An editorialist concludes that clinicians should no longer feel comfortable prescribing the diabetes drug.
In a recent meta-analysis, rosiglitazone was associated with increased risk for myocardial infarction and possibly cardiovascular mortality (Journal Watch May 24 2007). The authors noted that the industry-sponsored RECORD trial, specifically designed to examine cardiovascular outcomes associated with rosiglitazone, was still in progress. Because of the controversy sparked by the meta-analysis, the RECORD investigators conducted this interim analysis (about two thirds of the way through the trial).
Researchers in Europe and Australasia enrolled 4447 type 2 diabetic patients taking metformin or sulfonylurea monotherapy. Half the patients were randomized to receive add-on rosiglitazone; in the control group, metformin users received add-on sulfonylurea, and sulfonylurea users received add-on metformin. During an average follow-up of 3.75 years, the primary endpoint (cardiovascular death or cardiovascular hospitalization) occurred in 217 rosiglitazone patients and 202 controls (hazard ratio, 1.08; P=0.43). For secondary endpoints, the only statistically significant finding was an increased risk for heart failure in the rosiglitazone group compared with the control group (38 vs. 17 events; P=0.006). A slight excess of MIs in the rosiglitazone group was not significant (43 vs. 37 events; P=0.5).
Comment: By itself, this interim analysis doesn’t settle the question of whether rosiglitazone increases risk for MI or death. However, the results are not reassuring: The primary endpoint is in the "wrong direction" for rosiglitazone (although not statistically significantly so), and the increased risk for heart failure is striking. This report — superimposed on the earlier meta-analysis — convinces an editorialist that clinicians should no longer feel comfortable prescribing rosiglitazone. I agree.
— Allan S. Brett, MD
Published in Journal Watch General Medicine June 12, 2007