Investigate a strategy to minimize tourniquet-associated reperfusion injury. Sixteen healthy patients scheduled for repair of bimalleolar ankle fractures were randomized into two groups. In the standard release group (R, n1=6), the tourniquet was fully deflated at the end of surgery. In the staggered release group (SR, n2=10), the tourniquet was fully deflated for 30 s and subsequently reinflated to 300 mmHg. The procedure was repeated twice at 3 min intervals before complete removal. Haemodynamic and blood biochemistry measurements were obtained from an indwelling arterial catheter immediately before the initial tourniquet deflation and thereafter at 1, 4, 7 and 15 min. Serum Ca concentrations were less in group R at 4 min (1.027 ± 0.5 vs. 1.084 ± 0.07 mmol/l, P=0.046) and 7 min (1.045 ± 0.04 vs. 1.110 ± 0.06 mmol/l, P=0.013). The serum lactate concentration was higher in group R compared with group SR at 1 min (1.75 ± 0.19 vs. 1.33 ± 0.31 mmol/l, P=0.005) and 4 min (1.98 ± 0.23 vs. 1.48 ± 0.39 mmol/l, P=0.007), respectively. End-tidal CO2 was less in group SR compared with group R at 1 min (4.82 ± 0.45 vs. 5.68 ± 0.26 kPa, P=0.0004) and 4 min (5.01 ± 0.59 vs. 5.68 ± 0.35 kPa, P=0.01), respectively. At 15 min, less hypotension and bradycardia was noted in group SR. A staggered tourniquet release was associated with greater haemodynamic stability and reduced the rate of acute systemic metabolic changes associated with limb reperfusion. Reapplication of a tourniquet seemed to halt further reperfusion, providing a window period for patient evaluation and management.