0895-4356/91 $3.00+ 0.00 Copyrightc 1991PergamonPress plc
J
Ciin Epidemiol Vol. 44, No. 7, pp. 64-648, 1991 Printed in Great Britain.All rightsreserved
EPIDEMIOLOGICAL CHARACTERISTICS OF SUBARACHNOID HEMORRHAGE IN AN URBAN POPULATION RICHARD K. SIMPSON JR,* CHARLES F. CONTANT, DUNCAN K. FISCHER, DAVID A. CECH, CLAUDIA S. ROBERTSON and RAJ K. NARAYAN Department of Neurosurgery, Baylor College of Medicine Houston, TX 77030, U.S.A. (Received in revised form 19 December
1990)
Abstract-Several risk factors for unfavorable outcome from subarachnoid hemorrhage @AH) have been identified. The prevalence of such risk factors varies among ethnic groups and among men and women. The influence of ethnic background and gender as factors in the outcome after SAH has not been adequately studied and is the focus of the present investigation. Outcome in 145 consecutive patients was dichotomized as good and moderately disabled vs severely disabled, vegetative, and dead. A multiple logistic regression model was used to examine the factors of gender, ethnic group (white and non-white), age, admission neurological grade, pre-existing hypertension, and intravenous drug abuse. Our data reveal that hypertensive, white males, with a history of intravenous drug abuse, have a high risk of unfavorable outcome following SAH. These observations are important for the design and interpretation of future studies relating to SAH. Subarachnoid
hemorrhage
Neurologic outcome
INTRODUCTION The distribution of risk factors for many diseases and the access to medical care is known to vary with social and demographic conditions. It is recognized that caution must be exercised in drawing conclusions regarding the role of particular risk factors for a disease without making allowance for socioeconomic factors. Furthermore, once the patient has experienced the disease, the factors which led to having the disease may, or may not, be related to recovery from the disease. We will examine one such disease, subarachnoid hemorrhage (SAH), in an urban hospital setting. We will examine several *All correspondence should be addressed to: Richard K. Simpson Jr, M.D., Ph.D., Department of Neurosurgey, Bavlor Colleee of Medicine. 6560 Fannin. Suite 900.
Houston, TX”77030, U.S.A.’[Tel. (713) 798-4696; Fax (713) 798-37391.
Risk factors
of the reputed risk factors for SAH and how they relate to the probability of a favorable recovery. At the Ben Taub General Hospital, a major teaching affiliate of the Baylor College of Medicine and the major public hospital for the Houston-Harris County area, the vast majority of the patients are indigent. Furthermore, there is substantial representation from three major ethnic groups in the U.S.A.-Blacks, Hispanics, and Whites. This affords us the unique opportunity to examine the influence of purported risk factors for an unfavorable outcome from SAH within a relatively narrow socioeconomic stratum. “Stroke” consists of several distinct pathological entities: vascular occlusion, hypertensive hemorrhage, aneurysmal rupture, and rupture of arteriovenous malformations. Two of these entities: ruptured aneurysms and arteriovenous 641
RICHARDK. SIMPSONJR et
642 ETHNIC
DISTRIBUTION
49 %
Fig.
1
malformations comprise what we will term as subarachnoid hemorrhage (SAH). SAH, which accounts for between 3 and 9% of strokes depending on the community and diagnostic methods [l-3], is defined as the presence of blood in the subarachnoid space that surrounds the central nervous system and normally contains only cerebrospinal fluid [4]. Blood vessels traverse this space and can be damaged by trauma, ischemic infarction, aneurysm, arteriovenous malformation, or neoplastic disease. Under these conditions blood can be rapidly released into the subarachnoid space in massive volumes and cause devastating neurologic injury. Certain ethnic groups appear to be at higher risk for “stroke” than others and this increased risk may not simply be a reflection of differences in socioeconomic status [557]. Furthermore, chronic diseases that affect the cerebral vasculature such as hypertension (HTN), diabetes mellitus (DM), and “lifestyle” risk factors such as smoking and intravenous drug abuse (IVDA), have been found to be related to stroke [8-171. A recent report in this journal by Howard et al. [7] indicated that a fatal outcome from cerebral infarction is associated with HTN and DM. That study also revealed that strokerelated mortality was significantly higher in males and in non-whites. These conditions may also potentially govern neurologic outcome from SAH [18-211. Prior to this study, our clinical impression was that obese, hypertensive, diabetic, black females had the poorest outcomes after SAH. The “lifestyle” and medical risk factors for SAH are similar to those for stroke in general [l-3, 5-71. Hypertension has been implicated as a risk factor for SAH [3, 8, 12, 19, 221 but its role is controversial [23, 241. Diabetes mellitus, by virtue of the effects on the systemic vasculature,
al.
has long been viewed to be a potential risk factor [lo, 14,25-271; however, its contribution to the pathology of SAH has been disputed [lo, 181. Other risk factors for SAH that have been reported include cigarette smoking, alcohol consumption, and the use of oral contraceptives [28-301. Recent reports have also included IVDA, in particular cocaine, as an additional risk for SAH [21]. Factors associated with outcome following SAH have been age [20], clinical grade at admission [31], history of hypertension [ 19,231, elapsed time since ictus [24,32], and surgery [33]. Most of these analyses fail to consider the impact of chronic disease, along with ethnic background and gender, on subsequent outcome from a SAH. In an attempt to clarify these issues the present investigation was undertaken. METHODS
The data in this study were obtained from a retrospective review of the hospital records of 145 male and female patients belonging to Black, Hispanic, or White ethnic groups who were admitted to the Ben Taub General Hospital, Houston, Texas from 1980 to 1987 with the diagnosis of primary SAH. Patients were selected from the Neurosurgical Intensive Care roster based on the admission diagnosis of suspected SAH. From this list, only those with proven primary SAH were included in our study. SAH was defined as blood present in the subarachnoid space as determined by cerebrospinal fluid analysis, cranial computed tomography, or autopsy. All patients included in our study had either an aneurysm or arteriovenous malformation. Patients with SAH secondary to tumors, ischemic infarcts, or trauma, or patients with primary intracerebral hematomas were excluded from analysis. The majority of patients in our study had not been previously treated at our hospital for any medical or surgical condition, and had their initial presentation to the Emergency Center with severe headache, altered mental status, or coma. Some patients were incidentally found by family members in coma of uncertain duration. Admission neurological grade was based on the Hunt-Hess classification system for patients with SAH [3 11.In general, grade I patients have a slight headache and/or mild nuchal rigidity and are otherwise asymptomatic. Grade II patients have severe headache and nuchal rigidity, but no neurological deficit. Patients that are
Epidemiology of Subarachnoid Hemorrhage ADMISSION
NEUROLOGICAL
GRADE
60
50
t g 0 6
40
30
B f 2
20
10
0 I
in
II
Grade (Hunt-Hess
P
Is!
Classification1
Fig. 2
grade III are drowsy and confused and may have a mild focal neurologic deficit. Stupor, severe neurologic deficits, and early vegetative disturbances are found in grade IV patients. Those patients that are grade V are moribund. Pre-existing HTN and DM was defined as having a previous medical diagnosis of the disease and prescribed treatment as reported by the patient or the patient’s family. IVDA was defined as the use of injectable cocaine, by history, within 72 hours prior of admission. Obesity (OBES) was defined as being present when the patient’s weight was at least 20% greater than desired body weight. Outcome
DISTRIBUTION
643
grades, based on the Glasgow Outcome Score [34], were: Good (G), if only minor neurologic dysfunction was present; Moderately Disabled (MD) if a major deficit occurred but the patient functioned independently; Severely Disabled or Vegetative (SD/V) if the patient required constant nursing care or showed no evidence of higher mental function; or Dead (D). Outcome determinations were made at the time of discharge from the neurosurgical service (average length of stay = 25.6 days). For the following analyses, the outcome was dichotomized as G and MD (favorable) vs SD/V and D (unfavorable). Ethnicity was classified as White, Black or Hispanic. Age was categorized as < 30, 30-41, 42-55, and > 55. The admission Hunt-Hess grade was collapsed to two categories, 1 and 2 vs 3, 4 and 5. Bivariate analyses of the relationship between each potential risk factor (gender, ethnic group, HTN, DM, OBES, IVDA, entry grade, age) and the outcome were performed. Finally, a multiple logistic regression model was fit using gender, ethnic group, age, entry grade, and the presence of HTN, DM, OBES, and IVDA. A factor was considered statistically significant if the p-value associated with the regression coefficient was below 0.05. Variables were retained in the model if their p-values were ~0.20 or they were of a priori interest. The dichotomous outcome variable was used.
OF AGE AND GENDER
25
[
0
< 16
16-25
26-35
3645
Age Groups
Fig. 3
46-55
56-65
? ?Female
> 65
]
RICHARDK. SIMPSCIN JR et al.
644
RESULTS
One hundred and twenty-seven patients had cerebral aneurysms and 18 patients had arteriovenous malformations. The patient population comprised 71 Blacks (49%), 41 Whites (28%), and 33 Hispanics (23%). Eighty-three females and 62 males were included in the study. The average age was 44.7 years for females and 39.1 years for males. Below the age of 35 years there was a preponderance of males, while at 36 years of age and over there were more females. Admission neurological grades included 26 grade I, 28 grade II, 52 grade III, 24 grade IV, and 15 grade V patients. A history of pre-existing HTN, diagnosed and treated by a physician, was found in 75 (52%), DM in 22 (15%), OBES in 50 (34%), and IVDA in 17 (11 O/o> of the patients studied. Approximately 51% of patients had surgical treatment and 49% were managed conservatively. Overall, 5 1% of patients had a favorable outcome. Admission neurological grade was a very important factor in predicting outcome (p < 0.0001). Of the patients with good grades (l-2), 81% had favorable outcomes, whereas only 33% of those with poor grades (3-5) had favorable outcomes. In contrast, patient age alone did not predict outcome (p c 0.678). Favorable outcomes were seen in 55% of Table 1. Odds ratios, 95% confidence intervals, and p-values are presented for the individual risk factors under study for a poor outcome from SAH. These include HTN, age, gender, IVDA, OBES, grade, DM, and ethnic background Odds ratio
95% c1t
Probability value$
2.86
1.46; 5.59
0.002
1.22 0.99 1.66
0.49; 3.08 0.41: 2.42 0.65: 4.20
0.670 0.981 0.288
1.69
0.87; 3.29
0.120
9.62
2.12; 43.48
0.001
Obesity Yes vs no*
1.13
0.57; 2.25
0.725
Grade 3-s vs l-2’
8.93
3.97: 20.00
0.000
2.56
0.98; 6.71
0.05 1
1.08
0.50; 2.34 0.37; 1.56
0.845 0.716
Risk factor Hypertension
Present vs absent* Age 30-41 vs 55 41-55 3041 < 30*
*Reference group in logistic model. tCalculated from logistic regression coefficients.
to have higher risks than Blacks (OR = 2.28) while the Hispanics were approximately equal to the Blacks (OR = 1.02). While gender was not statistically significant (p < 0.056), an association between gender and outcome is suspected. A male’s odds ratio were 2.45 times that of a female’s odds ratio of an unfavorable outcome. Hypertensive patients had a much higher risk for unfavorable outcome (OR = 5.27, p < 0.0006). By comparison, the odds ratio for a poor entry grade was 12.08 (p < 0.0001). However, IVDA had the highest odds ratio, calculated to be 14.71 (p < 0.0045). Using the multiplicative model assumed in the logistic model, a hypertensive patient with a history of IVDA admitted with a poor entry grade following a SAH has approximately 5 x 12 x 14 = 840 times the odds of an unfavorable outcome than a non-hypertensive patient, without a history of IVDA and admitted with a higher grade. tended
DISCUSSION
We have observed large and statistically significant associations of poor outcome following SAH with history of hypertension, history of IVDA and admission status of the patient. Admission grade is a measure of the severity of
645
the hemorrhage’s immediate effect on the brain, and is a logical determinate of outcome. The odds ratio for males was elevated, and while not significant at the 0.05 cut-off, was highly suggestive of an association. There was no association of outcome with ethnicity/race or age, though the odds ratio for Whites was somewhat elevated. Diabetes, which in the bivariate analysis was significantly associated with outcome, and obesity were both found to have no significant association with outcome when adjusted for these other factors. The population based estimates of SAH incidence is reported at 13.4/100,000 for men and 15.8/100,000 for women (age standardized) in Auckland, New Zealand [22], 28/100,000 in Framingham, Massachusetts for all persons 30-88 years of age [3], and 15.9/100,000 in south Alabama for all ethnic groups and both genders over age 20 161. Ingall et al. [24] report the average annual incidence for the 40 years from 1945 to 1984 was 11.1 per 100,000 in the Rochester, Minnesota area. While the incidence of SAH has been shown to increase with age [3,22-241, Phillips’ [23] analyses did not indicate age was associated with outcome. Our data also indicated no strong or consistent effect of age. It is possible the effect of the physical damage caused by the hemorrhage, and measured here in the admission grade, overwhelms any effects intrinsically due to age. Females have been observed to have higher specific SAH rates than males age [3,4, 19,22-241, though the magnitude of the difference varies from study to study. Neither Pakarinen [4], nor Phillips et al. [23], found a relationship between gender and outcome. While our results do not contradict their results, they strongly suggest a gender difference in outcome may exist. The role of hypertension in the etiology of SAH is not certain. Sacco ef al. [3] found that both increases in blood pressure and clinical hypertension were associated with increased risk of SAH in the Framingham cohort study. Phillips et al. [23] and Ingall et al. [24] have argued that hypertension is not a risk factor, based on the failure of SAH to decline with the decline in other cerebrovascular diseases. Bonita et al. [22] noted that SAH rates in females were declining with the decline in prevalence in untreated hypertension, leading to their suggestion that hypertension is related to incidence. Phillips et al. [23], and later Ingall et al. [24], found
646
Rcmm
K. SIMPSON JR et al.
hypertension to be associated with outcome in the Mayo Clinic cohort study. Eskesen et al. [19] reported a similar strong association of hypertension with outcome. Sacco et al. [3] noted an excess of hypertension in the fatal cases of SAH but this was no statistically significant. Keller [20] found that survival from SAH was no different in hypertensive and normotensive patients. In large clinical series by Sahs et al. [35] and Graf et al. [36], hypertension was found to affect adversely neurological outcome. While we can shed no light on its etiological role, our results strongly implicate hypertension as a risk factor for unfavorable outcome from SAH. Evidence linking SAH to ethnic background is weak. Black females had higher rates than White females in Alabama, but the sample size was very small. No other studies have compared the ethnic specific incidence of SAH. Only indirect evidence based on prevalence of hypertension in ethnic groups leads to a possible role of ethnic background and outcome from SAH. Large epidemiological investigations have shown that Blacks have a significantly higher incidence of hypertension than Whites which is thought to be related, in part, to environmental and socioeconomic factors [5,6,37]. Data from early reports indicate that the incidence of SAH in Blacks is twice that of Whites [5,6]. No other non-White ethnic group in the U.S.A. has been extensively studied with regard to the epidemiological features of SAH. Specifically, very little data is available on outcome from SAH in the Hispanic population. Hypertension has been shown in a recent report by Samet et al. [38] to be less frequent in Hispanics as compared to Whites. Moreover, the rate of hypertension was found by LopezAqueres et al. [39] to be higher in Hispanic females than males. These observations might lead to the hypothesis that Blacks and Hispanics will have higher rates of SAH, and may have worse outcomes. However, we found no significant relationships between race/ethnic background and outcome from SAH. The odds ratio for Whites was elevated in the logistic model (2.28), but there was no overall difference among the odds ratios. In a recent review, Wier [33] suggested that disparities in the quality and availability of care renders the study of SAH in different ethnic groups of little value. We were fortunate that our data were obtained from one hospital serving a large urban population, yielding a relatively homogeneous socioeconomic group
with a broad ethnic distribution. The approach to management and the quality of care has been uniform for the study population. Several studies have suggested that diabetes is not a significant risk factor for SAH [lo, 14,18,26]. The incidence of DM in SAH patients has been reported by several large series to be between 2 and 3% [2, 181. The high incidence of diabetes in our sample of SAH patients may reflect the socioeconomic or ethnic characteristics of the patients treated at this hospital. Blacks and Hispanics have been reported to have a much higher incidence of diabetes than Whites [5,6,38]. Since diabetes has been found to be a significant risk factor in stroke, control of the disease has been suggested to be a means limiting the danger of intracerebral hemorrhage or ischemic infarction [5]. However, the effect of such management on the risk of SAH or its outcome is not clear. Our data not only did not indicate the “protective effect”, as reported by Adams et al. [ 181, of diabetes on outcome from aneurysmal rupture, but when adjusted for other factors, was not found related to outcome. The prevalence of obesity is high amongst the inner city poor, which is often seen coexisting with diabetes, and is particularly prevalent in the Hispanic population [38,40]. Like diabetes, obesity was not found to exert any obvious influence on post-SAH neurological outcome. Substance abuse, although not unique to urban centers, plays an important role in determining outcome from SAH. Alcohol, tobacco, and IVDA have been shown to have a detrimental effect on outcome [l 1, 16,211. Inner city hospitals have been treating complications from chronic and acute cocaine abuse with increasing frequency [13, 15, 17,211. The incidence of intracranial hemorrhage, ischemic cerebral infarctions, and SAH following cocaine abuse has also been increasing [13, 15, 17,211. That IVDA is prevalent in the inner city hospital population is well known. However, the rate of IVDA within different ethnic groups is unclear and no reliable data are currently available. Our data suggest that outcome from SAH is adversely affected by intravenous cocaine use. Nearly half of cocaine users died and all but two patients had poor functional recoveries. Cocaine users also had greater neurological dysfunction on admission to the hospital. Only two patients (both grade I) had good recoveries. Since the completion of our study, the use of inhaled cocaine or “crack” has replaced the
641
Epidemiology of Subarachnoid Hemorrhage intravenous form of the drug. The effect of crack on outcome from SAH has not been studied. Data concerning cigarette or tobacco use and alcohol consumption were not examined in this study. As both are likely to be associated with outcome, the maldistribution of these two factors could bias the results shown here. Ethnic differences in alcohol and tobacco consumption may well exist and be partly responsible for the racial/ethnic results we report [41]. Time from the onset of SAH to arrival at the hospital was not documented in these data. Phillips et al. [23] and Whisnant et al. [32] noted that about 8-10% of the SAH patients die before arrival in the hospital; mortality is extremely high in the 48 hours that follow the ictus. It is not known what differentials in seeking medical care may exist in our sample. We examined the relationship of grade to the other factors on the grounds that individuals of very poor grade were less likely to have behavioral factors that would influence their decision to seek care. Only age showed a suggestion of a relationship: there was an excess of grades 1 and 2 in the ~30 year old age group. All other variables were not related to grade. It is possible the methods of ascertainment of hypertension, diabetes, and IVDA could be inaccurate, resulting in misclassification. If this misclassification is random, the resulting bias will be towards an odds ratio of 1.0; such a bias would only increase the likelihood of not rejecting the null hypothesis. Non-random misclassification can bias the odds ratio in either direction. This bias might exist in our data, especially those related to history of IVDA, hypertension, and diabetes. The patient’s status was recorded at the time of admission without knowledge of the ultimate outcome. While this does not protect against some bias, it does reduce some of the possible effects. As none of these three factors were related to admission grade, recall bias due to the severity of the patient’s condition is not likely. It should be remembered that severely ill patients’ histories are routinely obtained from relatives and associates. Furthermore, in a public hospital setting, patients without medical histories are routine. As our findings regarding the influence of hypertension and diabetes on SAH are in basic agreement with other reports, we feel strong biases do not exist in these variables. Our initial clinical hypothesis was that nonWhite, obese, hypertensive women were at
greatest risk of poor outcome following SAH. We found no strong ethnic background effect in our sample. Though not strictly significant, there is a suggestion that males were having worse outcomes than females. Obesity was not found related to outcome. We have seen clear risks associated with hypertension, as were admission grade and IVDA. Age was not significantly associated with outcome, nor was diabetes mellitus. The effect of hypertension on the incidence of SAH is somewhat controversial. Our belief is that hypertension is likely a risk factor for SAH, based on the evidence presented by Sacco. Given the strong relationship to poor outcome, prevention of hypertension may be important in reducing post-SAH morbidity, if not the SAH itself. Use of cocaine in these patients had a devastating effect on outcome. While the effect on incidence is unclear, it is our belief that an unfavorable outcome from SAH is highly likely in the intravenous cocaine abuser. Acknowledgement-The authors wish to thank the reviewers for their comments which led to marked improvements in the paper.
REFERENCES I.
2.
3.
4.
5. 6.
I.
8.
9.
Alter M, Sobel E, McCoy RL, Francis ME, Shofer F, Levitt LP, Meehan EF. Stroke in the Lehigh Valley: Incidence based on a community-wide hospital register. Neuroepidemiology 1985; 4: l-15. Mohr JP, Caplan L-R, Melski JW, Goldstein RJ, Duncan GW. Kistler JP. Pessin MS. Bleich HL. The Harvard cooperative stioke registr;: A prospective registry. Neurology 1978; 28: 754-762. Sacco RL, Wolf PA, Bharucha NE, Meeks SL, Kannel WB, Charette LJ, McNamara PM, Palmer EP, D’Agostino R. Subarachnoid and intracerebral hemorrhage: Natural history, prognosis, and precursive factors in the Framingham study. Neurology 1984; 34: 847-854. Pakarinen S. Incidence, aetiology, and prognosis of primary subarachnoid hemorrhage: A study based on 589 cases diagnosed in a defined urban population during a defined period. Acta Neurol Scaud 1967; 743 (Suppl. 29): I-128. Cooper ES. Clinical cerebrovascular disease in hypertensive blacks. J Chin Hypertension 1987; 3: ,/9-84. Gross CR. Dase CS. Mohr JR. Cunnineham SC. Baker WE: Stroke in ‘South Alabama: IncGence and diagnostic features-a population based study. Stroke 1984; 15: 249-255. Howard G, Evans GW, Murros KE, Toole JF, Lefkowitz D, Truscott BL. Cause specific mortality following cerebral infarction. .I Clin Epidemiol 1989; 42: 45-51. Andrews RJ, Spiegel PK. Intracranial aneurysms: Age, sex, blood pressure, and multiplicity in an unselected series of patients. J Neurosurg 1979; 51: 27-32. Aronson JM. Intracranial vascular lesions in patients with diabetes mellitus. J Neuropathol Exp Neural 1973; 32: 183-196.
648
RICHARD
K.
SIMPSON
10. Biller J, Toffol GJ, Kassell NF, Adams HP, Beck DW, Boarini DJ. Spontaneous subarachnoid hemorrhage in young adults. Neurosurgery 1987; 21: 664-667. 11. Colditz GA, Bonita R, Stampfer MJ, Willett WC, Rosner B, Speizer FE, Hennekens CH. Cigarette smoking and risk of stroke in middle-aged women. N Engl J Med 1988; 318: 937-941. 12. Franks AJ. Prognostic factors in ruptured aneurysms of the circle of Willis: the significance of systemic hypertension. Neuropatb Appl Neuroblol 1978; 4: 61-70. 13. Golbe LI, Merkin MD. Cerebral infarction in a user of free-base cocaine (“crack”). Neurology 1986; 36: 1602-l 604. 14. Kane WC, Aronson SM. Cerebrovascular disease in an autopsy population. 3. Diminished frequency of cerebral hemorrhage in diabetics. Trans Am Neurol Assoc 1970; 95: 266-268. 15. Mangiardi JR, Daras M, Geller ME, Weitzner I, Tuchman AJ. Cocaine-related intracranial hemorrhage. Report of nine cases and review. Acta Neurol Stand 1988; 77: 177-180. 16. Stampfer MJ, Colditz GA, Willet WC, Speizer FE, Hennekens CH. A prospective study of moderate alcohol consumption and the risk of coronary disease and stroke in women. N Engl J Med 1988; 319: 267-273. 17. Wojak JC, Flamm ES. Intracranial hemorrhage and cocaine use. Stroke 1987: 18: 712-715. 18. Adams HP, Putman SF, Kassell NF, Torner JC. Prevalence of diabetes mellitus among patients with subarachnoid hemorrhage. Arch Neural 1984; 41: 1033-1035. 19. Eskeen V, Rosenorn J, Schmidt K, Ronde F. Pre-existing arterial hypertension in subarachnoid hemorrhage: An unfavourable prognostic factor. Br J Neurosurg 1987; 1: 455-461. 20. Keller AZ. Hypertension, age and residence in the survival with subarachnoid hemorrhage. Am J Epidemiol 1970; 91: 139-147. 21. Lichtenfeld PJ, Rubin DB, Feldman RS. Subarachnoid hemorrhage precipitated by cocaine snorting. Arch Neural 1984; 41: 223-224. 22. Bonita R, Beaglehole R, North JDK. Subarachnoid hemorrhage in New Zealand: An epidemiological study. Stroke 1983; 14: 342-347. 23. Phillips LH, Whisnant JP, O’Fallon WM, Sundt TM Jr. The unchanging pattern of subarachnoid hemorrhaee in a communitv. Neuroloev 1980: 30: 1034-1040. 24. Ingall TJ, Whisnant*JP, Wieb& DO,‘O’Fallon WM. Has there been a decline in subarachnoid hemorrhage mortality? Stroke 1989; 20: 718-724. 25. Kassell NF, Torner JC. The international cooperative study on timing of aneurysmal surgery-An update. Stroke 1984; 15: 566-570.
JR et al.
26. Greenhalgh RM, Laing S, Taylor GW. Risk factors in carotid artery stenosis and intracranial aneurysms. J Cardiovasc Surg 1980; 21: 559-567. 27. Kannel WB, Dawber TR, Cohen MS. Vascular disease of the brain--epidemiologic aspects: The Framingham study. Am J Public Health 1965; 55: 1355-1366. 28. Inman WHW. Oral contraceptives and fatal subarachnoid hemorrhage. Br Med J 1979: 2: 1468-1470. 29. Bell BA, SymonL. Smoking and subarachnoid hemorrhage. Br Med J 1979; 1: 577-578. 30. Hillbom M, Kaste M. Does alcohol intoxication precipitate aneurysmal subarachnoid hemorrhage? J Neurol Nemosurge Psycbiat 1981; 44: 523-526. 31. Hunt WE, Hess RM. Surgical risk as related to time of intervention in the repair of intracranial aneurysms, J Nenrosure 1968: 28: 14-20. 32. Whisnant J?, Phillips LH, Sundt TM Jr. Aneurysmal subarachnoid hemorrhage. Timing of surgery and mortality. Mayo Clin hoc 1982; 57: 471-475. 33. Weir B. Intracranial aneurysms and subarachnoid An overview. In: Wilkins RH, hemorrhage: Rengachary SS, Eds. Neurosurgery. New York: McGraw-Hill; 1985: Chap. 159, Vol. 2, 1308-1329. 34. Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet 1975; 1: 480-484. 35. Sahs AL, Nishioka H, Torner JC, Graf CJ, Kassell NF, Goettler LC. Cooperative study of intracranial aneurysms and subarachnoid hemorrhage: A long term prognostic study. Arch Neurol 1984; 41: 1147-1151. 36. Graf CJ, Nibbelink DW. Cooperative study on intracranial aneurysm and subarachnoid hemorrhage: Report on a randomized treatment study. Stroke 1974; 5: 559-601. 37. Sprafka JM, Folsom ARE, Burke GL, Edlavitch SA. Prevalence of coronary heart disease risk factors in an urban black population: The Minnesota heart survey. Prev Med 1988; 17: 321-334. 38. Samet JM, Coultas DB, Howard CA, Skipper BJ, Hanis CL. Diabetes, gallbladder disease, obesity and hypertension among Hispanics in New Mexico. Am J Epidemiol 1988; 128: 1302-1311. 39. Lopez-Aqueres W, Kemp B, Plopper M, Staples FR, Brummel-Smith K. Health care needs of the Hispanic elderly. J Am Ceriatr Sot 1984; 32: 191-198. 40. Hazuda HP, Haffner SM, Stern MP, Eifler CW. Effects of acculturation and socioeconomic status on obesity and diabetes in Mexican Americans. Am J Epidemiol 1988; 128: 1289-1301. 41. Otten MW, Teutsch SM, Williamson DF, Marks JS. The effect of known risk factors on the excess mortality of black adults in the United States. JAMA 1990; 263: 845-850.
J
Ciin Epidemiol Vol. 44, No. 7, pp. 64-648, 1991 Printed in Great Britain.All rightsreserved
EPIDEMIOLOGICAL CHARACTERISTICS OF SUBARACHNOID HEMORRHAGE IN AN URBAN POPULATION RICHARD K. SIMPSON JR,* CHARLES F. CONTANT, DUNCAN K. FISCHER, DAVID A. CECH, CLAUDIA S. ROBERTSON and RAJ K. NARAYAN Department of Neurosurgery, Baylor College of Medicine Houston, TX 77030, U.S.A. (Received in revised form 19 December
1990)
Abstract-Several risk factors for unfavorable outcome from subarachnoid hemorrhage @AH) have been identified. The prevalence of such risk factors varies among ethnic groups and among men and women. The influence of ethnic background and gender as factors in the outcome after SAH has not been adequately studied and is the focus of the present investigation. Outcome in 145 consecutive patients was dichotomized as good and moderately disabled vs severely disabled, vegetative, and dead. A multiple logistic regression model was used to examine the factors of gender, ethnic group (white and non-white), age, admission neurological grade, pre-existing hypertension, and intravenous drug abuse. Our data reveal that hypertensive, white males, with a history of intravenous drug abuse, have a high risk of unfavorable outcome following SAH. These observations are important for the design and interpretation of future studies relating to SAH. Subarachnoid
hemorrhage
Neurologic outcome
INTRODUCTION The distribution of risk factors for many diseases and the access to medical care is known to vary with social and demographic conditions. It is recognized that caution must be exercised in drawing conclusions regarding the role of particular risk factors for a disease without making allowance for socioeconomic factors. Furthermore, once the patient has experienced the disease, the factors which led to having the disease may, or may not, be related to recovery from the disease. We will examine one such disease, subarachnoid hemorrhage (SAH), in an urban hospital setting. We will examine several *All correspondence should be addressed to: Richard K. Simpson Jr, M.D., Ph.D., Department of Neurosurgey, Bavlor Colleee of Medicine. 6560 Fannin. Suite 900.
Houston, TX”77030, U.S.A.’[Tel. (713) 798-4696; Fax (713) 798-37391.
Risk factors
of the reputed risk factors for SAH and how they relate to the probability of a favorable recovery. At the Ben Taub General Hospital, a major teaching affiliate of the Baylor College of Medicine and the major public hospital for the Houston-Harris County area, the vast majority of the patients are indigent. Furthermore, there is substantial representation from three major ethnic groups in the U.S.A.-Blacks, Hispanics, and Whites. This affords us the unique opportunity to examine the influence of purported risk factors for an unfavorable outcome from SAH within a relatively narrow socioeconomic stratum. “Stroke” consists of several distinct pathological entities: vascular occlusion, hypertensive hemorrhage, aneurysmal rupture, and rupture of arteriovenous malformations. Two of these entities: ruptured aneurysms and arteriovenous 641
RICHARDK. SIMPSONJR et
642 ETHNIC
DISTRIBUTION
49 %
Fig.
1
malformations comprise what we will term as subarachnoid hemorrhage (SAH). SAH, which accounts for between 3 and 9% of strokes depending on the community and diagnostic methods [l-3], is defined as the presence of blood in the subarachnoid space that surrounds the central nervous system and normally contains only cerebrospinal fluid [4]. Blood vessels traverse this space and can be damaged by trauma, ischemic infarction, aneurysm, arteriovenous malformation, or neoplastic disease. Under these conditions blood can be rapidly released into the subarachnoid space in massive volumes and cause devastating neurologic injury. Certain ethnic groups appear to be at higher risk for “stroke” than others and this increased risk may not simply be a reflection of differences in socioeconomic status [557]. Furthermore, chronic diseases that affect the cerebral vasculature such as hypertension (HTN), diabetes mellitus (DM), and “lifestyle” risk factors such as smoking and intravenous drug abuse (IVDA), have been found to be related to stroke [8-171. A recent report in this journal by Howard et al. [7] indicated that a fatal outcome from cerebral infarction is associated with HTN and DM. That study also revealed that strokerelated mortality was significantly higher in males and in non-whites. These conditions may also potentially govern neurologic outcome from SAH [18-211. Prior to this study, our clinical impression was that obese, hypertensive, diabetic, black females had the poorest outcomes after SAH. The “lifestyle” and medical risk factors for SAH are similar to those for stroke in general [l-3, 5-71. Hypertension has been implicated as a risk factor for SAH [3, 8, 12, 19, 221 but its role is controversial [23, 241. Diabetes mellitus, by virtue of the effects on the systemic vasculature,
al.
has long been viewed to be a potential risk factor [lo, 14,25-271; however, its contribution to the pathology of SAH has been disputed [lo, 181. Other risk factors for SAH that have been reported include cigarette smoking, alcohol consumption, and the use of oral contraceptives [28-301. Recent reports have also included IVDA, in particular cocaine, as an additional risk for SAH [21]. Factors associated with outcome following SAH have been age [20], clinical grade at admission [31], history of hypertension [ 19,231, elapsed time since ictus [24,32], and surgery [33]. Most of these analyses fail to consider the impact of chronic disease, along with ethnic background and gender, on subsequent outcome from a SAH. In an attempt to clarify these issues the present investigation was undertaken. METHODS
The data in this study were obtained from a retrospective review of the hospital records of 145 male and female patients belonging to Black, Hispanic, or White ethnic groups who were admitted to the Ben Taub General Hospital, Houston, Texas from 1980 to 1987 with the diagnosis of primary SAH. Patients were selected from the Neurosurgical Intensive Care roster based on the admission diagnosis of suspected SAH. From this list, only those with proven primary SAH were included in our study. SAH was defined as blood present in the subarachnoid space as determined by cerebrospinal fluid analysis, cranial computed tomography, or autopsy. All patients included in our study had either an aneurysm or arteriovenous malformation. Patients with SAH secondary to tumors, ischemic infarcts, or trauma, or patients with primary intracerebral hematomas were excluded from analysis. The majority of patients in our study had not been previously treated at our hospital for any medical or surgical condition, and had their initial presentation to the Emergency Center with severe headache, altered mental status, or coma. Some patients were incidentally found by family members in coma of uncertain duration. Admission neurological grade was based on the Hunt-Hess classification system for patients with SAH [3 11.In general, grade I patients have a slight headache and/or mild nuchal rigidity and are otherwise asymptomatic. Grade II patients have severe headache and nuchal rigidity, but no neurological deficit. Patients that are
Epidemiology of Subarachnoid Hemorrhage ADMISSION
NEUROLOGICAL
GRADE
60
50
t g 0 6
40
30
B f 2
20
10
0 I
in
II
Grade (Hunt-Hess
P
Is!
Classification1
Fig. 2
grade III are drowsy and confused and may have a mild focal neurologic deficit. Stupor, severe neurologic deficits, and early vegetative disturbances are found in grade IV patients. Those patients that are grade V are moribund. Pre-existing HTN and DM was defined as having a previous medical diagnosis of the disease and prescribed treatment as reported by the patient or the patient’s family. IVDA was defined as the use of injectable cocaine, by history, within 72 hours prior of admission. Obesity (OBES) was defined as being present when the patient’s weight was at least 20% greater than desired body weight. Outcome
DISTRIBUTION
643
grades, based on the Glasgow Outcome Score [34], were: Good (G), if only minor neurologic dysfunction was present; Moderately Disabled (MD) if a major deficit occurred but the patient functioned independently; Severely Disabled or Vegetative (SD/V) if the patient required constant nursing care or showed no evidence of higher mental function; or Dead (D). Outcome determinations were made at the time of discharge from the neurosurgical service (average length of stay = 25.6 days). For the following analyses, the outcome was dichotomized as G and MD (favorable) vs SD/V and D (unfavorable). Ethnicity was classified as White, Black or Hispanic. Age was categorized as < 30, 30-41, 42-55, and > 55. The admission Hunt-Hess grade was collapsed to two categories, 1 and 2 vs 3, 4 and 5. Bivariate analyses of the relationship between each potential risk factor (gender, ethnic group, HTN, DM, OBES, IVDA, entry grade, age) and the outcome were performed. Finally, a multiple logistic regression model was fit using gender, ethnic group, age, entry grade, and the presence of HTN, DM, OBES, and IVDA. A factor was considered statistically significant if the p-value associated with the regression coefficient was below 0.05. Variables were retained in the model if their p-values were ~0.20 or they were of a priori interest. The dichotomous outcome variable was used.
OF AGE AND GENDER
25
[
0
< 16
16-25
26-35
3645
Age Groups
Fig. 3
46-55
56-65
? ?Female
> 65
]
RICHARDK. SIMPSCIN JR et al.
644
RESULTS
One hundred and twenty-seven patients had cerebral aneurysms and 18 patients had arteriovenous malformations. The patient population comprised 71 Blacks (49%), 41 Whites (28%), and 33 Hispanics (23%). Eighty-three females and 62 males were included in the study. The average age was 44.7 years for females and 39.1 years for males. Below the age of 35 years there was a preponderance of males, while at 36 years of age and over there were more females. Admission neurological grades included 26 grade I, 28 grade II, 52 grade III, 24 grade IV, and 15 grade V patients. A history of pre-existing HTN, diagnosed and treated by a physician, was found in 75 (52%), DM in 22 (15%), OBES in 50 (34%), and IVDA in 17 (11 O/o> of the patients studied. Approximately 51% of patients had surgical treatment and 49% were managed conservatively. Overall, 5 1% of patients had a favorable outcome. Admission neurological grade was a very important factor in predicting outcome (p < 0.0001). Of the patients with good grades (l-2), 81% had favorable outcomes, whereas only 33% of those with poor grades (3-5) had favorable outcomes. In contrast, patient age alone did not predict outcome (p c 0.678). Favorable outcomes were seen in 55% of Table 1. Odds ratios, 95% confidence intervals, and p-values are presented for the individual risk factors under study for a poor outcome from SAH. These include HTN, age, gender, IVDA, OBES, grade, DM, and ethnic background Odds ratio
95% c1t
Probability value$
2.86
1.46; 5.59
0.002
1.22 0.99 1.66
0.49; 3.08 0.41: 2.42 0.65: 4.20
0.670 0.981 0.288
1.69
0.87; 3.29
0.120
9.62
2.12; 43.48
0.001
Obesity Yes vs no*
1.13
0.57; 2.25
0.725
Grade 3-s vs l-2’
8.93
3.97: 20.00
0.000
2.56
0.98; 6.71
0.05 1
1.08
0.50; 2.34 0.37; 1.56
0.845 0.716
Risk factor Hypertension
Present vs absent* Age 30-41 vs 55 41-55 3041 < 30*
*Reference group in logistic model. tCalculated from logistic regression coefficients.
to have higher risks than Blacks (OR = 2.28) while the Hispanics were approximately equal to the Blacks (OR = 1.02). While gender was not statistically significant (p < 0.056), an association between gender and outcome is suspected. A male’s odds ratio were 2.45 times that of a female’s odds ratio of an unfavorable outcome. Hypertensive patients had a much higher risk for unfavorable outcome (OR = 5.27, p < 0.0006). By comparison, the odds ratio for a poor entry grade was 12.08 (p < 0.0001). However, IVDA had the highest odds ratio, calculated to be 14.71 (p < 0.0045). Using the multiplicative model assumed in the logistic model, a hypertensive patient with a history of IVDA admitted with a poor entry grade following a SAH has approximately 5 x 12 x 14 = 840 times the odds of an unfavorable outcome than a non-hypertensive patient, without a history of IVDA and admitted with a higher grade. tended
DISCUSSION
We have observed large and statistically significant associations of poor outcome following SAH with history of hypertension, history of IVDA and admission status of the patient. Admission grade is a measure of the severity of
645
the hemorrhage’s immediate effect on the brain, and is a logical determinate of outcome. The odds ratio for males was elevated, and while not significant at the 0.05 cut-off, was highly suggestive of an association. There was no association of outcome with ethnicity/race or age, though the odds ratio for Whites was somewhat elevated. Diabetes, which in the bivariate analysis was significantly associated with outcome, and obesity were both found to have no significant association with outcome when adjusted for these other factors. The population based estimates of SAH incidence is reported at 13.4/100,000 for men and 15.8/100,000 for women (age standardized) in Auckland, New Zealand [22], 28/100,000 in Framingham, Massachusetts for all persons 30-88 years of age [3], and 15.9/100,000 in south Alabama for all ethnic groups and both genders over age 20 161. Ingall et al. [24] report the average annual incidence for the 40 years from 1945 to 1984 was 11.1 per 100,000 in the Rochester, Minnesota area. While the incidence of SAH has been shown to increase with age [3,22-241, Phillips’ [23] analyses did not indicate age was associated with outcome. Our data also indicated no strong or consistent effect of age. It is possible the effect of the physical damage caused by the hemorrhage, and measured here in the admission grade, overwhelms any effects intrinsically due to age. Females have been observed to have higher specific SAH rates than males age [3,4, 19,22-241, though the magnitude of the difference varies from study to study. Neither Pakarinen [4], nor Phillips et al. [23], found a relationship between gender and outcome. While our results do not contradict their results, they strongly suggest a gender difference in outcome may exist. The role of hypertension in the etiology of SAH is not certain. Sacco ef al. [3] found that both increases in blood pressure and clinical hypertension were associated with increased risk of SAH in the Framingham cohort study. Phillips et al. [23] and Ingall et al. [24] have argued that hypertension is not a risk factor, based on the failure of SAH to decline with the decline in other cerebrovascular diseases. Bonita et al. [22] noted that SAH rates in females were declining with the decline in prevalence in untreated hypertension, leading to their suggestion that hypertension is related to incidence. Phillips et al. [23], and later Ingall et al. [24], found
646
Rcmm
K. SIMPSON JR et al.
hypertension to be associated with outcome in the Mayo Clinic cohort study. Eskesen et al. [19] reported a similar strong association of hypertension with outcome. Sacco et al. [3] noted an excess of hypertension in the fatal cases of SAH but this was no statistically significant. Keller [20] found that survival from SAH was no different in hypertensive and normotensive patients. In large clinical series by Sahs et al. [35] and Graf et al. [36], hypertension was found to affect adversely neurological outcome. While we can shed no light on its etiological role, our results strongly implicate hypertension as a risk factor for unfavorable outcome from SAH. Evidence linking SAH to ethnic background is weak. Black females had higher rates than White females in Alabama, but the sample size was very small. No other studies have compared the ethnic specific incidence of SAH. Only indirect evidence based on prevalence of hypertension in ethnic groups leads to a possible role of ethnic background and outcome from SAH. Large epidemiological investigations have shown that Blacks have a significantly higher incidence of hypertension than Whites which is thought to be related, in part, to environmental and socioeconomic factors [5,6,37]. Data from early reports indicate that the incidence of SAH in Blacks is twice that of Whites [5,6]. No other non-White ethnic group in the U.S.A. has been extensively studied with regard to the epidemiological features of SAH. Specifically, very little data is available on outcome from SAH in the Hispanic population. Hypertension has been shown in a recent report by Samet et al. [38] to be less frequent in Hispanics as compared to Whites. Moreover, the rate of hypertension was found by LopezAqueres et al. [39] to be higher in Hispanic females than males. These observations might lead to the hypothesis that Blacks and Hispanics will have higher rates of SAH, and may have worse outcomes. However, we found no significant relationships between race/ethnic background and outcome from SAH. The odds ratio for Whites was elevated in the logistic model (2.28), but there was no overall difference among the odds ratios. In a recent review, Wier [33] suggested that disparities in the quality and availability of care renders the study of SAH in different ethnic groups of little value. We were fortunate that our data were obtained from one hospital serving a large urban population, yielding a relatively homogeneous socioeconomic group
with a broad ethnic distribution. The approach to management and the quality of care has been uniform for the study population. Several studies have suggested that diabetes is not a significant risk factor for SAH [lo, 14,18,26]. The incidence of DM in SAH patients has been reported by several large series to be between 2 and 3% [2, 181. The high incidence of diabetes in our sample of SAH patients may reflect the socioeconomic or ethnic characteristics of the patients treated at this hospital. Blacks and Hispanics have been reported to have a much higher incidence of diabetes than Whites [5,6,38]. Since diabetes has been found to be a significant risk factor in stroke, control of the disease has been suggested to be a means limiting the danger of intracerebral hemorrhage or ischemic infarction [5]. However, the effect of such management on the risk of SAH or its outcome is not clear. Our data not only did not indicate the “protective effect”, as reported by Adams et al. [ 181, of diabetes on outcome from aneurysmal rupture, but when adjusted for other factors, was not found related to outcome. The prevalence of obesity is high amongst the inner city poor, which is often seen coexisting with diabetes, and is particularly prevalent in the Hispanic population [38,40]. Like diabetes, obesity was not found to exert any obvious influence on post-SAH neurological outcome. Substance abuse, although not unique to urban centers, plays an important role in determining outcome from SAH. Alcohol, tobacco, and IVDA have been shown to have a detrimental effect on outcome [l 1, 16,211. Inner city hospitals have been treating complications from chronic and acute cocaine abuse with increasing frequency [13, 15, 17,211. The incidence of intracranial hemorrhage, ischemic cerebral infarctions, and SAH following cocaine abuse has also been increasing [13, 15, 17,211. That IVDA is prevalent in the inner city hospital population is well known. However, the rate of IVDA within different ethnic groups is unclear and no reliable data are currently available. Our data suggest that outcome from SAH is adversely affected by intravenous cocaine use. Nearly half of cocaine users died and all but two patients had poor functional recoveries. Cocaine users also had greater neurological dysfunction on admission to the hospital. Only two patients (both grade I) had good recoveries. Since the completion of our study, the use of inhaled cocaine or “crack” has replaced the
641
Epidemiology of Subarachnoid Hemorrhage intravenous form of the drug. The effect of crack on outcome from SAH has not been studied. Data concerning cigarette or tobacco use and alcohol consumption were not examined in this study. As both are likely to be associated with outcome, the maldistribution of these two factors could bias the results shown here. Ethnic differences in alcohol and tobacco consumption may well exist and be partly responsible for the racial/ethnic results we report [41]. Time from the onset of SAH to arrival at the hospital was not documented in these data. Phillips et al. [23] and Whisnant et al. [32] noted that about 8-10% of the SAH patients die before arrival in the hospital; mortality is extremely high in the 48 hours that follow the ictus. It is not known what differentials in seeking medical care may exist in our sample. We examined the relationship of grade to the other factors on the grounds that individuals of very poor grade were less likely to have behavioral factors that would influence their decision to seek care. Only age showed a suggestion of a relationship: there was an excess of grades 1 and 2 in the ~30 year old age group. All other variables were not related to grade. It is possible the methods of ascertainment of hypertension, diabetes, and IVDA could be inaccurate, resulting in misclassification. If this misclassification is random, the resulting bias will be towards an odds ratio of 1.0; such a bias would only increase the likelihood of not rejecting the null hypothesis. Non-random misclassification can bias the odds ratio in either direction. This bias might exist in our data, especially those related to history of IVDA, hypertension, and diabetes. The patient’s status was recorded at the time of admission without knowledge of the ultimate outcome. While this does not protect against some bias, it does reduce some of the possible effects. As none of these three factors were related to admission grade, recall bias due to the severity of the patient’s condition is not likely. It should be remembered that severely ill patients’ histories are routinely obtained from relatives and associates. Furthermore, in a public hospital setting, patients without medical histories are routine. As our findings regarding the influence of hypertension and diabetes on SAH are in basic agreement with other reports, we feel strong biases do not exist in these variables. Our initial clinical hypothesis was that nonWhite, obese, hypertensive women were at
greatest risk of poor outcome following SAH. We found no strong ethnic background effect in our sample. Though not strictly significant, there is a suggestion that males were having worse outcomes than females. Obesity was not found related to outcome. We have seen clear risks associated with hypertension, as were admission grade and IVDA. Age was not significantly associated with outcome, nor was diabetes mellitus. The effect of hypertension on the incidence of SAH is somewhat controversial. Our belief is that hypertension is likely a risk factor for SAH, based on the evidence presented by Sacco. Given the strong relationship to poor outcome, prevention of hypertension may be important in reducing post-SAH morbidity, if not the SAH itself. Use of cocaine in these patients had a devastating effect on outcome. While the effect on incidence is unclear, it is our belief that an unfavorable outcome from SAH is highly likely in the intravenous cocaine abuser. Acknowledgement-The authors wish to thank the reviewers for their comments which led to marked improvements in the paper.
REFERENCES I.
2.
3.
4.
5. 6.
I.
8.
9.
Alter M, Sobel E, McCoy RL, Francis ME, Shofer F, Levitt LP, Meehan EF. Stroke in the Lehigh Valley: Incidence based on a community-wide hospital register. Neuroepidemiology 1985; 4: l-15. Mohr JP, Caplan L-R, Melski JW, Goldstein RJ, Duncan GW. Kistler JP. Pessin MS. Bleich HL. The Harvard cooperative stioke registr;: A prospective registry. Neurology 1978; 28: 754-762. Sacco RL, Wolf PA, Bharucha NE, Meeks SL, Kannel WB, Charette LJ, McNamara PM, Palmer EP, D’Agostino R. Subarachnoid and intracerebral hemorrhage: Natural history, prognosis, and precursive factors in the Framingham study. Neurology 1984; 34: 847-854. Pakarinen S. Incidence, aetiology, and prognosis of primary subarachnoid hemorrhage: A study based on 589 cases diagnosed in a defined urban population during a defined period. Acta Neurol Scaud 1967; 743 (Suppl. 29): I-128. Cooper ES. Clinical cerebrovascular disease in hypertensive blacks. J Chin Hypertension 1987; 3: ,/9-84. Gross CR. Dase CS. Mohr JR. Cunnineham SC. Baker WE: Stroke in ‘South Alabama: IncGence and diagnostic features-a population based study. Stroke 1984; 15: 249-255. Howard G, Evans GW, Murros KE, Toole JF, Lefkowitz D, Truscott BL. Cause specific mortality following cerebral infarction. .I Clin Epidemiol 1989; 42: 45-51. Andrews RJ, Spiegel PK. Intracranial aneurysms: Age, sex, blood pressure, and multiplicity in an unselected series of patients. J Neurosurg 1979; 51: 27-32. Aronson JM. Intracranial vascular lesions in patients with diabetes mellitus. J Neuropathol Exp Neural 1973; 32: 183-196.
648
RICHARD
K.
SIMPSON
10. Biller J, Toffol GJ, Kassell NF, Adams HP, Beck DW, Boarini DJ. Spontaneous subarachnoid hemorrhage in young adults. Neurosurgery 1987; 21: 664-667. 11. Colditz GA, Bonita R, Stampfer MJ, Willett WC, Rosner B, Speizer FE, Hennekens CH. Cigarette smoking and risk of stroke in middle-aged women. N Engl J Med 1988; 318: 937-941. 12. Franks AJ. Prognostic factors in ruptured aneurysms of the circle of Willis: the significance of systemic hypertension. Neuropatb Appl Neuroblol 1978; 4: 61-70. 13. Golbe LI, Merkin MD. Cerebral infarction in a user of free-base cocaine (“crack”). Neurology 1986; 36: 1602-l 604. 14. Kane WC, Aronson SM. Cerebrovascular disease in an autopsy population. 3. Diminished frequency of cerebral hemorrhage in diabetics. Trans Am Neurol Assoc 1970; 95: 266-268. 15. Mangiardi JR, Daras M, Geller ME, Weitzner I, Tuchman AJ. Cocaine-related intracranial hemorrhage. Report of nine cases and review. Acta Neurol Stand 1988; 77: 177-180. 16. Stampfer MJ, Colditz GA, Willet WC, Speizer FE, Hennekens CH. A prospective study of moderate alcohol consumption and the risk of coronary disease and stroke in women. N Engl J Med 1988; 319: 267-273. 17. Wojak JC, Flamm ES. Intracranial hemorrhage and cocaine use. Stroke 1987: 18: 712-715. 18. Adams HP, Putman SF, Kassell NF, Torner JC. Prevalence of diabetes mellitus among patients with subarachnoid hemorrhage. Arch Neural 1984; 41: 1033-1035. 19. Eskeen V, Rosenorn J, Schmidt K, Ronde F. Pre-existing arterial hypertension in subarachnoid hemorrhage: An unfavourable prognostic factor. Br J Neurosurg 1987; 1: 455-461. 20. Keller AZ. Hypertension, age and residence in the survival with subarachnoid hemorrhage. Am J Epidemiol 1970; 91: 139-147. 21. Lichtenfeld PJ, Rubin DB, Feldman RS. Subarachnoid hemorrhage precipitated by cocaine snorting. Arch Neural 1984; 41: 223-224. 22. Bonita R, Beaglehole R, North JDK. Subarachnoid hemorrhage in New Zealand: An epidemiological study. Stroke 1983; 14: 342-347. 23. Phillips LH, Whisnant JP, O’Fallon WM, Sundt TM Jr. The unchanging pattern of subarachnoid hemorrhaee in a communitv. Neuroloev 1980: 30: 1034-1040. 24. Ingall TJ, Whisnant*JP, Wieb& DO,‘O’Fallon WM. Has there been a decline in subarachnoid hemorrhage mortality? Stroke 1989; 20: 718-724. 25. Kassell NF, Torner JC. The international cooperative study on timing of aneurysmal surgery-An update. Stroke 1984; 15: 566-570.
JR et al.
26. Greenhalgh RM, Laing S, Taylor GW. Risk factors in carotid artery stenosis and intracranial aneurysms. J Cardiovasc Surg 1980; 21: 559-567. 27. Kannel WB, Dawber TR, Cohen MS. Vascular disease of the brain--epidemiologic aspects: The Framingham study. Am J Public Health 1965; 55: 1355-1366. 28. Inman WHW. Oral contraceptives and fatal subarachnoid hemorrhage. Br Med J 1979: 2: 1468-1470. 29. Bell BA, SymonL. Smoking and subarachnoid hemorrhage. Br Med J 1979; 1: 577-578. 30. Hillbom M, Kaste M. Does alcohol intoxication precipitate aneurysmal subarachnoid hemorrhage? J Neurol Nemosurge Psycbiat 1981; 44: 523-526. 31. Hunt WE, Hess RM. Surgical risk as related to time of intervention in the repair of intracranial aneurysms, J Nenrosure 1968: 28: 14-20. 32. Whisnant J?, Phillips LH, Sundt TM Jr. Aneurysmal subarachnoid hemorrhage. Timing of surgery and mortality. Mayo Clin hoc 1982; 57: 471-475. 33. Weir B. Intracranial aneurysms and subarachnoid An overview. In: Wilkins RH, hemorrhage: Rengachary SS, Eds. Neurosurgery. New York: McGraw-Hill; 1985: Chap. 159, Vol. 2, 1308-1329. 34. Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet 1975; 1: 480-484. 35. Sahs AL, Nishioka H, Torner JC, Graf CJ, Kassell NF, Goettler LC. Cooperative study of intracranial aneurysms and subarachnoid hemorrhage: A long term prognostic study. Arch Neurol 1984; 41: 1147-1151. 36. Graf CJ, Nibbelink DW. Cooperative study on intracranial aneurysm and subarachnoid hemorrhage: Report on a randomized treatment study. Stroke 1974; 5: 559-601. 37. Sprafka JM, Folsom ARE, Burke GL, Edlavitch SA. Prevalence of coronary heart disease risk factors in an urban black population: The Minnesota heart survey. Prev Med 1988; 17: 321-334. 38. Samet JM, Coultas DB, Howard CA, Skipper BJ, Hanis CL. Diabetes, gallbladder disease, obesity and hypertension among Hispanics in New Mexico. Am J Epidemiol 1988; 128: 1302-1311. 39. Lopez-Aqueres W, Kemp B, Plopper M, Staples FR, Brummel-Smith K. Health care needs of the Hispanic elderly. J Am Ceriatr Sot 1984; 32: 191-198. 40. Hazuda HP, Haffner SM, Stern MP, Eifler CW. Effects of acculturation and socioeconomic status on obesity and diabetes in Mexican Americans. Am J Epidemiol 1988; 128: 1289-1301. 41. Otten MW, Teutsch SM, Williamson DF, Marks JS. The effect of known risk factors on the excess mortality of black adults in the United States. JAMA 1990; 263: 845-850.
