Heart rate turbulence (HRT) is a promising marker for risk of mortality after acute myocardial infarction (AMI). We investigated HRT for risk stratification in high-risk patients after MI. HRT from 24-hour Holter monitoring in 481 hospitalized patients after AMI with heart failure and/or diabetes with left ventricular dysfunction before randomization in the Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study (EPHESUS). Over a 1-year follow-up, 55 died, 49 of cardiovascular causes. HRT onset (TO) and slope (TS) were calculated using previous and cohort-optimized cutpoints and their independent contribution to risk of cardiovascular death determined. Models were tested with <5 ventricular premature complexes (PVCs) categorized as normal (n = 452) and with <5 VPCs excluded (n = 342). In EPHESUS, optimal cutpoints were TS < or = 3.0 and TO > or = 0.0. The strongest model for predicting cardiovascular mortality used EPHESUS cutpoints excluding subjects with <5 VPCs. On 3-category HRT model multivariate analysis (TS and TO normal, TS or TO abnormal, TS and TO abnormal), both TS and TO abnormal (relative risk 3.64, 95% confidence interval 1.55 to 8.55, p = 0.003) and left ventricular ejection fraction < or =30% (relative risk 1.97, 95% confidence interval 1.04 to 3.73, p = 0.037) independently predicted cardiovascular death. In conclusion, HRT is an independent predictor of cardiovascular death in a high-risk population after AMI, with a possibly higher optimal cutpoint for HRT slope than previously reported.