The mechanism leading to the formation of antiphospholipid antibodies (aPL) is still unknown. Because an in vitro study suggested that aPL may derive from pro-oxidant conditions, we sought a relationship between aPL and isoprostanes, indices of lipid peroxidation in vivo. Thirty patients with systemic lupus erythematosus have been studied. Seventeen (56.6%) were positive for aPL because they had lupus anticoagulant and/or high titer of anticardiolipin antibodies (aCL). Plasma levels of tumor necrosis factor (TNF ) and urinary excretion of two isoprostanes, 8-epi-PGF2alpha and IPF2alpha -I, free radical catalyzed oxidation products of arachidonic acid, were measured. Patients with systemic lupus erythematosus had higher urinary excretion of 8-epi-PGF2alpha and IPF2alpha -I than controls; urinary excretion of the two isoprostanes was highly correlated (Rho = 0.74, P < .0001). Urinary 8-epi-PGF2alpha was highly correlated with both aCL titer (Rho = 0. 70, P < .0001) and TNF (Rho = 0.84, P < .0001), a measure of disease severity. Excretion of this isoprostane was also higher in those patients who exhibited aPL (P < .0001). Comparable correlations were observed with the isoprostane IPF2alpha -I. No difference of 8-epi-PGF2alpha was observed between patients with and without previous history of thrombosis. This study, showing the existence of a close association between aPL and increased in vivo lipid peroxidation, supports the hypothesis that these antibodies may result from pro-oxidative conditions and suggests that inflammation may play an important role.