Hunting for cardiovascular signals from diabetes drugs

Diabetes drug development entered a new era in 2007 when a meta-analysis of 42 studies that had compared rosiglitazone with other drugs in a total of nearly 28,000 patients suggested that rosiglitazone treatment led to significantly increased rates of cardiovascular death and myocardial infarction.

Based largely on that, the Food and Drug Administration in 2008 issued recommendations to companies developing new diabetes drugs to run large-scale trials aimed at assessing their cardiovascular effects.

^{Mitchel L. Zoler/IMNG Medical Media}

The new data show just how painstaking investigators have become in trying to parse out hints of cardiovascular danger lurking in masses of megatrial data.

In 2011, the FDA placed restrictions on prescribing rosiglitazone, and U.S. use of the drug plummeted. (An FDA panel voted in June to relax some of the restrictions the agency had imposed.)

Some consequences of the long shadow cast by the rosiglitazone experience played out in talks at the annual congress of the European Society for the Study of Diabetes in September, as well as at the annual congress of the European Society of Cardiology a few weeks before that. The new data show just how painstaking investigators have become in trying to parse out hints of cardiovascular danger lurking in masses of megatrial data.

Heart failure haunts DPP-4 inhibitors

At ESC, as well as in a pair of simultaneously published articles, researchers reported results from two large trials designed to follow the FDA recommendations and assess cardiovascular safety for two new selective dipeptidyl peptidase-4 (DPP-4) inhibitors, saxagliptin (Onglyza) and alogliptin, in patients with type 2 diabetes.

Saxagliptin was the drug with the red-flag signal, while alogliptin faced guilt by association. The saxagliptin study, SAVOR-TIMI 53, used patients with either an established history of cardiovascular disease (CVD) or multiple CVD risk factors. The study’s primary safety endpoint, the combined rate of CVD death, nonfatal myocardial infarction, or nonfatal stroke during 2 years of follow-up, was virtually identical in the two groups, 7.3% in the saxagliptin patients and 7.2% in the controls.

The study’s secondary safety endpoint, which included the primary combined plus other endpoints such as need for revascularization and hospitalization for heart failure, was also similar in the two study arms, with this expanded combined endpoint occurring in 12.8% of the saxagliptin patients and 12.4% of the controls (N. Engl. J. Med. 2013;369:1317-26).

The whisper of a saxagliptin problem was in just one secondary, individual safety endpoint, hospitalizations for heart failure. These occurred in 3.5% of the saxagliptin patients and in 2.8% of controls, a statistically significant difference. Further data presented at EASD showed that this difference clustered predominantly in the subgroup of patients who entered the study with presumed heart failure, based on having a blood level of NT-pro brain natriuretic peptide of at least 333, about 25% of patients in both treatment arms (although the investigators identified heart failure in only 13% of enrolled patients). Within this subgroup, heart failure–associated hospitalizations occurred in 10.9% of the saxagliptin patients and 8.9% of the controls, a statistically significant difference, reported Dr. Deepak L. Bhatt, one of the investigators for the SAVOR-TIMI 53 study, which randomized more than 16,000 patients.

In contrast, the alogliptin study, EXAMINE, which randomized more than 5,000 patients with type 2 diabetes and a recent acute coronary syndrome event to treatment with alogliptin or placebo, found no real confirmation of this problem (N. Engl. J. Med. 2013;369:1327-35).

After the early-September report of this saxagliptin and heart failure hospitalization signal, the EXAMINE researchers went back to their data to specifically look at this issue, Dr. William B. White, the lead investigator, said when he presented the new findings at EASD. During a median of 18 months of follow-up, the rate of hospitalizations for heart failure was 3.1% in the alogliptin group and 2.9% in the placebo group, a difference that was not statistically significant, although it trended just slightly toward an alogliptin problem. Among the subgroup of patients who entered EXAMINE with a history of heart failure, the rate of the combined cardiovascular safety endpoint of CV death, nonfatal MI, or nonfatal stroke was 18% in the alogliptin group and 22% in the placebo group, a difference that was not statistically significant.

Despite the absence of any apparent heart failure–hospitalization effect in EXAMINE, Dr. Naveed Sattar, the invited discussant for the two reports at EASD, meta-analyzed the heart failure–hospitalization results from both studies, and found a combined, 24% relative increase in this outcome for the two DPP-4 inhibitors combined, a statistically significant effect largely driven toward significance by the underlying statistical significance in the saxagliptin study; the alogliptin results couldn’t resolve this difference since they trended in the same direction. Dr. Sattar concluded that it raised enough doubt about the safety of this entire drug class in patients with pre-existing heart failure to recommend not using drugs from this class in patients with any indication of heart failure – regardless of severity – until the apparent link received further scrutiny. Dr. Bhatt promised a report with further information on the heart failure issue at the American Heart Association Scientific Sessions in November.