Background and Purpose Spontaneous intracerebral hemorrhage has remained a serious disease despite recent improvements in medical treatment. This study was designed to identify modifiable risk factors for intracerebral hemorrhage.
Methods Health habits, previous diseases, and medication of 156 consecutive patients with intracerebral hemorrhage aged 16 to 60 years (96 men and 60 women) were compared with those of 332 hospitalized control patients (192 men and 140 women) who did not differ from case subjects in respect to age, day of onset of symptoms, or acuteness of disease onset.
Results After adjustment for sex, age, hypertension, body mass index, smoking status, and alcohol consumption during the last week, patients who had consumed 1 to 40, 41 to 120, or >120 g of alcohol within the 24 hours preceding the onset of illness had a relative risk (95% confidence interval) of hemorrhage of 0.3 (0.2 to 0.7), 4.6 (2.2 to 9.4), and 11.3 (3.0 to 42.8), respectively, compared with those who had consumed 0 g. In addition, alcohol intake within 1 week before the onset of illness, excluding use within the last 24 hours, increased the risk of hemorrhage; adjusted risks were 2.0 (1.1 to 3.5) for 1 to 150 g, 4.3 (1.6 to 11.7) for 151 to 300 g, and 6.5 (2.4 to 17.7) for >300 g compared with 0 g. The adjusted risk of hypertension for hemorrhage was 6.6 (3.9 to 11.3). Previous heavy alcohol consumption and current cigarette smoking were not independent risk factors for hemorrhage, but anticoagulant treatment was (P<.01). Erythrocyte mean corpuscular volume and γ-glutamyl transferase values were also higher in patients with intracerebral hemorrhage than in control subjects.
Conclusions Recent moderate and heavy alcohol intake as well as hypertension and likely also anticoagulant treatment seem to be independent risk factors for intracerebral hemorrhage.
Spontaneous ICH has remained a serious disease despite attempts at improving outcome by medical and neurosurgical treatment.1 2 3 4 Outcome is still determined mainly by the severity of bleeding, and the overall mortality is approximately 40% to 50%.1 5 6 7 Hence, identification of modifiable risk factors for ICH is important to influence the incidence of ICH, which is higher than that of subarachnoid hemorrhage, accounting for at least 10% of all strokes.
Most spontaneous hematomas are attributed to chronic arterial hypertension.1 6 Other potential risk factors for ICH include alcohol consumption,8 9 10 11 12 13 anticoagulant treatment,1 6 14 and to a lesser extent, aspirin use,15 thrombolytic therapy,16 and use of amphetamines or cocaine.1 6 These may also synergize with each other, eg, ethanol potentiates aspirin-induced prolongation of the bleeding time,17 and anticoagulant treatment increases the risk associated with thrombolytic therapy.16
Alcohol consumption has been reported to increase the risk of both hemorrhagic stroke (ICH8 9 10 11 12 13 and subarachnoid hemorrhage11 18 19 20 ) and brain infarction,11 13 21 and hypertension the risk of ICH1 6 8 12 and brain infarction.21 Current smoking seems to increase the risk of subarachnoid hemorrhage19 20 22 and brain infarction21 22 but not the risk of ICH.8 12 22 23 Many of these risk factors are known to correlate with each other (eg, alcohol use with hypertension and cigarette smoking). This study was designed to evaluate the role of health habits, previous diseases, and use of antithrombotic medicines as risk factors for ICH.
Subjects and Methods
The study comprised 156 consecutive patients with first-ever ICH aged 16 to 60 years who were admitted on an emergency basis to the Helsinki University Central Hospital and 332 control subjects from the same hospital who were matched as a group with ICH patients by sex, age, day of onset of symptoms, and acuteness of disease onset. The control subjects were admitted to the emergency department because of acute appendicitis (n=133), nephrolithiasis (n=74), dyspnea (n=58), cholecystitis (n=27), spontaneous pneumothorax (n=9), or because of viral meningitis (n=31). These medical emergencies were considered to accumulate among neither heavy drinkers nor teetotalers.
The ICH patients of this hospital-based case-control study were gathered from the city of Helsinki and its neighborhood. Stroke patients aged younger than 60 years are referred and admitted as emergency cases to our hospital without any selection. Elderly people (>60 years) were excluded since their strokes were not routinely examined by CT scan in the emergency ward.
We also excluded patients with intracerebral hematomas, which were caused by head trauma, brain tumor, saccular arterial aneurysm, arteriovenous malformation, or moyamoya disease. Patients with a known severe bleeding tendency or coagulation disorder as well as those with severe disabling diseases were also excluded because their living habits (eg, alcohol consumption) are restricted and severe bleeding or coagulation disorders are relatively infrequent causes of ICH.1 5 6 7 Patients and their family members were interviewed to exclude the patients with traumatic hematoma.
The patients and the control subjects were personally interviewed within 48 hours of admission according to a structured questionnaire. For ICH patients who were confused, unconscious, dysphasic, or died soon after admission, family members or fellow workers were interviewed. Data were obtained from family members or fellow workers in 85 cases and from both patients and family members in 71 cases. Because underestimation is more likely for alcohol intake than overestimation,11 the larger amounts reported in these cases were used in analyses, although the values reported by patients and relatives were quite similar during careful interviews. None of the case subjects, their relatives, or control subjects refused the interview.
The questionnaire included questions of the exact time of disease onset; the height and weight of the subject; previous diseases and hospital visits; recent drug use including analgesics, stimulants, and narcotics; recent drinking of coffee, beer, wine, and spirits; and current and previous smoking status. Information on all patients was also collected from the previous medical records of other hospitals and general practitioners to check diseases, medications, and the blood pressure values of the patients.
The BMI, calculated as weight/(height)2 (kg/m2), was used as the index of relative weight. Patients were considered to have definite hypertension if their blood pressure readings had repeatedly exceeded 160 mm Hg systolic or 95 mm Hg diastolic before the illness or if they used antihypertensive medication.
Recent alcohol consumption was recorded as grams of absolute ethanol consumed within 24 hours and 1 week before the onset of the disease (standard drink=12 g of alcohol). Problem drinking was assessed by the use of the CAGE questionnaire (Have you ever felt you should Cut down on your drinking? Have people Annoyed you by criticizing your drinking? Have you ever felt bad or Guilty about your drinking? Have you ever had a drink first thing in the morning to steady your nerves or to get rid of a hangover [Eye-opener]?).20 24 The CAGE interview deals with abnormal drinking behavior (eg, drinking on waking) and alcohol-induced problems rather than merely with the amount of alcohol consumed. Patients with two or more positive answers to the four questions in the CAGE interview were considered CAGE positive. CAGE positiveness is a sensitive indicator of previous and current excessive drinking (sensitivity and specificity, 80% to 90%).20 24 Smoking was categorized as follows: never smoked, former (regular) cigarette smoker, and current cigarette smoker, with 10 and 20 cigarettes per day as cutoff points.
The MCV (normal range, 80 to 96 fL) was used as a laboratory marker of alcohol intake because it was routinely available for most patients (94% of case subjects and 81% of control subjects) soon after admission to the hospital. Besides heavy drinking, cigarette smoking has also been shown to elevate MCV.25 The GGT (normal range, 0 to 45 U/L) was also measured routinely in ICH patients (90%), but it was available in only 80 control patients (24%). Blood sampling frequency for MCV and GGT measurements was not associated with drinking and smoking habits with the exception of a more frequent GGT measurement (P=.069) among heavily drinking male control subjects than among other male control subjects.
ICH was verified on admission by a CT scan in 155 patients and by autopsy only in 1 patient. The sites of the hematomas were divided into the following groups: subcortical white matter (n=34), putaminal (n=57), thalamic (n=9), combined (n=31; extension of putaminal hematoma to thalamus and/or to subcortical white matter), caudate nucleus (n=5), cerebellar (n=13), or pontine (n=7). Angiograms were performed in 106 patients to exclude possible vascular anomalies. Autopsy was performed in 36 of 57 patients who died because of ICH.
The data were analyzed with the biomedical data package statistical programs (BMDP Statistical Software Inc, version 1993, University of California at Los Angeles). Categorical variables were compared by Fisher’s exact two-tailed test, Pearson χ2 test, test for linear trend, or Mantel-Haenszel statistics. Continuous variables, which are expressed as mean±SD or as median with the 25th and 75th percentiles (when the distribution was skewed), were compared by the Student’s t test, two-way ANOVA (after logarithmic transformation of variables if necessary), Mann-Whitney U test, or Spearman’s rank correlation coefficients (rs) when appropriate.
The odds ratios as estimates of univariate and multivariate RRs and 95% CIs before and after adjustment for possible confounding variables were calculated by logistic regression. Hypothesis testing and estimation of the 95% CI were carried out with the use of the standard error estimate for the logistic coefficient estimates. With the use of stepwise logistic regression, the variables for which the simultaneous risks of ICH were tested were as follows: sex, age, hypertension, BMI, previous cardiovascular or cerebrovascular disease, diabetes mellitus, hyperlipidemia, amount of alcohol consumed within 24 hours and 1 week before illness, CAGE positiveness, previous heavy drinking, smoking status, current smoking, and use of nonsteroidal anti-inflammatory drugs. Anticoagulant treatment could not be tested in logistic models because none of the control subjects had used it.
The day of onset of illness did not differ significantly between the groups; 30% of the male case subjects, 33% of the male control subjects, 28% of the female case subjects, and 31% of the female control subjects fell ill during weekends or holidays. Fifty-seven patients with ICH (37%) died because of hemorrhage.
The frequencies of baseline characteristics, health habits, and values of laboratory markers of case and control subjects by sex are shown in Table 1⇓. The drinking and smoking habits of the control patients did not differ by the cause of hospital admission.
As expected, hypertension was more common among ICH patients than among control subjects (Table 1⇑). Use of antihypertensive medication was also more frequent in ICH patients (24%) than in control subjects (13%) (P<.01). Hypertension was associated directly and independently (P<.05) with both age and BMI in all subgroups (case and control subjects by sex) but not with recent alcohol consumption.
Recent light drinking (1 to 40 g of ethanol within 24 hours or 1 to 150 g within 1 week) did not seem to increase the risk of ICH, but moderate (41 to 120 g within 24 hours or 151 to 300 g within 1 week) and heavy (>120 g within 24 hours or >300 g within 1 week) drinking increased the risk in a dose-dependent manner (Table 1⇑). The risk in women was associated mainly with moderate drinking since heavy alcohol intake as well as CAGE positiveness was infrequent among women.
Alcohol consumption within 24 hours or 1 week before the illness was somewhat greater in nonhypertensive patients than in patients with hypertension. Contrary to the direct association between age and hypertension, age was inversely associated with the amount of recent alcohol consumption (rs=−.06, P=NS for 24-hour use; rs=−.20, P<.01 for 1-week use) and the number of positive CAGE answers (rs=−.10, P<.05). In addition, MCV values were higher in nonhypertensive ICH patients (95.6±7.1) than in hypertensive ICH patients (93.3±7.5) (P=.039), although GGT values did not differ according to hypertension likely because, in addition to alcohol consumption, overweight increased GGT values (rs=.26, P<.01).
Both in men and women cigarette smoking status correlated significantly with the amount of alcohol consumed within 1 week before the illness (rs=.33, P<.001 in men; rs=.17, P<.05 in women) and the number of positive CAGE answers (rs=.29, P<.001 in men; rs=.16, P<.05 in women). Moreover, smoking status correlated inversely with age (rs=−.14, P<.05 in men; rs=−.27, P<.001 in women).
Univariate RRs of ICH for currently smoking men and women were 1.3 (95% CI, 0.8 to 2.1; P=NS) and 2.1 (95% CI, 1.1 to 3.9; P=.027), respectively, compared with those who were not current smokers. The risk of hemorrhage among all current smokers was 1.6 (95% CI, 1.1 to 2.3; P=.026). A nonsignificant increase in the risk of hemorrhage was observed among heavy smokers (Table 1⇑).
Two-way ANOVA showed that MCV values were significantly elevated in current smokers (P<.01) and CAGE-positive patients (P<.0001), without interaction, but only CAGE positiveness increased GGT values (P<.0001). MCV and GGT values were significantly (P<.01) higher in male ICH patients than in male control subjects, but the differences were not as great between female groups (Table 1⇑).
The occurrence of the following previous diseases was not significantly different between the case and control subjects: myocardial infarction (2% among case subjects versus 4% among control subjects), minor cerebral infarction (4% versus 2%), angina pectoris (6% versus 6%), transient ischemic attack (5% versus 3%), atrial fibrillation (2% versus 2%), diabetes mellitus (9% versus 7%), and hyperlipidemia (7% versus 9%).
Anticoagulant treatment was significantly associated with the occurrence of intracerebral hematoma (P=.015 for women; P=.073 for men; P=.0006 for both sexes together; and P=.0061 for association between warfarin treatment and occurrence of ICH) (Table 1⇑). Of the 7 ICH patients with anticoagulant therapy, 2 women were treated with low-dose heparin. Five others (3%) received warfarin for either previous minor ischemic cerebral infarction (n=3; none were in the location of present ICH) or deep-vein thrombosis (n=2). The median of the INR of the last blood sample before ICH was 4.3 (range, 2.1 to 4.8; INR in 3 of 5 patients was >4.0). The median time of warfarin treatment was 2 years (range, 1.5 months to 14 years).
Adjustment for hypertension did not abolish the significant association between any anticoagulant (P=.0008) or warfarin (P=.0054) treatment and occurrence of ICH. On the other hand, patients on anticoagulant treatment had used significantly (P<.05) less alcohol before ICH than other ICH patients: all patients with anticoagulant treatment were nondrinkers or light drinkers and CAGE negative. They were also older than other ICH patients (P<.05).
Use of nonsteroidal anti-inflammatory drugs (mainly acetylsalicylic acid) within 1 week before the onset of illness was nonsignificantly (RR, 1.5; 95% CI, 0.92 to 2.3) more common in patients with ICH than in control subjects; 22% of male case subjects versus 16% of male control subjects and 38% of female case subjects versus 29% of female control subjects had used these drugs. Use of these drugs in ICH patients was not associated with alcohol consumption, hypertension, smoking status, or anticoagulant treatment. Only one drug abuser was included in the study; a 22-year old female patient suffered a putaminal ICH shortly after an intravenous injection of amphetamine.
Stepwise logistic regression showed that significant independent risk factors for ICH were hypertension (P<.0001), the amount of alcohol consumed within 24 hours (P<.0001), and simultaneously also the amount of alcohol consumed within 1 week before the onset of illness (P<.001). When recent alcohol variables were in the same model (24-hour alcohol amount was subtracted from the 1-week amount), adjusted RRs remained significant, suggesting that the longer the recent moderate or heavy drinking the greater the risk of ICH (Table 2⇓).
Among men, significant independent risk factors for ICH were hypertension and the amounts of alcohol consumed within 24 hours and within 1 week, resembling the results obtained in whole patient population. After adjustment for age, hypertension, BMI, smoking status, and 1-week alcohol intake, the RRs of alcohol consumption within 24 hours were as follows: 0.2 (95% CI, 0.04 to 0.6) for 1 to 40 g, 5.1 (95% CI, 2.1 to 12.7) for 41 to 120 g, and 11.1 (95% CI, 2.7 to 46.2) for >120 g intake. Correspondingly, the adjusted RR of hypertension was 8.6 (95% CI, 3.9 to 18.7), and the RRs of 1-week alcohol intake were as follows: 2.4 (95% CI, 1.1 to 5.3) for 1 to 150 g, 4.1 (95% CI, 1.1 to 14.9) for 151 to 300 g, and 9.0 (95% CI, 2.8 to 29.7) for >300 g. RRs of hypertension and recent alcohol variables for ICH increased after adjustments for each other compared with univariate risks because recent alcohol consumption was inversely associated with hypertension in male ICH patients (P<.05) and because age was directly associated with hypertension but inversely associated with recent alcohol use.
Among women, independent risk factors for ICH were hypertension (RR, 4.5; 95% CI, 2.1 to 9.6) and the amount of alcohol consumed within 24 hours (RR, 0.7; 95% CI, 0.3 to 1.8 for 1 to 40 g; and RR, 5.3; 95% CI, 1.5 to 18.7 for >40 g intake) after adjustment for age, hypertension, BMI, and smoking status, but the amount of alcohol consumed within 1 week was not simultaneously a risk factor with 24-hour intake as it was in men. After adjustment for each other, the RRs of hypertension and 24-hour alcohol intake remained similar to those in univariate analysis since hypertension was not associated with recent drinking. When alcohol intake within 1 week was tested instead of alcohol intake within 24 hours, consumption of more than 150 g of ethanol within 1 week increased the risk of ICH (RR, 6.5; 95% CI, 1.9 to 22.7).
Inclusion of interaction terms of separate variables into the multivariate models did not change the magnitude of RRs or ranges of CIs of the significant main factors. Also, exclusion of those patients who had used anticoagulant drugs did not change the results. Restriction of the analysis to those 71 patients who were personally interviewed influenced the results only slightly.
When CAGE was tested instead of recent alcohol variables, CAGE positiveness (problem drinking) was associated with a significantly increased risk of ICH compared with CAGE negativeness (Table 3⇓). However, previous heavy drinking without recent heavy drinking (within the last week before the illness) did not increase the risk of ICH (adjusted RR, 0.5; 95% CI, 0.2 to 1.3).
The location of hematoma was significantly (P<.05) more often other than subcortical white matter in hypertensive patients (in 86% of cases), while in nonhypertensive patients the hematoma was located subcortically in 31% of cases. The location of hematoma (subcortical versus other site) did not differ significantly according to recent alcohol intake and CAGE positiveness.
The role of hypertension as a risk factor for ICH is well known,1 6 8 12 but the independent role of alcohol and anticoagulant treatment has not been as well established. According to this study, recent heavy and moderate alcohol consumption, hypertension, and likely also anticoagulant treatment are independent risk factors in both men and women.
Previous studies have suggested that alcohol consumption might increase the risk of ICH independent of hypertension.8 9 10 11 12 13 However, most studies have investigated hemorrhagic stroke by combining ICH with subarachnoid hemorrhage. This could yield biased results since ICH and subarachnoid hemorrhage seem to have different risk factors. ICH has been studied separately from other types of stroke during the CT era only in few studies.8 12 13 The control groups in these case-control studies have either included patients with chronic diseases or the control subjects were voluntary healthy working people but significantly younger than the case subjects. Such control subjects might have a different prevalence of hypertension and smoking and drinking habits from the general population with an age distribution similar to that of the case subjects. In addition, control subjects who are smokers or who have drinking problems are expected to refuse to participate in the study more often than others, which may overestimate the observed results.13 19
It has also remained unknown whether recent habitual or occasional drinking as well as previous habitual drinking increases the risk of ICH. Such a problem is difficult to resolve by mailed questionnaires. Recall bias cannot be avoided, and people usually report their drinking of alcohol as the number of drinks per day or week without mentioning either periodicity or peaks of their alcohol consumption.
In case-control studies the selection of the cases and control subjects is of crucial importance. In this study we tried to decrease the potential bias by as complete a collection of patients as possible and by matching the hospitalized control subjects by sex, age, and acuteness and day of disease onset. All those interviewed in this study agreed to participate, and the interview was performed within 48 hours after admission.
It is important to exclude subjects with a traumatic subcortical intracerebral hematoma, which expands as a result of bleeding and clotting disorders and hence may resemble spontaneous hematoma. The locations of hematomas in this study, which corresponded to previous studies of ICH, were different from traumatic hematomas.1 2 3 5 8
Although mortality rates in population-based studies have varied from 44% to 50%,5 7 the rate among patients 60 years or younger was lower (39%).26 The 6-month mortality of ICH seems to increase with age independent of other risk factors, especially in those older than 60 years.27 The high proportion of patients who succumbed to ICH in our hospital (37%) suggests that not very many patients from the regional ICH population remained unadmitted. In addition, amyloid angiopathy seems to be an important cause of ICH in elderly patients,1 6 while the role of alcohol (and possibly other health habits like smoking and use of illicit drugs) may decrease by age since heavy alcohol consumption is known to cause an excess of untimely deaths, leaving those with better health habits to reach old age.10 24
Alcohol consumption and cigarette smoking are more frequent in the city of Helsinki and its neighborhood than in other parts of the country.28 These habits are also more frequent among men and are inversely associated with age in Finland. This did not cause bias in the present study because the control subjects were from the same region. We also had to interview relatives instead of patients who were in poor condition. Analysis after exclusion of the patients whose family members or fellow workers were interviewed did not change the results.
One may ask whether alcohol exposure of the control patients was already decreased because of possible prodromal symptoms of the acute illness before admission since the onset of ICH might be considered to occur more acutely than the onset of the other illnesses. However, alcohol consumption of the control subjects was calculated before the first symptoms, including prodromal symptoms of the disease. Neither did comparing the amount of recent alcohol use within 24 hours with that during 1 week among control subjects suggest that alcohol use was decreased shortly before the onset of illness. On the other hand, symptoms of ICH also quite often progress within the 6 hours following the first symptoms.1 4 6
Recent alcohol consumption and current smoking of our hospital control subjects were similar to those of a randomly selected population control group29 30 and a general health survey among Finnish people28 with similar age and sex distributions, which were gathered during the same time period and from the same region. Likewise, the prevalence of elevated MCV values and CAGE positiveness of our control subjects did not deviate from that of the general population.29 30
A prior history of chronic hypertension is a well-known risk factor for ICH, and its occurrence among ICH patients varies from 45% to 70%.1 3 5 6 7 8 12 23 The prevalence of treated and untreated hypertension by age among our control subjects was similar to that reported in the general population from the southwestern part of Finland during the same period.31 This suggests that hypertension was not overrepresented in our control group.
Blood pressure is raised independently by age, BMI, pulse rate, amount of regular alcohol use, and sodium intake.31 32 33 34 35 36 Daily drinking of alcohol results in a dose-dependent gradual elevation of blood pressure within a few days to weeks.32 33 34 After cessation of drinking blood pressure is normalized. Although short-term occasional drinking may not consistently elevate blood pressure,34 it has been reported that, in normotensive subjects, acute consumption of beer can transiently raise both systolic and diastolic blood pressure.37 The maximum response occurs at peak blood alcohol concentration. A reactive form of hypertension associated with increased catecholamines and vascular hyperresponsiveness is a well-known effect during alcohol withdrawal.34 38 In heavy drinkers, a high variability of alcohol consumption (episodic drinking) seems to increase blood pressure values more than a low variability of alcohol intake.36
The pathophysiological mechanisms by which alcohol could contribute to ICH include hypertension, impaired hemostasis, decreased circulating levels of clotting factors produced by the liver, excessive fibrinolysis, and disseminated intravascular coagulation.10 It is unlikely that chronic hypertension is the main mechanism by which alcohol consumption increases the risk of ICH since this risk has remained significant after adjustment for hypertension in both this and previous studies.8 9 12 13 On the other hand, increase in blood pressure during alcohol drinking and withdrawal may be an important mechanism. A transient increase in blood pressure together with cerebral arteriolar vasoconstriction39 during alcohol exposure might cause rupture of small cerebral arteries. This theory is supported by the observation that recent drinking (within 24 hours of ICH) was more important as a risk factor than the amount of alcohol consumed within 1 week, and that previous drinking was not a risk factor. Because heavy regular drinkers are usually also recent drinkers, heavy drinking in general will appear as a risk factor, as has been found in previous studies.
The association of recent drinking and ICH was not influenced by sex, age, BMI, smoking status, hypertension or other previous diseases, or use of nonsteroidal anti-inflammatory drugs or anticoagulants. Although bleeding and clotting disorders have been suggested as mediators between alcohol use and ICH in a selected uncontrolled hospital-based patient population treated mainly surgically,40 the frequency of hemostatic disorders could explain only a minority of the association between alcohol and risk of ICH. On the other hand, heavy drinking with liver dysfunction or hemostatic disorders (thrombocytopenia, prolonged bleeding time, or prolonged prothrombin time) was found to be correlated with increased volume of putaminal ICH.41 Hence, hemostatic disorders may be more frequent among the neurosurgical ICH patient population than among the population-based ICH patient population.
Anticoagulant treatment has been shown to be a risk factor for ICH, especially in the elderly.14 42 Intracranial bleeding caused by anticoagulants is much more common in the elderly because they are treated more often by anticoagulants. The prevalence of anticoagulant treatment in a Finnish general population aged 40 to 60 years is approximately 0.5%.14 Therefore, it was not surprising that none of our control subjects used these medicines. Only 4% of our case subjects used anticoagulants (3% used warfarin). This number does not significantly differ from that of a recent Finnish report, in which 4 of 64 (6%) ICH patients aged 30 to 60 years used warfarin.14
The risk of hemorrhage is also directly associated with intensity of anticoagulation. Bleeding that occurs when the INR is less than 3.0 is frequently associated with an obvious underlying disease.42 Two of our 5 patients with warfarin treatment had an INR of less than 3.0; both of them were hypertensive and had putaminal ICH. None of our patients had used thrombolytic agents before ICH. The risk of ICH associated with the use of these drugs has been shown to be low after myocardial infarction. This low risk may be increased somewhat by the use of anticoagulants and among elderly people.16 The prevalence of thrombocytopenia did also not differ between our case and control subjects. These observations suggest that liver dysfunction or hemostatic disorders must be severe if bleeding or clotting disorders can be considered as the sole cause of ICH.
Only one of our patients had used an illicit drug (intravenous amphetamine). Although recreational drugs infrequently cause ICH,1 6 their use may be an important cause among young adults. These drugs likely cause ICH by abrupt increase in blood pressure and by vasculitis.
According to previous studies, current smoking does not increase the risk of ICH,22 23 although it is the most important modifiable risk factor for subarachnoid hemorrhage.19 20 22 This is not unexpected since blood pressure values are generally lower in smokers than in nonsmokers.22 43 However, smoking a cigarette transiently causes an acute increase in blood pressure for approximately 3 hours.19 43 Hence, awake heavy smokers may be hypertensive. In this study current smokers, particularly heavily smoking women, had an increased risk of ICH even after adjustment for hypertension but no longer after adjustment for alcohol intake.
In conclusion, heavy and moderate recent alcohol drinking and hypertension are the most important modifiable risk factors for ICH among adults aged 60 years or younger. The risk associated with hypertension seems to increase with age, while the risk associated with alcohol consumption is most clearly seen in young adults.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|INR||=||international normalized ratio|
|MCV||=||erythrocyte mean cell volume|
This study was supported in part by grants from the Paavo Nurmi Foundation and the Maire Taponen Foundation (Dr Juvela).
Reprint requests to S. Juvela, MD, Department of Neurosurgery, Helsinki University Central Hospital, Topeliuksenkatu 5, FIN-00260 Helsinki 26, Finland.
- Received February 8, 1995.
- Revision received June 1, 1995.
- Accepted June 1, 1995.
- Copyright © 1995 by American Heart Association
Ojemann RG, Heros RC. Spontaneous brain hemorrhage. Stroke. 1983;14:468-475.
Fogelholm R, Nuutila M, Vuorela A-L. Primary intracerebral haemorrhage in the Jyväskylä region, Central Finland, 1985-89: incidence, case fatality rate, and functional outcome. J Neurol Neurosurg Psychiatry. 1992;55:546-552.
Calandre L, Arnal C, Fernandez Ortega J, Bermejo F, Felgeroso B, del Ser T, Vallejo A. Risk factors for spontaneous cerebral hematomas: case-control study. Stroke. 1986;17:1126-1128.
Klatsky AL, Armstrong MA, Friedman GD. Alcohol use and subsequent cerebrovascular disease hospitalizations. Stroke. 1989;20:741-746.
Gorelick PB. The status of alcohol as a risk factor for stroke. Stroke. 1989;20:1607-1610.
Camargo CA. Moderate alcohol consumption and stroke: the epidemiologic evidence. Stroke. 1989;20:1611-1626.
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.
Fogelholm R, Eskola K, Kiminkinen T, Kunnamo I. Anticoagulant treatment as a risk factor for primary intracerebral haemorrhage. J Neurol Neurosurg Psychiatry. 1992;55:1121-1124.
Antiplatelet Trialists’ Collaboration. Collaborative overview of randomised trials of antiplatelet therapy, I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. BMJ. 1994;308:81-106.
Longstreth WT Jr, Nelson LM, Koepsell TD, van Belle G. Cigarette smoking, alcohol use, and subarachnoid hemorrhage. Stroke. 1992;23:1242-1249.
Juvela S, Hillbom M, Numminen H, Koskinen P. Cigarette smoking and alcohol consumption as risk factors for aneurysmal subarachnoid hemorrhage. Stroke. 1993;24:639-646.
Shaper AG, Phillips AN, Pocock SJ, Walker M, Macfarlane PW. Risk factors for stroke in middle aged British men. BMJ. 1991;302:1111-1115.
Shinton R, Beevers G. Meta-analysis of relation between cigarette smoking and stroke. BMJ. 1989;298:789-794.
Broderick JP, Brott TG, Duldner JE, Tomsick T, Huster G. Volume of intracerebral hemorrhage: a powerful and easy-to-use predictor of 30-day mortality. Stroke. 1993;24:987-993.
Daverat P, Castel JP, Dartigues JF, Orgogozo JM. Death and functional outcome after spontaneous intracerebral hemorrhage: a prospective study of 166 cases using multivariate analysis. Stroke. 1991;22:1-6.
Lahelma E, Berg M-A, Helakorpi S, Prättälä R, Rahkonen O, Puska P. Suomalaisten aikuisten koettu terveydentila ja terveyskäyttäytyminen 1979-93. Suom Lääkäril. 1994;49:2213-2227.
Puddey IB, Beilin LJ, Vandongen R, Rouse IL, Rogers P. Evidence for a direct effect of alcohol consumption in normotensive men: a randomized controlled trial. Hypertension. 1985;7:707-713.
Saunders JB. Alcohol: an important cause of hypertension. BMJ. 1987;294:1045-1046.
Smith WCS, Crombie IK, Tavendale RT, Gulland SK, Tunstall-Pedoe HD. Urinary electrolyte excretion, alcohol consumption, and blood pressure in the Scottish Heart Health Study. BMJ. 1988;297:329-330.
Marmot MG, Elliott P, Shipley MJ, Dyer AR, Ueshima H, Beevers DG, Stamler R, Kesteloot H, Rose G, Stamler J. Alcohol and blood pressure: the INTERSALT study. BMJ. 1994;308:1263-1267.
Potter JF, Watson RD, Skan W, Beevers DG. The pressor and metabolic effects of alcohol in normotensive subjects. Hypertension. 1986;8:625-631.
Altura BM, Altura BT, Gebrewold A. Alcohol-induced spasms of cerebral blood vessels: relation to cerebrovascular accidents and sudden death. Science. 1983;220:331-333.
Niizuma H, Suzuki J, Yonemitsu T, Otsuki T. Spontaneous intracerebral hemorrhage and liver dysfunction. Stroke. 1988;19:852-856.
Niizuma H, Shimitsu Y, Nakasato N, Jokura H, Suzuki J. Influence of liver dysfunction on volume of putaminal hemorrhage. Stroke. 1988;19:987-990.