To compare 1-, 6- and 24-hour ambulatory electrocardiograms for prediction of mortality after acute myocardial infarction (AMI), all patients with AMI hospitalized in Hamilton, Ontario during 1 year were identified. There were 683 patients discharged alive after AMI. One-, 6- and 24-hour ambulatory electrocardiographic results were available in 565 patients, and follow-up mortality data at 1 year was available in 560. Mean age of the patients was 64 years; 160 (29%) had previous AMI and 105 (19%) had had congestive heart failure. One hundred and fifty-two patients (27%) were receiving beta blockers, and 31 (6%) were receiving antiarrhythmic drugs. Regression modeling of survival times up to 1 year showed that all 3 durations of recording were univariate predictors of mortality. Using greater than 10 ventricular premature complexes/hour as the criterion of a positive test, neither the 6- nor 24-hour data contained statistically significant residual explanatory power after the 1-hour data were accounted for by the model. The longer durations of recording increased sensitivity at a cost of decreased specificity. The positive and negative predictive values of the 3 durations of recording were virtually identical. The presence of ventricular tachycardia was not a significant predictor of mortality in this population. There appears to be no benefit to ambulatory electrocardiographic recordings greater than 1 hour when they are to be used for prediction of 1-year mortality after AMI.