This multicenter study comprised a group of 900 patients (207 females and 693 males, aged 23-68 years, mean 53) with ischaemic heart disease. Go medications other than nitrates, nifedipine and diuretics were administered at the time of study. In all patients a simultaneous standard 12-lead ecg and a phonocardiogram was registered. QT and QS2 intervals were then measured, and the QT/QS2 index calculated. QT/QS2 ratio 1.0 was considered as a normal one. A 24-h Holter ecg monitoring was performed in each patients, and ectopic ventricular activity was graded according to the Lown's classification. For patients with each class of arrhythmia the mean value of QT/QS2 was calculated. All means were similar, with values 1.0. Proportion of patients with abnormal values of QT/QS2 index was similar in patients showing different Lown classes of arrhythmia. Since a 24-hour monitoring does not give a full information about the arrhythmic events, patients with the history of VT/VF were analyzed separately. In this group an increase of QT/QS2 index was observed significantly more frequently than in other patients (37% vs 19%, p = 0.016). It is concluded that no close relationship exists between QT/QS2 index and the type of ventricular arrhythmia found on the Holter monitoring. However, pathologic QT/QS2 values seem to characterize the patients with increase risk of VT/VF.