Context Studies have found that individuals who consume 1 alcoholic drink every
1 to 2 days have a lower risk of a first acute myocardial infarction (AMI)
than abstainers or heavy drinkers, but the effect of prior drinking on mortality
after AMI is uncertain.
Objective To determine the effect of prior alcohol consumption on long-term mortality
among early survivors of AMI.
Design and Setting Prospective inception cohort study conducted at 45 US community and
tertiary care hospitals between August 1989 and September 1994, with a median
follow-up of 3.8 years.
Patients A total of 1913 adults hospitalized with AMI between 1989 and 1994.
Main Outcome Measure All-cause mortality, compared by self-reported average weekly consumption
of beer, wine, and liquor during the year prior to AMI.
Results Of the 1913 patients, 896 (47%) abstained from alcohol, 696 (36%) consumed
less than 7 alcoholic drinks/wk, and 321 (17%) consumed 7 or more alcoholic
drinks/wk. Compared with abstainers, patients who consumed less than 7 drinks/wk
had a lower all-cause mortality rate (3.4 vs 6.3 deaths per 100 person-years;
hazard ratio [HR], 0.55; 95% confidence interval [CI], 0.43-0.71) as did those
who consumed 7 or more drinks/wk (2.4 vs 6.3 deaths per 100 person-years;
HR, 0.38; 95% CI, 0.25-0.55; P<.001 for trend).
After adjusting for propensity to drink and other potential confounders, increasing
alcohol consumption remained predictive of lower mortality for less than 7
drinks/wk, with an adjusted HR of 0.79 (95% CI, 0.60-1.03), and for 7 or more
drinks/wk, with an adjusted HR of 0.68 (95% CI, 0.45-1.05; P = .01 for trend). The association was similar for total and cardiovascular
mortality, among both men and women, and among different types of alcoholic
beverages.
Conclusion Self-reported moderate alcohol consumption in the year prior to AMI
is associated with reduced mortality following infarction.
In the general population, individuals who have 1 alcoholic drink every
1 to 2 days have lower rates of nonfatal acute myocardial infarction (AMI)
and coronary mortality compared with abstainers and heavier drinkers.1,2 Studies in patients with hypertension
or diabetes also show a U-shaped relationship between alcohol consumption
and coronary heart disease (CHD).3,4
However, the effect of drinking in patients with documented CHD is less certain.
Alcohol has diverse physiological actions. Although the balance of effects
appears to favor prevention of CHD in asymptomatic persons, it may differ
in those with established CHD. For example, alcohol lowers exercise tolerance
in patients with angina,5 causes dose-dependent
coronary vasoconstriction,6 raises heart rate
and systolic blood pressure,7,8
and increases levels of plasminogen activator inhibitor 1.9,10
These effects could adversely influence the prognosis of patients with CHD.
To our knowledge, no previous study has determined the effect of alcohol
consumption on mortality immediately following AMI. Two previous studies that
examined the effect of alcohol consumption among men with remote history of
MI produced differing conclusions.11,12
One of these studies found that consumption of 2 to 3 drinks/wk was associated
with lower mortality among male physicians who reported a previous MI,11 while the other study found no effect of consuming
1 to 14 drinks/wk among men with a previous MI.12
However, these earlier studies were limited by their restriction to men with
prevalent CHD and use of self-reported diagnoses.
We therefore studied mortality following AMI as a function of alcohol
consumption in the year prior to AMI in patients enrolled in the Determinants
of Myocardial Infarction Onset Study (the Onset study).13,14
This multicenter, prospective cohort study included chart reviews and face-to-face
interviews with patients who were hospitalized with confirmed AMI.
Enrollment and Data Collection
The Onset study was conducted in 45 community and tertiary care medical
centers. Between August 1989 and September 1994, 1935 patients (601 women
and 1334 men) were interviewed a median of 4 days after having an AMI. Trained
research interviewers identified eligible patients by reviewing coronary care
unit admission logs and patient charts. For inclusion, patients were required
to have a creatine kinase level higher than the upper limit of normal for
each center, positive MB isoenzymes, an identifiable onset of AMI symptoms,
and ability to complete a structured interview. We excluded patients with
missing information about usual alcohol consumption (n = 5) and those with
a history of alcoholism who reported current abstention (n = 17), leaving
1913 patients eligible for analysis. The institutional review board of each
center approved this protocol, and all participants provided informed consent.
Interviewers used a structured data abstraction and questionnaire form.
Participants reported average frequency during the past year of consumption
(and corresponding numbers of drinks) of wine, beer, and liquor individually.
Because the average ethanol content of a typical drink differs by beverage
type (ie, 13.2 g for beer, 10.8 g for wine, and 15.1 g for liquor), we determined
each patient's average weekly ethanol consumption from wine, beer, and liquor.
We defined a standard serving of alcohol as 15.0 g of ethanol and categorized
average alcohol consumption as none, less than 7 drinks/wk (<105 g of ethanol),
or 7 or more drinks/wk (≥105 g of ethanol). We classified the 2 groups
of drinkers as light and moderate drinkers, respectively. For beverage-specific
analyses, we considered patients to be drinkers of a single type of alcoholic
beverage (wine, beer, or liquor) if the other 2 beverages each represented
less than 5% of the total number of drinks.
Other information collected included age, sex, medical history, and
prescription and nonprescription medication use. During the chart review,
interviewers recorded complications of congestive heart failure or ventricular
arrhythmias based on clinical diagnoses documented in medical records, as
well as blood pressure on admission and all creatine kinase values available
at the time of chart review (median number of values = 4).
We defined initial hypotension and hypertension as systolic blood pressure
on admission of less than 90 mm Hg and more than 200 mm Hg, respectively.
We defined current aspirin use as the reported use of any aspirin or aspirin-containing
product in the 4 days prior to the index AMI. We used 1990 US census data
to derive median household income from ZIP codes (available for 1857 patients).15 We defined noncardiac comorbidity as any diagnosis
of cancer, respiratory disease, renal failure, or stroke recorded in medical
records.
We searched the National Death Index for deaths of Onset study participants
through December 31, 1995, and requested death certificates from state offices
of vital statistics records for all probable matches using a previously validated
algorithm.16 Three physicians independently
verified the determination of each death. Two physicians categorized the cause
of each death as cardiovascular disease or non–cardiovascular disease.
Disagreements among raters were resolved by discussion. All-cause mortality
was the primary outcome measure in all analyses, and cardiovascular mortality
was a secondary outcome measure.
To control for differences between groups of patients in factors other
than alcohol consumption, we calculated propensity scores.17
Each patient's score represents that individual's probability of consuming
a given amount of alcohol, relative to abstention, based on other demographic,
behavioral, and clinical characteristics. We created 2 sets of models: those
that compared abstainers with light drinkers (excluding moderate drinkers)
and those that compared abstainers with moderate drinkers (excluding light
drinkers). To do this, we used multivariable logistic regression models in
which the dependent variable was light or moderate alcohol consumption, defined
dichotomously, and the independent variables were covariates and interaction
terms that could influence the probability of consuming a given amount of
alcohol. Specifically, we included age and body mass index (as linear and
quadratic terms), sex, previous MI, previous congestive heart failure, previous
angina, diabetes mellitus, hypertension, noncardiac comorbidity, previous
medication use (aspirin, β-blockers, calcium channel blockers, digoxin,
and angiotensin-converting enzyme inhibitors individually), current/previous
smoking, frequency of exertion (in 3 categories), household income (in quartiles),
education (in 3 categories), and interaction terms of sex with age and race.
The area under the receiver operating characteristic curve was 0.72 for the
light-drinker model and 0.81 for the moderate-drinker model, indicating good
discrimination between drinkers and abstainers. We incorporated the actual
propensity score into subsequent models, although we obtained similar results
when we used indicator variables for quintiles or deciles of the propensity
score.
In the propensity regression analysis, we assigned indicator variables
to patients with missing data on education (n = 58) and household income (n
= 56). For all other variables, we assigned individuals with missing covariate
information (n<23 for any variable) mean levels of continuous covariates
and modal levels of binary covariates. Models that deleted individuals with
any missing covariate information (n = 177) yielded similar results (data
not shown).
We analyzed continuous and binary variables using t tests and Fisher exact tests, respectively. We compared unadjusted
Kaplan-Meier survival plots using the log-rank test. We used Cox proportional
hazards models to examine the independent effect of alcohol use on mortality.
In these models, we controlled for the actual propensity score, age, sex,
and measures of index AMI treatment and severity (peak creatine kinase level,
receipt of thrombolytic therapy, and congestive heart failure and ventricular
tachycardia during hospitalization). Models that incorporated the covariates
used in the propensity regression analysis directly into Cox models (without
use of a propensity score) with indicator variables for the 2 levels of alcohol
consumption gave similar results (data not shown). We additionally controlled
for Q-wave vs non–Q-wave AMI, initial hypotension, and initial hypertension
(with indicator variables for missing information) in sensitivity analyses.
The smaller numbers of patients in beverage-specific models precluded adjustment
for the entire group of covariates; the truncated (adjusted for age, sex,
smoking, and previous MI) and complete models gave similar results in analyses
of the complete data set.
We repeated adjusted analyses within strata of covariates to explore
possible effect modification. We tested hazard ratios (HRs) for linear trend
across categories of alcohol consumption using a random-effects model.18 We tested the proportionality of hazards using time-varying
covariates and found no significant violations. We present HRs from Cox models
with 95% confidence intervals (CIs). All probability values presented are
2-sided.
Characteristics of Onset study participants according to alcohol consumption
are shown in Table 1.14 Of the 1913 patients, 896 (47%) reported abstention
in the year prior to their MI, 696 (36%) reported consumption of less than
7 drinks/wk, and 321 (17%) reported consumption of 7 or more drinks/wk. The
median consumption in the heavier consumption group was 15 drinks/wk. Higher
alcohol consumption was associated with younger age, male sex, current/former
smoking, increased physical exertion, more frequent use of thrombolytic therapy,
higher household income, higher educational attainment, and white race. It
was inversely associated with hypertension, diabetes, previous CHD, use of
calcium channel blockers and angiotensin-converting enzyme inhibitors, and
congestive heart failure during index hospitalization.
Alcohol Consumption and Mortality
Figure 1 shows unadjusted
survival according to average weekly alcohol consumption. Survival was lowest
among abstainers and greatest among patients who consumed 7 or more drinks/wk
(P<.001).
Table 2 shows HRs for all-cause
mortality according to average weekly alcohol consumption. Alcohol consumption
was associated with lower mortality in unadjusted and adjusted models. Consumption
of 7 or more drinks/wk was associated with somewhat lower mortality than consumption
of less than 7 drinks/wk (HR, 0.83; 95% CI, 0.54-1.28). Table 2 also shows the adjusted HRs for mortality from cardiovascular
causes. Alcohol consumption was also inversely associated with cardiovascular
mortality, again with somewhat lower mortality among heavier drinkers than
among lighter drinkers.
To evaluate the robustness of our findings, we repeated our adjusted
analyses within prespecified patient subgroups. Among men, the adjusted HRs
for all-cause mortality were 0.83 (95% CI, 0.60-1.15) and 0.76 (95% CI, 0.48-1.19)
among patients who consumed less than 7 and 7 or more drinks/wk, respectively.
The corresponding HRs for women were 0.64 (95% CI, 0.38-1.07) and 0.37 (95%
CI, 0.04-3.21). These results were similar but more precise when modeled as
indicator variables in stratified Cox analyses, rather than in propensity
analyses. We also found similar associations between alcohol consumption and
survival among patients with a first or recurrent MI, with or without diabetes,
with or without hypertension, and with incomes higher or lower than the mean
for this study.
Models that additionally controlled for initial hypotension, initial
hypertension, and presence of Q-wave (vs non–Q-wave) AMI did not materially
change our results. Our results were also unchanged when we excluded hypertension
as a covariate in the full model and when we excluded patients with noncardiac
comorbidity. Finally, we examined the effect of alcohol consumption among
patients who consumed 1 type of alcoholic beverage exclusively, although our
power was limited by a smaller number of patients (Table 3). The effect of alcohol consumption was generally similar
among the 3 beverage types.
In this prospective cohort study of early survivors of AMI, moderate
alcohol consumption during the prior year, as measured at the time of index
AMI, was associated with lower subsequent mortality following AMI. This association
was present in unadjusted and adjusted analyses, extended to both men and
women, and was similar for cardiovascular and all-cause mortality.
We know of no prior studies that specifically studied alcohol use and
mortality immediately following AMI. Muntwyler et al11
studied the mortality of male physicians who reported history of MI and found
that moderate drinkers had a relative risk of death of 0.7, a result that
corresponds well to our findings. However, those men survived their MIs long
enough to be considered for enrollment in the Physicians' Health Study. Hence,
the authors could not exclude the possibility of differential mortality between
drinkers and abstainers after MI but prior to consideration for enrollment
in the study. Because we specifically included early survivors of AMI in the
Onset study, our results argue against that possibility; in addition, we extend
their findings to women and patients of lower socioeconomic status.
Shaper and Wannamethee,12 in a study
of 455 men, found no association between consumption of 1 to 14 drinks/wk
and survival in men with history of MI. However, the 95% CIs for the effect
of alcohol consumption were wide (0.77-1.56) and compatible with the results
of this study. Also, moderate drinkers were previously found to have a lower
case-fatality rate of AMI than abstainers in the same population,19 raising the possibility that abstainers at the highest
risk of death may have died before entry into that analysis.
In exploratory analyses, we found no substantial difference in survival
among patients who reported predominant consumption of wine, beer, or liquor.
Although some studies have found that wine consumption is associated with
a lower risk of CHD than is consumption of other alcoholic beverages,20,21 a systematic review found no consistent
effect of beverage type on CHD.22 Indeed, studies
in countries where wine consumption is uncommon report similar results to
studies done elsewhere.23,24
Several physiological effects of alcohol could possibly explain our
findings. In experimental studies, alcohol consumption increases high-density
lipoprotein cholesterol (HDL-C) levels25,26
and decreases levels of prothrombotic factors, particularly fibrinogen.27,28 Alcohol also reduces platelet aggregability
and ablates cyclic flow reductions in mechanically stenosed canine coronary
arteries.29,30 Other potentially
beneficial effects of alcohol include inhibition of ischemic-induced arrhythmias31 and lowering of pulmonary artery pressure among patients
with congestive heart failure.32
Whether the observed association between alcohol consumption and post-AMI
survival reflects the physiological effects of alcohol or confounding by other
factors associated with alcohol can only be answered in a long-term randomized
trial, although such a trial is unlikely to be performed in the near future.
As with any observational study, the associations we observed could
be accounted for, at least in part, by differences between alcohol consumers
and abstainers. For example, alcohol consumers tend to engage in more physical
activity and have higher socioeconomic status than abstainers.33
In fact, the association between alcohol consumption and lower mortality was
substantially attenuated after adjusting for an alcohol propensity score that
incorporates measures of physical activity and socioeconomic status (Table 2). Despite our inclusion of frequency
of physical exertion, educational attainment, race, and median household income
by ZIP code of residence in the propensity score, residual confounding by
these factors (as well as by unmeasured factors) may exist. Some abstainers
may be former drinkers who quit because of illness. However, several studies
have shown that former drinkers and lifelong abstainers have similarly elevated
risks of CHD,1,34 and we excluded
patients who reported history of alcohol abuse and had become abstainers.
Our results also remained robust when we excluded patients with noncardiac
comorbidity.
While controlling for possible confounding in our analyses, we may have
overadjusted for some covariates that are actually intermediates in the causal
pathway between alcohol consumption and lower mortality. For example, if alcohol
consumption lowers the risk of diabetes,35
which, in turn, lowers long-term mortality following AMI, then adjusting for
diabetes may have caused us to underestimate the actual survival difference
associated with alcohol consumption. However, given the important concerns
about confounding in observational studies of alcohol use, we believe this
conservative approach is most appropriate.
We asked patients to report their usual alcohol consumption prior to
the AMI that resulted in their hospitalization. If pre-AMI and post-AMI drinking
patterns do not correlate well, the true effect of post-AMI alcohol consumption
on survival could be different than that reported herein. However, among 711
men who had an AMI in the Health Professionals Follow-Up Study, the correlation
between alcohol consumption before and after AMI was high (r = 0.79), while the median absolute change in consumption was minimal
(Eric B. Rimm, ScD, written communication, March 2000). Similarly, alcohol
consumption following AMI was similar to, but lower than, consumption prior
to AMI among Dutch men.36
Our results could also be influenced by inaccuracies in identification
of deaths among Onset study participants. However, we used a validated method
to search the National Death Index,16 and 3
physicians blinded to alcohol consumption confirmed each death.
Our power to evaluate heavier alcohol consumption was limited. For example,
only 107 patients reported consumption of 21 or more drinks/wk, of whom only
8 died. Hence, we could not assess these 107 patients as a separate subgroup
and cannot generalize our results to individuals who consume that quantity
of alcohol or more.
Because individuals tend to underreport alcohol intake,37
the actual number of servings that Onset study participants consumed may differ
from the values reported here. However, such underreporting is unlikely to
affect the rank order of alcohol consumption among Onset study participants
and, therefore, should not affect the internal validity of our results.
Although we only interviewed early survivors of AMI who could complete
a structured interview, it is possible that AMI severity may have produced
differential recall of usual alcohol consumption. However, we have previously
shown that neither recent nor habitual alcohol consumption influences AMI
severity in this population.38 Moreover, in
a subset of 115 Onset study participants who had HDL-C levels measured during
hospitalization, estimated alcohol consumption and HDL-C level were correlated
to the same degree found in the Second National Health and Nutrition Examination
Survey (Pearson r = 0.21; P
= .02), confirming the validity of our instrument.39
We do not have information on how Onset study participants may have
changed their smoking habits following hospitalization. Numerous studies link
heavier alcohol consumption to lower success in smoking cessation and greater
likelihood of relapse among ex-smokers.40 If
alcohol consumption mediates these smoking behaviors directly, lack of control
for differences in posthospitalization smoking habits is appropriate. If not,
confounding by changes in smoking habits may actually have led us to underestimate
the true difference between drinkers and abstainers, because abstainers will
tend to disproportionately become and remain nonsmokers.
Our results have 2 somewhat different implications for clinical practice.
First, our results provide prognostic information for patients who survive
an AMI. Adults who abstained from alcohol prior to AMI appeared to be at particularly
high risk of long-term mortality; specialized strategies for secondary prevention
may be appropriate for these individuals. Second, our results are consistent
with the hypothesis that light or moderate alcohol use following AMI is safe,
although studies that formally assess post-MI consumption are needed to confirm
this. Even given the associations demonstrated in this study, individuals
should continue to consult their physicians regarding the advisability of
consuming alcohol following an AMI. Determination of the risks and benefits
of alcohol consumption for an individual requires consideration of numerous
personal, clinical, and social factors that cannot be addressed with aggregate-level
observational data such as ours.
In summary, we found that Onset study participants who reported alcohol
consumption prior to AMI had lower mortality following AMI than abstainers.
This finding was consistent across patient subgroups and beverage types, was
similar for cardiovascular and all-cause mortality, and persisted even after
controlling for potentially confounding factors in a rigorous propensity analysis.
Thus, taken together, the overall epidemiological evidence suggests that moderate
alcohol consumption is associated with a lower risk of AMI and a lower risk
of long-term mortality following AMI.
1.Maclure M. Demonstration of deductive meta-analysis: ethanol intake and risk of
myocardial infarction.
Epidemiol Rev.1993;15:328-351.Google Scholar 2.Thun MJ, Peto R, Lopez AD.
et al. Alcohol consumption and mortality among middle-aged and elderly US
adults.
N Engl J Med.1997;337:1705-1714.Google Scholar 3.Valmadrid CT, Klein R, Moss SE, Klein BE, Cruickshanks KJ. Alcohol intake and the risk of coronary heart disease mortality in
persons with older-onset diabetes mellitus.
JAMA.1999;282:239-246.Google Scholar 4.Palmer AJ, Fletcher AE, Bulpitt CJ.
et al. Alcohol intake and cardiovascular mortality in hypertensive patients.
J Hypertens.1995;13:957-964.Google Scholar 5.Orlando J, Aronow WS, Cassidy J, Prakash R. Effects of ethanol on angina pectoris.
Ann Intern Med.1976;84:652-655.Google Scholar 6.Hayes SN, Bove AA. Ethanol causes epicardial coronary artery vasoconstriction in the intact
dog.
Circulation.1988;78:165-170.Google Scholar 7.Abe H, Kawano Y, Kojima S.
et al. Biphasic effects of repeated alcohol intake on 24-hour blood pressure
in hypertensive patients.
Circulation.1994;89:2626-2633.Google Scholar 8.Regan TJ. Alcohol and the cardiovascular system.
JAMA.1990;264:377-381.Google Scholar 9.Dimmitt SB, Rakic V, Puddey IB.
et al. The effects of alcohol on coagulation and fibrinolytic factors: a controlled
trial.
Blood Coagul Fibrinolysis.1998;9:39-45.Google Scholar 10.Hendriks HF, Veenstra J, Velthuis-te Wierik EJ, Schaafsma G, Kluft C. Effect of moderate dose of alcohol with evening meal on fibrinolytic
factors.
BMJ.1994;308:1003-1006.Google Scholar 11.Muntwyler J, Hennekens CH, Buring J, Gaziano JM. Mortality and light to moderate alcohol consumption after myocardial
infarction.
Lancet.1998;352:1882-1885.Google Scholar 12.Shaper AG, Wannamethee SG. Alcohol intake and mortality in middle-aged men with diagnosed coronary
heart disease.
Heart.2000;83:394-399.Google Scholar 13.Mittleman MA, Maclure M, Tofler GH, Sherwood JB, Goldberg RJ, Muller JE.for the Determinants of Myocardial Infarction Onset Study Investigators. Triggering of acute myocardial infarction by heavy physical exertion.
N Engl J Med.1993;329:1677-1683.Google Scholar 14.Muller JE, Mittleman MA, Maclure M, Sherwood JB, Tofler GH.for the Determinants of Myocardial Infarction Onset Study Investigators. Triggering myocardial infarction by sexual activity.
JAMA.1996;275:1405-1409.Google Scholar 15. Census of Population and Housing, 1990. Washington, DC: Bureau of the Census; 1992. Summary tape file 3.
16.Rich-Edwards JW, Corsano KA, Stampfer MJ. Test of the National Death Index and Equifax Nationwide Death Search.
Am J Epidemiol.1994;140:1016-1019.Google Scholar 17.D'Agostino Jr RB. Propensity score methods for bias reduction in the comparison of a
treatment to a non-randomized control group.
Stat Med.1998;17:2265-2281.Google Scholar 18.Berlin JA, Longnecker MP, Greenland S. Meta-analysis of epidemiologic dose-response data.
Epidemiology.1993;4:218-228.Google Scholar 19.Wannamethee G, Whincup PH, Shaper AG.
et al. Factors determining case fatality in myocardial infarction.
Br Heart J.1995;74:324-331.Google Scholar 20.Renaud SC, Gueguen R, Siest G, Salamon R. Wine, beer, and mortality in middle-aged men from eastern France.
Arch Intern Med.1999;159:1865-1870.Google Scholar 21.Gronbaek M, Deis A, Sorensen TI.
et al. Mortality associated with moderate intakes of wine, beer, or spirits.
BMJ.1995;310:1165-1169.Google Scholar 22.Rimm EB, Klatsky A, Grobbee D, Stampfer MJ. Review of moderate alcohol consumption and reduced risk of coronary
heart disease.
BMJ.1996;312:731-736.Google Scholar 23.Keil U, Chambless LE, Doring A, Filipiak B, Stieber J. The relation of alcohol intake to coronary heart disease and all-cause
mortality in a beer-drinking population.
Epidemiology.1997;8:150-156.Google Scholar 24.Yuan JM, Ross RK, Gao YT, Henderson BE, Yu MC. Follow up study of moderate alcohol intake and mortality among middle
aged men in Shanghai, China.
BMJ.1997;314:18-23.Google Scholar 25.Rakic V, Puddey IB, Dimmitt SB, Burke V, Beilin LJ. A controlled trial of the effects of pattern of alcohol intake on serum
lipid levels in regular drinkers.
Atherosclerosis.1998;137:243-252.Google Scholar 26.Cox KL, Puddey IB, Morton AR, Beilin LJ, Vandongen R, Masarei JR. The combined effects of aerobic exercise and alcohol restriction on
blood pressure and serum lipids.
J Hypertens.1993;11:191-201.Google Scholar 27.Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. Moderate alcohol intake and lower risk of coronary heart disease.
BMJ.1999;319:1523-1528.Google Scholar 28.Conlan MG, Folsom AR, Finch A.
et al. Associations of factor VIII and von Willebrand factor with age, race,
sex, and risk factors for atherosclerosis.
Thromb Haemost.1993;70:380-385.Google Scholar 29.Renaud SC, Beswick AD, Fehily AM, Sharp DS, Elwood PC. Alcohol and platelet aggregation.
Am J Clin Nutr.1992;55:1012-1017.Google Scholar 30.Keller JW, Folts JD. Relative effects of cigarette smoke and ethanol on acute platelet thrombus
formation in stenosed canine coronary arteries.
Cardiovasc Res.1988;22:73-78.Google Scholar 31.Bernauer W. The effect of ethanol on arrhythmias and myocardial necrosis in rats
with coronary occlusion and reperfusion.
Eur J Pharmacol.1986;126:179-187.Google Scholar 32.Greenberg BH, Schutz R, Grunkmeier GL, Griswold H. Acute effects of alcohol in patients with congestive heart failure.
Ann Intern Med.1982;97:171-175.Google Scholar 33.Fagrell B, De Faire U, Bondy S.
et al. The effects of light to moderate drinking on cardiovascular diseases.
J Intern Med.1999;246:331-340.Google Scholar 34.Klatsky A, Armstrong MA, Friedman GD. Risk of cardiovascular mortality in alcohol drinkers, ex-drinkers,
and non-drinkers.
Am J Cardiol.1990;66:1237-1242.Google Scholar 35.Rimm EB, Chan J, Stampfer MJ.
et al. Prospective study of cigarette smoking, alcohol use, and the risk of
diabetes in men.
BMJ.1995;310:555-559.Google Scholar 36.Cleophas TJ, Tuinenberg E, van der Meulen J, Zwinderman KH. Wine consumption and other dietary variables in males under 60 before
and after acute myocardial infarction.
Angiology.1996;47:789-796.Google Scholar 37.Chick J, Kreitman N, Plant M. Saving face? survey respondents who claim their last week's drinking
was atypical.
Drug Alcohol Depend.1981;7:265-272.Google Scholar 38.Mukamal KJ, Muller JE, Maclure M, Sherwood JB, Mittleman MA. Lack of effect of recent alcohol consumption on the course of acute
myocardial infarction.
Am Heart J.1999;138:926-933.Google Scholar 39.Linn S, Carroll M, Johnson C.
et al. High-density lipoprotein cholesterol and alcohol consumption in US
white and black adults: data from NHANES II.
Am J Public Health.1993;83:811-816.Google Scholar 40.Shiffman S, Balabanis M. Associations between alcohol and tobacco. In: Fertig JB, Allen JP, eds. Alcohol and Tobacco:
From Basic Science to Clinical Practice. Bethesda, Md: National Institute
on Alcohol Abuse and Alcoholism; 1995. NIH publication 95-3931.