738 Diabetes Care Volume 39, May 2016

Robert J. Smith,1 Allison B. Goldfine,2 and Evaluating the Cardiovascular William R. Hiatt3 Safety of New Medications for CARDIOVASCULAR DISEASE AND DIABETES Type 2 Diabetes: Time to Reassess? Diabetes Care 2016;39:738–742 | DOI: 10.2337/dc15-2237

The U.S. Food and Drug Administration (FDA) issued a Guidance for Industry in 2008 defining preapproval and postapproval requirements for the demonstra- tion of cardiovascular safety for all new medications developed for glycemic management in type 2 diabetes. Seventeen large, prospective, randomized, controlled clinical trials involving more than 140,000 subjects thus far have been completed or are ongoing in accordance with this guidance. All five of the completed trials, involving three different drug classes, have met their primary objective to exclude an unacceptable level of ischemic cardiovascular risk as defined in the FDA guidance. Additionally, one trial found an increased risk of hospitalization for heart failure, and another demonstrated decreases in cardio- vascular mortality and hospitalization for heart failure. Given that a heightened risk of cardiovascular ischemic events has not been demonstrated across several classes of new diabetes drugs, we believe it is time for the scientificcommunityand the FDA to consider a more targeted approach to what is, in effect, a global cardiovascular safety trial requirement for all new type 2 diabetes medications in development.

Multiple new medications for glycemic management of type 2 diabetes have re- cently been approved by the U.S. Food and Drug Administration (FDA), with addi- tional drugs in development. In the regulatory process, all agents for treatment of hyperglycemia in diabetes are evaluated for efficacy in lowering blood glucose levels 1Department of Medicine, Alpert Medical School and overall safety and are subjected to particularly rigorous screening for cardio- and Department of Health Services, Policy & vascular safety. The approach to these cardiovascular studies is largely directed by a Practice, School of Public Health, Brown Univer- sity, and Ocean State Research Institute, Provi- Guidance for Industry issued by the FDA in 2008 stating that all new type 2 diabetes dence VA Medical Center, Providence, RI drug development programs should rule out unacceptable cardiovascular risk by 2Research Division, Joslin Diabetes Center, demonstrating an upper bound of the two-sided 95% CI of the risk ratio ,1.8 Harvard Medical School, Boston, MA preapproval for a composite end point of major adverse cardiac events (MACE), 3Division of Cardiology, University of Colorado consisting of at least cardiovascular death, nonfatal myocardial infarction, and School of Medicine, and CPC Clinical Research, Aurora, CO stroke (1). For drugs in which the upper bound of the CI is between 1.3 and 1.8, the guidance states that subsequent postapproval trials of the same composite Corresponding author: Robert J. Smith, robert_ [email protected]. MACE end point should be conducted to continue marketing. These recommendations Received 13 October 2015 and accepted 23 were motivated by the high prevalence of cardiovascular disease in diabetes (accounting February 2016. for approximately 70% of deaths) and, in part, by concerns about a signal of potentially © 2016 by the American Diabetes Association. increased cardiovascular risk for the thiazolidinedione rosiglitazone (2) occurring on a Readers may use this article as long as the work is background of additional cardiac safety concerns with antihyperglycemic medications properly cited, the use is educational and not for for type 2 diabetes. The latter included evidence for increased deaths and major profit, and the work is not altered. cardiovascular events during the development program of the peroxisome proliferator– See accompanying articles, pp. 664, 668, activated receptor (PPAR) agonist muraglitazar (3), persistent uncertainty about cardio- 677, 686, 694, 701, 709, 717, 726, and vascular safety with sulfonylureas (4), increased mortality with intense glucose control 735. care.diabetesjournals.org Smith, Goldfine, and Hiatt 739

in the Action to Control Cardiovascular Sitagliptin [TECOS] trial) (10), the sodium– In addition to concerns about ischemic Risk in Diabetes (ACCORD) trial (5), and glucose cotransporter 2 (SGLT2) inhibitor cardiovascular events, as addressed by increased risk for congestive heart failure empagliflozin (BI 10773 [Empagliflozin] the recommendations for a composite with pioglitazone (6) and rosiglitazone (7). Cardiovascular Outcome Event Trial in MACE in the FDA guidance, there is an Seventeen large, randomized, controlled, Type 2 Diabetes Mellitus Patients increased risk of heart failure in patients multicenter postapproval trials involving [EMPA-REG OUTCOME] study) (11), and with diabetes (13). Increased heart failure more than 140,000 participants thus far the glucagon-like peptide 1 (GLP-1) recep- risk has previously been demonstrated have been initiated in accordance with tor agonist lixisenatide (Evaluation of Lix- with thiazolidinediones (6,7) but was not the FDA guidance (Tables 1 and 2). Results isenatide in Acute Coronary Syndrome apparent during the initial develop- from five of these trials now have been [ELIXA] study) (12). The sitagliptin and lix- ment of any of the agents for which reported, providing an opportunity to ex- isenatide studies demonstrated a hazard postmarketing safety trials were subse- amine the utility of this approach to eval- ratio for a composite MACE end point ap- quently performed under the 2008 FDA uate the safety of new antihyperglycemic proximating unity and an upper bound of guidance. Unexpectedly, exploratory medications for type 2 diabetes. the 95% CI ,1.30 for both drugs, while analysis of hospitalization for heart failure The FDA Endocrinologic and Metabolic the empagliflozin results provided evi- as an individual clinical end point revealed Drugs Advisory Committee was convened dence for cardiovascular benefit with a an elevated hazard ratio for saxagliptin at in April 2015 to review the results of hazard ratio of 0.86 and upper bound of 1.27 (95% CI 1.07–1.51) and a nonsignifi- postmarketing trials of the dipeptidyl the 95% CI 0.99 (Table 1). Thus, all five cant numerical increase in risk for peptidase 4 (DPP-4) inhibitors saxagliptin trials involving medications from three alogliptin with a hazard ratio of 1.19 (Saxagliptin Assessment of Vascular Out- different drug classes have excluded an (95% CI 0.90–1.58). The hazard ratios for comes Recorded in Patients with Diabetes unacceptable level of ischemic cardiovas- hospitalization for heart failure as a Mellitus–Thrombolysis in Myocardial In- cular risk (which was the primary analysis) prespecified, adjudicated individual end farction 53 [SAVOR-TIMI 53] trial) and as defined in the FDA 2008 guidance point approximated unity in the sitagliptin alogliptin (Examination of Cardiovascular document. and lixisenatide trials (Table 1), and in- Outcomes with Alogliptin versus Standard All-cause mortality, another important, creased risk for an expanded MACE end of Care [EXAMINE] trial) in type 2 diabe- prespecified secondary end point, raised point that included hospitalization for tes. The study data reviewed at those focused concerns by demonstrating in- heart failure was not evident with any of meetings demonstrated hazard ratios consistent findings across the DPP-4 in- these incretin-modulating drugs (10,12). for a composite MACE end point (cardio- hibitor trials. For saxagliptin, all-cause By contrast, the empagliflozin trial vascular death, nonfatal myocardial in- mortality in the intent-to-treat analysis showed a decreased hazard ratio for hos- farction, and nonfatal stroke) of 1.00 for population showed a hazard ratio of pitalization for heart failure as an individual saxagliptin and 0.96 for alogliptin. The up- 1.11 (95% CI 0.96–1.27), which increased end point at 0.65 (95% CI 0.50–0.85) and per bound of the 95% CI for the two drugs to 1.23 (95% CI 1.02–1.48) when evalu- as a component of an expanded MACE was 1.12 and 1.16, respectively, which ated in the on-treatment population. It is composite outcome combined with CV was below a threshold for concern based important to note that, consistent with death but excluding stroke at 0.66 (95% on the FDA guidance (Table 1) (8,9). The the lack of increased risk of MACE, the CI 0.55–0.79) (11). Thus, hospitalization advisory committee vote on whether ac- increased mortality risk was primarily for heart failure as a stand-alone explor- ceptable cardiovascular risk profiles were driven by noncardiovascular deaths, atory end point demonstrated inconsistent demonstrated was 13 “yes,” 1 “no,” and 1 with no specific identified causal con- results for the DPP-4 inhibitors, with a po- abstention for saxagliptin and was unan- cerns. By contrast, all-cause mortality tential concern observed for saxagliptin, a imous at 16 “yes” for alogliptin (note that was not numerically increased with less evident trend for alogliptin, and no there was not complete overlap in mem- alogliptin, sitagliptin, or lixisenatide, and numeric or statistical difference for bers attending each meeting). Results of all-cause mortality was significantly de- sitagliptin. Hospitalization for heart failure three additional cardiovascular outcome creased with empagliflozin (Table 1). It is risk was either neutral or decreased with trials conducted under the FDA guidance not possible to directly compare agents as the other two agents studied. On the basis subsequently have been published on these trials had different inclusion criteria of the findings in the SAVOR-TIMI 53 and the DPP-4 inhibitor sitagliptin (Trial Eval- and participants were not randomized EXAMINE trials, the FDA Endocrinologic uating Cardiovascular Outcomes with across the various medications. and Metabolic Drugs Advisory Committee

Table 1—Completed cardiovascular outcome trials for new antihyperglycemic medications in type 2 diabetes Trial (drug), n of subjects MACE* Hospitalization for heart failure All-cause mortality SAVOR-TIMI 53 (saxagliptin), n = 16,492 1.00 (0.89–1.12) 1.27 (1.07–1.51) 1.11 (0.96–1.27) EXAMINE (alogliptin), n = 5,380 0.96 (0.80–1.16) 1.19 (0.90–1.58) 0.88 (0.71–1.09) TECOS (sitagliptin), n = 14,671** 0.98 (0.88–1.09) 1.00 (0.83–1.20) 1.01 (0.90–1.14) EMPA-REG OUTCOME (empagliflozin), n = 7,020 0.86 (0.74–0.99) 0.65 (0.50–0.80) 0.68 (0.57–0.82) ELIXA (lixisenatide), n = 6,068 1.02 (0.89–1.17) 0.96 (0.82–1.16) 0.94 (0.78–1.13) Data are hazard ratio (95% CI). *MACE components: SAVOR-TIMI 53, EXAMINE, and EMPA-REG OUTCOME trials: cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke; TECOS and ELIXA trials: cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, and hospitalization for unstable angina. **The TECOS trial was designed and initiated prior to the 2008 FDA guidance but is compliant with the guidance. 740 Cardiovascular Safety of Type 2 Diabetes Drugs Diabetes Care Volume 39, May 2016

Table 2—In-progress cardiovascular outcome trials for new antihyperglycemic medications in type 2 diabetes Trial (NCT number) Medication Planned enrollment Planned completion date LEADER (NCT01179048) Liraglutide 9,340 November 2015 SUSTAIN 6 (NCT01720446) Semaglutide 3,297 January 2016 CANVAS (NCT01032629) Canagliflozin 4,407 June 2017 CARMELINA (NCT01897532) Linagliptin 8,300 January 2018 EXSCEL (NCT01144338) Exenatide 14,000 January 2018 ITCA 650 (NCT01455896) Exenatide 4,000 July 2018 CAROLINA (NCT01243424) Linagliptin 6,000 September 2018 DECLARE-TIMI 58 (NCT01730534) Dapagliflozin 17,150 April 2019 REWIND (NCT01394952) Dulaglutide 9,622 April 2019 HARMONY Outcomes (NCT02465515) Albiglutide 9,400 May 2019 Cardiovascular Outcomes Following Treatment With Ertugliflozin in Participants With Type 2 Diabetes Mellitus and Established Vascular Disease (MK-8835-004) (NCT01986881) Ertugliflozin 3,900 October 2020 A Study to Assess Cardiovascular Outcomes Following Treatment With Omarigliptin (MK-3102) in Participants With Type 2 Diabetes Mellitus (MK-3102-018) (NCT01703208) Omarigliptin 4,202 December 2020 CANVAS, CANagliflozin cardioVascular Assessment Study; CARMELINA, Cardiovascular and Renal Microvascular Outcome Study With Linagliptin in Patients With Type 2 Diabetes Mellitus; CAROLINA, Cardiovascular Outcome Study of Linagliptin Versus Glimepiride in Patients With Type 2 Diabetes; DECLARE-TIMI 58, Multicenter Trial to Evaluate the Effect of Dapagliflozin on the Incidence of Cardiovascular Events; EXSCEL, Exenatide Study of Cardiovascular Event Lowering Trial; HARMONY Outcomes, Effect of Albiglutide, When Added to Standard Blood Glucose Lowering Therapies, on Major Cardiovascular Events in Subjects With Type 2 Diabetes Mellitus; ITCA 650, A Study to Evaluate Cardiovascular Outcomes in Patients With Type 2 Diabetes Treated With ITCA 650; NCT, National Clinical Trial; REWIND, Researching Cardiovascular Events With a Weekly Incretin in Diabetes; SUSTAIN 6, Trial to Evaluate Cardiovascular and Other Long-term Outcomes With Semaglutide in Subjects With Type 2 Diabetes.

recommended a label warning regarding glycemic differences in the placebo and cardiovascular safety evaluation but not theriskofheartfailureandfurtherheart usual care groups, cardiovascular bene- for assessment of cardiovascular out- failure safety monitoring for saxagliptin fit was only demonstrated for the come benefit, and the FDA does not oth- and alogliptin. It is important to note that SGLT2 inhibitor empagliflozin (11). erwise require new diabetes drugs with the evidence for increased heart failure This represents the first large-scale, the intended indication of improving risk with these two pharmacological prospective, randomized, controlled glycemic control to also show cardiovas- agents requires confirmation. Addition- cardiovascular outcome trial of any cular benefit. The current FDA guidance ally, as none of the recently reported trials antihyperglycemic medication for to exclude cardiovascular risk of new dia- conducted under the 2008 FDA guidance type 2 diabetes showing cardiovascular betes medications therefore leads to a focused on a population with a high risk benefit, although smaller studies have trial design that focuses on the risk of thatwouldgenerateanadequateevent seen benefit with other agents. The the drug per se and allows for use and rate for heart failure, the capacity of these mechanism and consistency of the ben- dose adjustments of multiple diabetes completed studies to detect heart failure efit across the SGLT2 inhibitor class re- drugs other than the specific agent under risk with certainty may be limited. mains to be defined and cannot be fully study. The cardiovascular outcome stud- All of these cardiovascular outcome established based on the results of a sin- ies occur on a backdrop of established trials (including the empagliflozin study) gle trial where the primary intent was to drug efficacy for glycemic control, and were primarily designed to exclude demonstrate exclusion of cardiovascular these safety trials thus are not designed unacceptable risk of cardiovascular risk. A recent press release for the to examine either glycemic efficacy or ischemic events by first testing for LEADER (Liraglutide Effect and Action in glycemic effects as potential components noninferiority relative to placebo on the Diabetes: Evaluation of Cardiovascular of drug superiority in lowering cardiovas- primary MACE or expanded MACE end Outcome Results) trial investigating the cular risk. Furthermore, to accrue suffi- point in accordance with the FDA guid- cardiovascular safety of the GLP-1 recep- cient events in a timely manner, these ance. Additionally, the analysis plans tor agonist liraglutide reports reduced trials generally enroll patients with more permitted secondary testing for supe- cardiovascular risk assessed by the com- advanced atherosclerotic cardiovascular riority over placebo if the noninferior- posite outcome of the first occurrence of risk or established cardiovascular disease. ity threshold was achieved, providing an cardiovascular death, nonfatal myocar- The typical trial observation periods are opportunity to show potential cardiovas- dial infarction, or nonfatal stroke (14). most effective for assessing short-term cular benefit. In the context of the trial However, the data from this study cardiovascular risk but do not permit as- designs, which included mean follow-up have not yet been made available. The sessment of long-term risks or benefits duration #3 years and modest or no 2008 FDA guidance defines a need for for these comorbid conditions. care.diabetesjournals.org Smith, Goldfine, and Hiatt 741

In reviewing the saxagliptin and financial costs that add directly or indi- cardiovascular safety trials for new diabe- alogliptin trials, multiple FDA advisory rectly to the overall health care budget tes medications. We support the over- committee members questioned the wis- in the U.S. and potentially divert funding all FDA process for evaluating new dom of continuing to impose the burden away from new research efforts. antihyperglycemic drugs, which requires of cardiovascular outcome safety trials on The argument to continue the broad adequate data to demonstrate efficacy in all new diabetes medications in order to requirement to perform large cardiovas- lowering blood glucose and to provide achieve full market approval. They noted cular outcome trials for all new diabetes reasonable assurance of overall safety that concern about the potential for ex- drugs to simply extend the patient-years prior to initial drug approval. For medi- cessive cardiovascular risk with diabetes of exposure beyond the phase 3 trial re- cations that achieve approval, the FDA therapies has been somewhat mitigated quirement for approval to identify all un- is empowered to mandate additional by the readjudication of the original signal foreseen safety concerns could be made targeted postapproval safety studies to from the rosiglitazone cardiovascular out- for any drug class. In the case of diabetes address concerning specificsafetysignals come trial, which demonstrated that drugs, the AleCardio trial was a study that are identified either during drug de- there was no imbalance in MACE events enrolling 7,226 patients with a recent velopment or postmarketing. This pro- despite earlier doubts raised from meta- acute coronary syndrome event to de- cess has been effective in identifying analyses of the phase 2 trial data (2). It is termine if the dual activator of PPAR-a and evaluating cardiovascular safety now apparent that the concern about in- and PPAR-g aleglitazar would reduce the concerns and also other safety issues, as creased ischemic cardiovascular disease risk of cardiovascular events (15). The exemplified by the recent required inves- risk with diabetes medications, which trial was stopped early due to futility tigation of potential increased risks of has been broadly applied by the FDA to (lack of superiority), and, at that time, pancreatitis and pancreatic cancer with all new antihyperglycemic type 2 diabetes several new safety concerns also were DPP-4 inhibitors and GLP-1 receptor ago- drugs in development regardless of drug identified. The latter included adverse nists. We question whether a continued class or prior signals of cardiovascular risk, events previously associated with other recommendation for large-scale cardiovas- has not been substantiated in large-scale PPAR-a or PPAR-g activators (heart fail- cularsafetytrialstoruleoutariskofische- trials on three DPP-4 inhibitors (saxagliptin, ure, decreased glomerular filtration, and mic events for all new antihyperglycemic alogliptin, and sitagliptin), the GLP-1 recep- bone fractures), plus evidence for in- medications, as detailed in the FDA 2008 tor agonist lixisenatide, and the SGLT2 in- creased risk of gastrointestinal bleeding, Guidance for Industry, is cost-effective and hibitor empagliflozin. This new information which was unanticipated. Of note, there justified.Webelieveitwouldbemoreef- should be incorporated into the totality of was a numerical increase in the risk of fective to conduct such studies only when evidence to assess the impact and unin- hospitalization from heart failure with a concerning safety data in the preapproval tended consequences of the 2008 FDA hazard ratio of 1.22, but this was not con- package, plausible mechanisms of risk, or a guidance being applied broadly to diabetes firmed to be significant at the time the known class effect support a large invest- drug development. study was terminated (95% CI 0.94–1.59) ment in a cardiovascular outcome trial. If The history leading up to the FDA and thus remains unresolved. The identi- the preapproval package or postapproval 2008 guidance, combined with the fication of gastrointestinal bleeding in the monitoring provides a signal of other car- new evidence on the DPP-4 inhibitor, AleCardio study as a risk not previously diac risks, such as heart failure, then trials GLP-1 receptor agonist, and SGLT2 in- associated with agents sharing a similar specifically designed to evaluate those hibitor drug classes, raises the issue of mechanism of action is of interest from a risks should be performed. Thus, we how to best assess all aspects of cardiac general safety consideration. However, support a more informed approach to safety in the development of new dia- the current FDA recommendation to per- safety trials that would direct a more in- betes medications. The additional 12 form cardiovascular safety trials for all dividualized strategy for addressing con- ongoing cardiovascular outcome studies new diabetes drugs is specifically intended cerns about the cardiac safety of new include three with DPP-4 inhibitors, six to identify specific cardiovascular risks, not diabetes medications. with GLP-1 receptor agonists, and three all unanticipated safety concerns. The FDA with SGLT2 inhibitors and in total in- has a variety of postmarketing surveillance volve more than 90,000 additional mechanisms to identify safety concerns patients (Table 2). The results of these that are appropriate in the context of Duality of Interest. R.J.S. has served on the FDA trials will undoubtedly expand the any drug requiring broadsafetyoversight. Endocrinologic and Metabolic Drugs Advisory Committee since 2011, has been chair of the knowledge of the cardiovascular and The 2008 FDA guidance on the evalu- committee since 2014, and chaired the 14 April heart failure risks of new diabetes drugs ation of cardiovascular safety in the de- 2015 advisory meeting on the SAVOR-TIMI 53 and will likely reveal a better under- velopment of new antihyperglycemic and EXAMINE trials. A.B.G. served on the FDA standing of the potential additional medications for type 2 diabetes was Endocrinologic and Metabolic Drugs Advisory safety concerns. However, we question thoughtfully designed based on the ev- Committee from 2007 to 2012 and previously served on a scientific advisory board for Novo whether continued broad application of idence available at that time. However, Nordisk with no ongoing obligation. She pro- the current FDA guidance on cardiovas- we believe the scientific community vided an invited, independent commentary on cular safety to all new diabetes drugs, and the FDA now should reconsider the the TECOS trial at the 2015 American Diabetes including additional members of al- totality of the current evidence and en- Association Scientific Sessions. W.R.H. has fi served on FDA advisory committees since 2003, ready studied drug classes, is justi ed. gage in a discussion on safety assessment has been an active member of the FDA Endo- Each new large-scale cardiovascular and monitoring strategies that would crinologic and Metabolic Drugs Advisory Com- outcome trial is accompanied by high represent a more targeted approach to mittee since 2013, and participated in the 14 742 Cardiovascular Safety of Type 2 Diabetes Drugs Diabetes Care Volume 39, May 2016

April 2015 advisory meeting on the SAVOR- 5. Gerstein HC, Miller ME, Byington RP, et al.; 11. Zinman B, Wanner C, Lachin JM, et al.; TIMI 53 and EXAMINE trials. No other potential Action to Control Cardiovascular Risk in Diabe- EMPA-REG OUTCOME Investigators. Empagliflozin, conflicts of interest relevant to this article were tes Study Group. Effects of intensive glucose cardiovascular outcomes, and mortality in reported. lowering in type 2 diabetes. N Engl J Med type 2 diabetes. N Engl J Med 2015;373: 2008;358:2545–2559 2117–2128 6. Lincoff AM, Wolski K, Nicholls SJ, Nissen SE. 12. Pfeffer MA, Claggett B, Diaz R, et al.; ELIXA References Pioglitazone and risk of cardiovascular events in Investigators. Lixisenatide in patients with 1. U.S. Food and Drug Administration. Guidance patients with type 2 diabetes mellitus: a meta- type 2 diabetes and acute coronary syndrome. for Industry. Diabetes MellitusdEvaluating Car- analysis of randomized trials. JAMA 2007;298: N Engl J Med 2015;373:2247–2257 diovascular Risk in New Antidiabetic Therapies 1180–1188 13. Nichols GA, Gullion CM, Koro CE, Ephross to Treat Type 2 Diabetes, December 2008. Avail- 7. Kahn SE, Haffner SM, Heise MA, et al.; ADOPT SA, Brown JB. The incidence of congestive heart able from www.fda.gov/downloads/Drugs/ Study Group. Glycemic durability of rosiglitazone, failure in type 2 diabetes: an update. Diabetes GuidanceComplianceRegulatoryInformation/ metformin, or glyburide monotherapy. N Engl Care 2004;27:1879–1884 Guidances/ucm071627.pdf. Accessed 12 October J Med 2006;355:2427–2443 14. Novo Nordisk. Victoza significantly re- 2015 8. Scirica BM, Bhatt DL, Braunwald E, et al.; duces the risk of major adverse cardiovascular 2. Hiatt WR, Kaul S, Smith RJ. The cardiovascu- SAVOR-TIMI 53 Steering Committee and Investiga- events in the LEADER trial [press release]. lar safety of diabetes drugs–insights from the tors. Saxagliptin and cardiovascular outcomes in Bagsværd, Denmark, 4 March 2016. Available rosiglitazone experience. N Engl J Med 2013; patients with type 2 diabetes mellitus. N Engl from www.novonordisk.com/media/news- 369:1285–1287 J Med 2013;369:1317–1326 details.1991879.html. Accessed 11 March 3. Nissen SE, Wolski K, Topol EJ. Effect of 9. White WB, Cannon CP, Heller SR, et al.; 2016 muraglitazar on death and major adverse car- EXAMINE Investigators. Alogliptin after acute 15. Lincoff AM, Tardif JC, Schwartz GG, et al.; diovascular events in patients with type 2 dia- coronary syndrome in patients with type 2 dia- AleCardio Investigators. Effect of aleglitazar betes mellitus. JAMA 2005;294:2581–2586 betes. N Engl J Med 2013;369:1327–1335 on cardiovascular outcomes after acute 4. Hemmingsen B, Schroll JB, Lund SS, et al. 10. Green JB, Bethel MA, Armstrong PW, et al.; coronary syndrome in patients with type 2 Sulphonylurea monotherapy for patients with TECOS Study Group. Effect of sitagliptin on car- diabetes mellitus: the AleCardio ran- type 2 diabetes mellitus. Cochrane Database diovascular outcomes in type 2 diabetes. N Engl domized clinical trial. JAMA 2014;311: Syst Rev 2013;4:CD009008 J Med 2015;373:232–242 1515–1525