点击显示 收起
the Department of Clinical Neurosciences (E.F., J.L.W.), Brown Medical School, Providence, RI
Department of Neurology (J.P.B.), University of Cincinnati, Ohio
Departments of Internal Medicine (W.N.K., C.M.V., L.M.B.), Neurology (L.M.B.), and Epidemiology and Public Health (L.M.B.), Yale University School of Medicine, New Haven, Conn
Stroke Program (L.B.M.), University of Michigan, Ann Arbor
Mayo Medical School (T.B.), Rochester, Minn
Case Western Reserve University School of Medicine (R.I.H.), Cleveland, Ohio.
Abstract
Background and Purpose— To identify risk factors for intracerebral hemorrhage (ICH), we examined data from the Hemorrhagic Stroke Project (HSP), a case-control study of hemorrhagic stroke among men and women aged 18 to 49 years.
Methods— Case subjects for the HSP were recruited from 44 hospitals in the United States. Eligibility criteria included an ICH within 30 days preceding enrollment, no history of stroke or known brain lesion. For this report, we focused on patients with primary ICH, defined as not associated with an aneurysm, arteriovenous malformation or other structural lesion. Two control subjects were sought for each case subject. A multivariate regression analysis was performed to determine risk factors for primary ICH.
Results— A total of 1714 patients with hemorrhagic stroke were identified for participation in the HSP. Of these, 217 cases met the criteria for primary ICH. Cases with primary ICH were matched to 419 controls. Independent risk factors for ICH included hypertension (adjusted odds ratio [OR], 5.71; 95% CI, 3.61 to 9.05), diabetes (adjusted OR, 2.40; 95% CI, 1.15 to 5.01), menopause (adjusted OR, 2.50; 95% CI, 1.06 to 5.88), current cigarette smoking (adjusted OR, 1.58; 95% CI, 1.02 to 2.44), alcoholic drinks2/day (adjusted OR, 2.23; 95% CI, 1.16 to 4.32), caffeinated drinks5/day (adjusted OR, 1.73; 95% CI, 1.08 to 2.79), and caffeine in drugs (adjusted OR, 3.55; 95% CI, 1.24 to 10.20).
Conclusions— Among young men and women, the major risk factors for primary ICH can be modified, suggesting that this type of stroke may be preventable. Our findings for caffeine and menopause warrant further study.
Key Words: cerebral hemorrhage risk factors
Introduction
Intracerebral hemorrhage (ICH) and subarachnoid hemorrhage (SAH) occur in 55 000 to 60 000 people in the United States every year.1 The mortality of hemorrhagic stroke is 40% to 50%.1 Therefore, prevention of hemorrhagic stroke remains the most important means of reducing its morbidity and mortality. Effective prevention requires understanding the factors underlying the occurrence of ICH and the population subgroups at greatest risk. We have previously reported on the risk factors for aneurysmal SAH in the young.2
Only 20% of ICHs have a clear cause of arterial rupture identified on brain imaging or cerebral angiography.1,3 Most ICHs result from rupture of a small penetrating artery or arteriole within the brain. ICHs in the deep subcortical regions, cerebellum, and brain stem are the most common locations for ICH and are thought to follow damage of small penetrating arteries from sustained hypertension.1,4
In contrast, only 50% to 60% of cases with a lobar or cortical hemorrhage have a history of hypertension.1,5 A small percentage of lobar hemorrhages are attributable to identified structural abnormalities such as vascular malformations or tumors. The identification of amyloid deposition in cortical arterial blood vessels in the 1970s identified amyloid angiopathy as associated with lobar hemorrhages in the elderly (aged >70 years).6 Amyloid deposition in leptomeningeal blood vessels before the age of 60 years is rare in autopsy studies.6
The Hemorrhagic Stroke Project (HSP) was a collaboration between investigators of 4 clinical stroke centers and their surrounding hospitals, the Food and Drug Administration, and manufacturers of products containing phenylpropanalamine (PPA). The main purpose of the project was to examine the association of PPA to risk of hemorrhagic stroke in persons aged 18 to 49 years. Primary results of the HSP have been reported elsewhere.7 The goal of this analysis was to study the risk factors for "primary ICH without a documented structural cause."
Methods
Between December 1994 and July 1999, we identified patients with symptomatic subarachnoid or ICH from 44 hospitals in Connecticut, Massachusetts, Ohio, Kentucky, Rhode Island, and Texas.7 Eligibility criteria for cases included: age 18 to 49 years, ability to communicate and complete the interview within 30 days of stroke event, no previously diagnosed brain lesion predisposing to hemorrhage (ie, arteriovenous malformation, tumor, aneurysm), and no previous stroke. Cases were recruited in-person or by telephone as soon as they were identified, provided their personal physician approved.
An ICH was diagnosed by appropriate symptoms plus a CT or MRI showing blood in the brain parenchyma. For the present analysis, we focused on cases of primary ICH, including only those cases without documented structural cause, such as an aneurysm or arteriovenous malformations. The primary analysis for this report includes patients with either negative angiograms or, if an angiogram was not performed, a CT scan showing no evidence for an underlying lesion. The HSP did not direct physicians in their diagnostic evaluation of these patients.
We attempted to identify 2 matched controls for each case using random digit dialing. Matching criteria included telephone exchange, gender, ethnic group (black versus nonblack), and age. All control subject interviews had to be completed within 30 days of the stroke of the case subject event to minimize seasonal differences in exposures.
Trained researchers used a structured questionnaire to obtain demographic, risk factor, behavioral, and pharmaceutical information from all subjects. Interviews were conducted in-person unless the subject refused or a meeting could not be arranged within 30 days of the focal time of the case subject.7 These interviews were essential to the design of the HSP. Thus, prospective ascertainment of risk factors studied in this analysis represents a prespecified secondary aim of the HSP.
We computed the odds ratio (OR) and associated probability value for the association between ICH and a subject characteristic or exposure using exact conditional logistic modeling for matched sets. To identify independent risk factors, dichotomous features with P<0.10 for the bivariate association with ICH were considered for inclusion in a multivariate logistic model using a forward selection algorithm, with criterion for entry set at P=0.05. The multivariate model was performed using asymptotic methods. Exact logistic models were estimated by the LogXact Program, version 2.1 (Cytel Software Corporation). Adjusted models were estimated using SAS Version 8.0 (SAS Corporation).
Results
Between 1994 and 1999, 1714 patients with hemorrhagic stroke were identified (Table 1). Of the 277 cases of ICH, 217 cases met the criteria for primary ICH. Two control subjects were enrolled for 202 case subjects (93%) and 1 control subject for 15 case subjects (7%). All control subjects were matched to their case subjects on gender and telephone exchange. Age matching was successful for 414 controls (99%), and ethnicity matching was achieved for 404 controls (96%).
In bivariate analyses (Table 2), compared with controls, cases of primary ICH were more likely to have lower educational achievement, have a history of hypertension, diabetes, menopause, and hypothyroidism, to smoke and use alcohol regularly, and to consume caffeine in beverages and drugs.
In the multivariate model of risk for primary ICH, the highest OR for the association with ICH was observed for hypertension (OR, 5.71; 95% CI, 3.61 to 9.05) (Table 3). Other significant independent risk factors included diabetes, menopause, current smoking, consumption of >2 alcohol beverages daily, and consumption of caffeine in beverages or drugs. A separate multivariate model considering packs/day (1/day versus none or <1/day) in model selection suggested a dose response for smoking 1 pack/day as a risk factor (OR, 2.52; 95% CI, 1.47 to 4.32).
When a stratified, adjusted analysis was performed for the variables displayed in Table 3, the ORs for men and women were of similar magnitude for diabetes, cigarette use, caffeine in drugs, and caffeine in drinks. The ORs were higher among men compared with women for hypertension (8.42 versus 3.44) and alcohol use (2.69 versus 1.64).
Discussion
Hypertension is the dominant risk factor for ICH among men and women aged 18 to 49 years in our study, although diabetes, menopause, current smoking, heavy alcohol use, and caffeine in beverages and drugs may also play a role. Because hypertension, smoking, and regular alcohol consumption are modifiable and common in the young, some of the severe morbidity and mortality associated with ICH in persons aged 18 to 49 years is preventable.
In previous population-based case-control studies, ICH has been consistently associated with advancing age, race (black and Asian populations), hypertension, anticoagulant use, and heavy alcohol use.1,8–12 Hypertension is the most important modifiable risk factor in these reports. Risk factors that have been variably related to ICH include prior ischemic stroke, diabetes, and prior aspirin use.12–14 More recently, smoking has been clearly associated with risk for ICH in men15 and women.16 Our data confirm the importance of hypertension and the additional contributions of diabetes, smoking, and regular alcohol use as risk factors for ICH in the young and middle-aged population. Because alcohol use, in moderation, may be protective for ischemic stroke, our data suggest that heavy alcohol use should be reduced, not eliminated. Unfortunately, the location of ICH in the HSP was not determined; therefore, we could not confirm whether the majority of these primary ICHs were in a deep location, classically associated with ICHs attributable to hypertensive arteriolar vasculopathy at autopsy.
We also could not confirm an association between ICH and use of warfarin, as compared with other studies,12,17 probably because of the younger age of our study population as compared with studies that include ICH cases of all ages. The prevalence of coumadin use in persons aged 18 to 49 years is very low: only 4 subjects were exposed to coumadin in our study. In addition, because the HSP excluded cases of ICH and SAH who had any prior stroke, we are unable to address the association between prior ischemic stroke and ICH.
Other studies have reported a higher risk of ICH among persons with hypocholesteremia18 or a lower risk with hypercholesteremia.19 We could not examine this association. The mechanism underlying this relationship is unknown.
Two very recent studies reported familial aggregation of spontaneous ICH in cases of ICH as compared with controls among stroke patients.12,20 We suspect that the lack of familial aggregation in our patients may be attributable to the relatively young age of our population. For example, in elderly patients with lobar ICH, apolipoprotein E4 and E2 alleles have been associated with lobar ICH secondary to presumed or documented amyloid angiopathy of cortical blood vessels.12 Yet, amyloid deposition in cortical blood vessels before the age of 50 years is extremely rare.6 Thus, our population would be unlikely to demonstrate any familial aggregation of ICH related to the genetics of amyloid angiopathy.
Our study demonstrates an association between menopause and primary ICH. There are sparse data on the risk of ICH and menopause outside of trials studying hormone replacement therapy (HRT). Menopausal women have been reported to have an increased risk of hemorrhagic stroke during the first 6 months of HRT21 but not with HRT after an acute myocardial infarction.22 More potent estrogens and the estrogen/progestin combination have been associated with a reduced risk of fatal ICH.23 In the Women’s Health Initiative, estrogen/progestin was not associated with ICH in women aged 50 to 79 years.24 The relationship between menopause itself and ICH is not well understood and warrants further study, particularly because treatment of menopausal symptoms may increase the risk of ischemic stroke. Also, this subgroup is small and may be affected by the young age of our cohort.
Our data also reveal an association of consumption of caffeine in beverages and pharmaceuticals with primary ICH. Our previous report of risk factors for aneurysmal SAH in the HSP reported an association between caffeine in pharmaceuticals and aneurysmal SAH, which we regarded as suggestive.2 Case reports have described stroke associated with pharmaceutical doses of caffeine in combination with PPA25 and studies involving hypertensive animals suggest that high doses of caffeine in combination with PPA may result in ICH.26 To our knowledge, this is the first report of ICH in humans associated with caffeine in beverages or pharmaceutical agents. Our finding is based on a secondary analysis of the HSP. Although suggestive, this should only be regarded as hypothesis generating.
Biases that might have affected this analysis of the HSP include selection bias and recall bias. Selection bias could be related to our exclusion of dead and noncommunicative case subjects. We excluded 34% of screened subjects because of death or inability to participate in the interview. It is possible that risk factors are different for ICH leading to death and impaired communication compared with less severe ICH. Another case-control study of risk factors for ICH is examining subjects with all degrees of severity.12,27 Information on risk factor exposure is being obtained from medical records and, when available, in-person structured interviews. Preliminary reports indicate that the distribution of risk factors among fatal cases of ICH are similar to that among nonfatal cases.12,27 These findings suggest that selection bias related to eligibility criteria may not be a substantial weakness of our study.
Recall bias refers to the tendency of case subjects, compared with controls, to have more or less accurate recall of exposures. Although recall bias is discussed, efforts to demonstrate that it has an important effect on measured associations have commonly failed.28 In the HSP, we adopted several safeguards against recall bias, including a highly structured interview. In addition, to overcome greater stimulation for recall among case subjects, we used a shorter interval between the focal time and interview dates for control subjects.
In summary, hypertension, smoking, and regular heavy alcohol use are important, modifiable risk factors for ICH in the young and middle-aged and present an opportunity for prevention of this devastating type of stroke. Our data should encourage physician identification and treatment of hypertension, smoking, and heavy alcohol use. The association of ICH with menopause and caffeine consumption deserve further investigation.
Acknowledgments
Supported by Novartis Consumer Health Inc, Thompson Medical Company Inc, and Chattem Inc.
References
Broderick J. Intracerebral hemorrhage. In: Gorelick P, Alter M, eds. Handbook of Neuroepidemiology. New York: Marcel Dekker Inc; 1994: 141–167.
Broderick JP, Viscoli CM, Brott T, Kernan WN, Brass LM, Feldmann E, Morgenstern LB, Wilterdink JL, Horwitz RI; for the Hemorrhagic Stroke Project Investigators. Major risk factors for aneurysmal subarachnoid hemorrhage in the young are modifiable. Stroke. 2003; 34: 1375–1381.
Broderick J, Brott T, Tomsick T, Huster G. Intracerebral hemorrhage is more than twice as common as subarachnoid hemorrhage. J Neurosurg. 1993; 78: 188–191.
Challa V, Moody D, Bell M. The Charcot-Bouchard aneurysm controversy: impact of a new histologic technique. J Neuropathol Exp Neurol. 1992; 51: 264–271.
Broderick J, Brott T, Tomsick T, Leach A. Lobar hemorrhage in the elderly. The undiminishing importance of hypertension. Stroke. 1993; 24: 49–51.
Vinters H, Gilbert J. Cerebral amyloid angiography: incidence and complications in the aging brain. II. The distribution of amyloid vascular changes. Stroke. 1983; 14: 924–928.
Kernan W, Viscoli C, Brass L, Broderick J, Brott T, Feldmann E, Morgenstern L, Wilterdink J, Horwitz R. Phenylpropanolamine and the risk of hemorrhagic stroke. N Engl J Med. 2000; 343: 1826–1832.
Thrift AG, McNeil JJ, Forbes A, Donnan GA. Three important subgroups of hypertensive persons at greater risk of intracerebral hemorrhage. Melbourne Risk Factor Study Group. Hypertension. 1998; 31: 1223–1229.
Tanaka H, Ueda Y, Hayashi M, Date C, Baba T, Yamashita H, Shoji H, Tanaka Y, Owada K, Detels R. Risk factors for cerebral hemorrhage and cerebral infarction in a Japanese rural community. Stroke. 1982; 13: 62–73.
Hylek EM, Singer DE. Risk factors for intracranial hemorrhage in outpatients taking warfarin. Ann Intern Med. 1994; 120: 897–902.
Monforte R, Estruch R, Graus F, Nicolas JM, Urbano-Marquez A. High ethanol consumption as risk factor for intracerebral hemorrhage in young and middle-aged people. Stroke. 1990; 21: 1529–1532.
Woo D, Sauerbeck LR, Kissela BM, Khoury JC, Szaflarski JP, Gebel J, Shukla R, Pancioli AM, Jauch EC, Menon AG, Deka R, Carrozzella JA, Moomaw CJ, Fontaine RN, Broderick JP. Genetic and environmental risk factors for intracerbral hemorrhage: preliminary results of a population based study. Stroke. 2002; 33: 1190–1195.
Brott T, Thalinger K, Hertzberg V. Hypertension as a risk factor for spontaneous intracerebral hemorrhage. Stroke. 1986; 17: 1078–1083.
Thrift AG, McNeil JJ, Forbes A, Donnan GA. Risk of primary intracerebral haemorrhage associated with aspirin and nonsteroidal anti-inflammatory drugs: case-control study. BMJ. 1999; 318: 759–764.
Kurth T, Kase CS, Berger K, Schaeffner ES, Buring JE, Gaziano JM. Smoking and the risk of hemorrhagic stroke in men. Stroke. 2003; 34: 1151–1155.
Kurth T, Kase CS, Berger K, Gaziano JM, Cook NR, Buring JE. Smoking and the risk of hemorrhagic stroke in women. Stroke. 2003; 34: 2792–2795.
Wintzen A, de Jonge H, Loeliger E, Bots G. The risk of intracerebral hemorrhage during oral anticoagulant treatment: a population study. Ann Neurol. 1984; 16: 553–558.
Ruiz-Sandoval JL, Cantu C, Barinagarrementeria F. Intracerebral hemorrhage in young people: analysis of risk factors, location, causes, and prognosis. Stroke. 1999; 30: 537–541.
Woo D, Kissela BM, Khoury JC, Sauerbeck L, Haverbusch MA, Szaflarski JP, Gebel JM, Pancioli AM, Jauch EC, Schneider A, Kleindorfer D, Broderick JP. Hypercholesterolemia, HMG-CoA reductase inhibitors, and risk of intracerebral hemorrhage. Stroke. 2004; 35: 1360–1364.
Woo D, Sauerbeck L, Khoury J, Carrozzella J, Moomaw C, Kissela B, Broderick J. Familial aggregation of stroke and intracerebral hemorrhage. Stroke. 2000; 31: 316. Abstract.
Lemaitre RN, Heckbert SR, Psaty BM, Smith NL, Kaplan RC, Longstreth WT Jr. Hormone replacement therapy and associated risk of stroke on postmenopausal women. Arch Intern Med. 2002; 162: 1954–1960.
Angeja BJ, Shlipak MG, Go AS, Johnston SC, Frederick PD, Canto JG, Barron HV, Grady D. National Registry of Myocardial Infarction 3 Investigators. Hormone therapy and the risk of stroke after acute myocardial infarction in postmenopausal women. J Am Coll Cardiol. 2001; 38: 1297–1301.
Schairer C, Adami HO, Hoover R, Persson I. Cause-specific mortality on women receiving hormone replacement therapy. Epidemiology. 1997; 8: 59–65.
Wassertheil-Smoller S, Hendrix SL, Limacher M, Heiss G, Kooperberg C, Baird A, Kotchen T, Curb JD, Black H, Rossouw JE, Aragaki A, Safford M, Stein E, Laowattana S, Mysiw WJ; WHI investigators. Effect of estrogen plus progestin in postmenopausal women: the Women’s Helath Initiative: a randomized trial. J Am Med Assoc. 2003; 289: 2673–2684.
Kikta DG, Devereaux MW, Chandar K. Intracranial hemorrhages due to phenylpropanolamine. Stroke. 1985; 16: 510–512.
Mueller SM, Muller J, Asdell SM. Cerebral hemorrhage associated with phenylpropanolamine in combination with caffeine. Stroke. 1984; 15: 119–123.
Kissela BM, Sauerbeck L, Woo D, Khoury J, Carrozella J, Pancioli A, Jauch E, Moomaw CJ, Shukla R, Gebel J, Fontaine R, Broderick J. Subarachnoid hemorrhage: a preventable disease with a heritable component. Stroke. 2002; 33: 1321–1326.
Aitken J, Bain C, Ward M, Siskind V, MacLennan R. How accurate is self-reported family history of colorectal cancer Am J Epidemiol. 1995; 141: 863–871.