The risk of coronary heart disease in type 2 diabetic patients exposed to thiazolidinediones compared to metformin and sulfonylurea therapy

Purpose To evaluate whether the risk of coronary heart disease (CHD) differs among adult diabetic patients treated with thiazolidinediones (TZDs) and similar patients treated with combined oral metformin and sulfonylurea (M + S) therapy. Methods We conducted a retrospective cohort study involving 25 140 diabetic patients aged 18 and older who had at least one pharmacy claim for a TZD or combined M + S therapy between 1 January 1999 and 30 June 2002. We used propensity score matching to adjust for observable differences between initiators of combined M + S therapy and TZD initiators. The data were analyzed in two ways: first based on the original matched groups, 'as balanced', without accounting for switching to another medication during follow-up, and second based on actual antidiabetic drug use during follow-up, 'as treated'. Cox proportional hazards regression and multivariable Poisson regression were performed to compare the risk of CHD events. Results In the 'as balanced' analysis, the risk for CHD among TZD users relative to combination drug users was close to the null value (adjusted hazard ratio: 1.02, 95% confidence intervals (CI): 0.87-1.20). In the 'as treated' analysis, the risk of CHD was similar for periods of current use of TZDs compared to periods of non-use (incidence rate ratio: 1.10, 95%CI: 0.96-1.25). Conclusions These results do not suggest a cardioprotective or deleterious effects of TZDs compared with combined M + S oral therapy on the short-term CHD event risk in persons with type 2 diabetes after accounting for the greater baseline CHD risk in TZD initiators. Copyright (c) 2007 John Wiley & Sons, Ltd.