The relation between both time and frequency domain analyses of RR variability and mortality was examined in a series of 226 consecutive patients with acute myocardial infarction admitted to 3 district hospitals in London. All patients underwent 24-hour Holter monitoring early after infarction (mean 83 hours, range 48 to 180), and time and frequency domain analyses of RR variability were performed using commercially available software. During an 8-month follow-up period (range 3 to 12 months), there were 19 cardiac deaths (8.4%). Time domain analysis confirmed reduced RR variability (SDRR, SDANN, SD) among nonsurvivors compared with survivors. However, there was no difference between the groups when the percentage of absolute differences between successive RR intervals > 50 ms (pNN50) and the root-mean-square of successive differences (RMSSD)--vagal measures of RR variability--were analyzed. Frequency domain analysis demonstrated a significant difference between those who died and the survivors when the low-frequency component--modulated by both vagal and sympathetic mechanisms--was analyzed; however, this was less marked when the high-frequency component--modulated by vagal activity--was analyzed. None of these measures of RR variability was related to infarct site or left ventricular ejection fraction. In conclusion, the data confirm the association between low RR variability and mortality after acute myocardial infarction. However, the mechanism does not appear to relate exclusively to decreased parasympathetic tone. The data suggest that the increased risk of early mortality associated with reduced RR variability reflects an imbalance in sympathovagal function that is unrelated to left ventricular function.