The action potential duration (APD) of ventricular muscles obtained from diabetic animals is reported to be prolonged. We studied the effect of varied periods of diabetes on APD prolongation using isolated ventricular papillary muscles from streptozocin-injected rats. We found that a diabetic period greater than 30 wk was necessary for the evolution of significant prolongation of APD. We then studied the effect of repeated hypoxia (PO2 40 mmHg) and normoxia (PO2 300 mmHg) on prolonged APD of diabetic muscles and compared the findings with those from control rats. Transmembrane potentials were recorded with conventional glass microelectrodes. Under normoxic conditions, the APDs of diabetic muscles were significantly prolonged, the maximum upstroke velocity of action potentials tended to be decreased, and the resting membrane potential was not changed significantly compared with controls. The first hypoxia (20 min) shortened the APD in both diabetic and control rats but more so in diabetic rats, thereby making the APD of diabetic rats virtually identical to the control rat APD during the hypoxia. On subsequent reoxygenation (30 min), these hypoxia-induced changes were almost recovered. The second hypoxia (20 min) produced more severe shortening of the APD in both muscle types, and the effect was again far greater in diabetic rats than controls, ending with a reversed sequence of APDs and the APD of diabetic muscles much shorter than the controls. Excessive shortening of APD in diabetic muscles during hypoxic conditions was speculated to be due to greater increases in the outward K+ current through ATP-regulated K+ channels that may be secondary to the more severe reduction of intracellular ATP concentrations in diabetic versus control hearts.