Clinical investigations have demonstrated a link between use of the sulfone cyclooxygenase-2 (COX-2) inhibitor, rofecoxib, and increased risk for atherothrombotic events. This increased risk was not observed for a sulfonamide COX-2 inhibitor (celecoxib), indicating a potential non-enzymatic mechanism for rofexocib. To test this hypothesis, we compared the independent effects of COX-2 inhibitors on human LDL oxidation, an important contributor to atherosclerotic cardiovascular disease. The results showed that rofecoxib (100 nM) significantly decreased (>40%, p<0.001) the lag time for LDL conjugated diene formation and increased levels of thiobarbituric-acid-reactive-substances (TBARS) in vitro. The pro-oxidant activity of rofecoxib was dose-dependent and attenuated by 70% (p<0.001) with the antioxidant, Trolox. Rofecoxib and etoricoxib (100 nM) also caused a marked increase (>35%, p<0.001) in non-enzymatic generation of isoprostanes, as measured by mass spectroscopy. Addition of rofecoxib to fresh human plasma reduced the oxygen radical antioxidant capacity (ORAC) by 34% (p<0.0001). By contrast, other selective (celecoxib, valdecoxib, meloxicam) and non-selective COX inhibitors (ibuprofen, naproxen, diclofenac) had no significant effect on LDL oxidation rates or plasma ORAC values, even at suprapharmacologic levels. X-ray diffraction analysis showed that sulfone COX-2 inhibitors interact differently with membrane phospholipids, suggesting a physico-chemical basis for the pro-oxidant activity. These results demonstrate that sulfone COX-2 inhibitors increase the susceptibility of biological lipids to oxidative modification through a non-enzymatic process. These findings may provide mechanistic insight into reported differences in cardiovascular risk for COX-2 inhibitors.