The isometric contraction (supramaximal tetanic stimulation) of anterior tibialis muscle was studied in 32 New Zealand white rabbits after 5 hr of ischemia. Reperfusion was achieved after systemic heparinization (100 U/kg) by removal of vascular clamps (normal reperfusion, NR, N = 10); isolated pump perfusion at 15 cc/min for 30 min followed by normal reperfusion (controlled reperfusion, CR, N = 8); CR with a Sepacell 500 filter in the circuit (leukopenic, thrombocytopenic, controlled reperfusion, L/TR, N = 9); or adding 25,000 U of urokinase to the initial reperfusate (UKR, N = 5). Experimental muscle is compared to control nonischemic contralateral muscle in each animal and expressed as percentage of control function. Specimens were studied by light microscopy. No significant difference in mean function at 2 hr was seen between the four groups, with NLR having 53% of control function, CR 55% of control function, L/TR 61% of control function, and UKR 48% of control function. "No reflow," as defined by the absence of Doppler flow signals over the muscle pedicle with no recovery of function during reperfusion and continued incidence of persistent ischemia, was seen in NLR 4/10, CR 5/8, and L/TR 6/9 preparations with arteriolar, capillary, and venule thrombi documented by light microscopy. In contrast, "no reflow" was not seen in UKR (0/5, P less than 0.05). Peak function at any interval (potential maximal recovery) in muscles that adequately reperfused was best in CR (73%) and L/TR (73%). No difference in the degree of injury in adequately reperfused muscles was seen between the four groups.(ABSTRACT TRUNCATED AT 250 WORDS)