Can We Now Dismiss Concerns Over Rosiglitazone?

Gregory A. Nichols, PhD


July 20, 2009

Rosiglitazone Evaluated for Cardiovascular Outcomes in Oral Agent Combination Therapy for Type 2 Diabetes (RECORD): A Multicentre, Randomized, Open-Label Trial

Home PD, Pocock SJ, Beck-Nielsen H,
Curtis PS,Gomis R, Hanefeld M, Jones NP,
Komajda M, McMurray JJV, for the RECORD Team
Lancet. 2009, published online June 5, 2009

Study Summary

The aim of the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes (RECORD) trial was to assess cardiac outcomes in patients receiving a combination of rosiglitazone with metformin plus sulfonylurea, compared with patients receiving metformin plus sulfonylurea alone for control of blood glucose. Patients with type 2 diabetes on monotherapy with either metformin or sulfonylurea and with hemoglobin A1c > 7% but ≤ 9% were randomly assigned to the addition of rosiglitazone (n = 1103) or metformin (if already on sulfonylurea, n = 1122) or to rosiglitazone (n = 1117) and sulfonylurea (if already on metformin, n = 1105). Patients were followed for a median of 6 years for the primary endpoint of time to first cardiovascular hospitalization or cardiovascular death. The study was designed to test noninferiority of rosiglitazone, thus requiring an upper limit of the confidence interval of the hazard ratio (HR) to be less than 1.20. With a target A1c of 7% or lower, a third oral agent (either metformin or sulfonylurea) was added to the rosiglitazone group followed by cessation of rosiglitazone and addition of insulin if A1c values reached 8.5%. In the metformin/sulfonylurea group, insulin was added if A1c reached 8.5%.

The primary outcome, cardiovascular hospitalization or cardiovascular death, occurred in 321 and 323 participants assigned to the rosiglitazone and metformin/sulfonylurea groups, respectively (HR, 0.99; 95% confidence interval [CI], 0.85-1.16). Although mean A1c declined more in the rosiglitazone group, 25% of patients in that group required additional medications, compared with 17% of patients in the metformin/sulfonylurea group. A sensitivity analysis that excluded patients after transfer from dual therapy yielded similar results for the primary outcome. A predefined composite secondary endpoint of cardiovascular death, myocardial infarction, and stroke was also not significant (HR, 0.93; 95% CI, 0.74-1.15) for rosiglitazone vs metformin/sulfonylurea. Individually, HRs for myocardial infarction were elevated (HR, 1.14; 95% CI, 0.80-1.63) and depressed for stroke (HR, 0.72; 95% CI, 0.49-1.06), but were not significantly different.


Although the European Medicines Agency approved the use of pioglitazone and rosiglitazone in 2000, post-marketing cardiovascular outcome studies of these 2 thiazolidinediones were requested because of concerns over the increased risk for heart failure and other cardiovascular effects. RECORD was commenced, and its importance was greatly enhanced in 2007 after a meta-analysis suggested that rosiglitazone increased the risk for myocardial infarction and possibly death.

On reading the current article, it might be tempting to conclude that rosiglitazone has been exonerated from these risks. After all, the HR for the primary outcome was close to unity, and the noninferiority criterion was clearly met. However, there were a number of hints in the data that raise concerns. Although the study was clearly not powered to examine myocardial infarction independently, a nonsignificant HR of 1.14 emerged. If this result was added to the aforementioned Nissen and Wolski meta-analysis, the overall risk of myocardial infarction with rosiglitazone would be even higher than previously reported. By the same argument, however, the favorable but nonsignificant RECORD finding would reduce the marginally significant result for cardiovascular death reported by Nissen and Wolski.

In addition to the elevated HR for myocardial infarction, the current study reported that there was an indication that primary endpoints were more common among patients with previous ischemic heart disease in the rosiglitazone group. This suggests that some of the earlier studies that raised concerns about rosiglitazone may need to be re-examined with stratification by existing ischemia. Because this may not be possible, and because ischemic heart disease is not always diagnosed, the increased risk associated with rosiglitazone in the presence of existing heart disease might not be easy to pre-identify.

On another note, it was puzzling that more patients in the rosiglitazone group required additional therapies to maintain glycemic control. Rosiglitazone has previously been found to be more durable, but now that advantage is in question.

In summary, the RECORD study suggests that rosiglitazone is probably safe and effective in carefully selected patients. However, concerns about this agent are far from alleviated. As with all therapies, one must ask whether the risks are worth the benefits, and whether another available option might be the wiser choice.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: