Cardiovascular magnetic resonance (CMR) using contrast enhancement allows exact determination of the site and transmural extent of myocardial infarction (MI). We evaluated whether 12-lead electrocardiography can differentiate transmural from non-transmural MI or determine the site of MI by comparing the findings with those of contrast-enhanced CMR. A total of 27 patients (59.5+/-12.9 years) with a history of MI (6.4+/-2.9 months) underwent CMR (Magnetom, Siemens, Erlangen, Germany). Cine images were acquired in the horizontal and vertical long axes and short axis by TrueFISP. Contrast-enhanced CMR images were acquired in the same axes by segmented FLASH 15 min after administration of gadolinium-DTPA (0.15 mmol/kg). This showed the MI to be transmural in 11 patients and non-transmural in 16. An electrocardiogram (ECG) was recorded in all patients before CMR. T-wave alterations, descending ST-depression, pathological Q-waves and absent R waves were more frequent in non-transmural MI than transmural MI, as defined by contrast-enhanced CMR (p> or =0.618). However, none of the differences were statistically significant. R-wave reduction, q waves and horizontal ST-depression were more frequent in transmural than in non-transmural MI (p> or =0.157). Again, the differences were not significant. The sensitivity of the ECG for MI localization was highest in inferior infarctions (85.71%), the specificity was highest in anterior infarctions (100%), the best positive predictive value (80%) was achieved for anterolateral infarctions, and the best negative predictive value for lateral infarctions (95.83%). Transmural and non-transmural MI cannot be differentiated by ECG. The ECG is most accurate in detecting anterolateral MI.