Ischemic preconditioning in the rat significantly improves functional recovery following global ischemia by undefined mechanisms. It has been suggested that preconditioning protects by altering the tissue metabolic milieu during ischemia, either by increasing ischemic tissue accumulation of a beneficial substance (e.g. adenosine), or inhibiting tissue accumulation of a malefic component (e.g. protons). If this is the case, we hypothesized that no protection should be afforded by preconditioning against a prolonged period of hypoxia, since the continued coronary flow would prevent the accumulation of any metabolic products in the myocardium. To test this hypothesis, isolated buffer-perfused rat hearts were preconditioned by 5 min of ischemia + 5 min of reperfusion and then subjected to 30 min of ischemia, or 25 min of substrate-free hypoxia, or 60 or 90 min of hypoxia with substrate. Function was re-assessed after reperfusion/reoxygenation for a further 30 min and compared to non-preconditioned controls. Ischemic preconditioning improved functional recovery following 30 min of global ischemia (% recovery of developed pressure (LVDP) in control v preconditioned hearts was 31 +/- 4 v 66 +/- 6%; P < 0.05). Importantly, this protection was achieved almost entirely via a better preservation of diastolic function (end diastolic pressure = 78 +/- 3 mmHg in control and 40 +/- 5 mmHg in preconditioned hearts following 30 min of reperfusion; P < 0.05). However, no preconditioning-induced protection was observed following either substrate-free hypoxia or hypoxia with substrate (% recovery of LVDP in control v preconditioned hearts was 31 +/- 4 v 34 +/- 4% after 25 min of substrate-free hypoxia, 48 +/- 3 v 53 +/- 6% after 60 min of hypoxia + substrate and 25 +/- 5 v 30 +/- 6% after 90 min of hypoxia + substrate respectively). Furthermore, no protection by preconditioning against hypoxia-induced diastolic dysfunction was observed. We conclude that preconditioning protects against ischemic injury, but not hypoxic injury. Although hypoxia-induced injury may differ from that induced by ischemia, the results are consistent with the hypothesis that the continued presence of flow with hypoxia abolishes the protective effect of preconditioning. Furthermore, the results support the concept that preconditioning of the ischemic myocardium requires the accumulation of a factor in the ischemic myocardium, either to exert the preconditioning protective effect, or as a factor of injury against which preconditioning affords protection.