superoxide was xanthine oxidase.17 Supportive evidence from other laboratories19-23 led to experimental attempts to reduce ischemic and reperfusion injury in animal models of myocardial infarction. In several studies, the addition of catalase was not felt to be necessary in vivo because of the presence of significant amounts in erythrocytes.
Myocardial Infarct Size: Independent VariablesTo test the hypothesis that SOD will reduce infarct size in an animal model, all other significant independent variables must be known and measured. The Animal Models for Protecting Ischemic Myocardium (AMPIM) cooperative study in dogs found that the anatomic area at risk, collateral blood flow, and hemodynamic determinants of myocardial oxygen demand (measured as the ratepressure product) could explain most of the variance in infarct size with the protocol used.24 However, indexes of contractility were not measured. There were small, unexplained differences in infarct size between participating laboratories, and there was lower collateral blood flow in anesthetized than in conscious dogs. The unexplained differences in infarct size and collateral flow suggest the existence of other unmeasured independent variables that may alter collateral flow and thereby explain some of the residual variance. Also, the preparatory surgery may have induced collateral development in the conscious model, in which 80% of the variability could be explained by the measured indexes (including collateral blood flow). The AMPIM study concluded that a 10% (in unconscious animals) to 13% (in conscious animals) reduction in infarct size (p<0.05) would be detected only 50% of the time with 15 animals each in the control and treatment groups.
Statistical ConsiderationsThe AMPIM study clearly shows that in the dog model considerable variation exists in area at risk, collateral flow, and hemodynamic indexes (rate pressure product); thus, any intervention trial should measure and include these variables in the analysis of the results. Several published SOD trials involving animals did not measure collateral blood flow,2,3 whereas in pigs5 and rabbits,13 flow measurements