Cardiovascular morbidity and mortality in patients with type 2 diabetes are a major problem in clinical practice. Thiazolidinediones (TZDs) are agonists of the peroxisome proliferator-activated receptor gamma which improve glycaemic control by reducing insulin resistance. TZDs also seem to have beneficial effects on various cardiovascular risk factors and consequently may have the potential to reduce the risk of cardiovascular disease (CVD). Although the first large-scale clinical trial evaluating the effect of a TZD on secondary prevention of major adverse cardiovascular outcomes supported this hypothesis, a recently published meta-analysis raised substantial uncertainty about the cardiovascular safety of rosiglitazone. This article summarises the evidence from completed and ongoing outcome trials with TZDs, as well as the recent meta-analytic data on their cardiovascular safety, aiming to provide an up-to-date and balanced view of a very important field. Data from clinical trials consistently indicate that treatment with glitazones significantly increase the risk of heart failure. Despite the fact that rosiglitazone and pioglitazone have much more similarities than differences with regards to their effects on cardiovascular risk factors, pioglitazone seems to have more favourable effects on major cardiovascular outcomes. This issue also highlights the potential hazards involved in using surrogate end-points for drug approval.