Rosiglitazone associated with increased risk of acute MI

Clinical Question

Does rosiglitazone increase the risk of adverse cardiovascular events?

Bottom Line

This study found that short-term use of rosiglitazone is associated with an increased risk of myocardial infarction (MI) and cardiovascular death. Although the absolute increase was small, the goal of diabetes treatment is to decrease cardiovascular disease, the largest cause of death in patients with diabetes. Longer and larger studies using cardiovascular outcomes as a prespecified outcome and meta-analyses using individual patient level data are needed to more definitively answer this question. A study of pioglitazone (Actos) found no increase in cardiovascular events but showed an increase in hospitalizations for heart failure. Aggressive attention to cardiovascular risk factors remains the most important thing we can do for our diabetic patients, more important even than strict blood sugar control, as we learned in the United Kingdom Prospective Diabetes Study. (LOE = 1a-)


Nissen SE, Wolski K. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. N Engl J Med. 2007;356:2457-2471.  [PMID:17517853]

Study Design

Meta-analysis (randomized controlled trials)


Self-funded or unfunded


Outpatient (any)


The thiazolidinediones have a checkered history: troglitazone caused hepatotoxicity and was removed from the market, while muraglitazar never made it to market because it increased the risk of cardiovascular events. Rosiglitazone, though, was thought to be safe based on the results of short, small studies submitted to the FDA. However, these studies were designed to detect changes in blood sugar, and not to detect changes in cardiovascular mortality, MI, or all-cause mortality. The authors of this meta-analysis have therefore pooled data from 42 smaller trials to determine if the use of the drug increases the risk of MI or cardiovascular death. Since the studies were not specifically designed to evaluate cardiac outcomes, most did not describe how cardiac end points were determined. Most studies were between 24 weeks' and 52 weeks' duration, with a typical dosage range for rosiglitazone of 4 mg to 8 mg per day. The average age of patients was 56 years and more than half were men; the mean hemoglobin A1C was 8.2%. The results showed a significant increase in the likelihood of MI (odds ratio [OR] = 1.43; 95% CI, 1.03 - 1.98) and a borderline increase in the risk of death from cardiovascular causes (OR = 1.64; 95% CI, 0.98 - 2.74). The absolute increase in risk of MI was small, approximately 0.4%. On the other hand, the studies were short, and most excluded patients with pre-existing heart disease, which explains the small total number of cardiovascular events in both groups. Results were similar whether the control group took placebo or an active comparator, although the number of patients in each comparator group was relatively small. The authors used a fixed effects model to combine the studies; a more conservative approach would have been to use a random effects model, which may not have shown a significant difference in outcomes.