A protocol was developed to compare prolonged heart preservation by hypothermic storage with prolonged hypothermic storage interrupted by a period of reperfusion. Hearts from adult mongrel dogs were excised after administration of 4 degrees C crystalloid cardioplegia. Group A hearts (N = 7) underwent 7.5 hours of ischemia at 4 degrees C followed by 1.5 hours of reperfusion and rewarming (A0). Group B hearts (N = 8) underwent 3 hours of ischemia at 4 degrees C, 1.5 hours of reperfusion and rewarming (B1), 3 additional hours of ischemia at 4 degrees C following repeat cardioplegia, and finally 1.5 hours of reperfusion and rewarming (B2). During reperfusion, hearts were defibrillated and left ventricular (LV) function was assessed by measuring isovolumic peak systolic pressure and maximum positive rate of rise of LV pressure (+dP/dtmax) with an intraventricular balloon. LV biopsy samples for adenosine triphosphate (ATP) assay were obtained before ischemia and after each rewarming period. One Group A heart could not be defibrillated and studied. All Group B hearts completed the protocol. LV function, as assessed by peak pressure and +dP/dtmax, at B1 and B2 exceeded values obtained at A0, but the differences were not statistically significant. The mean ATP level was 63.4 +/- 7.7% of baseline at B1 and 79.7 +/- 4.3% of baseline at B2 (p less than .03). The mean ATP level was 57.9 +/- 5.9% of baseline at A0 (p less than .007, B2 vs. A0). It is presumed that intermittent reperfusion allows repletion of substrate stores, which results in improved myocardial protection.(ABSTRACT TRUNCATED AT 250 WORDS)