This study was designed to compare the prognostic value of predischarge ambulatory electrocardiographic monitoring for 1, 6 and 24 hours in 188 patients surviving a first acute myocardial infarction. Ventricular premature complexes (VPCs) were considered as a mean hourly rate or classified using Lown and Moss grading systems. During the 1-year follow-up 20 cardiac deaths occurred. For all 3 monitoring times, a higher number of VPCs/hour and a higher Moss grade were associated with mortality, whereas a Lown grading system gave prognostic information only for the first hour of recording. Monitoring time did not influence specificity or sensitivity in predicting mortality; > or = 3 VPCs/hour showed a higher sensitivity than > or = 10 VPCs/hour (p < 0.05) with a comparable specificity. After 1-hour data entered the model, neither the 6- or the 24-hour data entry improved the overall likelihood ratio statistic, regardless of what VPC grading system was used. These results demonstrate that continuous electrocardiographic recordings of > 1 hour are unnecessary when they are to be used for detecting ventricular arrhythmia as a predictor of mortality in patients surviving a first acute myocardial infarction.