OBJECTIVE: To evaluate the value of QT interval as a cardiac risk factor in middle aged people. METHODS: The association between QT interval and cardiac risk factors and mortality in a middle aged Finnish population of 5598 men and 5119 women was evaluated over a 23 year follow up. To adjust the QT interval confidently for heart rate, a nomogram was constructed from the baseline electrocardiograms separately for men and women. RESULTS: Nomogram-corrected QT interval (QTNc) prolongation was associated with elevated blood pressure and signs of cardiovascular disease; QTNc shortening was associated with smoking. Over 10% prolongation of QTNc predicted death in men with heart disease: adjusted relative risk (RR) was 2.17 (95% confidence interval 0.67-7.45) for sudden death; 2.12 (1.25-3.59) for total cardiovascular mortality; and 1.92 (1.23-3.00) for all cause mortality. In healthy men the increase in RR was not significant: sudden death, 1.48 (0.67-3.25); total cardiovascular mortality, 1.25 (0.92-1.70); all cause mortality, 1.21 (0.96-1.53). However, healthy men with long QTNc in the lowest heart rate quartile exhibited an RR of 2.75 (1.00-7.40) for sudden death. Over 10% shortened QTNc predicted cardiovascular death in men with heart disease who smoked; RR 3.72 (1.45-9.54). Non-smoking men with short QTNc had low mortality risks irrespective of possible signs of cardiovascular disease. The trends in mortality risks were similar but weaker for women. CONCLUSIONS: In a middle aged population, prolonged QT interval predicts cardiac mortality in men with signs of cardiovascular disease. In women and healthy men this risk is weak and may reflect subclinical heart disease. A shortened QT interval predicts death in men with heart disease who smoke.