The European Commission–sponsored Safety Evaluation of Adverse Reactions in Diabetes (SAFEGUARD) project evaluates the pancreatic and cardiovascular safety of oral glucose-lowering drugs in persons with type 2 diabetes mellitus (T2DM). We present the results of the meta-analysis component. To combine results from observational studies comparing the risk of acute myocardial infarction (AMI) between rosiglitazone and pioglitazone. We searched PubMed, Embase, and the Cochrane library to identify observational studies of persons with T2DM reporting age-and-sex-adjusted relative risks (RR) on the associations of any oral glucose-lowering drug and AMI, heart failure, or cardiovascular mortality, published in 2000-2011. Of the 44 studies abstracted, 11 compared AMI risk for the use of rosiglitazone and pioglitazone. Estimates were pooled when there were at least 3 independent data points for the overall analysis and 2 for subgroup analyses, using random effects models. Heterogeneity of studies was evaluated using the chi-squared test. Nine cohort and 2 nested case-control studies, published between 2007 and 2010, were included in the meta-analysis; 3 studies reported estimates for monotherapy drug use. The summary RR (95% CI) for AMI for rosiglitazone versus pioglitazone monotherapy was 1.33 (0.98-1.79); the RR when cotreatment with other glucose-lowering drugs was allowed was 1.08 (1.02, 1.15). Based on 2 studies, the RR for incident AMI was 1.10 (1.00-1.21), for incident or prevalent AMI, 1.14 (1.00, 1.30). For new drug users, the RR was 1.10 (1.01, 1.20), whereas for new or prevalent users it was 1.24 (0.86-1.80). Significant and relevant heterogeneity was present in most analyses. No studies reported on dose or duration effects. Heterogeneity among estimates is present and needs to be investigated further. Overall estimates suggested a small increase in risk for rosiglitazone over pioglitazone but some estimates were compatible with a null effect. The new database study in the SAFEGUARD project will provide results to better elucidate the AMI risk of these drugs.
