AIMS--To examine the influence of age on the prediction of sudden death after acute myocardial infarction based on heart rate variability (HRv), left ventricular ejection fraction (LVEF), and the frequency of ventricular extrasystoles. BACKGROUND--Autonomic and left ventricular function and the frequency of ventricular extrasystoles change with age but the influence of age on the prediction of sudden death from these variables has not been examined. METHODS--The 477 patients who had been through an early postinfarction risk stratification protocol and followed up for a mean of 790 days were dichotomised at 60 years of age. RESULTS--Sudden deaths occurred with similar frequency in both age groups (12 (4.7%) of the 256 patients aged < 60 years and seven (3.2%) of the 221 older patients). Sudden death, however, accounted for 52% of all deaths in the young group but only 18.4% of all deaths in the older group (p < 0.01). An HRv index of < 20 units combined with an average of more than 10 ventricular extrasystoles an hour on Holter monitoring (VE10) had a sensitivity of 50%, a positive predictive accuracy of 33%, and a risk ratio of 18 in the young group (p < 0.001) but was not significantly predictive in older patients. The situation was similar when the combination of an LVEF < 40% with VE10 was considered. This combination had a sensitivity of 44%, positive predictive accuracy of 36.4%, and a risk ratio of 16.1 in young patients (p < 0.001), but was not significantly predictive in older patients. The combination of VE10 with either LVEF < 40% or HRv < 20 units gave a sensitivity of 75%, positive predictive accuracy of 30%, and a risk ratio of 30 in young patients (p < 0.001), but the relation between this combination and sudden death in older patients was not statistically significant. CONCLUSION--In postinfarction patients aged < 60 sudden death was a more predominant mode of death and was more reliably predicted from a depressed HRv index, an LVEF < 40%, and VE10 than in older postinfarction patients. These findings may have important implications for post-infarction risk stratification and management.