alcohol y acv

7/31/2019 Alcohol y Acv

1/12

REVIEW

Alcohol Consumption and Risk of StrokeA Meta-analysisKristi Reynolds, MPH

L. Brian Lewis, MPH

John David L. Nolen, MD, PhD, MSPH

Gregory L. Kinney, MPH

Bhavani Sathya, MPH

Jiang He, MD, PhD

STROKE IS THE THIRD LEADING

cause ofdeathand a major causeof disability in the UnitedStates.1,2 In 1999, 167366deaths

in theUnitedStatesresultedfrom stroke.1

Approximately 30% of stroke survivorsare permanently disabled and 20% re-quireinstitutionalizedcare.1 Strokeisalsoa huge financial burden forpatients, theirfamilies,andthehealth care system. Thecost ofstroke intheUnitedStates in 2002is estimated to be $49.4 billion, whichincludes direct health expenditures andlost productivity resulting from morbid-

ity and mortality.1Alcoholic beverages are consumed

widely throughout the world, and anassociation between alcohol consump-tionand stroke could haveconsiderablepublic health and clinical implications.Over the past 2 decades, many observa-tional epidemiologic studies3-37 haveexamined the role of alcohol as both ariskfactor anda potential protective fac-tor for stroke. Heavy alcohol consump-tion hasbeen linkedto an increasedriskoftotalstroke,23,32 ischemicstroke,29,33 and

hemorrhagicstroke.

3,7,33,35

However,stud-ies investigatingthe associationbetweenmoderate alcohol consumption andstroke havereported conflicting results.Some studies have reported that mod-erate alcohol consumption is inverselyrelated to risk of total stroke,31 ischemicstroke,27,31,37 andhemorrhagicstroke,27,31

while others found that moderate alco-

hol consumption is positively related torisk of stroke.3,25

We performeda meta-analysis of epi-

demiologic studies to examine the rela-tive risk of stroke at various levels ofalcohol consumption.

METHODS

Study Selection

A literaturesearch of the MEDLINE da-tabase (from January 1966 throughApril 2002) using the Medical Subject

Headings alcohol drinking, ethanol, cere-brovascular accident, cerebrovasculardis-orders, and intracranial embolism and

thrombosis and the keywordstroke wasperformed. The search was restricted to

Author Affiliations: Department of Epidemiology,Tu-lane University School of Public Health and TropicalMedicine, New Orleans, La.Corresponding Author and Reprints: Kristi Rey-nolds,MPH, Departmentof Epidemiology, Tulane Uni-versity Health SciencesCenter, School of Public Healthand Tropical Medicine, 1430 Tulane Ave SL18, NewOrleans, LA 70112 (e-mail: [email protected]).

Context Observational studies suggest that heavy alcohol consumption may in-crease the risk of stroke while moderate consumption may decrease the risk.

Objective To examine the association between alcohol consumption and relativerisk of stroke.

Data Sources Studies published in English-language journals were retrieved by search-ing MEDLINE (1966April 2002) using Medical Subject Headings alcohol drinking,ethanol, cerebrovascular accident, cerebrovascular disorders, and intracranial embo-lism and thrombosis and the key wordstroke; Dissertation Abstracts Online using thekeywords stroke and alcohol; and bibliographies of retrieved articles.

Study Selection From 122 relevant retrieved reports, 35 observational studies (co-hort or case control) in which total stroke, ischemic stroke, or hemorrhagic (intracerebralor total) stroke was an end point; the relative risk or relative odds and their variance (ordata to calculate them) of stroke associated with alcohol consumption were reported;alcohol consumption was quantified; and abstainers served as the reference group.

Data Extraction Information on study design, participant characteristics, level of alco-holconsumption, stroke outcome, control forpotentialconfoundingfactors, andrisk esti-mates was abstracted independently by 3 investigators using a standardized protocol.

Data Synthesis A random-effects model and meta-regression analysis were usedto pool data from individual studies. Compared with abstainers, consumption of morethan 60 g of alcohol per day was associated with an increased relative risk of totalstroke, 1.64 (95% confidence interval [CI], 1.39-1.93); ischemic stroke, 1.69 (95%CI, 1.34-2.15); and hemorrhagic stroke, 2.18 (95% CI, 1.48-3.20), while consump-tion of less than 12 g/d was associated with a reduced relative risk of total stroke,

0.83 (95%, CI, 0.75-0.91) and ischemic stroke, 0.80 (95% CI, 0.67-0.96), and con-sumption of 12 to 24 g/d was associated with a reduced relative risk of ischemic stroke,0.72 (95%, CI, 0.57-0.91). The meta-regression analysis revealed a significant non-linear relationship between alcohol consumption and total and ischemic stroke and alinear relationship between alcohol consumption and hemorrhagic stroke.

Conclusions These results indicate that heavy alcohol consumption increases therelative risk of stroke while light or moderate alcohol consumption may be protectiveagainst total and ischemic stroke.

JAMA. 2003;289:579-588 www.jama.com

2003 American Medical Association. All rights reserved. (Reprinted) JAMA, February 5, 2003Vol 289, No. 5 579

by guest on April 23, 2012jama.ama-assn.orgDownloaded from
http://jama.ama-assn.org/http://jama.ama-assn.org/http://jama.ama-assn.org/http://jama.ama-assn.org/


2/12

studies published in English-languagejournals and conducted in human sub-jects. We also conducted a searchof ab-stracts listed in Dissertation AbstractsOnline using the keywords stroke andalcohol, and we performed a manual

search of references cited in publishedoriginal study and relevant review ar-ticles.38-48 The contents of 122 ab-stracts or full-text manuscripts identi-fied during the literature search werereviewed independently by 2 investi-gators in duplicate to determinewhether they met the criteria for inclu-sion. When there were discrepancies be-tween investigators for inclusion or ex-clusion, other investigators conductedadditional evaluation of the study anddiscrepancies were resolved in confer-ence. To be included in our meta-

analysis, a publishedstudy had to meetthe following criteria: (1) observa-tional cohort or case-control study inwhich total stroke, ischemic stroke, orhemorrhagic (intracerebral or total)stroke was an endpoint; (2) relative riskor relative odds and their variance (ordata to calculate them) of stroke asso-ciated with alcohol consumption werereported; (3)alcohol consumption wasquantified;and (4)abstainers wereusedas the reference group.

Fifty-threestudies wereidentifiedand

abstracted. Four studies reported totalhemorrhagic stroke as the outcome,which includes intracerebral and sub-arachnoid hemorrhage.4,7,10,11 Noneofthestudies reported information on subdu-ral hemorrhagic strokes. We have usedthe term hemorrhagic stroke through-out thearticle. Two reports consisted ofthesame case patients but different con-trols and were treated as 2 separate stud-ies.23,24 From the53 studies, 18 were fur-ther excluded for various reasons. Twostudies were excluded because com-

bined risk estimates were reported formen and women but levels of alcoholconsumption were notthesame for menas for women.49,50 We excluded 5 stud-iesthat examined only theeffectof bingedrinking or acute alcohol consumption(within 24 hours before stroke)51-55 be-cause our study assessed habitual alco-hol consumption and relative risk of

stroke. Five studies that lacked suffi-cient data for calculation of relative riskestimates wereexcluded.56-60 Theremain-ing 6 excluded reports did not use ab-stainers as the reference group.61-66 Weincluded 19 cohort studies and 16 case-

control studies in our final analysis.Data Abstraction

All data were independently abstractedin triplicate by means of a standardizeddata-collectionform. Discrepancieswereresolved by discussion and referencingtheoriginal publication.We didnot con-tact authors to request additional infor-mation. Study characteristics recordedwere as follows: title, articles first au-thors name, year, and source of publi-cation, country of origin, study design(cohort study or case-control study),

characteristics of the study population(samplesize; samplingmethods; anddis-tribution of age, sex, and race), mea-sures of outcome and exposure, dura-tion of follow-up(for prospective cohortstudies), confounding factors con-trolled for by matching or adjustment,and the relative risk (or relative odds)of stroke associated with alcohol con-sumption and the corresponding con-fidence interval (or SE). Relative risksoverall and in each subgroup, accord-ing to sex, subtype of stroke, level of al-

cohol consumption, and type of alco-holic beverage, were abstracted.

Statistical Analysis

Relative risk was used as a measure ofthe relation between alcohol consump-tion and risk of stroke. For case-control studies, relative odds were usedas a surrogate measure of the corre-sponding relative risk. Because the ab-solute risk of stroke is low, the rela-tive odds approximate the relativerisk.Relative risks from individual studies

for each level of alcohol consumptionand the corresponding SEs were trans-formed to their natural logarithms tostabilize the variances and to normal-ize the distributions. The SEs werederived from the confidence intervalsprovided in each study.

The studies included in our meta-analysis often differed in the measure-

ment units of alcohol consumption (eg,grams, milliliters, ounces, or drinksconsumed every day, week, or month).Therefore, we first converted these dif-ferent units of alcohol consumption tograms per day. Among the 35 studies

included in our meta-analysis, 20 re-ported alcohol consumption as grams.We used the following conversion fac-tors for the 4 studies that reported al-cohol data as milliliters or ounces: 1 mL,0.785 g; 1 fl oz, 28.41 mL (United King-dom); and 1 fl oz, 29.58 mL (UnitedStates). Two of the 11 studies that re-ported alcohol data as drinks providedconversion factorsin their articles. Theother 9 used common conversion fac-tors.67 In the latter, a drink was de-fined as 12 g in the United States, 10 gin Australia and Europe, and 21.2 g in

Japan, which is the standard drink vol-ume in Japan.67

Alcohol consumption was reportedas categorical data with a range in allstudies. We assigned the mean of theupper and lower bounds in each cat-egory as the average alcohol consump-tion. An upper bound was not re-ported in many studies for the categoryof highest consumption, so we as-sumed it to be the same amplitude asthe preceding category for calculationof average alcohol consumption in this

category. In our meta-analysis, alco-hol consumption was categorized into5 groups: none (reference),less than 12,12 to 23, 24 to 60, and more than 60g/d. We assigned the level of alcoholconsumption from each study to thesegroups based on the calculated aver-age consumption of alcohol. In somestudies, the average alcohol consump-tion from more than 1 category fell intothe same group of alcohol consump-tion in our meta-analysis. When this oc-curred, we pooled the relative risks

within each category for each study andthen we pooled across all studies.Both fixed-effects and DerSimonian

and Laird random-effectsmodels68 wereused to calculate thepooled relative riskacross levels of alcohol consumption.Al-though bothmodelsyieldedsimilar find-ings, results from the random-effectsmodel are presented herein because

ALCOHOL CONSUMPTION AND RISK OF STROKE

580 JAMA, February 5, 2003Vol 289, No. 5 (Reprinted) 2003 American Medical Association. All rights reserved.



3/12

Table 1. Characteristics of 19 Cohort Studies of Alcohol Consumption and Risk of Stroke

SourceStudy

ParticipantsExposure

AssessmentDuration of

Follow-up, yFollow-up

ProcessOutcome

AssessmentNo. of

Stroke Cases Controlled Variables

Donahue et al,3

19867878 Men aged

45-69 y inHawaii

In-personinterview

12 Clinicalexaminationsat years 2 and6 andcontinued

surveillance

Hospital dischargediagnosis, clinicaldiagnosis, deathcertificate, orautopsy record

290 Age, BMI, cigarette smoking,hypertension, serumcholesterol, uric acid,glucose level, hematocrit

Kono et al,4

19865135 Men in

JapanSelf-administered

questionnaire19 Vital status

ascertainedby medicalassociation

Death certificate 230 Age, cigarette smoking

Gordon andDoyle,5

1987

1910 Men aged38-55 y inNew York

Self-administeredquestionnaire

29 Vital statisticsrecords,newspapers,or reportsfrom proxies

Proxy reports ordeath certificate

33 None

Stampfer et al,6

198887526 US

womenaged34-59 y


4 Biennialquestion-naires

Medical records 120 Age, cigarette smoking,hypertension, DM, serumcholesterol level, obesity,exercise, cholesterol intake,saturated andpolyunsaturated fat intake,parental history of MIbefore age 60 y,menopausal status,hormone use, study period

Klatsky et al,7

1989107137 US

men andwomenaged 50 y


6 Surveillance of hospitaldischarges

Clinical diagnosis 674 Age, sex, race, cigarettesmoking, SBP, coffeeconsumption, BMI,baseline disease

Shaper et al,8

19917735 UK men

aged40-59 y

In-personinterview

8 Death register Cl inical diagnosis ordeath certificate

110 Age, cigarette smoking, SBP

Goldberg et al,9

19946069 Men aged

51-75 y inHawaii

In-personinterview

15 Clinicalexaminationsat years 2 and6 andcontinuedsurveillance

Hospital dischargediagnosis, clinicaldiagnosis, ordeath certificate

70 Age; cigarette smoking; SBP;serum cholesterol, serumtriglyceride, and serum uricacid levels, coffeeconsumption, total caloricintake

Hansagi et al,10

199515 077 Men and

womenaged 40 yin Sweden


20 Death register Death certificate 769 Age, cigarette smoking

Iso et al,11 1995 2890 Men aged40-69 y inJapan

In-personinterview

10.5 Not specified Clinical diagnosis andCT scan

178 Age, cigarette smoking,hypertension, serum totalcholesterol level, DM

Kiyohara et al,12

19951621 Men and

womenaged 40 yin Japan

In-personinterview

26 Biennialexaminations,mail, ortelephone

Neurologicalexamination, CTscan,angiography,lumbar puncture,or autopsy

304 Age, sex, hypertension

Palmer et al,13

19956369 Men and

womenaged18-90 y inEngland

In-personinterview(1971-1976)

Self-administeredquestionnaire(after 1976)

22 Questionnaireevery 1-2 y

Death cert ificate 159 Age, sex, cigarette smoking,SBP

Yuan et al,14

199718 244 Men

aged45-64 y inChina

In-personinterview

9 Annual contact Death certificate 269 Age, cigarette smoking,educational level

Maskarinec etal,15 1998

27 678 Men andwomenaged 30 yin Hawaii

In-personinterview

20 Passive follow-up Death certificate 433 Age, BMI, cigarette smoking,ethnicity, educational level

Hart et al,16

19995766 Men aged

35-64 y inScotland

In-personinterview

21 NHS deathregister

Death cert ificate 133 Age, BMI, cigarette smoking,DBP, serum cholesterollevel, educational level,social class, fathers socialclass, car use, siblings,deprivation category,adjusted FEV, angina,ischemia on ECG,bronchitis

(continued)





4/12

significant heterogeneity was identifiedamong studies.68 A weighted meta-regressionanalysiswith no intercepttermwas performed to examine the associa-tion between alcohol consumption andthe natural logarithm of the relative risk

of stroke. We used the pool-firstmethod proposed by Greenland andLongnecker.69 This method was chosenbecause several studies reported find-ing a nonlinear, J- or U-shaped relation-ship between alcohol consumption andrelative risk of stroke. This methodis ad-vantageous because it can easily be ex-tended to test nonlinearity and identifyJ- or U-shapedcurves, or other relation-ships between exposure levels and rela-tive risks. For each included study, weperformed an initial fit of a quadratic

curve. When a nonsignificant term wasfound in the initial model, a subse-quent fit of a simpler model (linear orsolitary square term) was conducted.

Prestated subgroup analyses wereconducted by subtype of stroke and sexfor the different levels of alcohol con-sumption. Subgroup analyses were notperformed by type of alcoholic bever-

age due to the lack of such detailed in-formation in most studies.

To assess the potential for publica-tion bias, we constructed a funnel plotin which thelog relative risks were plot-ted against their SEs.70 In addition, a

rank correlation for the association be-tween standardized log relative risksand their SEs was conducted using theKendall correlation coefficient. Thecorrelation between sample size andrelative risk would be high if smallstud-ies with null results were less likely tobe published. A significant correlationbetween sample size and relative riskwould not exist in the absence of thistype of publication bias.70

RESULTS

The characteristics of the study sub-jects and design of the cohort studiesare presented in TABLE 1.Ofthe19co-hort studies, 8 were conducted in theUnited States. The number of subjectsin the cohort studies ranged from 1621in the study by Kiyohara et al12 to107137 in the study by Klatsky et al.7

Among the 19 cohort studies, 15 re-

ported total stroke as the outcome. Inaddition, 7 studies reported ischemicstroke, and 7 studies reported hemor-rhagic stroke as the outcome. The fol-low-up period ranged from 4 to 30years. The study population in 7 co-

hort studies consisted of men andwomen, 1 consisted entirely of women,and 11 consisted of only men.

Twelve of the 16 case-control stud-ies were conducted outside the UnitedStates (TABLE 2). The number of casesubjects enrolled in these studiesranged from 89 in the study by Hen-rich and Horwitz26 to 677 in the studyby Sacco et al,34 and the correspondingnumber of control subjects rangedfrom 153 in the study by Palomki etal29 to 1139 in the study by Sacco et

al.

34

Total stroke was the study out-come in 9 studies, whereas 8 studiescollected data on ischemic stroke and5 collected data on hemorrhagicstroke. Fourteen of the 16 case-controlstudies were composed of both menand women, 1 case-control study con-sisted of only women, and 1 case-control study consisted of only men.

Table 1. Characteristics of 19 Cohorts Studies of Alcohol Consumption and Risk of Stroke (cont)

Source, yStudy

ParticipantsExposure

AssessmentDuration of

Follow-up, yFollow-up

ProcessOutcome

AssessmentNo. of

Stroke Cases Controlled Variables

Leppl et al,17

199926 556 Men

aged50-69 y inFinland


6.1 National hospitaldischargeregister andnational deathregister

Clinical diagnosis ordeath certificate

960 Age, BMI, cigarette smoking,serum cholesterol level,DM, educational level,leisure time physicalactivity, heart disease,

supplementation with-tocopherol or betacarotene

Romelsj et al,18

199949 618 Men

aged17-45 y inSweden


25 Inpatient careregister anddeath register

Clinical diagnosis ordeath certificate

223 BMI, cigarette smoking, BP,fathers social class,running away from home,poor school well-being,parental divorce, pooremotional control, fewfriends, unemployment 3mo during life, poor health

Gaziano et al,19

200089 299 US men

aged40-84 y


5.5 National DeathIndex search

Death cert ificates 150 Age, BMI, cigarette smoking,DM, exercise

Jousilahti et al,20

200014 874 Men

and womenaged25-64 y inFinland


12 National hospitaldischargeregister orcentralstatisticaloffice ofFinland

Clinical diagnosis ordeath certificates

470 Age, BMI, cigarette smoking,serum total cholesterol,SBP, DBP, and study year

Djousse et al,21

20025209

Framing-ham, Mass,men andwomen

In-personinterview

30 Biennialexaminations

Clinical diagnosis andradiographicimages

441 Age, BMI, cigarette smoking,DM

Abbreviations: BMI, body mass index; CT, computed tomography; DBP, diastolic blood pressure; DM, diabetes mellitus; ECG, electrocardiogram; FEV, forced expiratoryvolume; MI, myocardial infarction; NHS, National Health Service; SBP, systolic blood pressure.





5/12

Table 2. Characteristics of 16 Case-Control Studies of Alcohol Consumption and Risk of Stroke

Source, y Stroke Cases Controls Case AssessmentExposure

Assessment Controlled Variables

Herman et al,22

1983132 Male and female

patients withincident strokeevent in 2 hospitalsin the Netherlands

239 Patients from thesame hospital

Clinical examination In-person interview Age, sex

Gill et al,23 1986 230 Male and femalepatients withstroke diagnosis inthe district hospitalin England

230 Hospital patients Clinical examination,CT scan,angiography, andpostmortemexaminations, orlumbar puncture

In-person interview Age, sex, race, cigarettesmoking, treatment ofhypertension, medication

Gill et at,24 1988 230 Male and femalepatients withstroke diagnosis inthe district hospitalin England

577 Male and femaleindustrial workers inthe samecommunity

Clinical examination,CT scan,angiography, andpostmortemexaminations, orlumbar puncture

In-person interview Age, race, cigarette smoking,treatment of hypertension,social class, drug therapy

Gorelick et al,25


patients withincident ischemicstroke in 3 medicalcenters in Chicago

410 Outpatient clinicpatients

Clinical diagnosis andCT scan

In-person interview Age, sex, race, cigarettesmoking, hypertension,method of hospitalpayment

Henrich andHorwitz,26

1989

89 Male and femalehospitalizedpatients withischemic stroke inConnecticut

178 Patientsdischarged from thesame hospital

Clinical examinationand CT scan Telephone interview None

Gill et al,27 1991 621 Male and femalehospitalizedpatients withstroke diagnosis in2 centers inEngland

573 Male and femaleindustrial workers inthe samecommunity

Clinical examination,CT scan,angiography andpostmortemexamination, orlumbar puncture

In-person interview Age, sex, race, cigarettesmoking, hypertension,social class, medication

Ben-Shlomoet al,28 1992

115 Male and femalehospitalizedpatients withincident stroke in 3hospitals in theUnited Kingdom

165 Generally matched,115 selectivelymatched, and 752community controls

Clinical examination,CT scan, orlumbar puncture

Cases, in-persontaped interview

Controls,self-administeredquestionnaire

General and selective controls:age, sex, cigarettesmoking, hypertension,DM, heart disease

Community controls: age,sex, cigarette smoking,hypertension, and socialclass

Palomki et al,29

1993156 Male hospitalized

patients withischemic stroke inFinland

153 Hospital patients Clinical diagnosis In-person interview Age, BMI, cigarette smoking,hypertension, DM,coronary heart disease,history of snoring

Shinton et al,30


patients withincident stroke in11 general practicepartnerships inEngland

198 Communitycontrols

Clinical examination,CT scan, orautopsy

Alcohol diary Age, sex, history ofcardiovascular disease

Jamrozik et al,31


patients withstroke diagnosis inAustralia

931 Communitycontrols from theelectoral roles

Clinical examination,CT scan, MRI, orautopsy

In-person interview Age, sex, cigarette smoking,hypertension, DM,previous stroke or TIA,previous MI, adding salt tofood, consumption of fish2 times/mo,claudication, use ofreduced fat or skim milk,

consumption of meat 4times/wk

Beghi et al,32


hospitalizedpatients withstroke in Italy

170 Patients in thesame hospital and202 communitycontrols

Clinical examination,CT scan, orneurologicalconsultation

In-person interview Age, sex

Caicoya et al,33


patients withincident stroke inSpain

477 Residents of thesame community

Clinical examination orCT scan

In-person interview Age, sex, cigarette smoking,hypertension, DM,hypercholesterolemia,cardiac disease

(continued)





6/12

The results from the random-effects model and the meta-regressionanalysis test for trend are presented inTABLE 3. The overall results indicate anonlinear association between alco-hol consumption and relative risk oftotal stroke (P =.002 for nonlinear

trend). Compared with the referencegroup of abstainers, alcohol consump-tion of less than 12 g/d, or less than 1drink per day based on US conver-sions, was significantly associated witha decreased relative risk of total stroke,while alcohol consumption of morethan 60 g/d, or more than 5 drinks perday, was significantly associated withan increasedrelative risk of total stroke.The association between alcohol con-sumption and relative risk of ischemicstroke was J-shaped with thelowest risk

among those consuming less than 12g/d, or less than 1 drink per day, or 12to 24 g/d, or 1 to 2 drinks per day, andthe highest risk among those consum-ing more than 60 g/d, or more than 5drinksper day, (FIGURE 1). Relativeriskof hemorrhagic stroke increased lin-early with increasingalcohol consump-tion, and those consuming more than

60 g/d, or more then 5 drinks per day,had the highest relative risk.

The associationbetweenalcohol con-sumption and relative risk of totalstroke was similar in men and women(Table 3 and FIGURE 2) although therelative risk was somewhat lower in

women consuming less than 12 g/d, orlessthan 1 drink per day, than in men.Likewise, the association was similar incase-control studies and cohort stud-ies, with alcohol consumption of lessthan 12g/d, or lessthan1 drink per day,among cohort studies and alcohol con-sumption of less than 24 g/d, or lessthan 2 drinks per day, among case-control studies associated with a sig-nificant reduced relative risk while al-cohol consumption of more than60 g/d,or more than 5 drinks per day, was as-

sociated with an increased relative risk.The findings from the sensitivityanalyses that excluded studies based ondifferent inclusion criteria are pre-sented in TABLE 4. Risk estimateschanged very little after the exclusionof outliers, studies without computedtomographic scans or other imagingmeasures, studies that did not adjust for

important confounders, or studies thatdid not exclude prevalent stroke casesat baseline.

There was no evidence of publica-tion bias in our study as indicated by afunnel plot (FIGURE 3) and the Ken-dall correlation coefficient. The Ken-

dall correlation coefficient for the SEand the standardized log relative riskwas 0.072 (P =.17) for all studies.When the outliers were excluded, theKendal correlation coefficient for theSE andthe standardized log relative riskbecame 0.053 (P=.32).

COMMENT

Several largeepidemiologicstudies thathave examined the effect of alcohol con-sumption on therisk of stroke have pro-vided inconsistent findings. In our cur-

rent meta-analysis,we found a J-shapedassociation between alcohol consump-tion and the relative risk of total andischemic stroke and a linear associa-tion between alcohol consumption andthe relative risk of hemorrhagic stroke.Moderate alcohol consumption was as-sociated with a reduced relative risk oftotal and ischemic stroke while heavy

Table 2. Characteristics of 16 Case-Control Studies of Alcohol Consumption and Risk of Stroke (cont)

Source Stroke Cases Controls Case AssessmentExposure

Assessment Controlled Variables

Sacco et al,34

1999677 Men and women with

incident cerebralinfarction in thecommunity inNew York

1139 Communitycontrols

Brain imaging andclinical diagnosis

In-person interview Age, sex, race, BMI, cigarettesmoking, hypertension,DM, cardiac disease,educational level

Thrift et al,35


patients with primaryhemorrhagic strokefrom 13 hospitals inMelbourne, Australia

331 Residents from thesame neighborhood

CT scan, MRI, orautopsy

In-person interview Age, sex, BMI, cigarettesmoking, DM, serumcholesterol level, SES,educational level, exercise,cardiovascular disease,hormone replacementtherapy

Zodpey et al,36


hospitalized patientswith incidenthemorrhagic strokein India

166 Patients from thesame hospital

CT scan In-person interview Age, sex

Malarcher et al,37

2001224 Female patients with

incident cerebralinfarction in 59hospitals inBaltimore-Washingtonregion in the UnitedStates

392 Female communityresidents

Hospital dischargediagnosis, clinicaldiagnosis,neuroimagingresults, or autopsyreports

In-person interview Age, race, BMI, cigarettesmoking, hypertension,DM, total cholesterol, HDLcholesterol level,geographic region ofresidence, educationallevel, coronary heartdisease

Abbreviations: BMI, Body mass index; CT, computed tomography; DM, diabetes mellitus; MI, myocardial infarction; MRI, magnetic resonance imaging; SES, socioeconomic status;TIA, transient ischemic attack.





7/12

alcohol consumption was associatedwith an increased relative risk of total,ischemic, and hemorrhagic stroke.

Therelationshipbetweenalcoholcon-sumption and stroke is believed to in-volve various mechanisms including

alcohol-induced hypertension, cardio-myopathy, coagulation disorders, atrialfibrillation, and reductions in cerebral

Figure 1. Scatterplot of Log Relative Risk and Meta-Regression Curve of Stroke Associated With Alcohol Consumption by Subtypes of Stroke

3

1

2

0

1

2

3 0 20 18040 60 80 100 120 140 160 0 20 18040 60 80 100 120 140 160

Alcohol Intake, g/d

NaturalLogarithmR

elativeRisk

Ischemic Stroke

Alcohol Intake, g/d

Hemorrhagic Stroke

Most studies provided more than 1 relative risk estimate for multiple levels of alcohol consumption.

Figure 2. Scatterplot of Log Relative Risk and Meta-Regression Curve of Stroke Associated With Alcohol Consumption by Sex

3

1

2

0

1

2

30 20 40 60 80 100 120 140

Alcohol Intake, g/d

NaturalLoga

rithmR

elativeRisk

Men

0 10 80706050403020

Alcohol Intake, g/d

Women

Most studies provided more than 1 relative risk estimate for multiple levels of alcohol consumption.

Table 3. Overall Relative Risk (95% Confidence Interval) of Stroke Associated With Alcohol Consumption and Test for Trend

No. ofStudies

Alcohol Intake, g/d

P Value

12 12-24 24-60 60Test for LinearAssociation*

Test for NonlinearAssociation

Overall 35 0.83 (0.75-0.91) 0.91 (0.78-1.06) 1.10 (0.97-1.24) 1.64 (1.39-1.93) .002

Type of strokeIschemic 15 0.80 (0.67-0.96) 0.72 (0.57-0.91) 0 .96 (0.79-1.18) 1.69 (1.34-2.15) .004

Hemorrhagic 12 0.79 (0.60-1.05) 0.98 (0.77-1.25) 1.19 (0.80-1.79) 2.18 (1.48-3.20) .004 .17

SexMen 27 0.89 (0.79-1. 01) 0. 94 ( 0. 84-1.05) 1.08 (0.96-1. 21) 1. 76 ( 1. 57-1.98) .001

Women 16 0.66 (0.61- 0.71) 0.79 (0.56-1.11) 0.80 (0.49- 1.30) 4.29 (1.30-14.14) .001

Study designCohort 19 0.82 (0.73-0.92) 0.94 (0.84-1.05) 1.06 (0.90-1.23) 1.63 (1.49-1.79) .02

Case control 16 0.80 (0.67-0.97) 0.65 (0.44-0.96) 1.12 (0.92-1.37) 1.98 (1.35-2.92) .03

*Tests for linear associations were performed only when nonlinear associations were not statistically significant.





8/12

blood flow.37,46,71,72 A plausible explana-tion of a reduced risk of ischemic stroke

with moderate alcohol consumption isthat alcohol increases high-density li-poprotein cholesterol levels and de-creases platelet aggregation and fibri-nolytic activity.6,71,72 Epidemiologicstudies also have consistently observeda protective effect of moderate alcoholconsumption on coronary heart dis-ease.73,74 Alcohol-induced hyperten-sion andcoagulation disordersare prob-able underlying mechanisms forhemorrhagic stroke.27,45,71 The antico-agulant effects of alcohol, althoughtheyappear to be beneficial for decreasing the

risk of ischemic stroke, may play an im-portant role in increasing the risk ofhemorrhagic stroke.71,74

There are several potential limita-tions in our study. First, our study is ameta-analysis of observational stud-ies. The quality of our study dependson data from original publications in-

cluded in our analysis. Our study mayinherit the problems of potential biasandconfounding effects associatedwithobservational studies. However, a ran-domizedcontrolled trial of alcohol con-sumption and stroke has not been per-

formed and is unlikely to be conductedin the future. Consequently, we mustrely on data from observational stud-ies to draw conclusions and make rec-ommendations.

Second, computed tomographicscans and other imaging techniqueswere not available for some early stud-ies. Furthermore, several studies onlyused death certificates or death regis-ter data for diagnosis of stroke out-come. However, our findings were un-likely due to misclassification of

outcome because the relative risks ofstroke associated with alcohol con-sumption did not change after exclu-sion of studies that did not use com-puted tomography or other imagingtechniques for diagnosis. Our find-ings were also unlikely due to con-founding effects because the relativerisks of stroke associated with alcoholconsumption were similar among allstudies and only those studies that con-trolled for important risk factors forstroke, such as cigarette smoking andhypertension. Additionally, our re-

sults were unlikely to result from pub-lication bias as demonstrated by the fun-nel plot and rank correlation analysis.

Several methodological issues re-garding epidemiologic research on thehealth impact of alcohol consumptionare worth considering. First, the selec-tion of the reference group may vary

among studies. For instance, some stud-ies used the lowest consumption levelas thereference group while others usedabstainers. In an effort to avoid com-bining studies that were not compa-rable, we chose to include only those

studies that used abstainers as the ref-erencegroup. It hasbeen suggested thatthe U- or J-shaped association be-tween alcohol consumption and mor-tality from cardiovascular disease maybe due to the inclusion of ex-drinkersin thereferencegroup of abstainers. Ex-drinkersmay havestopped alcoholcon-sumption due to health problems andthey are at increased risk for death fromcardiovascular disease.47,75,76 How-ever, several studies haveexaminedthispotential bias and concluded that the

J- or U-shaped relationship between al-cohol consumption and risk of cardio-v ascul ar di sease m o r tal i ty hel dtrue.6,13,27,77 Moreover, we conducted asensitivity analysis in which only pro-spective cohort studies that excludedprevalent stroke cases at baseline wereincluded, and we found that the shapeof association remained unchanged.Second, the health effects of bingedrinking may be differentthan those forregular drinkers. The failure to differ-entiate between these 2 groups couldpossibly obscure the observation of any

true association. Therefore, we only in-cluded studies that examined the effectof usual alcohol consumption ratherthanacutealcohol consumption. Third,the measurement units, especially thedefinition of an alcohol drink, variesamong studies. We attempted to over-come this problem by applying a com-

Figure 3. Funnel Plot of Log Relative Risk vsVariance of Log Relative Risks Among AllStudies

100

60

40

80

20

0

2.5 1.5 2.50.5 0.5 1.5

Natural Logarithm Relative Risk

Variance

Most studies provided more than 1 relative risk esti-mate for multiple levels of alcohol consumption.

Table 4. Overall Relative Risk (95% Confidence Interval) of Stroke Associated With Alcohol Consumption According to Different ExclusionCriteria*

Studies Included in Analysis

Alcohol Intake, g/d

12 12-24 24-60 60

All studies 0.83 (0.75-0.91) 0.91 (0.78-1.06) 1.10 (0.97-1.24) 1.64 (1.39-1.93)

All studies except outliers* 0.83 (0.75-0.91) 0.91 (0.78-1.06) 1.11 (0.98- 1.26) 1.62 (1.46- 1.81)

Studies that used computed tomography scans orother imaging measures as an outcome measure

0.84 (0.75-0.94) 0.86 (0.71-1.05) 1.14 (1.01- 1.35) 1.74 (1.37- 2.21)

Studies that controlled for important stroke risk factors 0.81 (0.71-0.92) 0.80 (0.64-1.00) 1.12 (0.94-1.33) 1.62 (1.19-2.21)

Cohort studies that used incident stroke events 0.83 (0.73-0.95) 0.91 (0.77-1.07) 1.02 (0.83-1.26) 1.58 (1.43-1.73)

*The 24 to 60 and 60 g/d levels in the study by Caicoya et al 33 were excluded because they were outliers.Studies that solely used death certificates or death registries for the outcome assessment were excluded. 4,5,10,13-16,19

Studies that did not control for age, cigarette smoking or hypertension were excluded.4,5,10,12,14,15,17-19,21,22,26,30,32,35,36

Cohort studies that did not exclude prevalent stroke events were excluded.4,5,8,10,13-16





9/12

monly used and validated method sug-gested by Turner.67 Finally, assessmentmethods for alcohol consumption mayalso vary among studies. The assess-ment of alcohol consumption is usu-ally based on self-reported alcohol hab-

its. Such data are subject to errors ofrecall.For example, heavy drinkersmaybe more likely to underreport their al-cohol consumption. The majority ofstudies in this meta-analysis used in-person interviews, while 11 studiesusedself-administered questionnaires, 1study conducted telephone inter-views, and 1 study used alcohol con-sumption diaries.

There are several advantages of ourstudy. The discrepancies among stud-ies regarding the association betweenalcohol consumption and relative risk

of stroke also may be attributable to asmallsample size in theindividualstud-ies, resulting in insufficient statisticalpower. This meta-analysis included alarge number of people from differentpopulations throughout the world. Ad-ditionally, we were able to assess thepatternof the association between levelof alcohol consumption andrelative riskof stroke with precision due to the largesample size. Finally, the association be-tween alcohol consumption and rela-tive risk of stroke was consistent among

subgroups by study design, sex, andstroke subtype.

Our findings have important clini-cal and public health implications. Inthe United States, 44% of adults, aged18 years or older, are current drinkerswho have consumed at least 12 drinksin the preceding year.78 Stroke is a ma-jor cause of death and disability in theUnited States and other countries.2 Inthe United States, there are approxi-mately 600000 new stroke cases eachyear.1 Given the widespread consump-

tion of alcohol in the general popula-tion andthe recognized health and eco-nomic burdens of stroke, our findingsare both important and timely. Ourstudy strongly suggests that reducingalcohol consumption in heavy drink-ers should be an important approachto prevention of stroke in the generalpopulation. Our study also suggests that

moderate alcohol consumption re-duces risk of ischemic stroke. How-ever, the implications of these find-ings should be examined cautiously.Any advice regarding the consump-tion of alcohol should be tailored to the

individual patients risks and poten-tial benefits.

Author Contributions: Study concept and design:Reynolds, Lewis, Nolen, Kinney, Sathya, He.Acquisition of data: Reynolds, Lewis, Nolen, Kinney,Sathya.Analysis and interpretation of data:Reynolds, Lewis,Nolen, Kinney, Sathya, He.Drafting of the manuscript: Reynolds, Nolen, Kinney,Sathya.Critical revision of the manuscript for important in-tellectual content: Reynolds, Lewis, Nolen, Kinney,Sathya, He.Statistical expertise: Reynolds, Nolen, Kinney, He.Obtained funding: He.Administrative, technical, or material support:Lewis,Kinney, Sathya.Study supervision: He.

Funding/Support: This study was supported in partby grant R01HL60300 fromthe National Heart, Lung,and Blood Institute.

REFERENCES

1. American Heart Association. 2002 Heart andStrokeStatistical Update. Dallas, Tex: American Heart As-sociation; 2001.2. Warlow CP. Epidemiologyof stroke. Lancet. 1998;352(suppl 3):1-4.3. Donahue RP, Abbott RD, Reed DM, Yano K. Al-coholand hemorrhagicstroke: the Honolulu HeartPro-gram. JAMA. 1986;255:2311-2314.4. Kono S, Ikeda M, Tokudome S, Nishizumi M, Ku-ratsune M. Alcohol and mortality: a cohort study ofmale Japanese physicians. Int J Epidemiol. 1986;15:527-532.5. Gordon T, Doyle JT. Drinking and mortality: the

Albany Study. Am J Epidemiol. 1987;125:263-270.6. Stampfer MJ, Colditz GA, Willett WC, Speizer FE,Hennekens CH. A prospective study of moderate al-cohol consumption and the risk of coronary diseaseand stroke in women. N Engl J Med. 1988;319:267-273.7. Klatsky AL, Armstrong MA, Friedman GD. Alco-hol use and subsequent cerebrovascular disease hos-pitalizations. Stroke. 1989;20:741-746.8. ShaperAG, PhillipsAN, Pocock SJ,Walker M, Mac-farlane PW.Riskfactorsfor stroke in middleagedBrit-ish men [comments]. BMJ. 1991;302:1111-1115.9. Goldberg RJ, Burchfiel CM, Reed DM, Wer-gowske G, Chiu D. A prospective study of the healtheffects of alcohol consumption in middle-aged andel-derly men: theHonoluluHeart Program. Circulation.1994;89:651-659.10. Hansagi H, Romelsjo A, Gerhardsson de VerdierM, Andreasson S, Leifman, A. Alcohol consumption

andstrokemortality: 20-yearfollow-up of 15077 menand women. Stroke. 1995;26:1768-1773.11. Iso H, Kitamura A, Shimamoto T, et al. Alcoholintake andthe riskof cardiovascular diseasein middle-aged Japanese men. Stroke. 1995;26:767-773.12. Kiyohara Y, Kato I, Iwamoto H, Nakayama K, Fu-jishima M. The impact of alcohol and hypertensiononstroke incidencein a general Japanesepopulation:theHisayama Study. Stroke. 1995;26:368-372.13. Palmer AJ, Fletcher AE, Bulpitt CJ, et al. Alcoholintakeand cardiovascularmortality in hypertensive pa-tients: report from the Department of Health Hyper-

tension Care Computing Project. J Hypertens. 1995;13:957-964.14. Yuan JM,Ross RK,Gao YT,Henderson BE,Yu MC.Followup study of moderate alcohol intakeand mor-tality among middle aged men in Shanghai, China[comments]. BMJ. 1997;314:18-23.15. Maskarinec G, Meng L, Kolonel L. Alcohol in-take, body weight, andmortality in a multiethnic pro-spective cohort. Epidemiology. 1998;9:654-661.

16. Hart CL,SmithGD, Hole DJ,HawthorneVM. Al-cohol consumption andmortalityfrom all causes,coro-naryheart disease,and stroke: results froma prospec-tive cohort study of Scottish men with 21 years offollow up [comments]. BMJ. 1999;318:1725-1729.17. Leppala JM, Paunio M, Virtamo J, et al. Alcoholconsumption and stroke incidence in male smokers.Circulation. 1999;100:1209-1214.18. Romelsjo A, Leifman A. Association between al-cohol consumption and mortality, myocardial infarc-tion, and strokein 25yearfollow upof 49618 youngSwedish men. BMJ. 1999;319:821-822.19. Gaziano JM, Gaziano TA, Glynn RJ, et al. Light-to-moderatealcoholconsumption andmortality in thePhysicians Health Study enrollment cohort.J Am CollCardiol. 2000;35:96-105.20. Jousilahti P, Rastenyte D, Tuomilehto J. Serumgamma-glutamyl transferase, self-reported alcoholdrinking, andthe riskof stroke.Stroke. 2000;31:1851-

1855.21. Djousse L, Ellison R, Beiser A, Scaramucci A,DAgostino R, Wolf P. Alcohol consumption and riskof ischemic stroke: the Framingham study. Stroke.2002;33:907-912.22. Herman B, Schmitz PI, Leyten AC, et al. Multi-variatelogistic analysis of riskfactors forstrokein Til-burg, the Netherlands. Am J Epidemiol. 1983;118:514-525.23. Gill JS, Zezulka AV, Shipley MJ, Gill SK, BeeversDG. Stroke and alcohol consumption. N Engl J Med.1986;315:1041-1046.24. Gill JS, Shipley MJ, Hornby RH, Gill SK, BeeversDG. A community case-control study of alcohol con-sumption in stroke. Int J Epidemiol. 1988;17:542-547.25. Gorelick PB, Rodin MB, Langenberg P, Hier DB,Costigan J. Weekly alcohol consumption, cigarettesmoking, and the risk of ischemic stroke: results of a

case-control study at three urban medical centers inChicago, Illinois. Neurology. 1989;39:339-343.26. Henrich JB, Horwitz RI. Evidence against the as-sociation betweenalcohol useand ischemicstrokerisk.Arch Intern Med. 1989;149:1413-1416.27. Gill JS, Shipley MJ, Tsementzis SA, et al. Alcoholconsumption: a risk factor for hemorrhagic and non-hemorrhagic stroke. Am J Med. 1991;90:489-497.28. Ben-Shlomo Y, Markowe H, Shipley M, MarmotMG. Stroke risk from alcohol consumption using dif-ferent control groups. Stroke. 1992;23:1093-1098.29 . Paloma ki H, Kaste M. R egular light-to-moderate intake of alcohol and the risk of ischemicstroke: is there a beneficial effect? Stroke. 1993;24:1828-1832.30. Shinton R, Sagar G, Beevers G. Therelationof al-cohol consumption to cardiovascular risk factors andstroke: the West Birmingham Stroke Project. J Neu-rol Neurosurg Psychiatry. 1993;56:458-462.

31. JamrozikK, Broadhurst RJ, AndersonCS, Stewart-Wynne EG. The role of lifestyle factors in the etiol-ogy of stroke: a population-based case-control studyin Perth, Western Australia. Stroke. 1994;25:51-59.32. Beghi E,Boglium G,Cosso P,et al.Stroke and al-cohol intake in a hospital population: a case-controlstudy. Stroke. 1995;26:1691-1696.33. Caicoya M, Rodriguez T, Corrales C, Cuello R,Lasheras C. Alcohol and stroke: a community case-control studyin Asturias, Spain.J Clin Epidemiol. 1999;52:677-684.34. Sacco RL, Elkind M, Boden-Albala B, et al. The





10/12

protective effect of moderatealcoholconsumptiononischemic stroke. JAMA. 1999;281:53-60.35. Thrift A, Donnan G, McNeil J. Heavy drinking, butnot moderate or intermediate drinking, increases therisk of intracerebral hemorrhage. Epidemiology. 1999;10:307-312.36. Zodpey SP, Tiwari RR, Kulkarni HR. Risk factorsfor haemorrhagic stroke: a case-controlstudy. PublicHealth. 2000;114:177-182.

37. MalarcherAM, Giles WH, Croft JB,et al.Alcoholintake, type of beverage, and the risk of cerebral in-farction in young women. Stroke. 2001;32:77-83.38. MazzagliaG, Britton A, AltmannD, ChenetL. Ex-ploring the relationship between alcohol consump-tion and non-fatal or fatal stroke: a systematic re-view. Addiction. 2001;96:1743-1756.39. Camargo CA Jr. Case-control and cohort studiesof moderate alcohol consumption and stroke. ClinChim Acta. 1996;246:107-119.40. Papadakis JA, Ganotakis ES, Mikhailidis DP. Ben-eficial effect of moderatealcohol consumptionon vas-cular disease: myth or reality? J R Soc Health. 2000;120:11-15.41. Eichner ER. Alcohol, stroke and coronary arterydisease. Am Fam Physician. 1988;37:217-221.42. HillbomM, Kaste M. Alcoholabuseand brain in-farction. Ann Med. 1990;22:347-352.43. Ashley MJ, Ferrence R, Room R, Bondy S, RehmJ, Single E. Moderate drinking and health: implica-tions of recent evidence. Can Fam Physician. 1997;43:687-694.44. Klatsky AL. Alcohol and cardiovascular diseases:a historical overview. Novartis Found Symp. 1998;216:2-12.45. Gorelick PB,KellyMA. Alcoholas a risk factorforstroke. Heart Dis Stroke. 1992;1:255-258.46. HillbomM. Alcoholconsumption andstroke: ben-efits and risks. Alcohol Clin Exp Res. 1998;22(suppl7):352S-358S.47. Wannamethee SG, Shaper AG. Alcohol, coronaryheart diseaseand stroke: anexaminationof theJ-shapedcurve. Neuroepidemiology. 1998;17:288-295.48. BeilinLJ, PuddeyIB, Burke V. Alcohol andhyper-tension: kill or cure? J Hum Hypertens. 1996;10(suppl 2):S1-S5.49. Monforte R, Estruch R, Graus F, Nicolas JM, Ur-bano-Marquez A. High ethanol consumption as riskfactor for intracerebral hemorrhage in young and

middle-aged people. Stroke. 1990;21:1529-1532.

50. Beghi E, Bogliun G, Cosso P, et al. Cerebrovas-cular disorders and alcohol intake: preliminary resultsof a case-control study. Ital J Neurol Sci. 1992;13:209-214.51. HillbomM, HaapaniemiH, JuvelaS, Palomaki H,Numminen H, Kaste M. Recent alcohol consump-tion,cigarette smoking,and cerebral infarctionin youngadults. Stroke. 1995;26:40-45.52. Haapaniemi H, Hillbom M, Juvela S. Lifestyle-

associated riskfactors foracute brain infarction amongpersons of working age. Stroke. 1997;28:26-30.53. Hillbom M,NumminenH, JuvelaS. Recentheavydrinking of alcohol andembolic stroke.Stroke. 1999;30:2307-2312.54. JuvelaS, HillbomM, Palomaki H. Riskfactors forspontaneous intracerebral hemorrhage.Stroke. 1995;26:1558-1564.55. Haapaniemi H, Hillbom M, JuvelaS. Weekendandholiday increasein theonset of ischemic strokein youngwomen. Stroke. 1996;27:1023-1027.56. Blackwelder WC, Yano K, Rhoads GG, et al. Al-cohol and mortality: the Honolulu Heart Study. Am JMed. 1980;68:164-169.57. Kagan A, Popper JS, Rhoads GG. Factors relatedto stroke incidence in Hawaii Japanese men: the Ho-nolulu Heart Study. Stroke. 1980;11:14-21.58. Rodgers H, Aitken PD, French JM, Curless RH,Bates D, James FW. Alcohol and stroke: a case-control study of drinking habits past and present.Stroke. 1993;24:1473-1477.59. Klatsky AL, Friedman GD, Siegelaub AB. Alcoholand mortality: a ten-year Kaiser-Permanente experi-ence. Ann Intern Med. 1981;95:139-145.60. Klatsky AL, Armstrong MA, Friedman GD. Riskof cardiovascular mortality in alcohol drinkers, ex-drinkers and nondrinkers. Am J Cardiol. 1990;66:1237-1242.61. Berger K, Ajani UA, Kase CS, et al. Light-to-moderate alcohol consumption and risk of strokeamong US male physicians. N EnglJ Med. 1999;341:1557-1564.62. Wannamethee SG, Shaper AG. Patterns of alco-holintake andriskof strokein middle-agedBritish men.Stroke. 1996;27:1033-1039.63. Camargo CA Jr, Hennekens CH, Gaziano JM,Glynn RJ, Manson JE, Stampfer MJ. Prospectivestudy of moderate alcohol consumption and mortal-ityin US male physicians. ArchIntern Med. 1997;157:

79-85.

64. Truelsen T, Gronbaek M, Schnohr P, Boysen G.Intakeof beer, wine, andspirits andrisk of stroke: theCopenhagencity heartstudy.Stroke. 1998;29:2467-2472.65. Kissela B, Sauerbeck L, Woo D, et al. Subarach-noid hemorrhage: a preventable disease with a heri-table component. Stroke. 2002;33:1321-1326.66. Klatsky A, Armstrong M, Friedman G, Sidney S.Alcohol drinking and risk of hospitalization for ische-

mic stroke. Am J Cardiol. 2001;88:703-706.67. Turner C. How much alcohol is in a standarddrink: an analysis of 125 studies. Br J Addict. 1990;85:1171-1175.68. DerSimonian R, Laird N. Meta-analysis in clinicaltrials. Control Clin Trials. 1986;7:177-188.69. GreenlandS, Longnecker MP. Methods for trendestimation from summarized dose-responsedata, withapplications to meta-analysis. Am J Epidemiol. 1992;135:1301-1309.70. Begg CB. Publication bias. In: Cooper H, HedgesLF, eds. The Handbook of Research Synthesis. NewYork, NY: Russell Sage Foundation; 1994:399-409.71. Zakhari S. Alcohol and the cardiovascular sys-tem: molecular mechanisms for beneficial and harm-ful action. Alcohol Health Research World. 1997;21:21-29.72. Gorelick P, Sacco R, Smith D, et al. Prevention ofa first stroke: a review of guidelines and a multidisci-plinary consensusstatement fromthe National StrokeAssociation. JAMA. 1999;281:1112-1120.73. Marmot MG. Alcohol and coronary heart dis-ease. Int J Epidemiol. 2001;30:724-729.74. Rimm EB, Klatsky A, Grobbee D, Stampfer MJ.Review of moderate alcohol consumption and re-duced risk of coronary heart disease: is theeffect dueto beer, wine, or spirits. BMJ. 1996;312:731-736.75. Wannamethee G, Shaper AG. Men who do notdrink: a report from the British Regional Heart Study.Int J Epidemiol. 1988;17:307-316.76. Shaper AG, Wannamethee G, Walker M. Alco-hol and mortality in British men: explaining the U-shaped curve. Lancet. 1988;2:1267-1273.77. Maclure M. Demonstration of deductive meta-analysis: ethanol intakeand riskof myocardial infarc-tion. Epidemiol Rev. 1993;15:328-351.78. National Institute on Alcohol Abuse and Alcohol-ism. 10th Special Report to the US Congress on Al-cohol and Health. Rockville, Md: US Dept of Health

and Human Services; 2000.





11/12

widely, with particularly high rates of use by internists andphysicians in the Northeast and the South.

Michael A. Steinman, MDC. Seth Landefeld, MDSan Francisco VA Medical CenterSan Francisco, Calif

Ralph Gonzales, MD, MSPHUniversity of California, San Francisco

1. Gonzales R, Steiner JF, Lum A, Barrett PH Jr. Decreasing antibiotic use in am-bulatory practice: impact of a multidimensional intervention on the treatment ofuncomplicated acute bronchitis in adults. JAMA. 1999;281:1512-1519.2. GonzalesR, Malone DC, MaselliJH, Sande MA. Excessive antibiotic usefor acuterespiratory infections in the United States. Clin Infect Dis. 2001;33:757-762.

CORRECTIONS

Name Omitted: In the Original Contribution entitled Combination TherapyWithHormone Replacement and Alendronate for Prevention of Bone Loss in ElderlyWomen:A Randomized ControlledTrial published in theMay 21,2003,issue ofTHE JOURNAL (2003;289:2525-2533), Michael McClurg, MD, should be added tothe list of members of the Data and Safety Monitoring Board on page 2532 after

Peggy A. Norton, MD.

Error in Authors Name: In theReview articleentitledAlcohol Consumption andRisk of Stroke: A Meta-analysis published in the February 5, 2003, issue of THEJOURNAL (2003;289:579-588) in the byline, the initial letter L. was incorrectlyplaced in front of the name of author Brian Lewis, MPH.

CME ANNOUNCEMENT

Online CME to Begin in Mid-2003

In mid-2003, online CME will be available for JAMA/Archives journalsand will offer many enhancements:

Article-specific questions Hypertext links from questions to the relevant content Online CME questionnaire Printable CME certificates and ability to access total CME credits

We apologize for the interruption in CME and hope that you willenjoy the improved online features that will be available in mid-2003.

LETTERS

2798 JAMA, June 4, 2003Vol 289, No. 21 (Reprinted) 2003 American Medical Association. All rights reserved.



12/12

widely, with particularly high rates of use by internists andphysicians in the Northeast and the South.

Michael A. Steinman, MDC. Seth Landefeld, MDSan Francisco VA Medical CenterSan Francisco, Calif

Ralph Gonzales, MD, MSPHUniversity of California, San Francisco

1. Gonzales R, Steiner JF, Lum A, Barrett PH Jr. Decreasing antibiotic use in am-bulatory practice: impact of a multidimensional intervention on the treatment ofuncomplicated acute bronchitis in adults. JAMA. 1999;281:1512-1519.2. GonzalesR, Malone DC, MaselliJH, Sande MA. Excessive antibiotic usefor acuterespiratory infections in the United States. Clin Infect Dis. 2001;33:757-762.

CORRECTIONS

Name Omitted: In the Original Contribution entitled Combination TherapyWithHormone Replacement and Alendronate for Prevention of Bone Loss in ElderlyWomen:A Randomized ControlledTrial published in theMay 21,2003,issue ofTHE JOURNAL (2003;289:2525-2533), Michael McClurg, MD, should be added tothe list of members of the Data and Safety Monitoring Board on page 2532 after

Peggy A. Norton, MD.

Error in Authors Name: In theReview articleentitledAlcohol Consumption andRisk of Stroke: A Meta-analysis published in the February 5, 2003, issue of THEJOURNAL (2003;289:579-588) in the byline, the initial letter L. was incorrectlyplaced in front of the name of author Brian Lewis, MPH.

CME ANNOUNCEMENT

Online CME to Begin in Mid-2003

In mid-2003, online CME will be available for JAMA/Archives journalsand will offer many enhancements:

Article-specific questions Hypertext links from questions to the relevant content Online CME questionnaire Printable CME certificates and ability to access total CME credits

We apologize for the interruption in CME and hope that you willenjoy the improved online features that will be available in mid-2003.

LETTERS

2798 JAMA, June 4, 2003Vol 289, No. 21 (Reprinted) 2003 American Medical Association. All rights reserved.

b t A il 23 2012jD l d d f

alcohol y acv

Documents