Abstract

The authors investigated the relation between alcohol consumption and cognitive function in a United Kingdom cohort study (4,272 men, 1,761 women) with median follow-up of 11 years. Measures of alcohol consumption were obtained at baseline (1985–1988) and four subsequent phases of data collection. Cognitive function (memory test, AH4, Mill-Hill, phonemic and semantic fluency) was assessed at phase 5 (1997–1999), when participants were aged 46–68 years. Of people who reported drinking alcohol in the past year, those who consumed at least one drink in the past week, compared with those who did not, were significantly less likely to have poor cognitive function. The beneficial effect extended to those drinking more than 240 g per week (approximately 30 drinks). The effect was stronger for women than men and was not confined to those with evidence of vascular disease. Similar associations were found in cross-sectional and longitudinal analyses. The relations were not explained by confounding by smoking and by physical and mental health and, to a large extent, were not mediated by cholesterol or blood pressure. However, the relations were weakened when social position was added to the model. The authors concluded that for middle-aged subjects, increasing levels of alcohol consumption were associated with better function regarding some aspects of cognition. Nonetheless, it is not proposed that these findings be used to encourage increased alcohol consumption.

Received for publication June 26, 2003; accepted for publication February 27, 2004.

Previous studies have shown that light-to-moderate drinkers have better cognitive function than nondrinkers (15), with several authors reporting stronger associations for women (3, 4, 6). A U-shaped relation has been reported, suggesting that drinking at heavier volumes is associated with worse cognitive function (25). Other studies have found no consistent relation between alcohol drinking and cognition (79).

Many questions remain. Most existing studies are based on older cohorts, and it is unclear whether a possible protective effect on cognitive function extends to younger, healthier cohorts. It has previously been shown that benefit was confined to those with evidence of cardiovascular disease and diabetes (10). The precise nature of exposure to alcohol remains unclear. It is not known whether pattern of drinking is important, as well as average volume consumed (2, 5), and whether effects are similar in cross-sectional and longitudinal analyses (11). Most existing studies’ analyses used nondrinkers as the comparison group. Doing so may confuse the interpretation because nondrinkers, in particular former drinkers, are usually different from drinkers in other, prognostically important ways. It is better to categorize drinking into several levels and use one of these groups as the reference.

Several possible mechanisms might explain an association between alcohol consumption and cognitive function. For example, moderate consumption may be a proxy marker for good mental and physical health and for high socioeconomic position, both of which are related to good cognitive performance (12, 13). Alternatively, alcohol may have a causal effect via improved vascular function, which is itself associated with good cognitive ability in the general population (1416).

The objective of the present study was to test whether alcohol consumption (both average volume consumed and frequency of drinking) was associated with cognitive function in a sample of middle-aged adults. We were able to explore whether cross-sectional relations differed from longitudinal analyses (mean follow-up: 11 years) and whether changes in drinking behavior influenced any association. The Whitehall II Study offers an important opportunity to explore several possible explanations, with its repeated measures of alcohol consumption, measures of social position, physical and mental health tests, and a comprehensive battery of cognitive function tests.

MATERIALS AND METHODS

The longitudinal Whitehall II Study was established in 1985 to examine the socioeconomic gradient in health and disease among 10,308 civil servants (6,895 men and 3,413 women). Full details of the cohort and its follow-up have been published elsewhere (17). All civil servants aged 35–55 years in 20 departments based in London, United Kingdom, were invited by letter to participate. In total, 73 percent of those invited agreed to take part in phase 1. The 1985–1988 baseline examination (phase 1) involved a clinical examination and a self-administered questionnaire containing sections on demographic characteristics, health, lifestyle factors, work characteristics, social support, and life events. Clinical examination included measures of blood pressure, anthropometry, biochemistry, neuroendocrine function, and subclinical markers of cardiovascular disease. Subsequent phases of data collection have alternated between a postal questionnaire alone (phases 2, 4, and 6) and a postal questionnaire accompanied by a clinical examination (phases 1, 3, and 5). The median length of follow-up from phase 1 to phase 5 was 11 years, and 355 subjects died during this period. The questionnaire at phase 5 was completed by 7,830 respondents; the clinical examination, in which cognitive tests were administered, was attended by 6,553 participants. Cognitive function and alcohol consumption data at phase 5 were available for 6,033 respondents. With respect to baseline, available data were influenced by age (p = 0.001) and employment grade (p = 0.001) but not by sex (p = 0.61). The attrition rate was higher among older respondents and low socioeconomic status groups. Evaluation of self-reported dementia using the International Classification of Diseases, Ninth Revision, to recode a question on long-standing illness, asked as part of the phase 5 questionnaire, revealed no cases in this sample of 6,033. The University College London ethics committee approved the study.

Alcohol consumption

At baseline and subsequent study phases, participants were asked to report the number of alcoholic drinks they had consumed in the last 7 days. This information was divided into “measures” of spirits, “glasses” of wine, and “pints” of beer. In the United Kingdom, a standard measure of spirits and a glass of wine are considered to contain 8 g of alcohol, while a pint of beer contains 16 g of alcohol. Participants were also asked to report the frequency of their drinking over the last 12 months by circling one of six specified options (twice a day or more, almost daily, once or twice a week, once or twice a month, special occasions only, none).

Tests of cognitive function

The cognitive test battery consisted of five standard tasks chosen to comprehensively evaluate cognitive functioning in middle-aged adults. The choice of this battery rather than a specific test for dementia was guided by an attempt to capture the entire range of cognitive ability and not be restricted by ceiling effects in this middle-aged cohort. The battery was composed of the following tests. The first was a 20-word free-recall test of short-term memory. Participants were presented with a list of 20 one- or two-syllable words at 2-second intervals and were then asked to recall in writing as many of the words as they could, in any order; they had 2 minutes to do so. The AH4 (18) is composed of a series of 65 verbal and mathematical reasoning items of increasing difficulty. This test of inductive reasoning measures the ability to identify patterns and infer principles and rules. Participants had 10 minutes to complete this section. The Mill Hill vocabulary test (19) assesses knowledge of verbal meaning and encompasses the ability to recognize and comprehend words. We used this test in its multiple format, which consists of a list of 33 stimulus words ordered by increasing difficulty and six response choices. Finally, we used two measures of verbal fluency: phonemic and semantic. Phonemic fluency was assessed via “s” words and semantic fluency via “animal” words. Subjects were asked to recall in writing as many words beginning with “s” and as many animal names as they could. One minute was allowed for each test.

Possible confounders and mediators measured at baseline

Age and smoking status

Participants were asked to report whether they currently smoked (cigarettes, hand-rolled tobacco, or cigars), the number of items they smoked per day, or when they stopped smoking.

Socioeconomic position

On the basis of salary and work role, the civil service defines a hierarchy of employment grades ranging from senior executive officers to clerical and support staff. This measure was used to control for socioeconomic position in the analyses.

Mental health

The 30-item standard General Health Questionnaire was used as an indicator of mental health (20).

Physical Component Score

The Short Form-36 General Health Survey is a 36-item questionnaire that covers issues relating to physical, psychological, and social functioning (21). It is coded into eight scales: physical functioning, social functioning, role limitations due to physical problems, role limitations due to emotional problems, vitality, bodily pain, general health perception, and general mental health. These eight scales can be summarized into physical and mental components scores by using factor analysis (22, 23); in this study, we used the Physical Component Score as a measure of general physical functioning. A low score implies poor functioning, with 50 being the mean in the general US population (possible score range: 0–100).

Vascular risk factors

A Hawksley random-zero sphygmomanometer (Hawksley, Lancing, Sussex, United Kingdom) was used to measure blood pressure twice, with the subject in the sitting position after 5 minutes of rest. Serum cholesterol was determined by the cholesterol oxidate peroxidase colorimetric method (BCLkit; Boehringer, Mannheim, Germany).

Cardiovascular disease and diabetes during follow-up

Potential cases of angina and myocardial infarction arose from positive responses to self-completed questions on symptoms of chest pain (24), physician diagnoses of angina or myocardial infarction, cardiac investigative items (exercise electrocardiography and angiography), and treatments (nitrate medications and revascularizations). Twelve-lead resting electrocardiograms (Siemens Mingorec, Siemens Medical Solutions, Erlington, Germany) were performed in the screening clinic. To identify a group with evidence of cardiovascular disease, we used an inclusive measure of coronary heart disease (including abnormal resting electrocardiogram, angina, and myocardial infarction). Cases of intermittent claudication, stroke, and diabetes were obtained from recall of a physician’s diagnosis.

Statistical methods

The risk of poor cognitive performance associated with alcohol consumption was estimated by using binary logistic regression analyses. For each test, poor cognitive function was indicated by inclusion in the worst quintile. Analyses were carried out separately for men and women to explore possible effect modification by gender, and all analyses were controlled for age and smoking.

Both average volume of alcohol consumed per week and reported frequency of drinking were considered. The reference group was set as those who reported drinking in the last year but not in the past week. These participants were chosen in preference to lifetime abstainers or former drinkers, who typically have worse health profiles than drinkers do.

We compared analyses in which current drinking (cross-sectional analyses) and drinking reported at baseline (longitudinal analyses) were used. In a sensitivity analysis, those who changed their consumption between phases 1 and 5 were excluded. Finally, the analyses were repeated among a cohort that excluded those participants with evidence of cardiovascular disease/diabetes during follow-up.

The General Health Questionnaire, Physical Component Score, and employment grade (measured at baseline) were included in the regression as possible confounders. The mediating roles of total blood cholesterol and of systolic and diastolic blood pressure were also explored by entering these variables into the model.

RESULTS

The majority of participants (97 percent of the men and 81 percent of the women) reported at baseline that they had consumed alcohol in the last year (table 1). The most common consumption level reported by men was an average of 1–80 g of alcohol per week (46 percent), while the most common consumption level for women was “none in the past week” (44 percent). Those drinking at the highest weekly levels (men: >241 g; women: >161 g) were more likely to be of high or intermediate employment grade, while abstainers were more likely to be of low employment grades and to be nonsmokers. Mean weekly alcohol consumption showed a linear trend by employment grade for men (p = 0.01) and women (p < 0.01) (data not shown). Most men reported a drinking frequency of once or twice per week, although 30 percent reported almost daily consumption. Women were more likely than men to report drinking on special occasions only.

When the cognitive tests were performed, men were a mean age of 55 years and women 56 years. The range and mean scores for each of the cognitive tests are shown in table 2, along with the cutoffs defining the lowest quintile.

Compared with those drinkers who consumed an average of less than 1 g of alcohol per week (“none in the past week”), those drinking 1 g or more of alcohol per week at baseline were significantly less likely to be in the lowest quintile of each of the cognitive tests at phase 5, after adjustment for age and smoking. There was no suggestion of a U-shaped relation; those participants drinking at the highest levels reported in this study typically had the lowest odds ratios. Never drinkers had approximately a twofold risk of being in the lowest quintile (table 3). Results for both men and women were almost all statistically significant; however, the effect sizes for women tended to be slightly larger than those for men. Alcohol consumption showed a relation with all measures of cognitive ability, although the association with memory was least robust.

Adjusting for physical and mental health made little difference in the results (data not shown). Adding employment grade to the model weakened the association between alcohol consumption and cognitive function (table 3). Additional adjustments for vascular risk factors (cholesterol and blood pressure) did not attenuate the findings further (data not shown). The cross-sectional analyses (phase-5 alcohol consumption and phase-5 cognition) gave results very similar to those for the longitudinal analyses (phase-1 alcohol consumption and phase-5 cognition). Excluding those participants who changed their consumption between phases 1 and 5 did not affect the results. When those with evidence of cardiovascular disease and diabetes were excluded from the analyses, similar associations between alcohol and cognitive function were observed (data not shown).

Adjusting for smoking and age only showed that the lowest odds ratios were found for those drinking almost daily and twice a day or more compared with those who drank only on special occasions (table 4). Never drinkers were more likely to be in the lowest quintiles of cognitive function. Additional adjustments for physical and mental health and for employment grade weakened these associations.

DISCUSSION

In our cohort of middle-aged men and women, we found that those drinking 1 g or more of alcohol per week performed significantly better on all cognitive function tests than drinkers who consumed less. The beneficial effect extended to those drinking more than 240 g (30 drinks) per week. The effect was stronger for women than for men and was not confined to those with evidence of vascular disease. Associations were seen when we used baseline alcohol consumption and consumption reported at the time of the cognitive tests. The relation was not explained by confounding factors of age, smoking, and physical and mental health and was not mediated by cholesterol or blood pressure to a large extent. However, there was evidence that some of the relation between alcohol consumption and cognitive function may be explained by social position, as measured here by employment grade.

Some previous studies have reported a U-shaped relation, suggesting that drinking larger quantities of alcohol is associated with worse cognitive function (25). Kalmijn et al. (4) reported that the most benefit was found for those consuming between one and four drinks per day, while Zuccala et al. (3) reported a nadir of 0.5 liters of wine per day for women and 1 liter per day for men. The Whitehall II cohort included relatively few heavier drinkers, and it was therefore not possible to explore the upper limits of the possible benefit attributable to alcohol. In terms of cognitive function, we found that frequent drinking may be more beneficial than drinking only on special occasions. This finding is compatible with those from the few previous studies that explored drinking patterns and cognitive function (11).

Several possible mechanisms might explain the relation between alcohol consumption and improved cognitive function. First, the association may be mediated by confounding factors, such as good physical and mental health or social position. We did find some evidence for the role of social position, and there may still be confounding by other factors, for example, engaging in leisure activities that maintain cognitive function (2527). Our results appear to suggest some specificity in the association between alcohol consumption and cognitive ability. After adjustment for the confounders, we found that alcohol had no statistically significant beneficial impact on memory. However, results of the Mill Hill test, which measures crystallized intelligence, showed an association with alcohol for both men and women. Crystallized intelligence measures accumulated information and vocabulary; thus, the association with alcohol consumption could have been due to residual confounding by social factors. However, the association with the fluency measures implicates other pathways. Verbal fluency (both phonemic and semantic) is a measure of executive function or a “meta” cognitive ability integrating other cognitive processes such as attention and speed of information processing (2830). Our results suggest an association between volume of alcohol consumed and executive function.

Second, the association may be causal. It is well recognized that alcohol is related to reduced levels of cardiovascular disease and atherosclerosis, mediated by beneficial effects on lipid levels, lipoprotein(a) levels, insulin sensitivity, fibrinogen levels and fibrinolytic activity, and platelet function. Alcohol also reduces the risk of cerebral arterial occlusion and increased cerebral blood flow, which improves cognitive function (4). We previously showed in this cohort that the presence of vascular disease was associated with diminished cognitive function (16). Launer et al. (10) reported that the beneficial effects of alcohol on cognitive function were restricted to only those with existing cardiovascular disease or diabetes. Compared with abstainers, persons drinking one or two glasses of alcohol per day had a significantly lower risk of poor cognitive function. In our study, excluding those with a history of cardiovascular disease or diabetes did not attenuate the associations with alcohol consumption and cognition. However, these diseases do not fully reflect subclinical atherosclerosis and small-vessel disease in the brain, which may be a stronger intermediate factor.

Stronger protective effects of alcohol consumption on cognitive function in women have been reported before (1, 4, 6). The mechanisms behind this stronger association remain unclear. It is possible that the explanation is biologic, linked to sex differences in alcohol metabolism (31, 32). The gender differences may operate through body weight, tissue saturation, stomach enzymes, and the proportion of fat to water in the body. It is also possible that the sociocultural patterns of drinking behavior and beverage choice differ for men and women, but, unfortunately, we were unable to look at these effects with our data.

The objective of this study was to examine the association between cognition and alcohol consumption among persons who were younger and consequently healthier than those examined previously. The study design (involving responding to a long questionnaire and traveling to a screening clinic for the medical examination), age of the respondents, and the question on long-standing illness make it likely that this association was examined in a group free of dementia. The findings from this large study provide evidence of the benefits of alcohol consumption on cognitive function for relatively young men and women. This evidence was found in both longitudinal (11 years of follow-up) and cross-sectional analyses, and, with the repeated measures of alcohol consumption, we were able to exclude those who had changed their consumption levels during follow-up. A battery of cognitive tests was used to examine this association, because a test used to screen for dementia (e.g., the Mini-Mental State Examination) would have resulted in little variability in the cognitive scores because of ceiling effects. The disadvantage with the tests used is the lack of standard criteria to judge poor cognitive performance. In keeping with other research in this area, we present analysis in which the worst quintile was used to denote poor performance (33). However, the association between alcohol consumption and cognition was similar when continuous measures of both variables were used (results not shown but available from the authors).

The analysis presented in the paper used very light drinkers as the reference group, in preference to abstainers, so that the comparison groups were similar regarding other characteristics. In addition, we were also able to control for the possible confounding factors of physical and mental health and of social position in the analyses. Further research is required to examine the relation between alcohol consumption and cognitive decline, both normative age-related and dementia-related decline. Unfortunately, we do not yet have repeat measures of cognitive function and were not able to look at decline. Previous research found that decline in function was slower among drinkers than abstainers (9, 11), while others found no relation (7, 10).

In addition to the cognitive benefits reported here, moderate levels of alcohol consumption have a protective effect against coronary heart disease (34) and possibly ischemic stroke (35). However, the benefits of alcohol drinking may be outweighed by an increased risk of other diseases (including cirrhosis of the liver, pancreatitis, upper aerodigestive cancers, and alcohol psychoses) and of violence and accidents. The balance of risk and benefits is different for different sections of society (36), and it is not proposed that the findings from this and other studies be used to encourage increased consumption.

Correspondence to Dr. Annie Britton, International Centre for Health and Society, Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London WC1E 7HT, United Kingdom (e-mail: a.britton@ucl.ac.uk).

TABLE 1.

Characteristics of participants in the Whitehall II cohort (United Kingdom) who completed cognitive tests during phase 5 of data collection (1997–1999), by volume of alcohol consumed

 Baseline alcohol consumption Total 
 None in the past year None in the past week 1–80 g/week 81–160 g/week 161–240 g/week >241 g/week  
Men        
No. (%) 123 (2.9) 381 (8.9) 1,979 (46.3) 559 (13.1) 838 (19.6) 392 (9.2) 4,272 (100) 
Mean age in years (standard deviation) 56.1 (6.3) 55.5 (6.1) 56.0 (6.1) 55.3 (5.9) 55.0 (6.0) 54.8 (5.6) 55.5 (6.0) 
Employment grade (no. (column %))         
High  44 (36.1) 134 (35.4) 1,035 (52.9) 319 (57.4) 476 (57.3) 193 (49.2) 2,201 (52.0) 
Intermediate 63 (51.6) 201 (53.2) 821 (42.0) 220 (39.6) 328 (39.5) 178 (45.4) 1,811 (42.8) 
Low 15 (12.3) 43 (11.4) 100 (5.1) 17 (3.1) 26 (3.1) 21 (5.4) 222 (5.2) 
Current smoker (no. (%)) 10 (2.7) 34 (9.1) 133 (35.6) 35 (9.4) 82 (21.9) 80 (21.4) 4,272 (100) 
Mean blood pressure (mmHg)        
Systolic 123.5 123.4 122.8 122.8 123.4 125.3 123.2 
Diastolic 78.3 78.6 78.3 78.2 79.1 78.4 78.5 
Mean cholesterol level (mmol/liter) 5.6 5.8 5.8 5.9 6.0 6.2 5.9 
 None in the past year None in the past week 1–48 g/week 49–80 g/week 81–160 g/week >161 g/week  
Women        
No. (%) 335 (19.0) 771 (43.8) 240 (13.6) 224 (12.7) 88 (5.0) 103 (5.8) 1,761 (100) 
Mean age in years (standard deviation) 55.9 (6.04) 56.7 (5.96) 56.3 (6.02) 55.2 (5.90) 55.3 (5.65) 54.3 (5.65) 56.0 (6.04) 
Employment grade (no. (column %))        
High  26 (7.9) 26 (7.9) 120 (15.7) 75 (31.3) 87 (39.2) 47 (54.0) 360 (20.7) 
Intermediate 148 (44.8) 148 (44.8) 370 (48.6) 115 (47.9) 108 (48.6) 30 (34.5) 807 (46.3) 
Low 156 (47.3) 156 (47.3) 272 (35.7) 50 (20.8) 27 (12.2) 10 (11.5) 576 (33.0) 
Current smoker (no. (%)) 2 (1.0) 36 (17.8) 79 (39.1) 29 (14.4) 40 (19.8) 16 (7.9) 202 (100) 
Mean blood pressure (mmHg)        
Systolic  121.7 122.3 122.3 118.6 121.2 119.2 125.1 
Diastolic 75.1 75.2 75.3 74.4 74.1 74.2 77.5 
Mean cholesterol level (mmol/liter) 5.8 6.1 6.0 6.0 5.9 6.0 6.0 
 Baseline alcohol consumption Total 
 None in the past year None in the past week 1–80 g/week 81–160 g/week 161–240 g/week >241 g/week  
Men        
No. (%) 123 (2.9) 381 (8.9) 1,979 (46.3) 559 (13.1) 838 (19.6) 392 (9.2) 4,272 (100) 
Mean age in years (standard deviation) 56.1 (6.3) 55.5 (6.1) 56.0 (6.1) 55.3 (5.9) 55.0 (6.0) 54.8 (5.6) 55.5 (6.0) 
Employment grade (no. (column %))         
High  44 (36.1) 134 (35.4) 1,035 (52.9) 319 (57.4) 476 (57.3) 193 (49.2) 2,201 (52.0) 
Intermediate 63 (51.6) 201 (53.2) 821 (42.0) 220 (39.6) 328 (39.5) 178 (45.4) 1,811 (42.8) 
Low 15 (12.3) 43 (11.4) 100 (5.1) 17 (3.1) 26 (3.1) 21 (5.4) 222 (5.2) 
Current smoker (no. (%)) 10 (2.7) 34 (9.1) 133 (35.6) 35 (9.4) 82 (21.9) 80 (21.4) 4,272 (100) 
Mean blood pressure (mmHg)        
Systolic 123.5 123.4 122.8 122.8 123.4 125.3 123.2 
Diastolic 78.3 78.6 78.3 78.2 79.1 78.4 78.5 
Mean cholesterol level (mmol/liter) 5.6 5.8 5.8 5.9 6.0 6.2 5.9 
 None in the past year None in the past week 1–48 g/week 49–80 g/week 81–160 g/week >161 g/week  
Women        
No. (%) 335 (19.0) 771 (43.8) 240 (13.6) 224 (12.7) 88 (5.0) 103 (5.8) 1,761 (100) 
Mean age in years (standard deviation) 55.9 (6.04) 56.7 (5.96) 56.3 (6.02) 55.2 (5.90) 55.3 (5.65) 54.3 (5.65) 56.0 (6.04) 
Employment grade (no. (column %))        
High  26 (7.9) 26 (7.9) 120 (15.7) 75 (31.3) 87 (39.2) 47 (54.0) 360 (20.7) 
Intermediate 148 (44.8) 148 (44.8) 370 (48.6) 115 (47.9) 108 (48.6) 30 (34.5) 807 (46.3) 
Low 156 (47.3) 156 (47.3) 272 (35.7) 50 (20.8) 27 (12.2) 10 (11.5) 576 (33.0) 
Current smoker (no. (%)) 2 (1.0) 36 (17.8) 79 (39.1) 29 (14.4) 40 (19.8) 16 (7.9) 202 (100) 
Mean blood pressure (mmHg)        
Systolic  121.7 122.3 122.3 118.6 121.2 119.2 125.1 
Diastolic 75.1 75.2 75.3 74.4 74.1 74.2 77.5 
Mean cholesterol level (mmol/liter) 5.8 6.1 6.0 6.0 5.9 6.0 6.0 
TABLE 2.

Cognitive function tests administered to the Whitehall II cohort (United Kingdom) during phase 5 of data collection (1997–1999)

 Score range Men   Women  
  Mean score (standard deviation) Cutoff for the worst quintile  Mean score (standard deviation) Cutoff for the worst quintile 
Memory  0–20 6.85 (2.33)  6.87 (2.72) 
AH4 0–65 48.61 (10.07) 41  41.21 (12.41) 30 
Mill Hill vocabulary 0–33 25.67 (3.89) 22  22.91 (5.64) 17 
Phonemic fluency 0–35 16.97 (4.33) 12  16.51 (4.93) 12 
Semantic fluency 0–30 16.63 (4.02) 12  15.77 (4.66) 11 
 Score range Men   Women  
  Mean score (standard deviation) Cutoff for the worst quintile  Mean score (standard deviation) Cutoff for the worst quintile 
Memory  0–20 6.85 (2.33)  6.87 (2.72) 
AH4 0–65 48.61 (10.07) 41  41.21 (12.41) 30 
Mill Hill vocabulary 0–33 25.67 (3.89) 22  22.91 (5.64) 17 
Phonemic fluency 0–35 16.97 (4.33) 12  16.51 (4.93) 12 
Semantic fluency 0–30 16.63 (4.02) 12  15.77 (4.66) 11 
TABLE 3.

Odds of being in the lowest cognitive function quintile (1997–1999), by baseline alcohol consumption group (1985–1988), Whitehall II cohort, United Kingdom

 Cognitive function test 
 Memory   AH4   Mill Hill vocabulary   Phonemic fluency   Semantic fluency  
 OR† 95% CI†  OR 95% CI  OR 95% CI  OR 95% CI  OR 95% CI 
Men               
Adjusted for age and smoking               
Never 2.1  1.3, 3.4*  2.0 1.3, 3.0*  2.0 1.3, 3.2*  1.0 0.6, 1.6  1.7 1.1, 2.8* 
None in the past week          
1–80 g/week 0.8  0.6, 1.1  0.5 0.4, 0.7*  0.6 0.4, 0.7*  0.6 0.5, 0.8*  0.6 0.5, 0.8* 
81–160 g/week 0.8  0.6, 1.2  0.4 0.3, 0.6*  0.5 0.3, 0.6*  0.4 0.3, 0.6*  0.5 0.3, 0.7* 
161–240 g/week 0.8  0.6, 1.1  0.4 0.3, 0.5*  0.5 0.3, 0.6*  0.5 0.4, 0.7*  0.5 0.3, 0.6* 
>241 g/week 0.5  0.3, 0.8*  0.4 0.3, 0.5*  0.4 0.3, 0.6*  0.5 0.3, 0.7*  0.5 0.3, 0.7* 
Adjusted for age, smoking, physical and mental health, and employment grade               
Never 2.1 1.2, 3.5 *  2.2 1.3, 3.7*  2.1 1.3, 3.6*  0.8 0.5, 1.5  1.4 0.9, 2.5 
None in the past week          
1–80 g/week 1.0 0.7, 1.4  0.8 0.6, 1.1  0.8 0.6, 1.1  0.8 0.6, 1.0  0.8 0.6, 1.1 
81–160 g/week 1.1 0.7, 1.6  0.7 0.5, 1.0  0.7 0.5, 1.1  0.6 0.4, 0.8 *  0.7 0.5, 1.0 
161–240 g/week 1.1 0.7, 1.5  0.6 0.4, 0.9*  0.8 0.6, 1.2  0.7 0.5, 1.0  0.7 0.5, 1.0 
>241 g/week 0.6 0.4, 1.0  0.5 0.3, 0.7*  0.5 0.3, 0.8*  0.6 0.4, 0.9*  0.6 0.4, 0.9* 
               
Women               
Adjusted for age and smoking               
Never 1.6 0.9, 2.7  2.3 1.4, 3.8 *  2.4 1.5, 3.9 *  1.3 0.8, 2.2  2.3 1.4, 3.7 * 
None in the past week          
1–48 g/week 0.7 0.5, 0.9*  0.6 0.5, 0.9*  0.4 0.3, 0.6*  0.7 0.5, 0.9*  0.9 0.7, 1.2 
49–80 g/week 0.5 0.3, 0.8*  0.3 0.2, 0.5*  0.2 0.1, 0.3*  0.3 0.2, 0.6*  0.5 0.3, 0.8* 
81–160 g/week 0.6 0.4, 1.0  0.2 0.1, 0.3*  0.1 0.04, 0.2*  0.3 0.2, 0.5*  0.3 0.2, 0.6* 
>161 g/week 0.5 0.2, 0.9*  0.1 0.03, 0.3*  0.1 0.04, 0.3*  0.2 0.1, 0.5*  0.1 0.02, 0.4* 
Adjusted for age, smoking, physical and mental health, and employment grade               
Never 1.5 0.9, 2.6  1.9 1.1, 3.2*  2.1 1.2, 3.6*  1.2 0.7, 2.0  2.3 1.3, 4.1* 
None in past week          
1–48 g/week 0.7 0.5, 1.0  0.8 0.6, 1.2  0.6 0.4, 0.8*  0.8 0.6, 1.1  1.4 0.9, 2.0 
49–80 0.7 0.4, 1.1  0.6 0.4, 1.1  0.3 0.2, 0.6*  0.5 0.3, 0.9*  1.0 0.6, 1.8 
81–160 1.0 0.6, 1.7  0.6 0.3, 1.2  0.3 0.1, 0.7*  0.7 0.4, 1.3  1.3 0.7, 2.4 
>161 g/week 0.7 0.3, 1.5  0.3 0.1, 1.1  0.4 0.1, 1.3  0.5 0.2, 1.2  0.3 0.1, 1.3 
 Cognitive function test 
 Memory   AH4   Mill Hill vocabulary   Phonemic fluency   Semantic fluency  
 OR† 95% CI†  OR 95% CI  OR 95% CI  OR 95% CI  OR 95% CI 
Men               
Adjusted for age and smoking               
Never 2.1  1.3, 3.4*  2.0 1.3, 3.0*  2.0 1.3, 3.2*  1.0 0.6, 1.6  1.7 1.1, 2.8* 
None in the past week          
1–80 g/week 0.8  0.6, 1.1  0.5 0.4, 0.7*  0.6 0.4, 0.7*  0.6 0.5, 0.8*  0.6 0.5, 0.8* 
81–160 g/week 0.8  0.6, 1.2  0.4 0.3, 0.6*  0.5 0.3, 0.6*  0.4 0.3, 0.6*  0.5 0.3, 0.7* 
161–240 g/week 0.8  0.6, 1.1  0.4 0.3, 0.5*  0.5 0.3, 0.6*  0.5 0.4, 0.7*  0.5 0.3, 0.6* 
>241 g/week 0.5  0.3, 0.8*  0.4 0.3, 0.5*  0.4 0.3, 0.6*  0.5 0.3, 0.7*  0.5 0.3, 0.7* 
Adjusted for age, smoking, physical and mental health, and employment grade               
Never 2.1 1.2, 3.5 *  2.2 1.3, 3.7*  2.1 1.3, 3.6*  0.8 0.5, 1.5  1.4 0.9, 2.5 
None in the past week          
1–80 g/week 1.0 0.7, 1.4  0.8 0.6, 1.1  0.8 0.6, 1.1  0.8 0.6, 1.0  0.8 0.6, 1.1 
81–160 g/week 1.1 0.7, 1.6  0.7 0.5, 1.0  0.7 0.5, 1.1  0.6 0.4, 0.8 *  0.7 0.5, 1.0 
161–240 g/week 1.1 0.7, 1.5  0.6 0.4, 0.9*  0.8 0.6, 1.2  0.7 0.5, 1.0  0.7 0.5, 1.0 
>241 g/week 0.6 0.4, 1.0  0.5 0.3, 0.7*  0.5 0.3, 0.8*  0.6 0.4, 0.9*  0.6 0.4, 0.9* 
               
Women               
Adjusted for age and smoking               
Never 1.6 0.9, 2.7  2.3 1.4, 3.8 *  2.4 1.5, 3.9 *  1.3 0.8, 2.2  2.3 1.4, 3.7 * 
None in the past week          
1–48 g/week 0.7 0.5, 0.9*  0.6 0.5, 0.9*  0.4 0.3, 0.6*  0.7 0.5, 0.9*  0.9 0.7, 1.2 
49–80 g/week 0.5 0.3, 0.8*  0.3 0.2, 0.5*  0.2 0.1, 0.3*  0.3 0.2, 0.6*  0.5 0.3, 0.8* 
81–160 g/week 0.6 0.4, 1.0  0.2 0.1, 0.3*  0.1 0.04, 0.2*  0.3 0.2, 0.5*  0.3 0.2, 0.6* 
>161 g/week 0.5 0.2, 0.9*  0.1 0.03, 0.3*  0.1 0.04, 0.3*  0.2 0.1, 0.5*  0.1 0.02, 0.4* 
Adjusted for age, smoking, physical and mental health, and employment grade               
Never 1.5 0.9, 2.6  1.9 1.1, 3.2*  2.1 1.2, 3.6*  1.2 0.7, 2.0  2.3 1.3, 4.1* 
None in past week          
1–48 g/week 0.7 0.5, 1.0  0.8 0.6, 1.2  0.6 0.4, 0.8*  0.8 0.6, 1.1  1.4 0.9, 2.0 
49–80 0.7 0.4, 1.1  0.6 0.4, 1.1  0.3 0.2, 0.6*  0.5 0.3, 0.9*  1.0 0.6, 1.8 
81–160 1.0 0.6, 1.7  0.6 0.3, 1.2  0.3 0.1, 0.7*  0.7 0.4, 1.3  1.3 0.7, 2.4 
>161 g/week 0.7 0.3, 1.5  0.3 0.1, 1.1  0.4 0.1, 1.3  0.5 0.2, 1.2  0.3 0.1, 1.3 

* p < 0.05.

† OR, odds ratio; CI, confidence interval.

TABLE 4.

Odds of being in the lowest cognitive function quintile (1997–1999), by baseline frequency of alcohol consumption (1985–1988), Whitehall II cohort, United Kingdom

 No.† Cognitive function test 
  Memory   AH4   Mill Hill vocabulary   Phonemic fluency   Semantic fluency  
  OR‡ 95% CI‡  OR 95% CI  OR 95% CI  OR 95% CI  OR 95% CI 
Men                
Adjusted for age and smoking                
Never 121 1.9 1.2, 3.1*  1.7 1.0, 2.6*  1.7 1.1, 2.6*  1.0 0.6, 1.7  1.7 1.0, 2.7* 
On special occasions only  327          
1–2 times/month  494 1.0 0.7, 1.4  0.6 0.5, 0.8*  0.6 0.4, 0.8*  0.7 0.5, 1.0   0.7 0.5, 1.1 
1–2 times/week  1,848 0.8 0.5, 1.0  0.4 0.3, 0.6*  0.5 0.4, 0.6*  0.6 0.5, 0.8*  0.6 0.5, 0.8* 
Almost daily 1,280 0.6 0.4, 0.8*  0.3 0.2, 0.4*  0.3 0.2, 0.4*  0.5 0.3, 0.6*  0.4 0.3, 0.6* 
Twice/day or more 165 0.8 0.5, 1.4  0.2 0.1, 0.4*  0.2 0.1, 0.4*  0.5 0.3, 0.9*  0.4 0.2, 0.6* 
Adjusted for age, smoking, physical and mental health, and employment grade                
Never 112 2.1 1.2, 3.5*  2.2 1.3, 3.6*  1.5 0.9, 2.4  1.0 0.6, 1.8  1.6 1.0, 2.8* 
On special occasions only 306          
1–2 times/month 476 1.3 0.9, 1.9  0.9 0.7, 1.3  0.9 0.6, 1.3  1.0 0.7, 1.4  1.1 0.7, 1.6 
1–2 times/week 1,746 1.0 0.7, 1.4  0.7  0.5, 1.0  0.7 0.5, 1.0  0.9 0.7, 1.3  0.9 0.6, 1.3 
Almost daily 1,229 0.8 0.6, 1.2  0.5  0.4, 0.7*  0.6  0.4, 0.8*  0.8  0.5, 1.1  0.7 0.5, 1.0 
Twice/day or more 158 1.3 0.7, 2.2  0.4 0.2, 0.8*  0.4 0.2, 0.7*  0.9 0.5, 1.6  0.5 0.3, 1.0 
                
Women                
Adjusted for age and smoking                
Never  1.4 0.8, 2.3  2.0 1.2, 3.2*  2.6 1.6, 4.1*  1.4 0.9, 2.3  2.0 1.2, 3.2* 
On special occasions only 364          
1–2 times/month 251 0.5 0.3, 0.7*  0.5 0.3, 0.7*  0.6 0.4, 0.9*  0.7 0.5, 1.1  0.6 0.4, 1.0 
1–2 times/week 625 0.5 0.4, 0.7*  0.4 0.3, 0.6*  0.4 0.3, 0.5*  0.6 0.4, 0.8*  0.7 0.5, 1.0 
Almost daily 364 0.5 0.3, 0.7*  0.2 0.1, 0.3*  0.2 0.1, 0.3*  0.4 0.3, 0.6*  0.2 0.1, 0.4* 
Twice/day or more 29 0.3 0.1, 1.0  0.0 0.0, 1491.6  0.0 0.0, 1738.2  0.1 0.02, 0.9*  0.1 0.02, 0.9* 
Adjusted for age, smoking, physical and mental health, and employment grade                
Never 90 1.3 0.8, 2.3  1.6 0.9, 2.8  2.3 1.3, 4.0*  1.3 0.8, 2.2  1.9 1.1, 3.3* 
On special occasions only  323          
1–2 times/month  220 0.4 0.3, 0.7*  0.5 0.3, 0.8*  0.7 0.4, 1.1  0.8 0.5, 1.3  0.8 0.5, 1.3 
1–2 times/week 572 0.6 0.4, 0.9*  0.7 0.5, 0.9*  0.6 0.4, 0.8*  0.8 0.6, 1.1  1.2 0.8, 1.7 
Almost daily 344 0.7 0.5, 1.2  0.6 0.3, 0.9*  0.5 0.3, 0.9*  0.9 0.6, 1.4  0.7 0.4, 1.1 
Twice/day or more 27 0.2 0.0, 1.6  0.0   0.0   0.3 0.03, 2.1  0.4 0.1, 3.7 
 No.† Cognitive function test 
  Memory   AH4   Mill Hill vocabulary   Phonemic fluency   Semantic fluency  
  OR‡ 95% CI‡  OR 95% CI  OR 95% CI  OR 95% CI  OR 95% CI 
Men                
Adjusted for age and smoking                
Never 121 1.9 1.2, 3.1*  1.7 1.0, 2.6*  1.7 1.1, 2.6*  1.0 0.6, 1.7  1.7 1.0, 2.7* 
On special occasions only  327          
1–2 times/month  494 1.0 0.7, 1.4  0.6 0.5, 0.8*  0.6 0.4, 0.8*  0.7 0.5, 1.0   0.7 0.5, 1.1 
1–2 times/week  1,848 0.8 0.5, 1.0  0.4 0.3, 0.6*  0.5 0.4, 0.6*  0.6 0.5, 0.8*  0.6 0.5, 0.8* 
Almost daily 1,280 0.6 0.4, 0.8*  0.3 0.2, 0.4*  0.3 0.2, 0.4*  0.5 0.3, 0.6*  0.4 0.3, 0.6* 
Twice/day or more 165 0.8 0.5, 1.4  0.2 0.1, 0.4*  0.2 0.1, 0.4*  0.5 0.3, 0.9*  0.4 0.2, 0.6* 
Adjusted for age, smoking, physical and mental health, and employment grade                
Never 112 2.1 1.2, 3.5*  2.2 1.3, 3.6*  1.5 0.9, 2.4  1.0 0.6, 1.8  1.6 1.0, 2.8* 
On special occasions only 306          
1–2 times/month 476 1.3 0.9, 1.9  0.9 0.7, 1.3  0.9 0.6, 1.3  1.0 0.7, 1.4  1.1 0.7, 1.6 
1–2 times/week 1,746 1.0 0.7, 1.4  0.7  0.5, 1.0  0.7 0.5, 1.0  0.9 0.7, 1.3  0.9 0.6, 1.3 
Almost daily 1,229 0.8 0.6, 1.2  0.5  0.4, 0.7*  0.6  0.4, 0.8*  0.8  0.5, 1.1  0.7 0.5, 1.0 
Twice/day or more 158 1.3 0.7, 2.2  0.4 0.2, 0.8*  0.4 0.2, 0.7*  0.9 0.5, 1.6  0.5 0.3, 1.0 
                
Women                
Adjusted for age and smoking                
Never  1.4 0.8, 2.3  2.0 1.2, 3.2*  2.6 1.6, 4.1*  1.4 0.9, 2.3  2.0 1.2, 3.2* 
On special occasions only 364          
1–2 times/month 251 0.5 0.3, 0.7*  0.5 0.3, 0.7*  0.6 0.4, 0.9*  0.7 0.5, 1.1  0.6 0.4, 1.0 
1–2 times/week 625 0.5 0.4, 0.7*  0.4 0.3, 0.6*  0.4 0.3, 0.5*  0.6 0.4, 0.8*  0.7 0.5, 1.0 
Almost daily 364 0.5 0.3, 0.7*  0.2 0.1, 0.3*  0.2 0.1, 0.3*  0.4 0.3, 0.6*  0.2 0.1, 0.4* 
Twice/day or more 29 0.3 0.1, 1.0  0.0 0.0, 1491.6  0.0 0.0, 1738.2  0.1 0.02, 0.9*  0.1 0.02, 0.9* 
Adjusted for age, smoking, physical and mental health, and employment grade                
Never 90 1.3 0.8, 2.3  1.6 0.9, 2.8  2.3 1.3, 4.0*  1.3 0.8, 2.2  1.9 1.1, 3.3* 
On special occasions only  323          
1–2 times/month  220 0.4 0.3, 0.7*  0.5 0.3, 0.8*  0.7 0.4, 1.1  0.8 0.5, 1.3  0.8 0.5, 1.3 
1–2 times/week 572 0.6 0.4, 0.9*  0.7 0.5, 0.9*  0.6 0.4, 0.8*  0.8 0.6, 1.1  1.2 0.8, 1.7 
Almost daily 344 0.7 0.5, 1.2  0.6 0.3, 0.9*  0.5 0.3, 0.9*  0.9 0.6, 1.4  0.7 0.4, 1.1 
Twice/day or more 27 0.2 0.0, 1.6  0.0   0.0   0.3 0.03, 2.1  0.4 0.1, 3.7 

* p < 0.05.

† Sample sizes are affected by missing data.

‡ OR, odds ratio; CI, confidence interval.

References

1.
Elias PK, Elias MF, D’Agnostino RB, et al. Alcohol consumption and cognitive performance in the Framingham Heart Study.
Am J Epidemiol
 
1999
;
150
:
580
–9.
2.
Mukamal KJ, Longstreth WT, Mittleman MA, et al. Alcohol consumption and subclinical findings on magnetic resonance imaging of the brain in older adults. The Cardiovascular Health Study.
Stroke
 
2001
;
32
:
1939
–46.
3.
Zuccala G, Onder G, Pedone C, et al. Dose-related impact of alcohol consumption on cognitive function in advanced age: results of a multicenter survey.
Alcohol Clin Exp Res
 
2001
;
25
:
1743
–8
4.
Kalmijn S, van Boxtel MPJ, Verschuren MWM, et al. Cigarette smoking and alcohol consumption in relation to cognitive performance in middle age.
Am J Epidemiol
 
2002
;
156
:
936
–44.
5.
Ruitenberg A, van Swieten JC, Witteman JCM, et al. Alcohol consumption and risk of dementia: the Rotterdam Study.
Lancet
 
2002
;
359
:
281
–6.
6.
Dufouil C, Ducimetière P, Alpérovitch A. Sex differences in the association between alcohol consumption and cognitive performance.
Am J Epidemiol
 
1997
;
146
:
405
–12.
7.
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.
8.
Elwood PC, Gallacher JEJ, Hopkinson CA, et al. Smoking, drinking, and other life style factors and cognitive function in men in the Caerphilly cohort.
J Epidemiol Community Health
 
1999
;
53
:
9
–14.
9.
Cervilla JA, Prince M, Mann A. Smoking, drinking, and incident cognitive impairment: a community based study included in the Gospel Oak project.
J Neurol Neurosurg Psychiatry
 
2000
;
68
:
622
–6.
10.
Launer LJ, Feskens EJM, Kalmijn S, et al. Smoking, drinking, and thinking: the Zutphen Elderly Study.
Am J Epidemiol
 
1996
;
143
:
219
–27.
11.
Leroi I, Sheppard JM, Lyketsos CG. Cognitive function after 11.5 years of alcohol use: relation to alcohol use.
Am J Epidemiol
 
2002
;
156
:
747
–52.
12.
Rabbitt P, Donlan C, Watson P, et al. Unique and interactive effects of depression, age, socioeconomic advantage, and gender on cognitive performance of normal healthy older people.
Psychol Aging
 
1995
;
10
:
307
–13.
13.
Holland CA, Rabbitt P. The course and causes of cognitive change with advancing age.
Rev Clin Gerontol
 
1991
;
1
:
81
–96.
14.
Desmond DW, Tatemichi TK, Paik M, et al. Risk factors for cerebrovascular disease as correlates of cognitive function in a stroke-free cohort.
Arch Neurol
 
1993
;
50
:
162
–6.
15.
Launer LJ, Masaki K, Petrovitch H, et al. The association between midlife blood pressure levels and late-life cognitive function.
JAMA
 
1995
;
274
:
1846
–51.
16.
Singh-Manoux A, Britton A, Marmot M. Vascular disease and cognitive function.
J Am Geriatr Soc
 
2003
;
51
:
1445
–50.
17.
Marmot M, Smith GD, Stansfeld S, et al. Health inequalities among British civil servants: the Whitehall II Study.
Lancet
 
1991
;
337
:
1387
–93.
18.
Heim AW. AH4 group test of general intelligence ASE. Windsor, United Kingdom: NFER-Nelson Publishing Company Ltd, 1970.
19.
Raven JC. Guide to using the Mill Hill vocabulary scale with progressive matrices. London, United Kingdom: H K Lewis, 1965.
20.
Goldberg DP. The detection of psychiatric illness by questionnaire. London, United Kingdom: Oxford University Press, 1972. (Institute of Psychiatry Maudsley monographs no. 21).
21.
Ware JE, Snow KK, Kosinski M, et al. SF-36 health survey: manual and interpretation guide. Boston, MA: The Health Institute, New England Medical Center, 1993.
22.
Ware JE, Kosinski M, Keller SD. SF-36 Physical and Mental Health Summary Scales: a user’s manual. Boston, MA: The Health Institute, New England Medical Center, 1994.
23.
Ware JE, Kosinski M, Bayliss MS, et al. Comparison of methods of scoring the statistical analysis of SF-36 health profile and summary measures: summary of results from the Medical Outcomes Study.
Med Care
 
1995
;
33
:
264
–79.
24.
Rose GA. The diagnosis of ischaemic heart pain and intermittent claudication in field surveys.
Bull World Health Organ
 
1962
;
27
:
645
–58.
25.
Richards M, Hardy R, Wadsworth MEJ. Does active leisure protect cognition? Evidence from a national birth cohort.
Soc Sci Med
 
2003
;
56
:
785
–92.
26.
Singh-Manoux A, Richards M, Marmot M. Leisure activities and cognitive function in middle age: evidence from the Whitehall II Study.
J Epidemiol Community Health
 
2003
;
57
:
907
–13.
27.
Stevens FCJ, Kaplan CD, Ponds RWHM, et al. How ageing and social factors affect memory.
Age Ageing
 
1999
;
28
:
379
–84.
28.
Benito-Cuadrado MM, Esteba-Castillo S, Bohm P, et al. Semantic verbal fluency of animals: a normative and predictive study in the Spanish population.
J Clin Exp Neuropsychol
 
2002
;
24
:
1117
–22.
29.
Bryan J, Luszcz MA. Measurement of executive function: considerations for detecting adult age differences.
J Clin Exp Neuropsychol
 
2000
;
22
:
40
–55.
30.
Chertkow H, Bub D. Semantic memory loss in dementia of Alzheimer’s type. What do various measures measure?
Brain
 
1990
;
113
:
397
–417.
31.
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.
32.
Weight FF. Cellular and molecular physiology of alcohol actions in the nervous system.
Int Rev Neurobiol
 
1992
;
33
:
289
–348.
33.
Anstey KJ, Luszcz MA, Giles LC, et al. Demographic, health, cognitive, and sensory variables as predictors of mortality in very old adults.
Psychol Aging
 
2001
;
16
:
3
–11.
34.
Doll R, Peto R, Hall E, et al. Mortality in relation to consumption of alcohol: 13 years’ observations on male British doctors.
BMJ
 
1994
;
309
:
911
–18.
35.
Mazzaglia G, Britton A, Altman D, et al. Exploring the relationship between alcohol consumption and non-fatal or fatal stroke: a systematic review.
Addiction
 
2001
;
96
:
1743
–56.
36.
Britton A, McPherson K. Mortality in England and Wales attributable to current alcohol consumption.
J Epidemiol Community Health
 
2001
;
55
:
383
–8.