Skip main navigation

Low Total Serum Cholesterol and Intracerebral Hemorrhagic Stroke: Is the Association Confined to Elderly Men?

The Kaiser Permanente Medical Care Program
Originally published 1996;27:1993–1998


    Background and Purpose Epidemiological studies indicate a higher incidence of intracerebral (but not subarachnoid) hemorrhagic stroke among persons with low total serum cholesterol levels. This report further examines the prospective relationship of total serum cholesterol with subsequent intracerebral hemorrhage in a large, well-defined population.

    Methods The cohort included 61 756 enrollees in a health plan from the San Francisco–Oakland metropolitan area (46% men, 63% white), aged 40 to 89 years and free of cardiovascular disease at baseline. Sixteen-year incidence of combined nonfatal and fatal intracerebral hemorrhagic stroke (International Classification of Diseases [ICD], 8th revision, code 431, or ICD, 9th revision, codes 431 and 432) was investigated in relation to serum cholesterol measured in multiphasic health checkups made in 1977 through 1985. Intracerebral hemorrhagic events were ascertained using hospital discharge records and as underlying cause of death by the California Mortality Linkage Information System.

    Results From 1978 through 1993 (average of 10.7 years), there were 386 events (201 in men, 29% fatal; 185 in women, 42% fatal). By multivariate proportional hazards life-table regression analysis, serum cholesterol level below the sex-specific 10th percentile (<4.62 mmol/L [178 mg/dL] in men), compared with higher cholesterol level, was associated with a significantly increased risk of intracerebral hemorrhage in men aged 65 years or older (relative risk, 2.7; 95% confidence interval, 1.4 to 5.0). An excess risk was also observed among elderly women at the lowest cholesterol range, but a chance finding could not be ruled out. No relationship was seen among men or women aged 40 to 64, and no statistical interaction of low serum cholesterol with hypertension was found in either sex.

    Conclusions In these data, the association between low serum cholesterol level and intracerebral hemorrhage was confined to elderly men.

    In prospective studies, hemorrhagic stroke has been found to occur at higher rates in persons with low levels of blood total cholesterol than in persons with higher levels. This peculiar lipid-stroke relationship has been reported in Japanese populations,1234567 in Japanese men living in Hawaii,89 and among white men101112 and women.13 Although the reasons behind this association are unknown, several animal experiments14151617 and a recent autopsy study18 have indicated that a very low circulating cholesterol concentration may influence arteriolosclerosis, the pathological process leading to hemorrhage in the intracerebral penetrating arterioles (200 μm in diameter).

    The primary aim of this study was to assess the risk relation between serum cholesterol level and subsequent incidence of ICH stroke in a large, well-defined, ethnically diverse population. A secondary aim was to examine whether the association between serum cholesterol and the outcome under investigation varied by age or by the presence of hypertension.

    Materials and Methods

    The setting for this investigation was the Northern California Kaiser Permanente Medical Care Program, a health plan that serves about 25% of the San Francisco Bay area population. The study was approved by the Institutional Review Board of the Kaiser Foundation Research Institute.

    The Kaiser enrollees are socioeconomically and ethnically diverse but, on average, more educated than the general population.19 A data file containing annually determined membership has been maintained since 1976. On enrollment, members underwent a voluntary multiphasic health checkup examination during which detailed information on sociodemographic factors (age, sex, race, educational attainment), personal medical history (self-reported preexisting medical conditions), and health habits (history of cigarette smoking and alcohol consumption) was collected. The alcohol questionnaire classified persons by the use of alcoholic beverages in the past year as abstainers (participants who responded that they never or almost never drank alcohol), former drinkers, and current drinkers. Current drinkers were categorized by the average number of drinks per day consumed during the past year (ie, special occasions only, <1 per day, 1 to 2 per day, 3 to 5 per day, 6 to 8 per day, and ≥9 per day). Physiological measurements such as weight, height, and systolic and diastolic blood pressures and blood chemistry tests were also performed according to standardized procedures.20 Total serum cholesterol was measured after subjects had fasted for at least 8 hours. All the laboratory analyses were performed by the Kaiser Permanente Regional Laboratory (Berkeley, Calif), using enzymatic methods on the American Monitor KDA machine from 1979 through 1982 and on the Abbott VP machine thereafter through 1985. Calibration standards were referenced to the Abell-Kendall method.21 Hypertension was defined as systolic blood pressure above 140 mm Hg and diastolic blood pressure above 90 mm Hg or self-reported history of hypertension.

    The original study population included 158 449 Kaiser members aged 20 to 89 years who had a nonmissing cholesterol determination. Because the incidence of ICH was negligible among those younger than 40 years (n=86 606; rate=0.03%), the “at-risk” population was defined as those aged 40 to 89 years (n=71 843). To avoid potential influence of the disease itself on the level of serum cholesterol directly or indirectly via lifestyle change, 4473 individuals were excluded because of a history of self-reported cerebrovascular disease (n=798) or at least one of the following nonmutually exclusive medical conditions: subdural hematoma of traumatic origin (ICD-9 codes 850 to 854, n=266), brain tumor (ICD-9 191 to 225, n=279), encephalitis (ICD-9 323, n=30), AIDS encephalopathy (ICD-9 348.3, n=66), toxoplasmosis (ICD-9 130, n=17), multiple sclerosis (ICD-9 340, n=84), diabetic coma (ICD-9 250.0, n=3516), hepatic coma (ICD-9 572.2, n=129), and uremic coma (ICD-9 586, n=17). These conditions were ascertained from corresponding ICD codes in computerized hospital discharge abstracts. Of the remaining 67 370 members, 5614 had missing values on race, weight, or height and were excluded. Thus, 61 756 individuals were included in the present study (28 452 men and 33 304 women). Subjects with missing data on race, weight, or height did not differ significantly in other risk factors from the rest of the cohort (data not shown).

    Nonfatal ICH strokes were ascertained from computer files of all overnight hospitalizations in the Northern California Kaiser Centers. Fatal ICH strokes were ascertained as underlying cause of death by CAMLIS, a probability linkage system based on name, date of birth, race, place of birth, and social security number, which has been satisfactorily validated against the National Death Index.22 Mortality out of the state of California in the Kaiser population has been estimated to be only 2.2%.22 The follow-up period of observation in the present study was 16 years (1978 through the end of 1993), the average±SD follow-up time being 10.7±4.7 years. The hospitalization file included the patient's miscellaneous personal information, admission date, discharge date, primary discharge diagnoses, and diagnostic procedures. Both discharge diagnoses and the CAMLIS underlying cause of death were coded according to the ICD-8 (code 431) through 1978. Starting in 1979, the ICD-9 (codes 431 and 432) was used. A validity study undertaken by one of us (C.I.) using 50 randomly selected hospital records of patients with an ICH event found that 91% (44 of 48, two charts not found) had a diagnosis of ICH confirmed by CT of the head. The four events without a CT diagnosis occurred before the introduction of this diagnostic procedure in the Kaiser-affiliated hospitals in 1983. The analysis presented here is based on all events because the exclusion of events before 1983 (n=39; 12 in men, 27 in women) did not materially alter the results.

    Possible confounding conditions, potentially associated with both low serum cholesterol level and hemorrhagic disease, were identified from the study inception in 1978 until occurrence of an event, end of membership, or termination of follow-up, whichever came first. These conditions were ascertained from the multiphasic health checkup questionnaire and hospital discharge records, whether there was a diagnosis of hemorrhagic stroke or not, and included liver disease (from the question “Do you now have liver trouble?”; n=563), bleeding disorders (congenital factor VIII [286.0], factor IX [286.1] and factor XI [286.2] disorders, congenital disorders of other clotting factors [286.3], von Willebrand's disease [286.4], hemorrhagic disorders due to circulating anticoagulants [286.5], defibrination syndrome [286.6], vitamin K deficiency [286.7], other and unspecified coagulation defects [286.9], purpura [287], and Rendu-Osler-Weber syndrome [448]; n=565), vasculitis (systemic lupus erythematosus [710.0], polyarteritis nodosa [446.0], and other collagen vascular diseases [710.9]; n=52), and fibromuscular dysplasia (447.8; n=12). We also searched for congenital malformations of the cerebral vessels (747.81), but no cases were found.

    Incidence rates of combined nonfatal and fatal ICH events were calculated by sex and level of total serum cholesterol in two age strata: 40 to 64 and 65 to 89 years old. Seven groups of serum cholesterol level, based on sex-specific distribution of serum cholesterol among the “at-risk” cohort, were considered: <4.27 (<5th percentile), 4.27 to 4.61 (5th to 10th percentile), 4.62 to 5.05 (11th to 20th percentile), 5.06 to 5.64 (21st to 40th percentile), 5.65 to 6.19 (41st to 60th percentile), 6.20 to 6.89 (61st to 80th percentile), and ≥6.90 mmol/L (≥81st percentile) in men; and <4.35, 4.35 to 4.69, 4.70 to 5.13, 5.14 to 5.77, 5.78 to 6.37, 6.38 to 7.12, and ≥7.13 mmol/L in women, respectively. These cut points were chosen so as to provide a comprehensive description of the risk relation under investigation and to assess the exposure to very low cholesterol levels while yielding a reasonable number of events per cholesterol group for analysis. The association of serum cholesterol (both as a linear and a quadratic trend variable and for less than the sex-specific 10th percentile versus higher levels) with ICH stroke was assessed by proportional hazards life-table regression,23 adjusting for age and for multiple covariates. After graphical inspection of the data and testing whether the hazard of low cholesterol changed over time, there was no evidence of violation of the proportionality assumption of the Cox model. Covariates in the multivariate analysis included race (black, Asian, other, relative to white), educational attainment (vocational school or less versus some college or higher), body mass index (weight in kilograms per height in meters squared), smoking status (former, current, relative to never-smokers), alcohol consumption (former, occasional-light-moderate [up to 3 drinks per day], high [≥3 drinks per day], relative to abstainers), systolic blood pressure, individual confounding conditions (liver disease, bleeding disorders, vasculitis, and fibromuscular dysplasia), and serum glucose concentration, separately for men and women. Participants leaving the health plan during 2 consecutive years or more were censored from the analysis, but 1-year gaps in membership were allowed. A possible interaction of low cholesterol with hypertension was investigated by including a linear cholesterol by hypertension status term in age-adjusted sex-specific models. We were unable to examine the association of cholesterol subfractions or apolipoproteins with ICH stroke because these parameters were not measured at the multiphasic health checkup. In addition, the hospital database did not provide information on confirmation of diagnosis by autopsy.


    Selected baseline characteristics of the cohort are presented in Table 1. Mean serum cholesterol concentrations were 5.98 mmol/L (231 mg/dL) in men and 6.16 mmol/L (237 mg/dL) in women. The level of cholesterol increased with age, more markedly in women: it was 6.06 mmol/L (234 mg/dL) among women aged 40 to 64 years and 6.63 mmol/L (256 mg/dL) among women aged 65 to 89 years. On the other hand, it was 5.98 mmol/L (231 mg/dL) among men aged 40 to 64 years and 6.03 mmol/L (233 mg/dL) among men aged 65 to 89 years. The percentages of whites (about 63), current smokers (about 26), and hypertensive persons (about 27) were similar across sexes. More men (14%) than women (5%) reported consuming an average of 3 alcoholic drinks or more per day.

    During the follow-up period (1978 to 1993), 386 ICH strokes (201 in men, 185 in women) were documented. Of those, 58 in men (29%) and 77 in women (42%) were fatal.

    Table 2 shows the rates (per 100 persons) of ICH by level of serum cholesterol, sex, and age. In men aged 40 to 64 years, no clear trend was discernible. By contrast, there was a statistically significant inverse linear trend (P=.02) in men aged 65 years and older, with higher rates at the lowest two cholesterol categories (among whom 14 events occurred). No trend was noted for ICH among women aged 20 to 64 years old. Although rates among women aged 65 to 89 years were highest in the low-cholesterol groups (as a result of 5 events), the tests for linear or quadratic trends were not statistically significant.

    Table 3 presents the results of the univariate and multivariate life-table regression of ICH on a categorical cholesterol variable (cholesterol below the 10th percentile versus higher level). In men, there was a significantly elevated risk of ICH stroke below the 10th percentile of total cholesterol in the older age group (65 to 89 years; RR, 2.56), whereas there was an inverse relation among those aged 40 to 64 years (RR, 0.73). Accordingly, there was a highly statistically significant interaction between serum cholesterol and age (P=.0008). Because of reported inverse associations between total serum cholesterol level and hemorrhagic stroke in previous studies among middle-aged men,810 we conducted a separate analysis for men aged 55 to 64 years. The adjusted hazard ratio for ICH among men aged 55 to 64 with serum cholesterol level below 4.62 mmol/L (178 mg/dL) compared with men of the same age group with higher serum cholesterol level was 1.24 (95% CI, 0.49 to 3.11).

    Among women, an excess risk of fatal ICH was also observed in those with cholesterol below the 10th percentile (4.70 mmol/L) who were aged 65 years and above, but this estimate fell short of statistical significance. Adjustment for confounders had little effect on these results. Furthermore, these findings were unaltered when the analysis was restricted to events that occurred after 1983, when CT diagnosis was widely introduced.

    We compared RRs for ICH associated with low total serum cholesterol level according to hypertension status in men and women (Table 3). The interaction between serum cholesterol and hypertension status was not statistically significant in either sex (P=.95 among men, P=.08 among women).

    The effects of serum cholesterol appeared to be similar among whites and blacks (data not shown). The statistical power was too small for meaningful interpretation of associations between serum cholesterol and ICH among Asians or persons of “other” ethnicity. However, the conclusions were unchanged when the analysis was restricted to whites (data not shown).

    Summary estimates of the association of other risk factors at baseline with ICH are provided in Table 4. In men, the significant predictors were age, black race, Asian race, other race, diastolic blood pressure, and bleeding disorders. Among women, the significant predictive factors of ICH were age, current smoking, high alcohol consumption, diastolic blood pressure, and bleeding disorders. In a model that excluded diastolic blood pressure, a 1 SD increase in systolic blood pressure was a significant risk factor in men (RR, 1.28; 95% CI, 111 to 1.47) and in women (RR, 1.34; 95% CI, 1.16 to 1.55).


    In this study of a large, ethnically diverse population, the association of serum cholesterol level with subsequent ICH stroke was inconsistent across age groups. Inverse relationships were found in the elderly (although a chance association could not be ruled out in elderly women), but no relation was observed among middle-aged adults (40 to 64 years). Alternatively, the not statistically significant relation in elderly women may be attributable to limited statistical power, since there were only 219 women with cholesterol levels below the 10th percentile point, and only five events occurred among these women.

    Thus, our findings suggest that low cholesterol concentrations may contribute to the risk of suffering an ICH vascular accident in elderly men. These results differ from prior studies linking low levels of serum cholesterol with hemorrhagic stroke in middle-aged Japanese men89 and with hypertensive middle-aged men (diastolic blood pressure ≥90 mm Hg) screened for the Multiple Risk Factor Intervention Trial.1011 One explanation for this dissonance is the ability of our study to identify and exclude persons with concomitant or prior medical conditions potentially related to cerebral hemorrhagic disease, such as traumatic subdural hematoma, brain tumors, encephalitis, multiple sclerosis, AIDS encephalopathy, toxoplasmosis, and diabetic, hepatic, or uremic coma. Additionally, the lack of a relation among those subjects aged 40 to 64 years may also be attributable to probable better health associated with the predominantly employed status of our insured population.

    A potential bias of this study is misclassification of outcome (ie, true ICH strokes being coded as “undefined” strokes and therefore not identified or nonhemorrhagic strokes being coded as such in the hospital ICD codes). However, the likelihood of this source of bias appears to be small, for two reasons. First, the results were virtually unchanged when the analysis was repeated excluding events before the widespread availability of CT diagnosis (ie, events before 1983). Second, the diagnosis was sufficiently accurate to show consistent and strong risk relations for another reasonable risk factors, namely, blood pressure and bleeding disorders.

    The crucial point is whether low levels of serum cholesterol are directly or indirectly implicated in the pathogenesis of hemorrhagic stroke in elderly men. The chain of events in this type of stroke is initiated by injury to the medial muscle cells, causing blood plasma insudation, histolysis, and intimal deposition of fibrinoid substance.1415 The process then leads to some intimal thickening with fibrous tissue. The resulting arteriolosclerosis, associated with high arterial pressure, promotes the formation of intracerebral microaneurysms and their rupture. The causes of the medial muscle cell damage, which was the primary change of the arteriolonecrosis, are considered to be hypertension,14 aging,14 and diets low in cholesterol,162425 animal protein,25 and potassium.26 Low serum cholesterol level may then be a true harbinger of the disease, a surrogate for nutritional deficiencies, or a nonspecific sign of a debilitating disease associated with aging. The existence of genetic traits simultaneously predisposing to hypocholesterolemia and coagulation disorders remains to be elucidated.

    The influence of dietary factors could not be explored in this analysis because no information on habitual food intake was routinely obtained. The fact that the inverse association observed in men at age 65 to 89 years was independent of high alcohol intake and identified chronic diseases (ie, liver disease, collagen diseases, and bleeding disorders) argues in favor of low cholesterol as an etiologic factor for intracranial bleeding in the elderly. However, there may be other unknown or unmeasured factors for which adjustment was not performed, and thus residual confounding remains likely.

    High alcohol intake was a statistically significant risk factor for ICH stroke in women. Alcohol consumption may contribute to the development of ICH by several pathways: increased blood pressure (via increased cortisol and catecholamine production and alteration of intracellular calcium),27 reduction of platelet aggregation (possibly via alteration of platelet composition),282930 deranged coagulation as a result of interaction of acetaldehyde (the first metabolite of ethanol) with coagulation proteins,31 low levels of fibrinogen and von Willebrand factor,32 cardiomyopathy,33 and changes in cerebral blood flow.34 It has also been proposed that liver dysfunction associated with alcohol abuse may be a causative factor in the development of intracranial hemorrhage via impairment of the hemostatic system (ie, platelet aggregability and α2-antiplasmin activity).35

    Although the association of cigarette smoking with thrombotic and subarachnoid hemorrhagic stroke is well established,36373839 the published data are generally not supportive of an association between smoking and ICH.40 Our study suggests an association between current smoking and ICH in women but not in men.

    The public health implications of low cholesterol levels in the context of hemorrhagic stroke risk are probably not large for American men and women. First, arterial bleeding into the brain parenchyma accounts for only 10% of all strokes.40 Second, in these data, the relation is confined to men with advanced age (65 to 89 years). Moreover, increased blood pressure, high alcohol consumption, cigarette smoking (particularly among women), and bleeding disorders seem to play an important role. Thus, public health efforts to prevent the sequela or death by ICH stroke should concentrate on the detection and treatment of modifiable causes, namely, hypertension, alcohol abuse, and cigarette smoking.

    Selected Abbreviations and Acronyms

    CAMLIS=California Automated Mortality Linkage System
    CI=confidence interval
    ICD-8=International Classification of Diseases, 8th revision
    ICD-9=International Classification of Diseases, 9th revision
    ICH=intracerebral hemorrhage
    RR=relative risk

    Presented at the 29th Annual Meeting of the Society for Epidemiologic Research (SER), Boston, Mass, June 13-15, 1996.

    Table 1. Selected Baseline Characteristics of the Study Population of the Kaiser Permanente Medical Care Program

    CharacteristicsMen (n=28 452)Women (n=33 304)
    Age, y53.9 (10.0)54.5 (10.2)
    Education level, % vocational school or less31.136.3
    Race, %
    Smoking, %
    Alcohol, %
     <3 drinks/d68.868.7
     ≥3 drinks/d14.24.7
    Body mass index, kg/m226.0 (3.7)25.4 (5.0)
    Serum cholesterol, mmol/L5.98 (1.45)6.16 (1.24)
    Systolic blood pressure, mm Hg132 (18)133 (19)
    Diastolic blood pressure, mm Hg80 (10)79 (11)
    Hypertension, %†26.826.9
    Blood glucose, mg/dL103 (28)100 (25)
    Bleeding disorders, %‡1.00.6

    For continuous variables, means (SDs) are presented.

    *Hispanic non-white, American Indian, or Alaskan native.

    †Systolic blood pressure ≥160 mm Hg and diastolic blood pressure ≥90 mm Hg, or self-reported history of hypertension.

    ‡Congenital factor VIII (286.0), factor IX (286.1), and factor XI (286.2) disorders; congenital disorders of other clotting factors (286.3); von Willebrand's disease (286.4); hemorrhagic disorders due to circulating anticoagulants (286.5); defibrination syndrome (286.6); vitamin K deficiency (286.7); other and unspecified coagulation defects (286.9); purpura (287); and Rendu-Osler-Weber syndrome (448).

    Table 2. Rates (Per 100 Persons) of ICH by Level of Total Serum Cholesterol, Sex, and Age at Baseline: Kaiser Permanente Medical Care Program Follow-up 1978-1993

    Age 40-64 yAge 65-89 y
    Total Serum Cholesterol, mmol/LAt RiskNo.RateAt RiskNo.Rate
     Total23 9031170.494549841.84
    P, linear trend.46.02
    P, quadratic trend.18.62
     Total27 6641000.365640851.51
     P, linear trend.89.66
     P, quadratic trend.53.28

    Table 3. Age-Adjusted and Risk Factor–Adjusted RRs and Associated 95% CIs of ICH for Men and Women, According to Serum Cholesterol (Levels Below 10th Percentile Versus Higher Levels), Age, and Hypertension Status at Baseline: Kaiser Permanente Medical Care Program Follow-up 1978-1993

    Stratifying VariableMen, RR (95% CI)Women, RR (95% CI)
      40-64 y0.73 (0.36-1.51)0.92 (0.44-1.89)
      65-89 y2.56 (1.44-4.54)2.12 (0.85-5.23)
     Risk factor–adjusted*
      40-64 y0.71 (0.33-1.54)1.04 (0.45-2.40)
      65-89 y2.72 (1.46-5.02)2.23 (0.90-5.54)
    Hypertension status
      No1.47 (0.80-2.69)1.82 (0.91-3.62)
      Yes†1.49 (0.78-2.88)1.33 (0.48-3.67)
     Risk factor–adjusted*
      No1.49 (0.79-2.81)1.96 (0.94-4.08)
      Yes†1.32 (0.63-2.73)1.13 (0.35-3.62)

    The 10th percentile of serum cholesterol was 4.62 mmol/L (178 mg/dL) in men and 4.70 mmol/L (181 mg/dL) in women.

    *Adjusted for age, race, education level, body mass index, systolic blood pressure, smoking status (former, current), alcohol consumption, blood glucose, and prevalent medical conditions (liver disease, vasculitis, bleeding disorders, and fibromuscular dysplasia).

    †Systolic blood pressure above 140 mm Hg and diastolic blood pressure above 90 mm Hg and/or self-reported history of hypertension.

    Table 4. Multivariate RRs and Associated 95% CIs of ICH by Sex: Kaiser Permanente Medical Care Program Follow-up 1978-1993

    Risk Factors (1 SD Change or Reference Category)Men, RR (95% CI)Women, RR (95% CI)
    Age (10 y)2.06 (1.74-2.43)2.14 (1.78-2.56)
    Vocational school or less (higher education)1.33 (0.97-1.82)1.29 (0.94-1.78)
    Black race (vs white)1.84 (1.29-2.62)1.40 (0.96-2.06)
    Asian race (vs white)2.00 (1.16-3.44)1.38 (0.70-2.72)
    Other race (vs white)*2.85 (1.65-4.92)1.99 (0.98-4.05)
    Former smoking (vs never)1.27 (0.90-1.78)1.06 (0.69-1.61)
    Current smoking (vs never)1.19 (0.79-1.78)1.64 (1.13-2.37)
    High alcohol consumption (vs never)†1.26 (0.64-2.49)2.27 (1.10-4.70)
    Body mass index (3.7 kg/m2 in men; 5.0 kg/m2 in women)0.98 (0.84-1.15)0.91 (0.77-1.08)
    Systolic blood pressure (18 mm Hg in men; 19 mm Hg in women)1.14 (0.96-1.36)1.17 (0.98-1.39)
    Diastolic blood pressure (10 mm Hg in men; 11 mm Hg in women)1.20 (1.01-1.43)1.29 (1.08-1.55)
    Blood glucose (28 mg/dL in men; 25 mg/dL in women)1.07 (0.94-1.21)1.02 (0.88-1.18)
    Serum cholesterol (1.45 mmol/L in men; 1.24 mmol/L in women)0.84 (0.69-1.02)0.92 (0.79-1.08)
    Bleeding disorders (vs no bleeding disorders)‡2.34 (1.09-5.02)4.69 (2.27-9.69)

    Parameter estimates for liver disease, vasculitis, and fibromuscular dysplasia were nonsignificant and are not shown.

    *Hispanic non-white, American Indian, or Alaskan native.

    †≥3 drinks per day.

    ‡Congenital factor VIII (ICD 286.0), factor IX (286.1) and factor XI (286.2) disorders, congenital disorders of other clotting factors (286.3), von Willebrand's disease (286.4), hemorrhagic disorders due to circulating anticoagulants (286.5), defibrination syndrome (286.6), vitamin K deficiency (286.7), other and unspecified coagulation defects (286.9), purpura (287), and Rendu-Osler-Weber syndrome (448).

    This study was supported by contract RO1-AG-12264-01A1 from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Md.


    Correspondence to Dr Iribarren, Division of Epidemiology, School of Public Health, University of Minnesota, 1300 S Second St, Suite 300, Minneapolis, MN 55454-1015. E-mail [email protected].


    • 1 Ueshima H, Iida M, Shimamoto T, Konishi M, Tanigaki M, Nakanishi N, Ozawa H, Kojima S, Komachi Y. Multivariate analysis of risk factors for stroke: eight year follow-up study of farming villages in Akita, Japan. Prev Med.1980; 9:722-740.CrossrefMedlineGoogle Scholar
    • 2 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.CrossrefMedlineGoogle Scholar
    • 3 Lin CH, Shimizu Y, Kato H, Robertson TL, Furonaka H, Kodama K, Fukunaga Y. Cerebrovascular diseases in a fixed population of Hiroshima and Nagasaki, with special reference to relationships between type and risk factors. Stroke.1984; 15:653-660.CrossrefMedlineGoogle Scholar
    • 4 Ueda K, Hasuo Y, Kiyohara, Wada J, Kawano H, Kato I, Fujii I, Yanai T, Omae T, Fujishima M. Intracerebral hemorrhage in a Japanese community, Hysayama: incidence, changing pattern during follow-up, and related factors. Stroke.1988; 19:48-52.CrossrefMedlineGoogle Scholar
    • 5 Okada H, Horibe H, Ohno Y, Hayakawa N, Aoki N. A prospective study of cerebrovascular disease in Japanese rural communities, Akabane and Asahi, I: evaluation of risk factors in the occurrence of cerebral hemorrhage and thrombosis. Stroke.1976; 7:600-607.Google Scholar
    • 6 Kimura N, Toshima H, Nakayama Y, Mizuguchi T, Takayama K, Yoshinaga M, Fukami T, Tashiro H, Katayama F, Arima T, Yokota Y, Minagawa E, Tanaka R, Akiyoshi T, Soejima K, Yamada K, Mizunoe A, Nakamura K, Oshima F, Tanaka K, Akasu K, Niizaki T, Ikeda H, Nakamichi E, Ageta M, Miike Y, Inoue T, Nakagawa T, Nanbu S, Tanioka T, Shimada S. Population survey on cerebrovascular and cardiovascular diseases: the ten years experience in the farming village of Tanushimaru and the fishing village of Ushibuka. Jpn Heart J.1972; 13:118-127.CrossrefMedlineGoogle Scholar
    • 7 Shimamoto T, Komachi Y, Inada H, Doi M, Iso H, Sato S, Kitamura A, Iida M, Konishi M, Nakanishi N. Trends for coronary heart disease and stroke and their risk factors in Japan. Circulation.1989; 79:503-515.CrossrefMedlineGoogle Scholar
    • 8 Yano K, Reed DM, MacLean CJ. Serum cholesterol and hemorrhagic stroke in the Honolulu Heart Program. Stroke.1989; 20:1460-1465.CrossrefMedlineGoogle Scholar
    • 9 Iribarren C, Reed DM, Burchfiel CM, Dwyer JH. Serum total cholesterol and mortality: confounding and risk modification in Japanese-American men. JAMA.1995; 273:1926-1932.CrossrefMedlineGoogle Scholar
    • 10 Iso H, Jacobs DR Jr, Wentworth D, Neaton JD, Cohen J. Serum cholesterol levels and six-year mortality from stroke in 350,977 men screened for the Multiple Risk Factor Intervention Trial. N Engl J Med.1989; 320:904-910.CrossrefMedlineGoogle Scholar
    • 11 Neaton JD, Blackburn H, Jacobs D, Kuller L, Lee DJ, Sherwin R, Shih J, Stamler J, Wentworth D. Serum cholesterol level and mortality findings for men screened in the Multiple Risk Factor Intervention Trial. Arch Intern Med.1992; 152:1490-1500.CrossrefMedlineGoogle Scholar
    • 12 Giroud M, Creisson E, Fayolle H, Andre N, Becker F, Martin D, Dumas R. Risk factors for primary cerebral hemorrhage: a population-based study—the Stroke Registry of Dijon. Neuroepidemiology.1995; 14:20-26.CrossrefMedlineGoogle Scholar
    • 13 Gordon T, Kannel WB, Castelli WP, Dawber TR. Lipoproteins, cardiovascular disease, and death: the Framingham Study. Arch Intern Med.1981; 141:1128-1131.CrossrefMedlineGoogle Scholar
    • 14 Ooneda G, Yoshida Y, Suzuki K, Sekiguchi T. Morphogenesis of plasmatic arterionecrosis as the cause of hypertensive intracerebral hemorrhage. Virchows Arch A Pathol Anat.1973; 361:31-38.CrossrefMedlineGoogle Scholar
    • 15 Konishi M, So H, Komachi Y, Iida M, Shimamoto T, Jacobs DR Jr, Terao A, Baba S, Sankai T, Ito M. Associations of serum total cholesterol, different types of stroke, and stenosis distribution of cerebral arteries. Stroke.1993; 24:954-964.CrossrefMedlineGoogle Scholar
    • 16 Kojimahara M, Sekiya K, Ooneda G. Age-induced changes or cerebral arteries in rats: an electron microscope study. Virchows Arch A Pathol Anat.1973; 361:11-18.CrossrefMedlineGoogle Scholar
    • 17 Kobori K, Suzuki K, Yoshida Y, Ooneda G. Light and electron microscopic studies on rat arterial lesions induced by experimental arterial contraction. Virchows Arch A Pathol Anat Histol.1979; 385:29-39.CrossrefMedlineGoogle Scholar
    • 18 Ooneda G, Yoshida Y, Suzuki K, Shinkai H, Hori S, Kobori K, Takayama Y, Sekiguchi M. Smooth muscle cells in the development of plasmatic arterionecrosis, arteriosclerosis, and arterial contraction. Blood Vessels.1978; 15:148-156.MedlineGoogle Scholar
    • 19 Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health.1992; 82:703-710.CrossrefMedlineGoogle Scholar
    • 20 Collen MF, ed. Multiphasic Health Testing Services. New York, NY: John Wiley & Sons; 1978.Google Scholar
    • 21 Abell LL, Levy BB, Brodie BB, Kendall FE. Simplified method for the estimation of total cholesterol in serum and demonstration of its specificity. J Biol Chem.1952; 195:357-366.CrossrefMedlineGoogle Scholar
    • 22 Arellano M, Peterson G, Petitti DB, Smith R. The California automated mortality linkage system. Am J Public Health.1984; 74:1324-1330.CrossrefMedlineGoogle Scholar
    • 23 PHREG Procedure, SAS/STAT User's Guide, Edition 6.09. Cary, NC: SAS Institute Inc; 1994.Google Scholar
    • 24 Yamori Y, Horie R, Nara Y, Ohtaka M. Prophylactic trials for stroke-prone SHR, II: effects of fat, protein, and amino acid. Jpn Heart J.1977; 18:551-553.CrossrefMedlineGoogle Scholar
    • 25 Komachi Y, Iida M, Ozawa H. Interrelationships of food and stroke in Japan. Annu Rep Center Adult Dis.1975; 15:82-93.Google Scholar
    • 26 Tobian L. High potassium diets markedly protect against stroke deaths and kidney disease in hypertensive rats, a possible legacy from prehistoric times. Can J Physiol Pharmacol.1986; 64:840-848.CrossrefMedlineGoogle Scholar
    • 27 Potter JF, Beevers DG. Pressor effect of alcohol in hypertension. Lancet.1984; 1:119-122.CrossrefMedlineGoogle Scholar
    • 28 Haut MJ, Cowan DH. The effect of ethanol on hemostatic properties of human blood platelets. Am J Med.1974; 56:22-23.CrossrefMedlineGoogle Scholar
    • 29 Ruf JC, Berger JL, Renaud S. Platelet rebound effect of alcohol withdrawal and wine drinking in rats: relation to tannins and lipid peroxidation. Arterioscler Thromb Vasc Biol.1995; 15:140-144.CrossrefMedlineGoogle Scholar
    • 30 Hojnacki JL, Cluette-Brown JE, Deschenes RN, Mulligan JJ, Osmolski TV, Rencricca NJ, Barboriak JJ, Jakubowski JA. Effect of ethanol on low density lipoprotein and platelet composition. Lipids.1991; 26:884-890.CrossrefMedlineGoogle Scholar
    • 31 Basista MH, Joseph A, Smolen S, Koterba A, Brecher AS. Acetaldehyde alters coagulation protein function. Dig Dis Sci.1994; 39:2421-2425.CrossrefMedlineGoogle Scholar
    • 32 Elwood PC, Beswick AD, O'Brien JR, Yarnell JW, Layzell JC, Limb ES. Interrelationships between haemostatic tests and the effects of some dietary determinants in the Caerphilly cohort of older men. Blood Coagul Fibrinolysis.1993; 4:529-536.CrossrefMedlineGoogle Scholar
    • 33 Bridgen W, Robinson J. Alcoholic heart disease. Br Med J.1964; 2:1283-1289.CrossrefMedlineGoogle Scholar
    • 34 Berglund M, Risberg J. Regional cerebral blood flow during alcohol withdrawal. Arch Gen Psychiatry.1981; 38:351-355.CrossrefMedlineGoogle Scholar
    • 35 Fujii Y, Takeuchi S, Tanaka R, Koike T, Sasaki O, Minakawa T. Liver dysfunction in spontaneous intracerebral hemorrhage. Neurosurgery.1994; 35:592-596.CrossrefMedlineGoogle Scholar
    • 36 Kagan A, Popper JS, Rhoads GG, Yano K. Dietary and other risk factors for stroke in Hawaiian Japanese men. Stroke.1985; 16:390-396.CrossrefMedlineGoogle Scholar
    • 37 Abbott RD, Yin Y, Reed DM, Yano K. Risk of stroke in male cigarette smokers. N Engl J Med.1986; 315:717-720.CrossrefMedlineGoogle Scholar
    • 38 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.CrossrefMedlineGoogle Scholar
    • 39 Bell BA, Symon L. Smoking and subarachnoid haemorrhage. Br Med J.1979; 1:577-578.CrossrefMedlineGoogle Scholar
    • 40 Thrift AG, Donnan GA, McNeil JJ. Epidemiology of intracerebral hemorrhage. Epidemiol Rev.1995; 17:361-381.CrossrefMedlineGoogle Scholar


    eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

    Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.