AIMS: Serial forecasts of final myocardial infarct (MI) size during fibrinolytic treatment (Rx) of ST-elevation MI would allow the identification of high-risk patients with a predicted major loss of viable myocardium, at a point when treatment may still be modified. We investigated a model for such forecasting, using time and the ECG. METHODS AND RESULTS: We collected 234 patients with ST-elevation MI, without signs of previous MI, bundle branch block, or hypertrophy. MI size was determined by the Selvester score and was "forecasted" at: admission with patients stratified by delay time and an ECG acuteness score into three groups (EARLY, DISCORDANT, and LATE); 90 min after Rx by > or =70% ST-recovery or not and occurrence of "reperfusion peaks"; 4 h after Rx by ST re-elevations. EARLY patients had smaller final infarct sizes than LATE (9.4 vs. 20%, P=0.01). EARLY patients with > or =70% ST-recovery without a reperfusion peak had smaller infarct sizes than those with (3.1 vs. 12.5%, P=0.001). EARLY patients without ST re-elevations had smaller infarct sizes (1.5%) than those with some (9%) or many re-elevations (12%), P<0.001. CONCLUSION: Final infarct size can be forecasted using delay time and serial ECGs. Serially updated forecasts seem especially important when both clock-time and initial ECG- signs indicate earliness.