Clinical and experimental data published to date suggest several possible mechanisms by which cocaine may result in acute myocardial infarction. In individuals with preexisting, high-grade coronary arterial narrowing, acute myocardial infarction may result from an increase in myocardial oxygen demand associated with cocaine-induced increase in rate-pressure product. In other individuals with no underlying atherosclerotic obstruction, coronary occlusion may be due to spasm, thrombus, or both. With regard to spasm, the clinical findings are largely circumstantial, and the locus of cocaine-induced vasoconstriction remains speculative. Although certain clinical and experimental findings support the hypothesis that spasm involves the epicardial, medium-size vessels, other data suggest intramural vasoconstriction. Diffuse intramural vasoconstriction is not consistent with reports of segmental, discrete infarction. Whereas certain in vivo data suggest that these effects are alpha-mediated, other in vitro data suggest the opposite. The finding of cocaine-induced vasoconstriction in segments of (noninnervated) human umbilical artery suggests that the presence or absence of intact innervation is not sufficient to explain the discrepant data involving the possibility of alpha-mediated effects. Finally, the contribution of a primary, thrombotic effect of cocaine has not been excluded.