Abstract

In an attempt to develop improved methods of prediction of infarct size by enzymatic methods, Shell's original algorithm has been critically evaluated in an unselected series of patients. Poor performance of the model is partly the result of a systematic source of error associated with its mathematical formulation. A new model devoid of such limitations has therefore been developed. Residual deviations between predicted and observed CK release seem to be related to frequent and unpredictable extensions of infarction which could be verified by independent clinical, electrocardiographic, and enzymatic criteria. The modified model may possibly be applied to the evaluation of agents aimed at limiting the spread of irreversible injury.