The present study provided a model with which the kinetics of CK release in the early phase of reperfusion was investigated. By using Langendroff method the isolated rat heart was first perfused for 10 min for establishing equilibrium, then stopped for 10 min to establish global ischemia, and finally followed by reperfusion for sample collection in every 15 s for the measurement of CK activity (U/L) as an index of cellular damage. A characteristic biphasic release of CK was shown under condition of 3 min reperfusion with Krebs-Henseleit (K-H) solution without glucose. The 1st peak of CK release appeared abruptly in the first 15 s of reperfusion and the 2nd one, during 120-180 s of reperfusion. The appearance of the 2nd peak was shifted to 30-75 s by adding glucose (11.1 mmol/L) into the perfusate. The 1st peak mainly reflects ischemic injury while the 2nd represents reperfusion injury. Anoxia (95% N2 + 5% CO2) or glucose addition may delay or decrease both peaks, but low Ca2+ (0.05 mmol/L) only delays the appearance of the 2nd peak to 3 min. The results suggest that the oxygen paradox rather than calcium paradox is involved in both phases of CK release. As for low Ca2+ decreasing the 2nd peak may be attributed to its effect of reducing Ca2+ inflow and overload injury secondary to oxygen paradox.