Electrocardiographic (ECG) abnormalities are frequent in patients with intracranial insult. In this study, we evaluated the factors predisposing to the repolarization abnormalities, i.e., prolonged corrected QT (QTc) interval, ischemic-like ECG changes and morphologic end-repolarization abnormalities, and examined the prognostic value of these abnormalities in patients with subarachnoid and intracerebral hemorrhages requiring intensive care. This was a prospective, observational clinical study in a university-level intensive care unit. Clinical characteristics, the level of consciousness, and findings in primary head computed tomography were recorded on admission. The study period was divided into three 2-day sections. In each section, a 12-lead ECG, transthoracic echocardiography, the results of standard blood electrolytes and cardiac troponin I, as well as the rate of vasoactive and sedative drug infusions were recorded. Repolarization abnormalities such as prolongation of the QTc interval (millisecond), ischemic-like ECG changes, and morphologic end-repolarization abnormalities (present/absent) were evaluated and analyzed. The 1-year functional outcome was determined using the Glasgow Outcome Score. During the 2-year study period, 108 patients were included in the study. Different repolarization abnormalities were frequent in both types of hemorrhage. Prolongation of the QTc interval was predisposed by female gender (β, 24.5; P = 0.010) and the use of propofol (β, 30.5; P = 0.001). The predisposing factor for ischemic-like ECG changes were male gender (odds ratio [OR], 5.9; P = 0.003) and for morphological end-repolarization abnormalities aneurysmatic bleeding (OR, 13.0; P = 0.002). Ischemic-like ECG changes were common, in 87/108 patients during the study period, and were associated with a poorer 1-year functional outcome (OR, 4.7; lower 95% confidence interval, 1.5; P = 0.010). Each repolarization abnormality has characteristic predisposing factors. Ischemic-like ECG changes are common and are associated with a poorer 1-year functional outcome.