The purpose of this study was to determine the mechanisms by which high physiologic concentrations of glucose (11 mmol/L) were protective while even higher concentrations (20 or 50 mmol/L) were harmful when added to St. Thomas' Hospital No. 2 cardioplegic solution. We evaluated the recovery of isolated working rat hearts subjected to 3 hours of hypothermic multidose cardioplegic arrest. The addition of glucose 11 mmol/L was associated with better aortic flow (79.2% +/- 1.3%) than the addition of glucose 1 mmol/L (61.7% +/- 2.7%), 20 mmol/L (73.6% +/- 1.1%), or 50 mmol/L (66.0% +/- 3.2%) (p < 0.01 versus glucose 1 and 50 mmol/L). An increase in glucose concentration from 1 to 50 mmol/L progressively augmented glucose flux from 2.2 +/- 0.33 to 10.4 +/- 0.79 mumol/gm per 3 hours (p < 0.01), but higher glucose concentrations of 20 and 50 mmol/L inhibited glycogenolysis (p < 0.05 versus glucose 1 and 11 mmol/L), so that total glycolysis was decreased and consequently glycolytic adenosine triphosphate production was reduced from 35.9 +/- 0.47 (glucose 11 mmol/L) to 27.5 +/- 1.25 mumol/gm per 3 hours (glucose 50 mmol/L) (p < 0.01). The end products of glycolysis (lactate and protons) did not appear to affect the recovery of the hearts, because both lactate efflux and tissue lactate were highest in the presence of glucose 11 mmol/L and the pH of the cardioplegic effluent was more alkalotic in glucose 11 and 20 mmol/L. Thus a high physiologic concentration of glucose (11 mmol/L) in the cardioplegic solution improved recovery because of an increased glycolytic adenosine triphosphate production during cardioplegic arrest, whereas even higher concentrations of glucose inhibited these effects.