The aim of our work was to study (1) the antioxidant properties of lipoic acid (LA) and its reduced metabolite dihydrolipoic acid (DHLA) formed by reduction of LA and (2) the effects of treatment with LA and DHLA on (a) K(+) efflux from human red blood cells and (b) post-ischemic recovery and oxidative stress in isolated perfused rat hearts challenged with an ischemia-reperfusion (IR) sequence. In vitro, we used xanthine and xanthine oxidase to generate superoxide anion, which is not directly measurable by electron paramagnetic resonance (EPR), but specifically oxidizes the spin probe CPH into an EPR-detectable long lasting CP(*) nitroxide radical. While 5 mM of LA was ineffective in reducing the kinetics of CP(*) nitroxide formation, DHLA was shown to lessen this rate in a dose-dependent manner and at 30 mM was even more efficient than 300 UI/ml SOD. These results are in agreement with the fact that DHLA is able to directly scavenge superoxide anion. Red cells are a good model to investigate oxidative damage in biological membranes; hence, we used a suspension of erythrocytes incubated with 2,2(')-azobis(2-amidinopropane) hydrochloride (AAPH) which generates in vitro free radicals. DHLA provided more effective protection of red cells membranes than LA; DHLA was comparable to Trolox for its antioxidant potency. In vivo, treatment of rats (50 mg/kg/day i.p. for 7 days) with LA induced a slight increase in coronary flow (CF) in isolated perfused hearts, after 30 min of global total ischemia. This effect was not associated with an improvement in contractile function and reduction of myocardial oxidative stress. In conclusion, because of their ability to scavenge free radicals, LA and to an even greater degree DHLA were able to protect the membranes of red blood cells. This finding suggests that LA and DHLA might be useful in the treatment of diseases associated with oxidative stress such as diabetes.