Serum CPK-MB isoenzyme activity was measured serially after mercury embolization of the left circumflex coronary artery in five baboons and after clinical acute myocardial infarction (AMI) in 20 patients. The calculated amount of enzyme released into the baboons' circulation (CPK-MB(R)) correlated well (r=0.991) with the amount of MB isoenzyme depleted from the myocardium (CPK-MB(D)) when a previously determined decay constant (Kd) was used, but not when Kd was calculated from individual curves or when CPK-MM values were used. In clinical AMI, CPK-MMR averaged 97% (0 to 350%) greater than CPK-MB(R), probably because of release of MM isoenzyme from nonmyocardial sources. The mean Kd for CPK-MB in patients (0.0012 min-1) was significantly (P <0.01) lower than that obtained in the baboon following bolus injections (0.0018 min-1), probably reflecting delayed myocardial release of enzyme. Therefore, in both experimental and clinical AMI, serial samples for CPK-MB activity, but not total or CPK-MM activity, could provide an accurate index of myocardial enzyme depletion.