Clinical investigations have demonstrated a relationship between the extended use of rofecoxib and increased risk for atherothrombotic events. This has led to the removal of rofecoxib from the market and explicit cardiovascular safety warnings for other COX-2 selective and non-selective agents that remain on the market. Early explanations for the cardiotoxicity of rofecoxib, such as the relative cardioprotective effect of comparator agents (naproxen) or an "imbalance" between thromboxane and prostacyclin biosynthesis due to an absence of concomitant aspirin use, have not been substantiated by the evidence. New experimental findings indicate that the cardiotoxicity of rofecoxib is not a general class effect but may be due to its intrinsic chemical structure and unique primary metabolism. Specifically, rofecoxib has been shown to increase the susceptibility of human LDL and cell membrane lipids to oxidative modification, a hallmark feature of atherosclerosis. Rofecoxib was also found to promote the non-enzymatic formation of isoprostanes from biological lipids, which act as important mediators of inflammation in the atherosclerotic plaque. The explanation for such cardiotoxicity is that rofecoxib forms a reactive maleic anhydride in the presence of oxygen due to its chemical structure and primary metabolism (cytoplasmic reductase). By contrast, adverse effects on rates of LDL and membrane lipid oxidation were not observed with other chemically distinct (sulfonamide) COX-2 inhibitors under identical conditions. These findings provide a compelling rationale for distinguishing the differences in cardiovascular risk among COX-selective inhibitors on the basis of their intrinsic physico-chemical properties.