it has been proposed that brief hypoxia can substitute for ischemia in the preconditioning of cardiac tissue and salvage of ischemic myocardium. To elucidate a possible mechanism isolated rabbit hearts were subjected to a 30-min period of regional ischemia by occluding a previously snared coronary artery. Following 2 h of reperfusion infarct size was measured by staining left ventricular slices with triphenyltetrazolium chloride. In control hearts infarction averaged 28.7 +/- 1.9% of the risk zone, If the hearts were preconditioned with 5 min global ischemia/10 min reperfusion prior to the regional ischemia, then infarction was significantly reduced to 7.2 +/- 2.0% (P<0.01). When global hypoxia (pO(2) of perfusate 42.0 +/- 2.1 mmHg) for ten min substituted for the five min period of global ischemia, protection was comparable to that observed after ischemic preconditioning (10.2 +/- 1.5% infarction, P<0.01 v central). During hypoxic perfusion adenosine release increased 16-fold over baseline levels. This protection could not be blacked by adding either the adenosine antagonist 8-(p-sulfophenyl)theophylline or the alpha(1)-adrenergic blocker phenoxybenzamine to the hypoxic perfusate. However, co-administration of both agents to the hypoxic perfusate successfully aborted protection (22.6 +/- 2.9% infarction, P N.s. v control). Therefore, 10 min of hypoxia releases both norepinephrine and adenosine in sufficient quantities such that either can completely precondition the heart. (C) 1995 Academic Press Limited