American Journal of Epidemiology
Copyright © 1997 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 146, No. 5
Printed in U.S.A.
Sex Differences in the Association between Alcohol Consumption and
Cognitive Performance
Carole Dufouil,1 Pierre Ducimetiere,2 and Annick Alperovitch1 for the EVA Study Group
Heavy alcohol drinking is associated with cognitive impairment, but little is known concerning the cognitive
effect of moderate alcohol consumption. The Epidemiology of Vascular Aging (EVA) Study is a longitudinal
study of cognitive and vascular aging. Between June 1991 and June 1993, 1,389 subjects aged 59-71 years
(574 men and 815 women) who had been recruited from the electoral rolls of Nantes, France, were examined.
Trained psychologists administered a battery of 10 neuropsychological tests assessing most areas of cognitive functioning. Detailed information on alcohol intake was collected during a structured interview. A multiple
linear regression model was used to examine the relations between test scores and alcohol consumption at
baseline in men and women separately, controlling for age, education, income, depressive symptoms, and
smoking status. Odds ratios for being a high cognitive performer (i.e., being in the top 10% of the distribution
of summary scores from the neuropsychological battery) were calculated. Among men, neuropsychological
test scores were not associated with alcohol consumption in either univariate or multivariate analysis; nor did
the proportion of high cognitive performers vary by alcohol consumption. In contrast, among women,
significant positive associations between alcohol consumption and cognitive performance were observed for
most tests in multivariate analysis. For example, for the Mini-Mental State Examination, the adjusted mean
score for female nondrinkers was 27.5 and that for the heaviest drinkers 28.2 (p < 0.001). The odds ratio for
being a high cognitive performer was 2.5 (95% confidence interval 1.1-5.7) for women who usually consumed
two or more drinks per day in comparison with nondrinkers. These findings suggest that, among women,
moderate alcohol consumption may have a beneficial effect on cognitive function. Am J Epidemiol 1997;146:
405-12.
aged; alcohol drinking; cognition; sex factors
Many studies have demonstrated that alcohol abuse
is associated with both the incidence of dementia and
cognitive decline (1-3), but the effect of regular moderate alcohol consumption on cognitive performance
remains unclear (4-7). Two recent studies carried out
in elderly men (8, 9) suggested that moderate alcohol
consumption was associated with higher cognitive
scores. In a study of 333 Dutch men, Launer et al. (8)
found that men with cardiovascular disease or diabetes
mellitus and low to moderate alcohol intake had a
significantly lower risk for poor cognitive functioning;
an opposite trend was found in men without cardiovascular disease or diabetes, in whom alcohol intake
was associated with a slightly increased risk of a low
cognitive score. Alcohol did not influence subsequent
cognitive decline in either of the two subgroups (8). In
contrast, results reported by Christian et al. (9) from a
prospective study of 4,739 US veteran twins suggested
a small protective effect of past moderate alcohol
intake on cognitive function.
In the present study, the association between alcohol
consumption and cognitive functioning was examined
in 1,389 elderly men and women participating in the
Epidemiology of Vascular Aging (EVA) Study. Cognitive status was assessed by a battery of neuropsychological tests covering different areas of cognitive
functioning.
MATERIALS AND METHODS
Study design
Received for publication July 23, 1996, and in final form April 7,
1997.
Abbreviations: EVA, Epidemiology of Vascular Aging; SD, standard deviation.
1
INSERM U360, Recherches Epidemiologiques en Neurologie et
Psychopathologie, Paris, France.
2
INSERM U258, Epidemiologie Cardiovasculaire, Paris, France.
Reprint requests to Dr. Carole Dufouil, INSERM U360, Hopital La
Salpetriere, 75651 Paris, Cedex 13, France.
The EVA Study is a longitudinal study of vascular
and cognitive aging. The study population is composed of volunteers born between 1922 and 1932 who
were recruited from the electoral rolls of the city of
Nantes in western France. When a subject was recruited, his or her spouse was systematically asked to
405
406
Dufouil et al.
participate in the study if he or she was in the required
age range. During the baseline visit, which took place
between June 1991 and July 1993, 1,389 subjects were
enrolled in the study. The study protocol was approved
by the Comite d'Ethique du Centre Hospitalier
Universitaire de Kremlin-Bicetre (Kremlin-Bicetre,
France), and written consent was obtained from all
participants.
The EVA cohort is composed of volunteers who
differ from the general population aged 60-70 years in
terms of several characteristics. Fourteen percent of
the EVA participants have had 12 or more years of
schooling, as compared with 7 percent of the general
population; they also have a higher socioeconomic
status. Because of the study design, the proportion of
persons living alone, especially widowed women, is
low in the EVA cohort. Overall, the proportion of
women (58.9 percent) is similar to that in the general
population aged 60-70 years (60.1 percent).
Baseline interviews and examinations were conducted at the study center. Data on demographic background, occupation, medical history, medication use,
and personal habits were obtained using a standardized
questionnaire during a face-to-face interview. In the
present analysis, baseline data were used to examine
the cross-sectional relation between cognitive performance and alcohol consumption.
Neuropsychological tests
All participants were given a battery of 10 neuropsychological tests designed to cover multiple areas of
cognitive functioning. The tests were administered by
trained psychologists. The Mini-Mental State Examination provided a global assessment of cognitive functioning (10). Attentiveness was assessed via three
tests: two tests for visual attention—the Trail Making
Test, part B (11), and the Digit Symbol Substitution
portion of the Weschler Adult Intelligence ScaleRevised (12)—and one test for auditory attention, the
Paced Auditory Serial Addition Test (13). The battery
also included assessment of visual memory (the Benton Visual Retention Test (14)); of visual and perceptive processing (the Benton Facial Recognition Test
(15), short form); of verbal memory (the Auditory
Verbal Learning Test (16)); and of general skills such
as logical intelligence and reasoning (Raven Progressive Matrices (17)) and verbal fluency (the Word
Fluency Test (18)). Assessment of psychomotor speed
(the Finger Tapping Test (19)) was also included in
this battery. The amount of time allotted for completion of the battery ranged from 60 minutes to 90
minutes.
Alcohol
Usual daily intake of alcohol was estimated from a
questionnaire on food habits. Subjects were requested
to specify the type (wine, beer, aperitifs, spirits, liquor)
and quantity (in number of glasses) of alcoholic beverages they usually consumed at different times of the
day (at breakfast, in the morning, at lunch, during
snack time, at dinner, and in the evening). Total daily
intake was estimated in milliliters according to the
average number of milliliters of ethanol in one glass of
each type of alcoholic beverage (wine = 13 ml,
beer = 18 ml, liquor = 4 ml, aperitifs = 10 ml,
spirits = 24 ml). Total daily intake of ethanol was then
converted into the equivalent number of drinks per day
by | dividing total consumption in milliliters by 13,
which is the ethanol content of one glass of wine. In
view of the different alcohol intake habits of men and
women, frequency of daily alcohol consumption was
categorized differently by sex. For men, frequency
was categorized as never drinking, <2 drinks per day,
2-5 drinks per day, and ^ 5 drinks per day. For
women, frequency was grouped as never drinking, < 2
drinks per day, and S 2 drinks per day.
Covariates
The following covariates were selected on the basis
of available evidence suggesting their possible association with cognitive function, alcohol consumption, or
both: age, education, monthly income, smoking, hypertension, history of vascular disease, and depression.
Monthly income was categorized as low (<$l,000),
medium ($l,000-<$2,000), or high (>$2,O00). With
respect to smoking behavior, a subject was classified
as a never smoker, former smoker, or current smoker.
Two independent measurements of systolic and diastolic blood pressure were made with a digital electronic tensiometer after a 10-minute rest, and the mean
value was used in the analysis. Hypertension was
defined as systolic blood pressure a 160 mmHg or
diastolic blood pressure &95 mmHg, or use of antihypertensive medication. History of vascular disease
was defined as a self-reported history of stroke, myocardial infarction, angina pectoris, or arteritis. Depressive symptomatology was assessed by means of the
Center for Epidemiologic Studies Depression Scale
(20). This scale consists of 20 self-report items concerning symptoms and feelings experienced during the
preceding week. Each item is scored from 0 to 3
according to the frequency of the symptom. The
higher the score, the higher the level of depressive
symptomatology. Evaluation of the scale in a French
population showed that men and women scoring >16
or >22, respectively, should be considered at high risk
for clinical depression (21).
Am J Epidemiol
Vol. 146, No. 5, 1997
Alcohol Consumption and Cognitive Performance
Statistical methods
Because of sex differences in alcohol drinking habits, all analyses were conducted separately in men and
women. Relations between alcohol consumption and
the covariates were tested by the chi-square test for
categorical variables and by analysis of variance for
continuous variables.
We built a composite score from the 10 cognitive
test scores as follows. Each test score distribution was
standardized (mean = 0, standard deviation (SD) = 1)
so that all tests were equally weighted in the total
score. The composite score was computed as the mean
of the sum of the standardized scores; the composite
score increased with better overall cognitive performance. Subjects with scores over the 90th percentile
limit of the composite score distribution were classified as high cognitive performers.
TABLE 1.
The cross-sectional association between cognitive
function and alcohol intake was examined in several
ways. Analysis of variance and linear regression models were used to assess the relation of each of the 10
test scores and the composite score with alcohol consumption. Crude and adjusted analyses, controlling for
age, education, income, smoking, and depressive
symptoms, were performed. Logistic regression modeling provided estimates of the relative risk of high
overall cognitive performance, as defined above, associated with different levels of alcohol intake. Analyses were conducted using the SAS statistical software
package (22).
RESULTS
The characteristics of the study sample are shown in
table 1. Mean age was 65.0 years (SD 3.1) for both
Characteristics of subjects in the EVA* Study, by sex, Nantes, France, 1991-1993
Men
(n = 574)
Women
(n=815)
Age (years) (%)
59-62
63-66
67-71
22.8
43.2
34.0
24.4
41.8
33.7
Mean years of education
11.3(4.3)$
10.3 (3.2)
<0.001
Monthly income (%)
Low (<$1,000)
Medium ($1,00O-<$2,000)
High (>$2,000)
3.3
30.8
65.9
10.6
44.6
44.8
<0.001
Smoking status (%)
Never smoker
Former smoker
Current smoker
23.2
62.2
14.6
82.8
12.8
4.4
< 0.001
Equivalent§ no. of drinks of alcohol per day (%)
0
<2
2-5
>5
12.9
37.2
40.6
9.2
38.0
51.6
10.4
< 0.001
Type of alcoholic beverage (% of total alcohol intake)
Wine
Cider
Beer
Aperitif
Spirits
83.2
0.8
2.6
13.1
0.2
75.9
0.9
0.8
21.8
0.5
0.59
P
valuet
0.77
Mean systolic blood pressure (mmHg)
137 (18)
128 (16)
<0.001
Mean diastolic blood pressure (mmHg)
82 (12)
78 (10)
<0.001
History of vascular disease (%)
12.5
7.9
<0.01
High level of depressive symptomatology^] (%)
14.2
14.2
0.99
* EVA, Epidemiology of Vascular Aging.
t Based on chi-square test for qualitative variables and analysis of variance for continuous variables (k-1 df).
$ Numbers in parentheses, standard deviation.
§ Calculated according to ethanol content.
H Center for Epidemiologic Studies Depression Scale (20).
Am J Epidemiol
Vol. 146, No. 5, 1997
407
408
Dufouil et al.
men and women. The proportion of subjects with a
high monthly income was significantly higher in men
than in women. Most men reported past or present
smoking (76.8 percent) as compared with only 17.2
percent of women. Overall, mean daily alcohol consumption was 28.7 ml (SD 26.6) for men and 8.1 ml
(SD 12.4) for women; among drinkers, mean alcohol
consumption was 33.0 ml (SD 26.0) and 13.1 ml (SD
13.6) for men and women, respectively. The proportion of nondrinkers was higher in women than in men.
For men, as for women, most of the total alcohol
intake resulted from wine drinking. Hypertension was
found in 39.0 percent of men and 28.6 percent of
women (p < 0.001); similarly, a history of cardiovascular disease was more frequently reported by men
than by women. The proportion of subjects reporting a
high level of depressive symptoms was exactly the
same in men and women.
Factors associated with alcohol consumption
Men and women differed in terms of factors associated with alcohol consumption, except for age,
which was not related to alcohol consumption (table
2). Number of years of schooling and income were
significantly associated with alcohol intake in women
but not in men. Tobacco use and alcohol consumption
were significantly correlated in men but not in women.
In men, nondrinkers and very light drinkers (<2
drinks/day) had lower systolic and diastolic blood
pressures than persons who drank more heavily. There
was a nonsignificant relation between hypertension
and alcohol consumption in men, although hypertension was more frequent among those reporting two or
more drinks of alcohol per day (42.7 percent vs. 34.3
percent; p < 0.05). Blood pressure was independent of
alcohol intake in women, but the proportion of hypertensive women decreased with increasing alcohol consumption. Data suggested that the relation between
drinking and self-reported history of vascular disease
differed in men and women (table 2), but no significant association was found in either sex. Nevertheless,
in men, the lowest proportion of subjects reporting a
history of vascular disease was found among those
with the highest daily alcohol intake ( s 5 drinks/day),
whereas in women the proportion with a history of
vascular disease was higher in nondrinkers than in
drinkers.
The proportion of women with a high level of depressive symptomatology decreased—but not significantly—with higher alcohol intake. In men, there was
a significant relation between alcohol consumption
and depressive symptomatology, with data taking an
apparent U-shaped pattern.
Cognitive function and alcohol intake
Univariate (data not shown) and multivariate (table
3) analysis showed that the association between alcohol intake and cognitive functioning differed in men
and women. In men, both univariate and multivariate
analysis showed that none of the 10 test scores, nor the
composite score, was significantly associated with
drinking habits.
Overall, we found that women with low to moderate
alcohol intake had higher cognitive scores. Multiple
linear regression models adjusting for age, education,
income, smoking, and depression showed a significant
linear performance improvement with increasing alcohol intake for six cognitive measures in women (table
3); borderline improvements were found for two additional tests. Our findings did not suggest that specific cognitive areas were more susceptible to an alcohol effect than others.
Logistic regression analyses controlling for age, education, income, depression, and smoking were used
to estimate odds ratios for high cognitive performance
associated with different levels of alcohol intake (table
4). High performers were defined as subjects scoring
in the 90th percentile of the composite score. (There
were 58 men and 81 women in this category.) Nondrinkers were the reference category. In men, alcohol
consumption was not associated with high cognitive
scores; on the contrary, high alcohol consumption
(^5 drinks/day) was associated with poorer cognitive performance. In women, analyses showed a
dose-response association between alcohol intake and
high cognitive scores. In comparison with nondrinking
women, the odds for good cognitive functioning were
multiplied by 1.7 and 2.5 for women who consumed
< 2 drinks per day and ^ 2 drinks per day, respectively. No significant association was found between
smoking and high cognitive performance (table 4).
Both men and women who were current smokers had
higher odds of being high performers, as compared
with nonsmokers and former smokers, but odds ratios
were not significantly greater than 1.
Because of the cross-sectional nature of our data,
potential bias related to changes in drinking habits due
to hypertension or a history of vascular disease should
be considered. This bias might confound the relation
between alcohol consumption and cognitive functioning. When subjects with hypertension or a history of
vascular disease were excluded from the analysis, the
associations between alcohol consumption and cognitive functioning remained unchanged in both women
and men (data not shown).
Am J Epidemiol
Vol. 146, No. 5, 1997
CO
CD
i
p
en
9
!
27.6
58.1
14.3
134(17)
31.5
52.1
16.4
132 (16)
80(11)
34.3
12.3
20.3
Smoking status (%)
Never smoker
Former smoker
Current smoker
Mean systolic blood pressure (mmHg)
Mean diastolic blood pressure (mmHg)
Hypertension (%)
History of vascular disease (%)
High level of depressive symptomatology!! (%)
10.4
14.0
44.1
83 (12)
140 (18)
20.1
66.8
13.1
3.2
31.8
65.0
20.0
5.8
36.5
82(8)
139(18)
9.6
71.2
19.2
0.0
26.5
73.5
11.6(6.1)
&5
(n = 52)
64.6 (2.8)
df).
<0.05
0.45
0.15
<0.01
<0.001
<0.05
0.80
0.63
0.23
valued
P
15.7
9.5
34.1
78(11)
129 (16)
87.5
8.9
3.6
14.4
44.4
41.2
10.0 (3.3)
64.9 (2.9)
None
(n = 305)
Equivalentt no. of drinks of alcohol per day
* EVA, Epidemiology of Vascular Aging.
t Calculated according to ethanol content.
i Based on chi-square test for qualitative variables and analysis of variance for continuous variables (k-1
§ Numbers in parentheses, standard deviation.
U Center for Epidemiologic Studies Depression Scale (20).
15.1
12.4
34.3
79 (10)
4.0
29.7
66.3
2.9
33.3
63.8
Monthly income (%)
Low(<$1,000)
Medium ($1,000-<$2,000)
High (£$2,000)
11.1 (4.0)
11.6(4.2)
11.4(4.3)
Mean years of education
65.2 (2.9)
2-5
(n = 229)
<2
(n = 210)
64.9(3.1)
65.5 (3.0)§
None
(n = 73)
Men
Correlates of alcohol drinking habits among men and women in the EVA* Study, Nantes, France, 1991-1993
Mean age (years)
TABLE 2.
13.0
6.8
27.5
13.4
7.2
15.7
78 (10)
128(15)
128 (17)
78 (10)
79.5
16.9
3.6
5.1
33.3
61.5
11.1 (3.4)
80.4
14.5
5.1
47.4
43.7
8.9
10.4 (3.0)
64.7 (2.6)
£2
(n = 83)
Women
65.0 (3.0)
<2
(n = 414)
0.58
0.39
<0.05
0.78
0.99
0.09
<0.01
<0.05
0.80
P
value}
CO
n
01
l
§'
3:
CO
D
O
o
a
B>
5'
c
"D
CO
g.
O
o
8
410
Dufouil et al.
TABLE 3. Association* between alcohol intake and cognitive scores among men and women in the EVAf Study, by sex, Nantes,
France, 1991-1993
Equivalent no. of drinks of alcohol per day
Men
None
(n = 73)
Mini-Mental State Examination
Word Fluency Test
Digit Symbol Substitution Task
Benton Visual Retention Test
Benton Facial Recognition Test
Paced Auditory Serial Addition Test
Trail Making Test, part B#
Finger Tapping Test
Auditory Verbal Learning Test
Raven Progressive Matrices
Composite score
<2
(n = 210)
27.9
34.9
41.2
11.8
21.1
13.6
2.23
130
38.2
15.5
28.0
33.4
41.1
11.3
21.2
-0.14
2-5
(n = 229)
Women
a5
(n = 52)
14.1
2.32
130
39.0
15.1
28.2
35.6
41.3
11.6
21.2
14.0
2.36
129
40.2
15.3
28.1
34.0
40.0
11.5
20.7
13.5
2.34
-0.18
-0.09
p
value§
39.2
15.6
0.44
0.29
0.86
0.18
0.56
0.89
0.82
0.78
0.44
0.44
-0.15
0.55
127
None
(n = 305)
a2
<2
(n = 414)
(n = 83)
p
valueH
28.2
37.6
45.4
11.3
21.6
14.7
2.10
44.1
13.6
28.0
35.1
44.1
11.2
21.8
13.7
2.33
114
44.8
14.4
46.0
14.5
<0.001
0.03
0.005
0.69
0.45
0.06
<0.001
0.24
0.06
0.003
-0.15
-0.01
0.07
<0.001
27.5
34.1
42.2
11.2
21.9
13.2
2.50
113
115
* Mean scores after data were controlled for age, education, level of income, depressive symptoms, and smoking status,
t EVA, Epidemiology of Vascular Aging.
\ Calculated according to ethanol content.
§ Based on analysis of variance (3 df).
11 Based on multiple linear regression analysis (1 df).
# Lower scores indicate better cognitive function.
DISCUSSION
This study, conducted in a French population sample aged 59-71 years, suggests that the association
between alcohol consumption and cognitive abilities is
sex-dependent. In men, we found no significant relation between alcohol consumption and cognitive
scores. In contrast, among women, with a range of
daily alcohol consumption between zero and approximately four drinks, we found an overall positive linear association between cognitive scores and alcohol
consumption. This relation persisted when data were
controlled for age, income, education, and depressive
symptomatology, and could not be attributed to a
possible change in drinking habits due to vascular
disease or high blood pressure.
A few studies have examined the relation between
alcohol intake and cognitive function. Among four
cross-sectional studies that included both men and
women, three were inconclusive and only one, conducted in East Boston, Massachusetts, by Hebert et al.
(23), found an overall significant positive association
between alcohol consumption and cognitive functioning. Results of separate analyses stratified by sex were
not reported. In a sample of elderly men, Launer et al.
(8) reported that moderate alcohol consumption was
associated with better cognitive performance. In nei-
TABLE 4. Odds ratios for high cognitive performance in the EVA* Study, by level of alcohol consumption and smoking status,
Nantes, France, 1991-1993
Women
Men
high
Odds
ratio!
73
210
229
52
8.2
9.5
11.8
5.8
1.0
1.1
1.2
0.6
133
357
84
9.0
10.4
10.7
1.0
No.
Equivalent^ no. of drinks of alcohol
per day
0
<2
2-5
£5
Smoking status
Never smoker
Former smoker
Current smoker
%
1.0
1.3
95% Cl*
0.3-2.2
0.5-3.0
0.1-2.4
0.5-2.1
0.4-2.9
No.
%
high
Odds
ratio!
305
414
83
7.2
10.4
18.1
1.0
1.7
2.5
1.0-2.9
1.1-5.7
675
104
36
8.6
13.9
17.3
1.0
1.1
1.8
0.6-2.9
0.6-5.3
95% Cl
* EVA, Epidemiology of Vascular. Aging; Cl, confidence interval.
t Controlled for age, education, level of income, and depressive symptoms.
X Calculated according to ethanol content.
Am J Epidemiol
Vol. 146, No. 5, 1997
Alcohol Consumption and Cognitive Performance
ther Hebert et al.'s study (23) nor Launer et al.'s study
(8) did analysis of longitudinal data confirm the crosssectional relation found between alcohol consumption
and cognitive performance. In contrast, results reported by Christian et al. (9) from a study of 4,739
twin men in the United States indicated a small protective effect of past moderate alcohol intake on cognitive functioning.
The rationale for conducting analyses separately by
sex is that men and women usually have different
patterns of alcohol consumption, at least in France.
These differences concern not only the level of consumption but also the factors associated with it. In this
study, we found that associations between education,
income, depression, medication use, and alcohol consumption were stronger in women than in men. Analyses of relations between blood pressure, hypertension, and history of vascular disease also showed
differences between men and women.
Volunteers who participate in the EVA Study have
a higher educational level, socioeconomic status, and
cognitive performance than the general population of
the same age (24). However, the selection bias affects
men and women in the same way and cannot explain
our findings. Nevertheless, the possibility cannot be
excluded that unknown factors related to alcohol consumption influence the participation of men and
women differently.
The validity of data on alcohol consumption is another important methodological issue. It is not known
whether bias due to underestimation of self-reported
alcohol consumption was of a similar nature and magnitude in both sexes. It is likely that nondrinkers were
a more heterogeneous subgroup in men, since this
subgroup might include a higher proportion of exdrinkers. In contrast, compliance with medical advice
is usually better among women, who may be more
prone to stop drinking alcohol because of health problems. Since information on past drinking and drinking
history is not available in the EVA Study, we could
not verify these hypotheses. Thus, sex-dependent bias
and measurement errors might partly explain the different patterns of association between alcohol consumption and cognitive function seen in men and
women.
Studies have suggested that chronic moderate alcohol consumption may be associated with risk of cerebrovascular disease (25-28). Numerous biologic
mechanisms have been proposed as explanations for
this association, including blood pressure, cerebral
blood flow, and large cerebral artery arteriosclerosis.
In the EVA Study, we have found that the presence of
carotid plaques is associated with lower cognitive
scores in men only (29). We found no relation between
Am J Epidemiol
Vol. 146, No. 5, 1997
411
alcohol consumption and the presence of carotid
plaques or carotid artery wall thickness in either sex
(30, 31).
Numerous studies have indicated that acute exposures to alcohol in various concentrations have a positive or negative effect on neuronal activities (32, 33).
To our knowledge, the neuronal consequences of lowlevel chronic brain exposure to alcohol have not been
investigated. Sex differences in alcohol metabolism
(34, 35) and interactions between alcohol metabolism
and hormonal status might partly explain why the
relations between alcohol consumption and cognitive
functioning are not similar in men and women.
Even if biologic mechanisms cannot be totally excluded, the most likely interpretation for our findings
is that, in women—but not in men—low to moderate
alcohol consumption is a marker of general well-being
or of a behavioral pattern associated with good cognitive functioning. The sociocultural determinants of
male drinking habits might be less related to cognitive
abilities.
In conclusion, this analysis did not suggest any
negative effect of low to moderate alcohol intake on
cognitive functioning in older adults. Moreover, we
found that alcohol consumption was associated with
better cognitive scores in women. The cross-sectional
nature of our data and the recognition of possible
unmeasured confounding factors and bias limit the
interpretation of our findings.
ACKNOWLEDGMENTS
The EVA Study was organized under an agreement between INSERM (l'lnstitut National de la Sante et de la
Recherche M6dicale) and the Merck, Sharp and DohmeChibret Company (West Point, Pennsylvania). Dr. Carole
Dufouil was supported by a doctoral research grant from the
French Ministry of Research.
The authors thank C. Delanoe and S. Bachelier for secretarial and technical support. The authors are also grateful
to Drs. Rebecca Fuhrer and Christophe Tzourio for their
helpful advice.
REFERENCES
1. Saunders PA, Copeland JR, Dewey ME, et al. Heavy drinking
as a risk factor for depression and dementia in elderly men:
findings from the Liverpool longitudinal community study.
Br J Psychiatry 1991;159:213-16.
2. Colsher PL, Wallace RB. Elderly men with histories of heavy
drinking: correlates and consequences. J Stud Alcohol 1990;
51:528-35.
3. O'Hare T. Mental health problems and alcohol abuse: cooccurrence and gender differences. Health Soc Work 1995;
20:207-14.
412
Dufouil et al.
4. Baum-Baicker C. The psychological benefits of moderate
alcohol consumption: a review of the literature. Drug Alcohol
Depend 1985;15:305-22.
5. Scherr PA, Albeit MS, Funkenstein HH, et al. Correlates of
cognitive function in an elderly community population. Am J
Epidemiol 1988;128:1084-101.
6. Graves AB, van Duijn CM, Chandra V, et al. Alcohol and
tobacco consumption as risk factors for Alzheimer's disease:
a collaborative re-analysis of case-control studies.
EURODEM Risk Factors Research Group. Int J Epidemiol
1991;20(suppl 2):S48-57.
7. Farmer ME, White LR, Abbott RD, et al. Blood pressure and
cognitive performance: The Framingham Study. Am J Epidemiol 1987;126:1103-14.
8. Launer LJ, Feskens EJ, Kalmjin S, et al. Smoking, drinking,
and thinking: The Zutphen Elderly Study. Am J Epidemiol
1996;143:219-27.
9. Christian JC, Reed T, Carmelli D, et al. Self-reported alcohol
intake and cognition in aging twins. J Stud Alcohol 1995;56:
414-16.
10. Folstein MF, Folstein SE, McHugh PR. "Mini-Mental State":
a practical method for grading the cognitive state of patients
for the clinician. J Psychiatr Res 1975;12:189-98.
11. Reitan RM. Validity of the Trail Making Test as an indicator
of organic brain damage. Percept Mot Skills 1958;8:271-6.
12. Wechsler D. Manual for the Wechsler Adult Intelligence
Scale-Revised. New York, NY: Psychological Corporation,
1981.
13. Gronwall DM. Paced-auditory serial-addition task: a measure
of recovery from concussion. Percept Mot Skills 1977;44:
367-73.
14. Benton AL. Manuel du test de retention visuelle: applications
cliniques et expe'rimentales. 2tmc ed. francaise. Paris, France:
Centre de Psychologie Appliqu6e, 1965.
15. Benton AL, Hamsher K, Varney NR, et al. Contributions to
neurological assessment. New York, NY: Oxford University
Press, 1978.
16. Rey A. L'examen clinique en psychologie. Paris, France:
Presses Universitaires de France, 1964:141-93.
17. Raven J. Progressive matrices. Manuel PM-C. Issy-LesMoulineaux, France: Editions Scientifiques et Psychologiques, 1981.
18. Borkowski JG, Benton AL, Spreen O. Word fluency and brain
damage. Neuropsychologia 1967;5:135-40.
19. Reitan RM, Davison LA. Clinical neuropsychology: current
status and applications. New York, NY: Hemisphere Publications, 1974.
20. Radloff LS. The CES-D Scale: a self-report depression scale
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
for research in the general population. Appl Psychol Meas
1977;l:385-401.
Fuhrer R, Rouillon F. La version franchise de l'echelle CES-D
(Center for Epidemiologic Studies Depression Scale): description et traduction de l'echelle d'auto-evaluation. Psychiatr
Psychobiol 1989;4:163-6.
SAS Institute, Inc. SAS user's guide, version 6.11. Cary, NC:
SAS Institute, Inc, 1991.
Hebert LE, Scherr PA, Beckett LA, et al. Relation of smoking
and low-to-moderate alcohol consumption to change in cognitive function: a longitudinal study in a defined community of
older persons. Am J Epidemiol 1993,137:881-91.
Berr C, Dufouil C, Brousseau T, et al. Early effect of ApoE-e4
allele on cognitive results in a group of highly performing
subjects: The EVA Study. [Etude sur le Vieillissement Arteriel]. Neurosci Lett 1996;218:9-12.
Gorelick PB. Alcohol and stroke. Stroke 1987;18:268-71.
Camargo CA Jr. Moderate alcohol consumption and stroke:
the epidemiologic evidence. Stroke 1989;20:1611-26.
Marmot M, Brunner E. Alcohol and cardiovascular disease:
the status of the U shaped curve. BMJ 1991;303:565-8.
Gaziano JM, Buring JE, Breslow JL, et al. Moderate alcohol
intake, increased levels of high-density lipoprotein and its
subfractions, and decreased risk of myocardial infarction.
N Engl J Med 1993;329:1829-34.
Auperin A, Berr C, Bonithon-Kopp C, et al. Ultrasonographic
assessment of carotid wall characteristics and cognitive functions in a community sample of 59- to 71-year-olds. The EVA
Study Group. Stroke 1996;27:1290-5.
Bonithon-Kopp C, Touboul PJ, Berr C, et al. Relation of
intima-media thickness to atherosclerotic plaques in carotid
arteries: The Vascular Aging (EVA) Study. Arterioscler
Thromb Vase Biol 1996;16:310-16.
Bonithon-Kopp C, Touboul PJ, Berr C, et al. Factors of carotid
arterial enlargement in a population aged 59 to 71 years: The
EVA Study. Stroke 1996;27:654-60.
Littleton J, Little H, Laverty R. Role of neuronal calcium
channels in ethanol dependence: from cell cultures to the
intact animal. Ann N Y Acad Sci 1992;654:324-34.
Weight FF. Cellular and molecular physiology of alcohol
actions in the nervous system. Int Rev Neurobiol 1992;33:
289-348.
Desroches D, Orevillo C, Verina D. Sex- and strain-related
differences in first-pass alcohol metabolism in mice. Alcohol
1995;12:221-6.
Thomasson HR. Gender differences in alcohol metabolism:
physiological responses to ethanol. Recent Dev Alcohol 1995;
12:163-79.
Am J Epidemiol
Vol. 146, No. 5, 1997