Submaximal exercise systolic blood pressure (ExSBP) is a recognised predictor of cardiac mortality. This study examined the possibility that this might be related to increased QT dispersion or prolonged maximum QT(c) interval (QTcmax). Twenty-nine untreated hypertensive subjects were assessed. Each subject underwent an echocardiographic examination and a 24-h ambulatory blood pressure (ABP). ExSBP was measured during a 3-min lightweight submaximal Dundee step test. In multiple regressional analyses, only left ventricular mass index significantly predicted QT dispersion (R2 = 22.4%, P = 0.018) and QT(c) dispersion (R(2) = 25.3%, P = 0.012). However, with respect to QTcmax, ExSBP (R2 = 21.6%, P = 0.02) emerged as the sole significant predictor of this index. Five (17.2%) out of the 29 subjects had prolonged QTcmax > or = 430 ms and these subjects were not differentiated by 24-h ABP (146 (s.d. 21)/83 (16) vs 140 (14)/88 (11) mm Hg, P = Ns) but by ExSBP (226 (15) vs 188 (24) mm Hg, P = 0.002). In conclusion, systolic blood pressure measured during exercise correlated with QT(c) max in hypertension. This finding may partly explain the prognostic value of exercise blood pressure.