Inhibition of cyclooxygenase (COX) might favour non-enzymatic formation of cardiodepressive isoprostanes from arachidonic acid by radicals generated during reperfusion. This could explain deleterious effects of acetylsalicylic acid (ASA) on cardiac function. We examined the influence of COX inhibition on myocardial function after low-flow ischaemia and reperfusion, employing either ASA (100 micromol/l), the partially selective COX-2 inhibitor meloxicam (0.3 micromol/l and 3.0 micromol/l), or the highly selective COX-2 inhibitor SC 58125 (1.0 micromol/l and 3.0 microgmol/l). Isolated, buffer-perfused guinea pig hearts, performing pressure-volume work before and after consecutive low-flow ischaemia and reperfusion, were used for the study. Measurement of coronary and aortic flow, ejection time and heart rate served to calculate external heart work (EHW), before and after ischaemia. Additionally, release of prostacyclin and thromboxane A2, production of lactate, consumption of pyruvate and tissue concentration of the isoprostane 8-iso-PGF2alpha were measured. ASA significantly reduced recovery of EHW (46+/-18% vs. 82+/-15% for controls), whereas meloxicam and SC 58125 did not (64+/-15% and 74+/-13% recovery, respectively). Paradoxically, ASA increased reactive hyperaemia and consumption of pyruvate in the early reperfusion phase in comparison to all other groups, while lactate production did not differ. Prostacyclin production did not increase during reperfusion and was not significantly different between groups at any time point. In contrast, thromboxane A2 release increased about fivefold in the 2nd min of reperfusion under control conditions and in the presence of SC 58125, but was inhibited by ASA and by meloxicam in both concentrations. Isoprostane content of heart tissue was not detectably influenced under the mild reperfusion conditions used here. We conclude that ASA can aggravate postischaemic cardiac dysfunction, independent of COX inhibition. The deleterious effect in the present model might be due to uncoupling of mitochondrial oxidative phosphorylation rather than to direct effects of reduced eicosanoid release or radical induced formation of isoprostanes.