TECOS: No CVD Risks or Heart Failure With Sitagliptin in High-Risk Diabetic Patients

Michael O'Riordan

June 08, 2015

BOSTON, MA ( updated June 9, 2015 ) — In patients with type 2 diabetes and cardiovascular disease, the use of sitagliptin (Januvia, Merck), a dipeptidyl peptidase-4 (DPP-4) inhibitor used to lower HbA1c levels, is not associated with an increased risk of cardiovascular events, according to a large outcomes trial designed specifically to answer this question[1].

Most important, investigators did not observe any increase in the number of patients hospitalized for heart failure with sitagliptin, a critically watched secondary end point, given that other drugs in the class, most notably saxagliptin (Onglyza, AstraZeneca) in the SAVOR TIMI- 53 study, showed an increase in heart-failure events. There was also a trend toward an increased risk of heart-failure events among type 2 diabetic patients treated with alogliptin (Nesina, Takeda Pharmaceuticals) in the EXAMINE study.

Study investigators said the TECOS trial, presented today here at the American Diabetes Association (ADA) 2015 Scientific Sessions and published concurrently in the New England Journal of Medicine, should reassure physicians about the cardiovascular safety of the agent when given to high-risk diabetic patients.

"In this study, we specifically looked at the rates of heart-failure events and found there were no differences," Dr Eric Peterson (Duke Clinical Research Institute, Durham, NC), cochair of the TECOS executive committee, told the media during a press conference announcing the results. "As a matter of fact, there was no hint of heart failure seen in the sitagliptin-treated patients relative to placebo. Given the size of our study, the long duration of follow-up, as well as the higher risk of our population, we feel that this very adequately addresses and puts to bed the question that there is any risk for heart failure with this drug."

Answering the FDA's Questions

Speaking during the ADA press conference, cochair of the executive committee Dr Rury Holman (Oxford University, UK) said the TECOS investigators enrolled their first patient in 2008, right around the time the US Food and Drug Administration (FDA) mandated that all new diabetes drugs be subjected to more stringent clinical trials. As reported previously, the FDA now requires that all new drugs approved for diabetes rule out cardiovascular toxicity rather than simply showing the agent lowers HbA1c levels.

TECOS was a large-scale cardiovascular-outcomes study that enrolled 14,671 patients with type 2 diabetes and established cardiovascular disease to sitagliptin or placebo on top of their existing therapy. At baseline, the mean HbA1c level was 7.2%—the trial enrolled patients with a baseline HbA1c level ranging from 6.5% to 8.0%—and patients had been living with their diabetes for 11.6 years.

Regarding the primary end point—a composite of cardiovascular death, nonfatal MI, nonfatal stroke, or hospitalization for unstable angina—there was no significant difference among those treated with sitagliptin and those treated with placebo. In excluding unstable angina from the primary end point, there was also no significant increase in the risk of cardiovascular death, nonfatal MI, or nonfatal stroke. Overall, there was no signal of cardiovascular risk among any of the end points when analyzed separately or in combination.

TECOS: Intention-to-Treat Analysis

Outcome Sitagliptin (n=7332) Placebo (n=7339) Hazard ratio (95% CI)
Primary outcome (cardiovascular death, nonfatal MI, nonfatal stroke, or unstable angina) 11.4 11.6 0.98 (0.89–1.08)
Cardiovascular death, nonfatal MI, or nonfatal stroke 10.2 10.2 0.99 (0.89–1.10)
Cardiovascular death 5.2 5.0 1.03 (0.89–1.19)
Hospitalization for unstable angina 1.6 1.8 0.90 (0.70–1.16)
Fatal or nonfatal MI 4.1 4.3 0.95 (0.81–1.11)
Fatal or nonfatal stroke 2.4 2.5 0.97 (0.79–1.19)
Hospitalization for heart failure 3.1 3.1 1.00 (0.83–1.20)
Hospitalization for heart failure or cardiovascular death 7.3 7.2 1.02 (0.90–1.15)

The ADA meeting devoted a 2-hour block for the TECOS study, with investigators given time to explain the study rationale, the design, and the clinical outcomes. During the presentation, Holman provided data from multiple subgroups, including patients stratified by age, race, baseline renal function, and geographic region. In all these analyses, there was no signal of cardiovascular toxicity with sitagliptin. There was an interaction with body-mass index—with heavier patients faring better—but Holman said not to make too much of the finding, given how many subgroups were analyzed.

In the trial, approximately 18% of patients in both treatment arms had a prior diagnosis of congestive heart failure. Even among these high-risk patients, the use of sitagliptin did not increase the risk of cardiovascular events, nor did its use increase hospitalizations for heart failure, report investigators.

Speaking with the media, Holman said the study was designed as a noninferiority study but had a sufficient number of patients and statistical power for a superiority analysis. "At the time of the design of the study, there was general belief that, maybe, there might be [cardiovascular] advantages with this class of drug," said Holman.

Those advantages, however, did not pan out in TECOS, nor did they pan out in SAVOR or EXAMINE, despite previous observational studies and meta-analyses suggesting there might be a reduction in cardiovascular events with the DPP-4-inhibitor drug class. "We now have three studies with virtually the same result," said Holman. "I think we can safely say these drugs neither harm nor provide benefit for cardiovascular events."

Regarding the safety of sitagliptin, there was no significant difference in the overall incidence of infections, cancer, renal failure, or severe hypoglycemia. Numerically, there were more confirmed cases of acute pancreatitis among patients treated with sitagliptin—23 events vs 12 events in the placebo arm—but the difference was not statistically significant. The rates of pancreatic cancer were very low in both treatment arms, but it was numerically less frequent among those who received sitagliptin.

Not a Glucose-Lowering Study

Dr Rury Holman

During the press conference, Holman stressed the study was designed to achieve "glycemic equipoise" between the two treatment arms. Open-label use of antihyperglycemic therapy was encouraged as required to achieve an appropriate HbA1c level.

"This is a glucose-lowering drug being tested for cardiovascular benefit or harm," said Holman. "The idea was not to see a glucose difference. All patients were treated to similar targets, led by their physicians. At the end of the study, overall, there was less than a 0.3% difference in hemoglobin A1c. That's important because we didn't want a glucose difference that might have influenced the cardiovascular outcomes one way or the other."

Peterson added that TECOS was not designed to address the effectiveness of long-term glucose lowering in patients with type 2 diabetes. That would have required a significantly larger study of longer duration. Holman said the TECOS trial does prove the utility of sitagliptin, noting that patients randomized to the DPP-4 inhibitor required significantly fewer antihyperglycemic agents than those in the placebo arm and were less likely to start long-term insulin therapy.

Speaking during the ADA session, study investigator Dr John Buse (University of North Carolina School of Medicine, Chapel Hill) said the DPP-4 inhibitors have been "embraced" by patients and physicians alike. For physicians, the drug class is effective in lowering HbA1c when used in combination with metformin. For patients, the drugs are among the best-tolerated antihyperglycemic agents on the market, he added. The ADA and the European Association for the Study of Diabetes both recommend DPP-4 inhibitors as a second-line drug therapy for patients with diabetes.

Regarding the safety of the drugs, Buse said all three studies—SAVOR, EXAMINE, and TECOS—have shown a numerically higher number of subjects affected with pancreatitis, translating into roughly one extra case of pancreatitis per 1000 patient-years in TECOS. "At worst, acute pancreatitis is a rare complication of DPP-4 inhibition therapy," said Buse.

Clinical Implications of TECOS

In April 2015, the FDA's Endocrinologic and Metabolic Drugs Advisory Committee voted overwhelmingly in favor of updating labeling information for saxagliptin and alogliptin to highlight the potential for heart-failure risks in diabetic patients treated with the drugs.

As for why researchers observed a statistically significant 27% increase in the risk of heart failure with saxagliptin in SAVOR-TIMI, a nonsignificant trend toward increased heart failure in EXAMINE with alogliptin, yet no signal of harm in TECOS, Peterson said investigators really do not know the answer. He said there was very little difference in the type of patient enrolled in the three trials and little difference in how they were treated, including for cardiovascular-disease prevention.

That would leave either a difference between the three DPP-4 inhibitors or the possibility the signal of harm in SAVOR and EXAMINE was a play of chance.

TECOS, Holman added, was a significantly larger study with longer follow-up than the previous trials and is better suited to address whether or not the heart-failure risk with the DPP-4 inhibitors is real or a false signal. "The fact that heart failure was unexpected and we don't have a plausible biological mechanism, you always have to worry about the play of chance," he said. All that can be said at this point is that there was absolutely no hint of heart failure with sitagliptin.

"If the costs between the DPP-4 inhibitors were the same, it would be hard to make a case to go with anything other than sitagliptin," Dr Darren McGuire (University of Texas Southwestern Medical Center, Dallas), member of the TECOS trial executive committee, told heartwire . The reason is that the rates of MACE seem to be identical across trials and the possible increase in heart failure with the other agents. He noted the Kaplan Meier curves for hospitalized heart failure were virtually superimposable in TECOS, and at no time point was there any difference in the event rate between the placebo and sitagliptin arms.

Regarding the heart-failure signal in SAVOR, McGuire suspects the finding is a real finding, given the high number of events—289 hospitalizations for heart failure in the saxagliptin arm vs 228 in the placebo arm—their prospective collection and central adjudication, and the observation of separation in the event curves early in the course of study treatment. He added that while cross-study comparisons are difficult, the patients in the three trials "were more alike than they were different."

To heartwire , Holman said that if physicians are looking at the data and wanting to choose an agent, the risk-averse doctor treating type 2 diabetic patients would presumably "vote with their feet" and select sitagliptin.

In the UK, DPP-4 inhibitors are third-line agents, mainly because of their high cost, but the concern about heart failure has not altered how they are used in clinical practice. Sitagliptin is the most widely used DPP-4 inhibitor in the UK—it was the first to market and is on most formularies. Sitagliptin, said Holman, is on his formulary, so the results from TECOS will not alter his practice.

Dr Eric Peterson

"As cardiologists, we're driven by the evidence," added Peterson. "We can look only at the evidence we have from clinical studies. These are not cross-comparative studies, so the degree to which we extrapolate from one drug to another is done with caution. That said, if you look at sitagliptin, it is remarkably clean with just about every signal we looked at, every subgroup we looked at, and in every way we can look at the data. We teased this data backward and forward. Everything comes up looking like the drug does not have a cardiovascular safety signal."

From an endocrinologist's perspective, Buse said there are remaining questions, however. One of the biggest questions is the comparative effectiveness of the competing DPP-4 inhibitors. Studies are needed to determine whether there are meaningful differences in safety and efficacy between saxagliptin, alogliptin, and sitagliptin when tested in a head-to-head study. As for the FDA-mandated cardiovascular-outcomes studies, Buse said all three clinical trials were effective in assessing the short-term safety of the drugs, "but we really don't have adequate data on really long-term exposure that many patients may experience in their lifetime."

Dr Allison Goldfine (Joslin Diabetes Center, Boston, MA) said that given the FDA's requirements, there will be an abundance of cardiovascular-outcomes data in the years to come. These large-scale but relatively short-term clinical trials will be designed to show cardiovascular safety only and are not designed to address whether lowering HbA1c with a specific drug/mechanism provides an advantage in terms of cardiovascular risk reduction. "I think we do need to ask whether the multiple, ongoing safety trials ask and provide answers to the most important clinical questions," she said.

The study was supported by Merck, Sharpe, & Dohme. Peterson reports grant support from Eli Lilly, grant support and personal fees from Janssen, and personal fees from AstraZeneca, Bayer, and Sanofi. Holman reports grant support from Merck during the conduct of the study, grant support from AstraZeneca and Bristol-Myers Squibb, grant support and personal fees from Bayer, personal fees from Amgen, Intarcia, Novartis, Novo Nordisk, and Owen Green Mumford, and other support from GlaxoSmithKline, Janssen, and Takeda. Disclosures for the coauthors are listed on the journal website.


  1. Green JB, Bethel MA, Armstrong PW, et al. Effect of sitagliptin on cardiovascular outcomes in type 2 diabetes. N Engl J Med 2015; DOI:10.1056/NEJMoa1501352. Article