The Risk of Subarachnoid and Intracerebral Hemorrhages in Blacks as Compared with Whites

The medical records of all patients who had a possible intracerebral or subarachnoid hemorrhage in the Greater Cincinnati metropolitan area during 1988 were reviewed, including those in the record systems of all the acute care hospitals and coroners' offices in the five-county region, which includes two counties in Ohio and three counties in Kentucky. The catchment area for the 20 hospitals exceeds the geographic area of the five counties, ensuring that all cases of stroke that involved hospitalization were identified. Some cases of intracerebral or subarachnoid hemorrhage may have eluded identification, such as any that occurred in a nursing home or at home without hospitalization and without autopsy. The International Classification of Disease, 9th Revision, Clinical.Modification (ICD-9-CM) codes used for retrieving medical records included those for intracerebral hemorrhage (431 and 432.9), subarachnoid hemorrhage (430), cerebral aneurysm (437.3), arteriovenous malformation (747.81), and cerebrovascular accident (436). The primary and all secondary discharge diagnoses were included as part of the screening process. Since cases were often coded according to discharge date, we screened the records of potential case patients between January 1, 1988, and March 1, 1989, but we included only those with onset of hemorrhage from January 1 through December 31, 1988. All medical records were reviewed by a trained nurseabstractor under the close supervision of a neurologist. The abstracted clinical data and all available CT and magnetic resonance imaging films were evaluated by a neurologist. Films were unavailable for 25 of the 253 hospitalized patients; in these cases the CT report in the medical record was used. Except for those identified by autopsy records alone, all the patients had at least one CT scan performed. A neuroradiologist reviewed the CT films when both intracerebral and subarachnoid hemorrhages were present.

Intracerebral hemorrhage was defined as a clinical history and examination consistent with an intracerebral hemorrhage (sudden onset of headache, change in level of consciousness, or focal neurologic deficit), accompanied by a focal collection of blood within the brain parenchyma that was detected by CT scanning or at autopsy. Traumatic intracerebral hemorrhage and hemorrhagic cerebral infarction were excluded. Subarachnoid hemorrhage was defined as blood in the subarachnoid spaces, not caused by trauma, detected by CT scanning or at autopsy, or a clinical history and examination consistent with subarachnoid hemorrhage (sudden onset of severe headache or change in level of consciousness), with xanthochromia and many red cells in the cerebrospinal fluid. Cases of both intracerebral and subarachnoid hemorrhage were categorized as subarachnoid if the source of bleeding was an aneurysm or if the neuroradiologist's review indicated a subarachnoid origin of bleeding, and as intracerebral if a parenchymal source of bleeding was more likely. Cases of brain hemorrhage limited to the ventricles were classified as intracerebral (five cases). Inclusion in the study required that the patient's residence at the onset of hemorrhage have a ZIP Code in the five-county region. The numbers of blacks (171,718) and whites (1,086,462) in the five counties, according to the 1980 U.S. census, served as the denominators for calculating age- and sex-adjusted incidence rates. Although breakdowns according to age were not yet available from the 1990 census, the population of the five-county region increased by only 2.5 percent between 1980 and 1990, and the black proportion rose from 13.5 percent to 14.4 percent. Confidence intervals for the incidence rates, as adjusted for age and sex to the 1980 U.S. population, were calculated with a Poisson distribution.¹⁰

The overall risks of subarachnoid hemorrhage and intracerebral hemorrhage for blacks and whites were determined with MantelHaenszel tests.^{11 , 12} The overall risks for blacks and whites, as adjusted for age and sex to the 1980 U.S. population, were calculated and compared.¹¹ The rates of survival for 30 days after intracerebral or subarachnoid hemorrhage were calculated for blacks and whites by Kaplan–Meier life-table analysis¹³ and compared by a log-rank test. The number of years of potential life lost before the age of 65,¹⁴ a measure of premature mortality, was calculated for intracerebral and subarachnoid hemorrhage for blacks and whites. All P values were two-tailed.

Examination of 3233 medical records detected 266 cases of first spontaneous intracranial hemorrhage (186 intracerebral and 80 subarachnoid hemorrhages) among whites and blacks during 1988. Two intracerebral hemorrhages and 11 subarachnoid hemorrhages were identified by autopsy reports alone, and 2 subarachnoid hemorrhages were identified through clinical evaluation and lumbar puncture. All the other hemorrhages were documented by CT scanning.

Table 1. Table 1. Incidence of Subarachnoid and Intracerebral Hemorrhage among Blacks and Whites during 1988.*

The 266 patients consisted of 221 whites (159 with intracerebral and 62 with subarachnoid hemorrhages) and 45 blacks (27 with intracerebral and 18 with subarachnoid hemorrhages). The age- and sex-adjusted incidence of subarachnoid hemorrhage per 100,000 population was 12 for blacks (95 percent confidence interval, 6 to 17) and 6 for whites (95 percent confidence interval, 4 to 7). Blacks had a higher incidence of subarachnoid hemorrhage than whites in every adult age group (Table 1). The age- and sex-adjusted incidence of intracerebral hemorrhage per 100,000 population was 19 for blacks (95 percent confidence interval, 12 to 26) and 15 for whites (95 percent confidence interval, 12 to 17). The age-specific incidence rates for intracerebral hemorrhage were higher among adult blacks than adult whites up to the age of 75, after which the rates in whites were substantially higher (Table 1). Overall, the incidence rates for intracerebral hemorrhage increased markedly with age, whereas the rates for subarachnoid hemorrhage increased slightly after the age of 35 (Table 1). The 95 percent confidence intervals for the difference in incidence rates between blacks and whites were 0.3 to 11.5 for subarachnoid hemorrhage, -3.3 to 11.6 for intracerebral hemorrhage, and 0.8 to 19.4 for all hemorrhages combined.

Table 2. Table 2. Risk of Subarachnoid or Intracerebral Hemorrhage among Blacks as Compared with Whites.

Considering all age groups together, blacks had twice the risk of subarachnoid hemorrhage of whites (Table 2). Although the risk of intracerebral hemorrhage for all age groups together was not significantly different among blacks and whites, the risk of intracerebral hemorrhage among blacks under the age of 75 was 2.3 times that for whites (95 percent confidence interval, 1.5 to 3.6). By contrast, for those 75 or older, blacks had approximately one fourth the risk of intracerebral hemorrhage of whites (95 percent confidence interval, 0.1 to 0.8).

Since 95 percent of the black population in the five-county study area lives in Hamilton County, we checked whether white patients with stroke from the area's perimeters could have been treated outside the study region. The nearest hospitals outside the region had no admissions for intracerebral or subarachnoid hemorrhage from our five-county area. Of the 74 blacks who died of cerebrovascular disease in Hamilton County during 1988, 8 (11 percent) had an autopsy, as compared with 7 of 221 whites (3 percent). To determine whether the use of CT scanning and therefore the rate of identification of intracranial hemorrhage differed between blacks and whites, we reviewed the records of all the patients admitted to one of the largest urban study hospitals during 1988 with a diagnosis of "unspecified cerebrovascular disease or cerebrovascular accident" (ICD-9-CM code 436). Of 79 blacks, 67 (85 percent) underwent CT scanning or magnetic resonance imaging, as compared with 52 of 62 whites (84 percent).

The 30-day mortality from subarachnoid hemorrhage was not significantly different among blacks (50 percent) and whites (44 percent), and for intracerebral hemorrhage it was identical (44 percent). The amount of potential life lost because of subarachnoid or intracerebral hemorrhage was 1.9 years per 1000 blacks under 65 years of age, as compared with 0.5 year per 1000 whites.

Two causes of the 80 cases of subarachnoid hemorrhage were verifiable: ruptured aneurysm, as documented by angiography (31 cases), at autopsy (9), or during the operative removal of an associated parenchymal hemorrhage (3); and arteriovenous malformation (2 cases, both documented at autopsy). Of the remaining 35 hospitalized patients with subarachnoid hemorrhage, 11 had normal angiograms. The poor clinical condition of the rest precluded cerebral angiography. No aneurysm was seen in 11 of the 42 patients who underwent angiography. Of the 80 patients with subarachnoid hemorrhage, 41 (51 percent) had a history of hypertension (10 of 18 blacks [56 percent] and 31 of 62 whites [50 percent]).

Causes of the 186 cases of intracerebral hemorrhage included a ruptured arteriovenous malformation (10) and hemorrhage into a tumor (3). Other associated factors included the use of anticoagulant agents (nine), thrombolytic therapy for a pulmonary embolism (one) or myocardial infarction (one), and acute cocaine ingestion (one). Of the 186 patients, 128 (69 percent) had a history of hypertension (110 of 159 whites [69 percent] and 18 of 27 blacks [67 percent]).

This population-based study demonstrates that blacks have a greater risk of subarachnoid hemorrhage than whites. This observation may reflect different distributions of risk factors for subarachnoid hemorrhage,^{15 , 16} such as the increased prevalence of hypertension among blacks¹⁷ and of cigarette smoking among black men.¹⁸ To investigate whether black race is an independent risk factor for subarachnoid hemorrhage¹⁹ would require a large case–control or cohort study.

We expected to find a significantly increased risk of intracerebral hemorrhage among blacks because of their higher prevalence of hypertension,¹⁸ a condition closely associated with hemorrhage.^{4 , 19 20 21 22} Although the overall risk was not significantly different, blacks under the age of 75 proved to have a 2.3 times greater risk of intracerebral hemorrhage than whites. Among blacks 75 or older, the risk of intracerebral hemorrhage fell to only one fourth that for whites. Elderly black patients with intracerebral hemorrhage may be less likely than whites to be evaluated at an acute care hospital and to have the diagnosis corroborated by CT scanning. Alternatively, most blacks at risk for intracerebral hemorrhage may have hemorrhages at earlier ages, leaving those who survive into their 70s and 80s at lower risk. The importance of hypertension, or other risk factors, in the pathogenesis of intracerebral hemorrhage may also vary with age and race.⁴

The increased risk of brain hemorrhage among young and middle-aged blacks mirrors the excess mortality from stroke among blacks in the United States during 1985.²³ In the National Health and Nutrition Examination Survey I Epidemiologic Follow-up Study, cerebrovascular disease accounted for 28 percent of the excess mortality among blacks as compared with whites 35 to 77 years old.¹ In the present study, deaths within 30 days of intracerebral or subarachnoid hemorrhage accounted for 1.9 years of potential life lost per 1000 blacks under the age of 65, as compared with 0.5 year per 1000 whites. Subarachnoid and intracerebral hemorrhages are thus a major cause of excess mortality among young and middle-aged blacks.

Our study suggests that advancing age is more important as a risk factor for intracerebral hemorrhage than for subarachnoid hemorrhage. The incidence of intracerebral hemorrhage increased markedly with advancing age, as reported in previous incidence studies,^{2 3 4 , 8} whereas the rate of subarachnoid hemorrhage increased only slightly after the age of 35. The incidence of both subarachnoid and intracerebral hemorrhage may have been underestimated in the older groups, since the autopsy rate during 1988 for those under 65 years of age (30 percent) was much higher than the rate for those 65 and older (7 percent). However, even with undercounting of subarachnoid and intracerebral hemorrhages in the elderly, the marked age differential between intracerebral and subarachnoid hemorrhages would not have been altered.

Potential biases are unlikely to have confounded our findings. Blacks are reported to be less likely to seek medical attention for chest pain than whites.²⁴ If this behavior also applies to stroke symptoms, the difference we report between blacks and whites in the risk of subarachnoid and intracerebral hemorrhage would be an underestimate of the actual difference. The use of CT imaging among black and white patients hospitalized for stroke was similar in the present study and an earlier three-community study²⁵ and is unlikely to account for the differences between blacks and whites. The lower autopsy rates among whites in the present study may have resulted in a slight underestimate of the actual number of cases among whites. The coding of discharge diagnoses of stroke is performed by medical-record personnel without attention to race and does not explain the observed racial differences. Although unlikely, undercounting of blacks in the census could have led to a slight overestimation of the risk of brain hemorrhage among blacks.