, MD, e and Rury R. Holman, MB, ChB, FRCP b Durham, NC; Oxford, United Kingdom; Brisbane, Australia; and Whitehouse Station, NJ

Sitagliptin, an oral dipeptidyl peptidase-4 inhibitor, lowers blood glucose when administered as monotherapy or in combination with other antihyperglycemic agents. TECOS will evaluate the effects of adding sitagliptin to usual diabetes care on cardiovascular outcomes and clinical safety. TECOS is a pragmatic, academically run, multinational, randomized, doubleblind, placebo-controlled, event-driven trial recruiting approximately 14,000 patients in 38 countries who have type 2 diabetes (T2DM), are at least 50 years old, have cardiovascular disease, and have an hemoglobin A1c value between 6.5% and 8.0%. Eligible participants will be receiving stable mono- or dual therapy with metformin, sulfonylurea, or pioglitazone, or insulin alone or in combination with metformin. Randomization is 1:1 to double-blind sitagliptin or matching placebo, in addition to existing therapy in a usual care setting. Follow-up occurs at 4-month intervals in year 1 and then twice yearly until 1300 confirmed primary end points have occurred. Glycemic equipoise between randomized groups is a desired aim. The primary composite cardiovascular endpoint is time to the first occurrence of cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, or hospitalization for unstable angina, with cardiovascular events adjudicated by an independent committee blinded to study therapy. TECOS is a pragmatic-design cardiovascular outcome trial assessing the cardiovascular effects of sitagliptin when added to usual T2DM management. (Am Heart J 2013;166:983-989.e7.)

Type 2 diabetes mellitus (T2DM) is characterized by hyperglycemia and multiple metabolic derangements including insulin resistance, excess glucagon secretion, and progressive loss of insulin secretory ability. Microvascular complications are hallmarks of this condition, but patients with T2DM also experience high rates of macrovascular disease. 1 This excess risk of cardiovascular disease (CVD) in T2DM is associated with a higher occurrence of hypertension, obesity, dyslipidemia, hypercoagulability, and a proinflammatory state. 2 A doubling of CVD risk remains after adjustment for classical risk

From the aDivision of Endocrinology, Duke University Medical Center, Durham NC, b Diabetes Trials Unit, Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, Oxford, United Kingdom, cClinical Trials & Biostatistics Unit, QIMR Berghofer Medical Research Institute, Brisbane, Australia, dMerck Research Laboratories, Merck Sharp & Dohme Corp., Whitehouse Station NJ, and eDuke Translational Medicine Institute, Durham, NC. James A. de Lemos, MD, served as guest editor for this article. Submitted November 7, 2012; accepted September 4, 2013. Reprint requests: Rury R. Holman, MB, ChB, FRCP, University of Oxford Diabetes Trials Unit Oxford Centre for Diabetes, Endocrinology and Metabolism Oxford, United Kingdom. E-mail: [email protected] 0002-8703/$ - see front matter © 2013, Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2013.09.003

factors. 3 Given the emerging worldwide epidemic of diabetes, 4 the excess CVD mortality and morbidity is of major concern. Intensive glycemic control can reduce the risk of microvascular complications in type 1 diabetes (T1DM) and T2DM, 5,6 but the impact of glucose reduction upon cardiovascular complications is less clear. Trials of intensive glycemic control in patients with both T1DM and T2DM have identified a reduction in cardiovascular events over the longer term, and meta-analyses of data from trials of glucose lowering in T2DM estimate a 15% reduction in risk of fatal or non-fatal myocardial infarction (MI) conveyed by a 0.9% decrement in hemoglobin A1c (HbA1c). 7-9 However, concerns about the optimal intensity of glucose-lowering and the means of achieving this were raised by the ACCORD trial, which showed increased mortality among participants allocated to a target HbA1c of b6%. 10 Scrutiny of the cardiovascular safety of antihyperglycemic drugs in general increased following concerns raised about rosiglitazone and other drugs in development, 11 leading to new US Food and Drug Administration (FDA) 12 and European Medicines Agency (EMA) 13 requirements for cardiovascular safety evaluations of new glucose-lowering agents.

American Heart Journal December 2013

984 Green et al

Traditionally, diabetes medications have been approved based upon efficacy in lowering HbA1c, a biomarker accepted by most as a surrogate for reduction in risk of microvascular complications and hyperglycemic symptoms. The new FDA and EMA guidelines 12,13 emphasize that for new medications to gain approval, clinical trials data must exclude a clinically important excess of drugassociated cardiovascular risk. Trials should include persons at sufficiently high cardiovascular risk, eg, patients who are elderly, have multiple cardiovascular risk factors, and/or CVD, or some degree of renal dysfunction; these patients should be representative of populations likely to receive the medication in clinical practice. Studies should also be of sufficient duration to accumulate an adequate number of cardiovascular events for analysis. The FDA has further specified that, prior to submission of a new drug application, the sponsor must demonstrate in randomized controlled trials of the investigational drug that the risk ratio for important cardiovascular events is less than 1.8 (based on the upper bound of the 95% CI), with a reassuring point estimate. If the upper bound of the 95% CI for estimated cardiovascular risk in data from premarketing trials is between 1.3 and 1.8 and the overall risk-benefit balance is otherwise acceptable, a postmarketing trial to demonstrate a relative risk less than 1.3 will generally be required. 12

Methods Trial design TECOS is a randomized, double-blind, controlled trial comparing sitagliptin and placebo when added to best usual care in patients with T2DM and cardiovascular disease. Participants, healthcare providers, data collectors, event adjudicators, the sponsor, and the sponsor’s academic partners are blinded to participants’ assigned therapy.

Trial population TECOS will enroll approximately 14,000 participants from 38 countries. Sites are distributed with the aim of enrolling one third of participants from each of Europe, Australasia, and the Americas. The trial will enroll patients aged ≥50 years with T2DM, cardiovascular disease, and HbA1c values of 6.5% to 8.0%. Patients are eligible if their diabetes has been managed with stable-dose monotherapy or dual combination therapy with metformin, pioglitazone, or a sulfonylurea for ≥3 months prior to enrollment. As per the first protocol amendment adopted in September 2010, patients taking stable doses of insulin (±20% of total daily insulin dose), either alone or in combination with a stable dose of metformin for ≥3 months prior, are also eligible. Enrollment criteria are listed in online Appendix A. In response to a regulatory requirement, the protocol requires recruitment of ≥2000 patients receiving metformin monotherapy at baseline. The second protocol amendment in April 2012 removed a previous regulatory requirement to enroll a minimum number of patients receiving pioglitazone.

TECOS: an overview

Randomization

TECOS will assess the balance of risk and benefit from a cardiovascular perspective for sitagliptin, an orally administered dipeptidyl peptidase-4 (DPP-4) inhibitor approved for the treatment of T2DM. 14 DPP-4 inhibitors block the metabolism and inactivation of the incretin hormones glucagon-like peptide-1 (GLP-1) and glucosedependent insulinotropic polypeptide, thereby increasing the concentrations of active forms of these hormones, resulting in increased insulin secretion and suppression of glucagon secretion in a glucose-dependent fashion. Sitagliptin lowers blood glucose when administered as monotherapy, when combined with other oral antihyperglycemic agents, and when added to a regimen of insulin or insulin plus metformin. 14,15 TECOS will primarily test the hypothesis that sitagliptin, when used as part of usual diabetes care, is non-inferior to usual care without sitagliptin with regard to the risk of significant confirmed cardiovascular outcomes. If sitagliptin is found to be non-inferior in this respect, an assessment of superiority will be performed. Although TECOS was planned prior to the new FDA/EMA guidances, its conduct and planned analyses are consistent with the agencies' recommendations. TECOS, managed via an academic collaboration between the Duke Clinical Research Institute (DCRI) and the University of Oxford Diabetes Trials Unit (DTU), began enrollment in December 2008.

Eligible participants are randomly allocated via an interactive voice-response system to treatment with sitagliptin or placebo in a 1:1 ratio. Randomization codes are generated by the sponsor and are provided in strict confidence only to the facility packaging study medication, the group responsible for running the interactive voice-response system (independent of the sponsor, DTU, DCRI, and all operational teams), and the independent Data and Safety Monitoring Board (DSMB).

Drug administration and dosing Blinded study drug (sitagliptin 100 mg or placebo) will be administered in accordance with product labeling as a tablet taken orally once daily in the morning. Because sitagliptin is cleared primarily through the kidneys, a starting dose of 50 mg/ day will be used for patients with an estimated glomerular filtration rate (eGFR) of 30 to b50 mL/min per 1.73 m 2. During the study, eGFR values will be assessed at least annually to determine whether dose adjustment is necessary. If, based on 2 consecutive serum creatinine determinations, a participant's eGFR falls to 30 to b50 mL/min per 1.73 m 2, the dose will be reduced to 50 mg/day. If eGFR decreases to b30 mL/min per 1.73 m 2, the dose will be reduced to 25 mg/day. The dose will be up-titrated should a sustained recovery of eGFR occur.

Participant follow-up To perform the trial in pragmatic fashion, study visits and procedures are integrated into each participant's usual diabetes care schedule. Participants will continue to see their usual-care

American Heart Journal Volume 166, Number 6

physicians, as all interventions other than administration of study drug will be conducted by the usual-care provider. Participants' antihyperglycemic regimens may be adjusted as deemed necessary by the usual-care physician in order to achieve an appropriate individualized glycemic goal according to local guidelines. After enrollment, the addition of any therapies other than GLP-1 analogues or DPP-4 inhibitors is permitted, although rosiglitazone use is discouraged. Physicians will be asked to avoid changing antihyperglycemic therapy during the first 4 months of the study unless clinically essential, in order to assess the initial HbA1c-lowering effect of study drug. In the absence of contraindications, usual-care physicians will be encouraged to continue baseline antihyperglycemic agents in addition to study medication. Individuals with a history of repeated severe hypoglycemia in the prior year are excluded from participation (online Appendix A), minimizing enrollment of patients with hypoglycemia unawareness. Participants experiencing ≥2 episodes of severe hypoglycemia between study visits (episodes in which assistance was required, despite down-titration or discontinuation of other antihyperglycemic therapy) will discontinue study medication. Participants should receive counseling regarding appropriate lifestyle modification prior to enrollment as part of usual diabetes care; this will be reinforced during study visits. Investigators will communicate with usual-care physicians to reinforce participants' progress toward appropriate glycemic goals, as well as locally appropriate guidelines for cardiovascular risk reduction. Utilization of individually and locally appropriate guidelines in lieu of study-specific guidelines is most appropriate for the large, diverse, and globally-dispersed TECOS population. Ideally, this strategy will also yield glycemic equipoise between treatment groups. Study visits will occur as outlined in Figure. The study will continue until 1300 participants experience confirmed (by adjudication) primary composite cardiovascular events and ≥36 months have elapsed since the enrolment of 2000 patients receiving metformin monotherapy. All participants will be followed until the conclusion of the study with an expected median follow-up of approximately 4 years.

Measurements Medical history will be obtained at baseline, including each participant’s duration of diabetes, diabetes-related complications, and history of hypertension, dyslipidemia, cardiovascular complications, and smoking. Concomitant antihyperglycemic, lipid-lowering, antihypertensive, and anticoagulant/antithrombotic medications will be recorded at all visits, as will the number of episodes of severe hypoglycemia (hypoglycemia requiring the assistance of another individual), deaths, hospitalizations, cardiovascular events and interventions, other serious adverse events, adverse events resulting in study drug discontinuation, and physician visits. In addition, a specific list of expected clinical events, including study endpoints, diabetes-related endpoints, and known side effects of sitagliptin (online Appendix B) will be collected from randomization through trial conclusion. Data will also be captured systematically for all suspected or confirmed occurrences of malignancies and episodes of pancreatitis. Body weight, blood pressure, and heart rate will be recorded at all visits. Blood samples for HbA1c performed at designated

Green et al 985

local laboratories will be obtained at enrollment, months 4 and 8, annually, at study end, and at discontinuation visits. Samples will be drawn for future use (biomarkers) at enrollment and month 12, and sampling for genetic analysis will occur at any time that consent is obtained to do so. The strategy for biomarker measurement will be determined by the Executive Committee prior to study completion and the consent forms were designed to receive permission for flexibility in this rapidly changing field. The remaining laboratory information collected will be as available through usual care. Participants will be encouraged to bring laboratory reports to all in-person visits; otherwise, results of interest will be obtained from the usual-care provider. These include serum creatinine, urine albumin/creatinine ratio, hemoglobin, high-sensitivity C-reactive protein, and levels of lipoproteins and triglycerides.

Primary outcome The TECOS primary composite cardiovascular endpoint is defined as the time from randomization to the first confirmed cardiovascular-related death, nonfatal MI, nonfatal stroke, or unstable angina requiring hospitalization.

Secondary outcomes The TECOS secondary outcomes are: a composite endpoint defined as time to first confirmed cardiovascular-related death, nonfatal MI, or nonfatal stroke; time to occurrence of the individual components of the primary endpoint; time to allcause mortality; and time to hospital admission for congestive heart failure. Other predetermined outcomes include changes in urinary albumin:creatinine ratio, eGFR, body weight, and HbA1c from baseline; time to initiation of additional antihyperglycemic therapy and/or initiation of chronic insulin therapy; time to noncardiovascular death; time to first cardiovascular or peripheral revascularization procedure; frequency of severe hypoglycemia; and counts of outpatient visits and hospitalizations.

Endpoint adjudication A clinical events committee (CEC), blinded to treatment allocation and independent of the sponsor, will adjudicate events including cardiovascular-related death, non-fatal MI, nonfatal stroke, unstable angina requiring hospitalization, congestive heart failure requiring hospitalization, and acute pancreatitis. The adjudication of pancreatitis events will better define the relationship between sitagliptin exposure and this comorbid condition commonly associated with diabetes. 16 CEC members include board-certified or -eligible endocrinologists, cardiologists, neurologists, gastroenterologists, or physicians with other relevant expertise and CEC experience.

Statistical considerations Sample size and power calculations. For noninferiority, assuming a true underlying HR of 1.00, 611 patients with a confirmed primary composite cardiovascular endpoint are required to provide 90% power to yield the upper limit of the adjusted 95% CI for a HRb1.30 at a one-sided α level of .025. To provide consistency with regulatory guidance issued subsequent to trial design, the initial protocol non-inferiority margin of 1.20 has been increased to 1.30.

American Heart Journal December 2013

986 Green et al

Figure

Summary of study design.

For superiority, assuming a true underlying HR of 0.85 (or 0.80), 1300 patients with a confirmed primary cardiovascular composite endpoint are required to provide approximately 81% (or 98%) power at a one-sided α level of 0.025; i.e., upper limit of the adjusted 95% CI for a HRb1.00. Accordingly, to yield a minimum of 1300 events within a 6year interval from first participant entered until last participant completed (inclusive of the 3-year enrollment period), approximately 14,000 participants need to be randomized, assuming an annual primary composite cardiovascular endpoint rate of 2.5%3.0% 17 and an annual rate of loss to follow-up of ≤1%. Sample size calculations for the non-inferiority test assume also that ≤12% of patients will discontinue study medication or violate the protocol in year one, with a ≤10% annual rate thereafter.

Analysis plan The HR for the between-treatment difference in time to first primary composite cardiovascular endpoint will be calculated

using a Cox proportional hazards model stratified by region and prior diabetes pharmacotherapy at randomization with treatment included as an explanatory factor. If the upper bound of the 2-sided 95% CI for the HR is b1.30, then sitagliptin as part of usual care will be declared non-inferior to usual care without sitagliptin. A per-protocol (PP) approach will be used for the primary assessment of non-inferiority with an intention-to-treat (ITT) approach used as a supportive analysis. The validity of the proportional hazard assumption underlying the Cox survival analysis will be tested using an appropriate method. 18 If sitagliptin is found non-inferior to placebo with respect to the risk of the primary composite cardiovascular endpoint, the possible superiority with sitagliptin in relation to this endpoint will be evaluated. If the upper bound of the 95% CI for the HR from the Cox model is b1.00, then sitagliptin as part of usual care will be declared superior to usual care without sitagliptin (type I error adjusted at P b .025). The superiority assessment will be based primarily on ITT population, with PP approach based supportive analysis.

American Heart Journal Volume 166, Number 6

The PP approach will only include events occurring while participants comply with key protocol requirements because specific factors, such as use of open-label sitagliptin or other DPP-4 inhibitors and the observation of events after discontinuation of study medication, could confound the treatment effect and artificially increase chances of finding noninferiority. This approach is consistent with ICH Guideline E-9 19 and CONSORT statement. 20 However, the planned ITT approach may be of particular importance as it is often the analysis preferred by regulatory agencies and the clinical community in the assessment of noninferiority. Prespecified subgroup analyses for the primary composite cardiovascular endpoint in the PP population will include region, antihyperglycemic therapy at entry, race, sex, age, baseline HbA1c, duration of diabetes, baseline renal function, and history of previous cardiovascular events. Subgroups will be divided by tertiles for continuous variables. Appropriate summary measures along with HR (95% CI) for subgroups and their possible interactions (by treatment group) will be provided. Missing values for continuous clinical and biochemical study variables will be imputed using a multiple imputation approach for key secondary analyses. The non-parametric propensity score method will be used to impute missing values where data show a monotone missing pattern (if an observation is missing at time point t, then it is missing at all subsequent time points). If there is a random missing pattern, then the Bayesian Markov chain Monte Carlo method will be used.

Trial organization and funding TECOS is sponsored by Merck Sharp & Dohme Corp and managed collaboratively by the DCRI (Durham, NC) and the DTU (Oxford, UK). Prior to database lock, the sponsor will retain the randomization codes but all data (except for key safety data) will be held by the DCRI and DTU and analyzed only in a blinded fashion. The independent, unblinded DSMB and DSMB statistician will have full access to all data and randomization codes throughout the trial. The sponsor will have access to all data only after the primary results have been presented at a public conference or a manuscript has been accepted for publication. The CEC is responsible for adjudication of all cardiovascular and pancreatitis events. The DSMB will regularly review unblinded outcomes and safety data during the study. An Executive Committee (EC) comprising 9 academics with expertise in cardiology, endocrinology, or biostatistics and 4 representatives from Merck Research Laboratories is responsible for overall study design, management and oversight. The EC will supervise use of the biomarker and genetic samples as it is not prespecified. An Operations Committee including investigators from each participating country serves as the interface between the EC and study sites and reinforces the need for appropriate metabolic and cardiovascular risk management among participants. Individuals comprising these groups are listed in online Appendix C. The authors of this manuscript are responsible for the drafting and editing of the manuscript and its final contents.

Discussion TECOS is a pragmatic-design, randomized, international, double-blind, placebo-controlled clinical outcomes

Green et al 987

trial assessing the cardiovascular impact of sitagliptin when added to the usual management of T2DM. Given cardiovascular safety concerns raised in relation to some other classes of antihyperglycemic therapies, this is a timely assessment of the first-in-class DPP-4 inhibitor in accordance with recent FDA and EMA guidances. 12,13 Although TECOS will first assess sitagliptin for noninferiority with regard to cardiovascular outcomes, it is adequately powered to assess sitagliptin for superiority if noninferiority is shown. Evidence that incretin-based therapies may be cardioprotective is growing, with studies performed primarily in animals suggesting that GLP-1 has direct cardiovascular benefits, including protection of cardiomyocytes from ischemic damage as well as positive inotropic and chronotropic effects. 21 Sitagliptin is considered weight- and lipid-neutral but its administration has been found to significantly reduce 24hour systolic and diastolic blood pressures compared with placebo in patients with hypertension. 22 A pooled analysis of major adverse cardiovascular events reported in 25 sitagliptin trials found that the adjusted incidence rate ratio for such events was 0.83 (95% CI, 0.53-1.30) in sitagliptin-exposed versus nonexposed patients. 23 Favorable risk ratios were also reported from analyses of cardiovascular events in trials of other DPP-4 inhibitors, and a meta-analysis of data for the entire drug class. 24-27 These analyses, however, are limited by the generally short duration of drug exposure, small numbers of cardiovascular events, and inconsistent event adjudication in the included trials. Recently, “topline” reports from a large cardiovascular outcomes trial with saxagliptin have been reported and shown noninferiority, but not superiority. 28 The TECOS primary composite cardiovascular endpoint comprises cardiovascular death, nonfatal MI, nonfatal stroke and unstable angina requiring hospitalization. Prespecified secondary analyses will also examine a composite consisting only of cardiovascular death, nonfatal MI, and non-fatal stroke. The choice of endpoint was a major topic of discussion during trial design. Recognizing that many trials have included only the “hard” outcomes of cardiovascular death, MI, and stroke, 29 in accord with a recent review, 30 we included the broader outcome as primary because freedom from hospitalization is a major outcome of interest in chronic disease management. Furthermore, unstable angina requiring hospitalization is indicative of severe coronary atherosclerosis and an increased risk of MI. We believe the trial also has adequate power to demonstrate non-inferiority for this secondary analysis given data suggesting a favorable cardiovascular effect of DPP-4 inhibition. 27,31 Entry-level HbA1c values are limited to 6.5% to 8.0% to ensure only modest hyperglycemia at enrollment. Therefore, any initial glycemic improvement seen in the sitagliptin-treated group will likely be modest compared with the placebo group 32, and will optimally permit rapid

American Heart Journal December 2013

988 Green et al

equalization of glycemic control between groups subsequent to the month 4 visit. Glycemic equipoise between randomized groups is a desired outcome as it will permit assessment of the cardiovascular effects of sitagliptin independently of its glucose-lowering effects. Although the trial does not have a stated HbA1c goal, the need for individually-appropriate glycemic control consistent with regional guidelines is emphasized to ensure that trial activities occur in a setting of otherwise appropriate care. This approach will also permit relaxation of strict glycemic control for patients judged unlikely to benefit from it. The pragmatic design of TECOS, incorporating study procedures into existing usual care, may provide a model for future studies. Greater integration of large-scale trials into usual care will permit their execution in a more financially feasible manner. However, the ability to systematically collect all desired laboratory data from external sources is uncertain, and the degree to which the Operations Committee can influence investigators to achieve regionally-appropriate glycemic, lipid, and blood pressure control of participants is also unknown. Nonetheless, the performance of TECOS in the context of usual care should yield results that reflect the use of sitagliptin in current clinical practice. Enrolling and retaining 14,000 patients with T2DM and established CVD is challenging, particularly given the number of large diabetes trials recruiting similar patients. Since TECOS enrolled its first participant, clinical practice has seen a marked increase in the uptake of incretinbased therapies, a decrease in the use of thiazolidinediones due to safety concerns, 17,33 and a move toward earlier use of insulin. Additional data have also accumulated regarding the use of sitagliptin in combination with other oral agents and insulin. As a result, the TECOS protocol was amended in September 2010 to include participants receiving insulin therapy at baseline and to reduce the requirement for the number of participants receiving pioglitazone at baseline from 2000 to 1000. A second amendment in April of 2012 revised the noninferiority margin from 1.20 to 1.30 in keeping with existing regulatory guidance and eliminated the requirement for a minimum number of patients receiving pioglitazone. These adaptations allow TECOS to reflect significant changes in clinical practice, thus producing more generalizable results. Trials evaluating cardiovascular safety prior to approval of new diabetes drugs will require funding from the pharmaceutical industry. Other organizations are unlikely to sponsor such trials, given significant expenditures associated with long-term, large-scale outcomes trials and the number of agents that should be individually tested. Collaboration between industry and academia to perform pragmatic cardiovascular outcomes trials provides an efficient and cost-effective way to deliver results with independent management and analysis of data.

The enrollment of a diverse global population with T2DM receiving a wide range of other glucoselowering medications means that the results will help inform clinical management for the majority of patients with this condition who also have established cardiovascular disease.

Disclosures Conflicts of Interest. JBG reports receiving institutional research support from Amylin (now BMS) and Merck and compensation for lectures and consulting from Merck and Takeda. MAB reports research support from Merck, Amylin, Lilly, Novartis, and Bayer. S.P. has acted as consultant and speaker for Novartis and Amylin and received grants supporting clinical studies from Merck, Novo Nordisk, Pfizer, Amylin and Hospira. AR reports no potential conflicts of interest. KDK and DRS are employees of and own stock in Merck Sharp & Dohme Corp. RRH reports receiving research support from Amylin, Bayer, Merck, and Novartis; participating in advisory boards for Amylin, Lilly, Merck, Novartis, and Novo Nordisk; and receiving compensation for lectures from Bayer, Lilly, Merck, and Merck Serono. A complete listing of disclosure information for R.M.C. is available at https://dcri.org/about-us/conflict-of-interest.

References 1. Stamler J, Vaccaro O, Neaton J, et al. Diabetes, other risk factors, and 12-year cardiovascular mortality for men screened in the Multiple Risk Factor Interventional Trial. Diabetes Care 1993;16: 434-44. 2. Turner R, Millins H, Neil H, et al. Risk factors for coronary artery disease in non-insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS: 23). BMJ 1998;316:823-8. 3. Emerging Risk Factors Collaboration. Diabetes mellitus, fasting blood glucose concentration, and risk of vascular disease: a collaborative meta-analysis of 102 prospective studies. Lancet 2010;375: 2215-22. 4. International Diabetes Federation. IDF diabetes atlas (5th ed). 2011. http://www.idf.org/diabetesatlas (accessed August 1, 2013). 5. Diabetes Control and Complications Trial Research Group. The effect of intensive diabetes treatment on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86. 6. UK Prospective Diabetes Study (UKPDS) Group. Intensive bloodglucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352:837-53. 7. Nathan D, Cleary P, Backlund J, et al. for the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Study Research Group. Intensive diabetes treatment and cardiovascular disease in type 1 diabetes mellitus. N Engl J Med 2005;353:2643-53. 8. Holman R, Paul S, Bethel A, et al. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med 2008;359:1577-89.

American Heart Journal Volume 166, Number 6

9. Turnbull FM, Abraira C, Anderson RJ, et al. Intensive glucose control and macrovascular outcomes in type 2 diabetes. Diabetologia 2009; 52:2288-98. 10. Gerstein H, Miller M, Byington R, et al. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med 2008;358: 2545-59. 11. Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med 2007;356:2457-71. 12. Guidance for Industry: Diabetes Mellitus – Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2 Diabetes. (2008) http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ ucm071627.pdf (accessed August 1, 2013). 13. European Medicines Agency. Guideline on clinical investigation of medicinal products in the treatment of diabetes mellitus. 2010, http://www.ema.europa.eu/docs/en_GB/document_library/ Scientific_guideline/2012/06/WC500129256.pdf. (accessed July 18, 2012). 14. Januvia (sitagliptin) prescribing information. http://www.merck. com/product/usa/pi_circulars/j/januvia/januvia_pi.pdf (accessed August 1, 2013). 15. Vilsbøll T, Rosenstock J, Yki-Järvinen H, et al. Efficacy and safety of sitagliptin when added to insulin therapy in patients with type 2 diabetes. Diabetes Obes Metab 2010;12:167. 16. Drucker DJ. Incretin action in the pancreas: potential promise, possible perils, and pathological pitfalls. Diabetes 2013;62: 3316-23. 17. Dormandy J, Bhattacharya M, van Troostenburg de Bruyn AR. PROactive investigators. Safety and tolerability of pioglitazone in high-risk patients with type 2 diabetes: an overview of data from PROactive. Drug Saf 2009;32:187-202. 18. Survival analysis: A self learning text. 2nd edition. David G. Kleinbaum and Mitchel Klein. 2005, Springer Science + Business Media, LLC. Chapter 4 pp.131-57. 19. International Conference on Harmonisation. ICH harmonised tripartite guideline. Statistical principles for clinical trials (E9). http://www. ich.org/products/guidelines/efficacy/article/efficacy-guidelines. html. (accessed August 1, 2013). 20. Piaggio G, Elbourne DR, Altman DG, et al. Reporting of noninferiority and equivalence randomized trials: an extension of the CONSORT statement. JAMA 2006;295:1152-60.

Green et al 989

21. Mannucci E, Rotella CM. Future perspectives on glucagon-like peptide-1, diabetes and cardiovascular risk. J Nutrition Metab Cardiovasc Dis 2008;18:639-45. 22. Mistry G, Maes A, Lesseter K, et al. Effect of sitagliptin, a dipeptidyl peptidase-4 inhibitor, on blood pressure in nondiabetic patients with mild to moderate hypertension. J Clin Pharmacol 2008;48:592-8. 23. Engel SS, Round E, Golm GT, et al. Safety and tolerability of sitagliptin in type 2 diabetes: pooled analysis of 25 clinical studies. Diabetes Ther 2013;4:119-45. 24. Frederich R, Alexander J, Fiedorek F, et al. A systematic assessment of cardiovascular outcomes in the saxagliptin drug development program for type 2 diabetes. Postgrad Med 2010;122:16-27. 25. Schweizer A, Dejager S, Foley J, et al. Assessing the cardio-cerebral safety of vildagliptin: meta-analysis of adjudicated events from a large phase III type 2 diabetes population. Diab Obes Metab 2010; 12:485-94. 26. White W, Gorelick P, Fleck P, et al. Cardiovascular events in patients receiving alogliptin: a pooled analysis of randomized clinical trials. ADA Abstract No. 0391-PP, 2010. 27. Monami M, Iacomelli I, Marchionni N, et al. Dipeptidyl peptidase-4 inhibitors in type 2 diabetes: a meta-analysis of randomized clinical trials. J Nutr Metab Cardiovasc Dis 2010;20:224-35. 28. Information available at: http://news.bms.com/press-release/ astrazeneca-and-bristol-myers-squibb-announce-top-line-resultssavor-timi-53-cardiovas (accessed August 1, 2013). 29. Joffe HV, Parks MH, Temple R. Impact of cardiovascular outcomes on the development and approval of medications for the treatment of diabetes mellitus. Rev Endocr Metab Disord 2010;11:21-30. 30. Marcinak JF, Viswanathan P, Arora V, et al. Shift from surrogate end point to outcome trials: implications for cardiovascular safety assessment in development programs for antidiabetic drugs. Clin Pharmacol Ther 2012;91:514-20. 31. Lamanna C, Monami M, Bartoli N, et al. Dipeptidyl peptidase-4 inhibitors and cardiovascular events: A protective effect? European Association for the Study of Diabetes 2011 Meeting; September 16, 2011; Lisbon, Portugal. Abstract 244; 2011. 32. DeFronzo R, Stonehouse A, Han J, et al. Relationship of baseline HbA1c and efficacy of current glucose-lowering therapies: a metaanalysis of randomized clinical trials. Diabet Med 2010;27:309-17. 33. 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;34:916-22.

American Heart Journal Volume 166, Number 6

Appendix A. TECOS major inclusion and exclusion criteria Protocol/Amendment Number: 082–02 03 April, 2012 Inclusion criteria. 1. Patient has T2DM with HbA1c of ≥6.5% (48 mmol/mol) and ≤8.0% (64 mmol/mol). HbA1c must be documented within 3 months prior to study enrollment, while receiving: • Metformin, pioglitazone, or a sulfonylurea as monotherapy or any dual combination of metformin, pioglitazone, or a sulfonylurea continuously without alteration in dose for at least 3 months. Note: patients who have received insulin for only a short period (i.e. less than 14 days) during a hospitalization or for the management of acute illness will not be excluded for that reason. OR • A stable dose of insulin (±20% of the scheduled total daily insulin dose) either alone or in combination with a stable dose of metformin for at least 3 months. The use of supplemental/sliding scale insulin during the prior 3 months is permissible, as long as the total daily insulin dose is within ±20% of the scheduled total daily insulin dose. Note: Patients who have required modification of their usual daily insulin dose for a short period (b14 days) during a hospitalization or for the management of acute illness will not be excluded for that reason. 2. Patient is able to see a usual care provider at least twice per year. 3. Patient is ≥50 years of age with pre-existing vascular disease, defined as having any one of the following: • History of a major clinical manifestation of coronary artery disease (i.e., MI, surgical or percutaneous [balloon and/or stent] coronary revascularization procedure, or coronary angiography showing at least one stenosis ≥50% in a major epicardial artery or branch vessel); • Ischemic cerebrovascular disease, including: ◦ History of ischemic stroke. Strokes not known to be hemorrhagic will be allowed as part of this criterion; ◦ History of carotid arterial disease as documented by ≥50% stenosis documented by carotid ultrasound, magnetic resonance imaging (MRI), or angiography, with or without symptoms of neurologic deficit. • Atherosclerotic peripheral arterial disease, as documented by objective evidence such as amputation due to vascular disease, current symptoms of intermittent claudication confirmed by an anklebrachial pressure index or toe brachial pressure index less than 0.9, or history of surgical or percutaneous revascularization procedure.

Green et al 989.e1

4. Female patients agree to use an effective method of contraception or must not otherwise be at risk of becoming pregnant. 5. Patient understands the study procedures, alternative treatments available, and the risks involved with the study, and voluntarily agrees to participate by providing written informed consent. 6. Patient agrees to provide permission to obtain all medical records necessary for complete data ascertainment during the follow-up period. Exclusion criteria. 1. Patient has a history of type 1 diabetes mellitus or a history of ketoacidosis. 2. Patient has a history of ≥2 episodes of severe hypoglycemia during the 12 months prior to enrollment. Severe hypoglycemia (hypoglycemia requiring assistance) refers to instances in which the patient was sufficiently disoriented or incapacitated as to require help from another individual or from medical personnel (whether or not this assistance was actually provided). 3. Patient has taken an approved or investigational DPP-4 inhibitor agent (eg, sitagliptin, alogliptin, saxagliptin, or vildagliptin), GLP-1 analogues (eg, exenatide, exenatide LAR, or liraglutide), or a thiazolidinedione other than pioglitazone within the past 3 months. 4. Patient has cirrhosis of the liver, as assessed by medical history. 5. Patient is enrolled in another experimental protocol which involves the use of an investigational drug or device, or an intervention that would interfere with the conduct of the trial. 6. Patient has a planned or anticipated revascularization procedure. 7. Pregnancy or planned pregnancy during the trial period. 8. Patient has medical history that indicates a life expectancy of b2 years or might limit the individual’s ability to take trial treatments for the duration of the study. 9. Patient has a history or current evidence of any condition, therapy, lab abnormality, or other circumstance which, in the opinion of the investigator or coordinator, might pose a risk to the patient, make participation not in the patient’s best interest, confound the results of the study (eg, if patient cannot comply with requirements of the study), or interfere with the patient’s participation for the full duration of the study. 10. Patient has an estimated GFR (calculated based on serum creatinine via the MDRD formula) of b30 mL/ min per 1.73 m 2. 11. Patient has a known allergy or intolerance to sitagliptin. 12. Patient has previously been enrolled in the trial.

American Heart Journal December 2013

989.e2 Green et al

• • • • • •

Appendix B. Clinical events list Primary or secondary study endpoints A. Obvious study endpoints (These events will prompt the investigator to complete an endpoint package which will then be adjudicated according to the Clinical Events Classification [CEC] charter) Death • Cardiovascular (CV) death (i.e., fatal MI/cerebrovascular accident [CVA]/congestive heart failure [CHF]/arrhythmia, cardiac arrest, death following CV intervention) • Non-CV death Nonfatal MI Unstable angina Nonfatal CVA CHF requiring hospitalization B. Cardiovascular Events of Interest (some of these events will result in prompts to answer additional questions in the eCRF – these questions will be designed to determine whether or not a primary or secondary outcome of interest has occurred) Atrial fibrillation/atrial flutter Ventricular fibrillation/tachycardia requiring intervention Deep vein thrombosis (DVT) Pulmonary embolism Percutaneous Coronary Intervention (PCI) Coronary artery bypass graft (CABG) Coronary catheterization Stress test Abdominal aortic aneurysm/repair Carotid endarterectomy/carotid angioplasty and/or stenting Any hospitalization due to cardiovascular events (i.e., whether or not the hospitalization was for an obvious study endpoint) Shock/hypotension Accelerated or malignant hypertension/hypertensive urgency Transient ischemic attack (TIA) Syncope Renal artery angioplasty and/or stenting Other arterial angioplasty and/or stenting

Expected events and diabetic complications (subcategories indicate potential additional information to be captured, usually as an indication of severity) A. Peripheral vascular disease (PVD) • Limb PCI • Vascular surgery • Amputation • Surgical debridement of ulcer B. Gangrene C. Hypoglycemia/hyperglycemia/diabetic ketoacidosis/hyperosmolar hyperglycemic nonketotic coma D. Diabetic eye disease

E.

F.

G.

H. I.

J.

K.

Photocoagulation or other laser therapy Cataract extraction Blindness Enucleation Steroid/bevacizumab injection Scleral buckling or other retinal fixation procedure Diabetic neuropathy (including distal sensorimotor, focal/multifocal, or autonomic) • Foot ulcer Diabetic nephropathy • Microalbuminuria • Proteinuria Renal failure/peritoneal or hemodialysis/renal transplant (including creation of fistula or other vascular access for hemodialysis) Any hospitalization due to complications of DM Infections • Osteomyelitis ◦ IV antibiotic therapy vs. debridement • Cellulitis ◦ Oral vs. IV antibiotic therapy • Mucormycosis • Pneumonia ◦ Community acquired vs. hospital acquired ◦ Oral vs. IV antibiotic therapy • Bacteremia • Sepsis • Infected joints ◦ Prosthetic joint • Complicated or serious urinary tract infection (UTI)/pyelonephritis ◦ Requiring hospitalization • Malignant external otitis Gastrointestinal (GI) conditions • Abdominal pain • Nausea/vomiting • Diarrhea • Fatty liver disease/Nonalcoholic steatohepatitis (NASH) • Pancreatitis • Cholecystitis/cholelithiasis Metabolic conditions associated with diabetes • Hyperlipidemia/dyslipidemia • Hypertension • Gout

Terms listed in the sitagliptin product circular A. Allergic reactions • Hypersensitivity reactions • Anaphylaxis • Angioedema • Urticaria • Exfoliative skin reaction • Stevens-Johnson syndrome • Rash B. For all other terms please refer to the current local product circular

American Heart Journal Volume 166, Number 6

Appendix C. TECOS committee members Writing group Jennifer B. Green, M. Angelyn Bethel, Sanjoy K. Paul Arne Ring, Keith D. Kaufman, Deborah R. Shapiro, Robert M. Califf, and Rury R. Holman Executive committee Robert M. Califf (Co-Chair), Duke Translational Medicine Institute, Duke University, Durham, NC, USA; Rury R. Holman (Co-Chair), Diabetes Trials Unit, University of Oxford, Oxford, United Kingdom; Paul W. Armstrong, Canadian VIGOUR Centre, University of Alberta, Edmonton, Alberta, Canada; John B. Buse, Diabetes Care Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA; Barry J. Goldstein, Merck Research Laboratories, Merck Sharp & Dohme Corp., Whitehouse Station, NJ, USA; Robert G. Josse, St. Michael's Hospital, University of Toronto, Toronto, ON, Canada; Joerg Koglin, Merck Research Laboratories, Merck Sharp & Dohme Corp., Whitehouse Station, NJ, USA; John M. Lachin, Biostatistics Center, George Washington University, Rockville, MD, USA; Darren K. McGuire, University of Texas-Southwestern Medical Center, Dallas, TX, USA; Deborah R. Shapiro, Merck Research Laboratories, Merck Sharp & Dohme Corp., Whitehouse Station, NJ, USA; Robert Silverman, Merck Sharp & Dohme Corp., Whitehouse Station, NJ, USA; Eberhard Standl, Munich Diabetes Research Institute, Munich, Germany; Frans Van de Werf, University of Leuven, Leuven, Belgium Data and safety monitoring board Marc A. Pfeffer (Chair), Brigham and Women’s Hospital, Boston, MA, USA; Tim Clayton, London School of Hygiene and Tropical Medicine, London, United Kingdom; Hertzel C. Gerstein, McMaster University, Hamilton, ON, Canada; Leif Groop, Lund University, Malmö, Sweden; John J. McMurray, Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, Scotland, United Kingdom; Stuart J. Pocock, London School of Hygiene and Tropical Medicine, London, United Kingdom Clinical events classification committee – Duke Clinical Research Institute, Duke University, Durham, NC, USA Jennifer Green (Chair); Bryan Batch; Jodi Hawes; Paul Hess; Masaya Koshizaka; Valentina Lazzarini; Kenneth Mahaffey; Rob Matthews; Hans Tillman; Adriano Truffa. Coordinators: Christine Henshaw; Diana McFarron Operations committee Argentina: Isaac Sinay, Instituto Cardiovascular de Buenos Aires, Buenos Aires Australia: David Brieger, Concord Hospital, Concord, NSW; Steve Stranks, Repatriation General Hospital, Daw Park, South Australia

Green et al 989.e3

Belgium: Andre Scheen, Centre Hospitalier Universitaire de Liege, Liege Bulgaria: Tsvetalina Tankova, University Specialized Hospital for Active Treatment of Endocrinology, Sofia Brazil: Renato Lopes, Antonio Chacra, Federal University of Sao Paulo, Sao Paulo Canada: Irene Hramiak, St. Joseph’s Health Care, London, Ontario Chile: Carlos Raffo Grado, Centro de Estudios Clinicos V Region CINVEC, Vina del Mar China: Junbo Ge, Shongshan Hospital, Fudan University; Yang Wen Ying, China-Japan Friendship Hospital, Beijing Colombia: Pablo Aschner, Javeriana University School of Medicine, Bogota Czech Republic: Jan Skrha, Charles University, Prague Estonia: Anu Ambos, North Estonia Medical Center, Tallinn Finland: Timo Strandberg, Geri-Med Ltd., Helsinki France: Michel Marre, Hôpital Bichat, Paris Germany: Markof Hanefeld, GWT-TUD GmbH, Dresden; Axel Riefflin, Gemeinschaftspraxis Husby, Husby Great Britain: Amanda Adler, Addenbrooke’s Hospital, Cambridge Hong Kong: Juliana Chan, The Chinese University of Hong Kong, Shatin Hungary: Peter Ofner, Gottsegen Gyorgy Hungarian Institute of Cardiology, Budapest India: Sanjay Basumatary, Escorts Heart Institute & Research Center, New Delhi; Mukund Karmalkar, CARE Hospital, Hyderabad; Mukul Manchanda, Indraprastha Apollo Hospital, New Delhi: Sanjay Mittal, the Medicity, Guragon Israel: Julio Wainstein, The E. Wolfson Medical Center Italy: Giuseppe Ambrosio, University of Perugia, Perugia Lithuania: Neli Jakuboniene, Kaunas Medical Hospital, Kaunas Latvia: Valdis Pirags, Universitates Slimnica, Riga Malaysia: Mafauzy Mohamed, Hospital Universiti Sains, Kelantan Netherlands: Jan H. Cornel, Medical Center Alkmaar, Alkmaar Norway: Sigrun Halvorsen, Oslo University Hospital, Oslo New Zealand: Russell Scott, Christchurch Hospital, Christchurch; Harvey White, Auckland City Hospital, Auckland Poland: Andrzej Tykarski, Szpitali Kliniczny Przemienienia Panskiego Uniwersytetu Medycznego, Poznan Romania: Ioan Andrei Veresiu, “Iuliu Hatieganu” University of Medicine, Cluj Napoca Russia: Alexander V. Dreval, Moscow Regional Scientific Research Clinical Institute, Moscow Singapore: E Shyong Tai, National University Hospital Slovakia: Boris Krahulec, Comenius University, Bratislava

989.e4 Green et al

South Africa: Larry Distiller, Centre for Diabetes and Endocrinology, Houghton South Korea: Park Yong Soo, Hanyang University Guri Hospital, Gyeonggi-do, Seoul Spain: Adela Rovira, Fundación Jimenez Diaz – UTE, Madrid Sweden: Michael Alvarsson, Karolinska Institute, Stockholm Taiwan: Lee-Ming Chuang, National Taiwan University Hospital, Taipei Turkey: Tuncay Delibasi, Ankara Diskapi Training and Research Hospital, Ankara Ukraine: Alexander Parkhomenko, Natsionalnyi Naukovyi Tsentr Instytut Cardiologiyi, Kyiv United States: Helena Rodbard, Endocrine and Metabolic Consultants, Rockville; David Goff, Wake Forest University School of Medicine, Winston-Salem

Appendix D. Endpoint definitions Stroke Stroke is defined as a new, sudden, focal neurological deficit resulting from a presumed cerebrovascular cause that is not reversible within 24 hours or results in death and is not due to a readily identifiable cause (eg, tumor or trauma). The CEC will consider all clinically relevant information and imaging studies to classify all strokes as: • Primary hemorrhagic: Stroke with focal collections of intraparenchymal blood. Events of subarachnoid, subdural, and intraventricular hemorrhage will be recorded, but these events will not be considered part of the primary endpoint. • Primary ischemic infarction: Stroke without focal collections of intracranial blood. The occurrence of hemorrhagic conversion of a primary ischemic infarction will be recorded including whether it was symptomatic or asymptomatic. Stroke subtype will be assessed as cardioembolic, noncardioembolic (eg, atherothrombotic, lacunar, other known cause) and uncertain. • Uncertain: No imaging or autopsy data available

Myocardial infarction* The adjudication of MI as a clinical endpoint will consider the occurrence relative to a PCI or coronary artery bypass graft surgery. (It will also be consistent with the revised ACC/ESC MI definition document.) In the absence of a PCI or CABG, MI is defined as: Clinical symptoms consistent with MI and cardiac biomarker elevation (Troponin I or T or creatine kinase- muscle brain subunit [CK-MB]) greater than the site’s ULN or development of new pathological Q

American Heart Journal December 2013

waves (N0.04 seconds) in at least 2 contiguous leads in the absence of LBBB on the electrocardiogram or autopsy confirmation. For subjects having a PCI, a MI will be defined as: CK-MB (or troponin I/T or CK in the absence of CKMB) N3 × ULN for samples obtained within 24 hours of the procedure if the baseline values were normal or at least a 50% increase over elevated baseline values that were stable or decreasing or development of new pathological Q waves (N0.04 seconds) in at least 2 contiguous leads in the absence of LBBB on the electrocardiogram. Symptoms of cardiac ischemia are not required. After coronary artery bypass graft surgery, a MI is defined as either: CK-MB (or CK in the absence of CK-MB) N5 × ULN for samples obtained within 24 hours of the procedure with development of new pathological Q waves in at least 2 contiguous leads in the absence of LBBB on the electrocardiogram OR CK-MB (or CK in the absence of CK-MB) N10 x ULN for samples obtained within 24 hours of the procedure with or without development of new pathological Q waves in at least 2 contiguous leads in the absence of LBBB on the electrocardiogram. Note: For institutions that report an ‘intermediate’ or ‘equivocal’ range for biomarker elevation that is not definitively associated with myocardial necrosis or infarction then the ULN should be the lower value for the necrosis or infarct range, not the ‘equivocal’ or ‘intermediate’ range lower value.

Unstable angina requiring hospitalization* 1. Symptoms of myocardial ischemia at rest (chest pain or equivalent) or an accelerating pattern of angina with frequent episodes associated with progressively decreased exercise capacity. AND 2. Prompting an unscheduled hospitalization within 24 hours of the most recent symptoms. Hospitalization is defined as an admission to an inpatient unit or a visit to an emergency department that results in at least a 12 hour stay (or a date change if the time of admission/ discharge is not available). AND 3. At least one of the following: a. New or worsening ST or T wave changes on resting ECG (in the absence of confounders, such as LBBB or LVH) • ST elevation New ST elevation at the J point in two anatomically contiguous leads with the cut-off points: ≥0.2 mV in men (N0.25 mV in men b40 years) or ≥0.15 mV

American Heart Journal Volume 166, Number 6

in women in leads V2-V3 and/or ≥0.1 mV in other leads. • ST depression and T-wave changes New horizontal or down-sloping ST depression ≥0.05 mV in two contiguous leads, and/or new T inversion ≥0.3 mV in two contiguous leads. b. Definite evidence of inducible myocardial ischemia as demonstrated by: • an early positive exercise stress test, defined as ST elevation or ≥2 mm ST depression prior to 5 mets OR • stress echocardiography (reversible wall motion abnormality) OR • myocardial scintigraphy (reversible perfusion defect), OR • MRI (myocardial perfusion deficit under pharmacologic stress). c. Angiographic evidence of new or worsened ≥70% lesion and/or thrombus in an epicardial coronary artery that is believed to be responsible for the myocardial ischemic symptoms/signs. d. Need for coronary revascularization procedure (PCI or CABG) for the presumed culprit lesion(s). This criterion would be fulfilled if revascularization was undertaken during the unscheduled hospitalization, or subsequent to transfer to another institution without interceding home discharge. AND 4. Negative cardiac biomarkers and no evidence of acute MI General considerations. 1. Escalation of pharmacotherapy for ischemia, such as intravenous nitrates or increasing dosages of βblockers, should be considered supportive of the diagnosis of unstable angina. However, a typical presentation and admission to the hospital with escalation of pharmacotherapy, without any of the additional findings listed under category 3, would be insufficient to support classification as hospitalization for unstable angina. 2. If subjects are admitted with suspected unstable angina, and subsequent testing reveals a non-cardiac or non-ischemic etiology, this event should not be recorded as hospitalization for unstable angina. Potential ischemic events meeting the criteria for MI should not be adjudicated as unstable angina. 3. Planned hospitalization or rehospitalization for performance of an elective revascularization in patients who do not fulfill the criteria for unstable angina should not be considered a hospitalization for unstable angina. For example, • Hospitalization of a patient with stable exertional angina for coronary angiography and PCI that is prompted by a positive outpatient stress test should not be considered hospitalization for unstable angina.

Green et al 989.e5

• Rehospitalization of a patient meeting the criteria for unstable angina who was stabilized, discharged, and subsequently readmitted for revascularization, does not constitute a second hospitalization for unstable angina. 4. A patient who undergoes an elective catheterization where incidental coronary artery disease is found and who subsequently undergoes coronary revascularization will not be considered as meeting the hospitalization for unstable angina end point.

Death All deaths will be considered cardiovascular unless an unequivocal non-cardiovascular cause of death can be established. Sudden cardiac death (includes death due to arrhythmia). This refers to death that occurs unexpectedly in a previously stable patient and will include the following deaths: i. Witnessed and instantaneous without new or worsening symptoms and also in the absence of progressive circulatory failure, the latter lasting for 60 minutes or more. ii. Witnessed within 60 minutes of the onset of new or worsening symptoms unless a cause other than cardiac is obvious. iii. Witnessed and attributed to an identified arrhythmia (eg, captured on an ECG recording or witnessed on a monitor by either a medic or paramedic). iv. Patients resuscitated from cardiac arrest in the absence of pre-existing circulatory failure or other causes of death, including MI, and who die within 24 hours or without gaining consciousness; similar patients who die during an attempted resuscitation. v. Unwitnessed death in the absence of pre-existing progressive circulatory failure or other causes of death (information regarding the patient’s clinical status within the week preceding death should be present or the “presumed CV death” classification should be used). Death due to acute MI. Death occurring up to 7 days after a documented acute MI (verified either by the diagnostic criteria outlined above for acute MI or by autopsy findings showing recent MI or recent coronary thrombus) and where there is no conclusive evidence of another cause of death. NOTE: If death occurs before biochemical confirmation of myocardial necrosis can be obtained, the CECC will adjudicate based on clinical presentation and ECG evidence. Death due to a MI that occurs as a direct consequence of a cardiovascular investigation/procedure/operation will be classified as death due to other cardiovascular

989.e6 Green et al

cause (see definition for death due to other cardiovascular cause, below). Death due to heart failure. Death occurring in the context of clinically worsening symptoms and/or signs of heart failure without evidence of another cause of death: Any of the following: i. New or increasing symptoms and/or signs of heart failure requiring the initiation of, or an increase in, treatment directed at heart failure or occurring in a patient already receiving maximal therapy for heart failure. ii. Heart failure symptoms or signs requiring continuous intravenous therapy or oxygen administration. iii. Confinement to bed but only if this is due entirely to heart failure symptoms. iv. Pulmonary edema sufficient to cause tachypnea and distress not occurring in the context of an acute MI or as the consequence of an arrhythmia occurring in the absence of worsening heart failure. v. Cardiogenic shock (defined as hypotension resulting in failure to maintain normal renal or cerebral function for N60 minutes prior to death) not occurring in the context of an acute MI or as the consequence of an arrhythmia occurring in the absence of worsening heart failure. This category will include sudden death occurring during an admission for worsening heart failure. Death due to other cardiovascular causes. Death must be due to a fully documented cardiovascular cause not included above (eg, ruptured aortic aneurysm, pulmonary embolism, or cardiovascular intervention). As guidance: • Within 30 days of CV surgery • Within 7 days of a Catheterization, Ablation, CABG, or PTCA Death due to stroke. Death occurring within 30 days of a confirmed stroke Presumed cardiac death. All deaths not attributed to the above categories of cardiovascular death and not attributed to a non-cardiovascular cause. Non-cardiovascular deaths Non-cardiovascular death. A death will be considered noncardiovascular only if an unequivocal and documented non-cardiovascular cause can be established. This category includes deaths related to non-cardiovascular procedures. As guidance: • Within 30 days of Non-CV procedures

Congestive heart failure (CHF) All episodes of suspected CHF requiring hospitalization will be reported by the site investigator on the eCRF. CHF events that meet the following criteria will be study endpoints

American Heart Journal December 2013

1. Requires hospitalization defined as an admission to an inpatient unit or a visit to an emergency department that result in at least a 12 hour stay (or a date change if the time of admission/discharge is not available). AND 2. Clinical manifestation of CHF including at least one of the following: New or worsening: dyspnea, orthopnea, paroxysmal nocturnal dyspnea, edema, pulmonary basilar crackles, jugular venous distension, or radiological evidence of worsening heart failure. AND 3. Additional/increased therapy a. Intravenous treatment with diuretic, inotrope, or vasodilator therapy OR b. Mechanical or surgical intervention (mechanical circulatory support, heart transplantation or ventricular pacing to improve cardiac function,) or the use of ultrafiltration, hemofiltration or dialysis that is specifically directed at treatment of heart failure. Congestive Heart Failure (CHF) events will be reviewed using the following approach: (Please note that validation of the algorithm will be done by a clinician (CEC coordinator or CEC MD) 1. Data will be collected on the eCRF based on the endpoint definition 2. A Computer Algorithm will be created to assess the data on the eCRF 3. The algorithm will be a series of logic statements evaluating the eCRF data to determine if the data provided is consistent and supports the endpoint criteria. 4. If the data are consistent and data are complete on the eCRF then the algorithm will determine if an endpoint CHF event did or did not occur. If the data are non consistent or data are not complete despite query process, then the algorithm will not make a determination and the event will be reviewed by one CEC MD and a determination made about whether an endpoint CHF event did or did not occur. 5. The algorithm will be validated by a clinician. The clinician will review the algorithmic determinations about whether a CHF event did or did not occur for those that a decision is made for the first 25 events. 6. If there are discrepancies found in the algorithm the CEC coordinator will note these to the data manager and have all CHF events reviewed manually until the discrepancy is resolved. 7. A clinician will validate the algorithm on the next 10 CHF events after the discrepancy has been resolved to ensure that it supports the endpoint criteria.

Acute pancreatitis Reported events of acute pancreatitis will be adjudicated based on clinical symptoms, serologic markers and radiographic findings. Confirmed cases of acute pancreatitis

American Heart Journal Volume 166, Number 6

will be further categorized as mild or severe based on evidence of organ failure or radiographic criteria. To confirm a diagnosis of acute pancreatitis, the following criteria must be present: i. Symptoms of abdominal pain OR vomiting AND ii. Objective evidence of pancreatic inflammation 1. Elevated pancreatic enzymes, defined by a. Amylase OR Lipase N3× the upper limit normal. b. In patients with chronic pancreatitis, enzyme elevations N2× the upper limit normal OR 2. Evidence of pancreatitis documented by imaging a. Abdominal CT, MRI or ultrasound showing focal, diffuse and inhomogeneous gland enlargement Confirmed acute pancreatitis will be graded as mild, severe or unknown. To confirm severe pancreatitis, the following criteria must be present:

Green et al 989.e7

1. Evidence of organ failure (at least one of the following) a. Shock: systolic BP b90 mm Hg b. Pulmonary insufficiency: Pa02 b60 mm Hg c. Renal failure: serum creatinine N2 mg/dL after rehydration d. Gastrointestinal bleeding: N500 mL/24 hours OR 2. Local complications demonstrated on abdominal CT, MRI, or ultrasound (at least one of the following) a. Pancreatic necrosis (note: hemorrhagic pancreatitis is a pathologic term often used synonymously with pancreatic necrosis) b. Pancreatic abscess c. (Acute) Pancreatic pseudocyst Confirmed cases of acute pancreatitis not meeting criteria for severe pancreatitis will be classified as mild pancreatitis. *Definitions of Myocardial Infarction and Unstable Angina Requiring Hospitalization may change over time in response to evolving standards.