To evaluate the potential prognostic value of the circadian variation of QT intervals in predicting sudden cardiac death (SCD) in patients after myocardial infarction (MI), 15 pairs of post-MI patients (15 died suddenly within 1 year after MI [SCD victims] and 15 remained event-free [MI survivors]) were studied (mean age 60 +/- 8 years; 24 men and 6 women). The pairs were matched for age, gender, infarct site, presence of Q wave, left ventricular ejection fraction, thrombolytic and beta-blocker therapy. Fourteen normal subjects served as controls (mean age 55 +/- 9 years; 12 men). A 24-hour Holter electrocardiographic (ECG) recording was obtained from each subject. All recordings were analyzed using a Holter ECG analyser. QT, RR, and heart rate-corrected QT intervals (QTc) were automatically calculated by the analyzer, and hourly and 24-hour mean values of each measurement were derived from each recording. There was a pronounced circadian variation in the QT interval in parallel with the trend in the RR interval in normal subjects and in MI survivors. Circadian variation in both indexes was blunted in SCD victims. The QT interval was significantly longer at night than during the day in normal subjects (388 +/- 28 vs 355 +/- 21 ms, p = 0.001) and in MI survivors (358 +/- 25 vs 346 +/- 15 ms, p = 0.008), but not in SCD victims (357 +/- 32 vs 350 +/- 31 ms, p = 0.6). The 24-hour mean value of the QT interval in SCD victims did not differ significantly from that in normal subjects or MI survivors. The QT interval at night was significantly shorter in SCD victims than in normal subjects (357 +/- 32 vs 388 +/- 28 ms, p = 0.02), but daytime values were similar. The QT interval in SCD victims did not differ significantly from that of MI survivors at any time. The QTc interval exhibited a small circadian variation in normal subjects. This variation was abolished in SCD victims and MI survivors. The 24-hour mean value of QTc was significantly longer in SCD victims than in normal subjects (424 +/- 25 vs 402 +/- 21 ms, p = 0.02), and in MI survivors (424 +/- 25 vs 404 +/- 32 ms, p < 0.05). The QTc interval of SCD victims differed from that of normal subjects during both the day (421 +/- 25 vs 400 +/- 17 ms, p = 0.02) and night (424 +/- 26 vs 403 +/- 23 ms, p = 0.03). Thus, blunted circadian variation in QT intervals, abolished circadian variation in QTc intervals, and prolonged QTc intervals may suggest an increased risk of SCD in patients after MI.