The effects of a xanthine oxidase-mediated free radical-generating system containing purine and iron-loaded transferrin or solutions containing hydrogen peroxide and iron-loaded transferrin on substrate utilization and high-energy phosphates were evaluated by nuclear magnetic resonance (NMR) spectroscopy in isolated perfused rat hearts. Hearts were supplied with lactate, acetate, and glucose, and the contribution of each substrate to acetyl coenzyme A was measured in control hearts and in the presence of a free radical-generating system. Perfused hearts were monitored by 31P NMR, and tissue extracts were analyzed by 13C NMR. Free radicals decreased the phosphocreatine and beta-ATP peak areas and reduced contractile function. Under control conditions, lactate, acetate, and endogenous sources were the major contributors of acetyl coenzyme A units, with only 5% originating from glucose. In the presence of a xanthine oxidase-mediated free radical-generating system, the glucose contribution increased to 54%, while contributions from acetate and endogenous sources were significantly reduced. Both 13C and 31P NMR analyses showed no significant accumulation of glycolytic sugar phosphates, suggesting little inhibition of glyceraldehyde-3-phosphate dehydrogenase. The increased contribution of glucose to the tricarboxylic acid cycle relative to acetate and endogenous sources is consistent with activation of pyruvate dehydrogenase. In contrast, hearts exposed to a hydrogen peroxide-based free radical-generating system showed an increase in lactate utilization, a decrease in acetate utilization, and no change in glucose utilization compared with control hearts. Glycolytic sugar phosphates were found to accumulate, suggesting possible inhibition of glyceraldehyde-3-phosphate. Thus, different radicals or their metabolites may have varying effects on myocardial metabolism.