SUMMARY To better define the etiologic importance of hypertension for spontaneous intracerebral hemorrhage, hospital records were studied for all patients sustaining intracerebral hemorrhage during 1982 in the Cincinnati metropolitan area. Hypertension pre-dating the hemorrhage was present in 45% (69 of 154), as determined by history. A more inclusive definition of hypertension, combining those with a positive history with those found to have left ventricular hypertrophy by electrocardiogram or cardiomegaly by chest radiography, applied in 56% (87 of 154). The cases were compared to controls with and without hypertension derived from the NHANESII study of blood pressure (n = 16,204) to determine relative risk. For the presence of hypertension by history, the relative risk of intracerebral hemorrhage was 3.9 (95% confidence interval, 2.7 to 5.7). For the inclusive definition of hypertension, the relative risk was 5.4 (3.7 to 7.9). Relative risk was also determined for hypertension in blacks ( = 4.4), age > 70 ( = 7), prior cerebral infarction ( = 22), and diabetes ( = 3).We conclude that the term "hypertensive hemorrhage" should be used very selectively, particularly in whites, and propose that hypertension be viewed as one of several important risk factors for spontaneous intracerebral hemorrhage. Stroke Vol 17, No 6, 1986 THE TERM HYPERTENSIVE INTRACEREBRAL HEMORRHAGE is firmly imbedded in the medical literature.'" 3 When the characteristic high density lesion is seen in the computed tomographic scan (CT) critical differential diagnosis in the setting of acute stroke is too often truncated. Hypertension is quickly assumed to be causal, if present on the admission exam or if detected by history.We have been impressed at the bedside that an adequate history of hypertension is often lacking in patients with spontaneous intracerebral hemorrhage. When an adequate history for hypertension is established, other disorders of potential etiologic importance may be identified but are seldom given serious consideration. In prior studies of intracerebral hemorrhage, the blood pressure measurements (BP) were often those obtained during the acute event, 4 ' 5 or they were not provided. 6 Potential risk factors such as prior cerebral infarction or diabetes mellitus, were infrequently documented. 7 In 2 large studies of intracerebral hemorrhage in which criteria for the definition of hypertension were carefully specified, the results differed sharply regarding the importance of hypertension in pathogenesis. 89 A prospective epidemiologic study of intracerebral hemorrhage and potential risk factors is impractical. A ten-year study would require follow-up of approximately one million subjects to obtain 100 or more index cases. A longer study would require fewer subFrom the