Serum samples for creatine kinase (CK) activity were obtained serially during 48 h in patients with acute myocardial infarction (AMI) diagnosed according to conventional criteria. Infarct size was estimated as peak enzyme value (CK<sub>max</sub>), area under the enzyme curve measured planimetrically (AUC) and was also calculated according to Sobel et al. (IS) by a computer. In 16 patients with enzyme curves judged to fit the computer method, there was a good correlation between AUC and IS (r = 0.97), between AUC and CK<sub>max</sub> (r = 0.99) and accordingly also between IS and CK<sub>max</sub> (r = 0.96). The results from 34 consecutive patients with AMI were plotted in the graphs obtained from the 16 patients with ‘ideal’ enzyme curves. 15 % fell outside ± 2 SD when AUC was correlated to IS, 3 % when AUC was correlated to CK<sub>max</sub> and 12 % when IS was correlated to CK<sub>max</sub>. There was no difference between enzyme curves falling outside or within ± 2 SD. The explanation for the high correlation between these three methods may be that all these ways of estimating infarct size are imprecise and only give a rough estimation of the size of the infarction.