Electrocardiographic (ECG) changes are often associated with subarachnoid hemorrhage (SAH), but it is not well known whether these have prognostic value. The present study retrospectively investigated 122 consecutive patients with SAH caused by ruptured aneurysms. The patients were classified based on the in-hospital outcome into 80 survivors and 42 nonsurvivors. In nonsurvivors, abnormalities often observed on the 12-lead ECG on arrival at hospital were abnormal Q wave, ST depression, and T wave inversion. The ECG score was defined as the total number of leads that had any of these 3 ECG abnormalities. Univariate analysis revealed a strong correlation of in-hospital death with the ECG score, the neurological status estimated by the grading of Hunt and Kosnik, age, and QTc interval. In age- and sex-adjusted multiple logistic regression analysis, the ECG score was the most powerful risk stratifier (ECG score ≥6 vs ECG score <6; p=0.0026, odds ratio 14.39, 95% confidence interval 2.54-81.71). The ECG score, a new and simple method of semi-quantification, was a powerful risk predictor in the present patients with SAH. (Circ J 2002; 66: 567 - 570)