Mortensen EL, Jensen HH, Sanders SA, Reinisch JM. Better Psychological Functioning and Higher Social Status May Largely Explain the Apparent Health Benefits of WineA Study of Wine and Beer Drinking in Young Danish Adults. Arch Intern Med. 2001;161(15):1844-1848. doi:10.1001/archinte.161.15.1844
Findings from a recent series of Danish studies suggest that moderate wine drinkers are healthier than those who drink other alcoholic beverages or those who abstain.
To identify possible explanatory factors associated with the health benefits of wine consumption through the examination of a wide spectrum of social, cognitive, and personality characteristics related to both beverage choice and health in young Danish adults.
Subjects and Methods
Descriptive cross-sectional study of characteristics associated with beverage choice in a sample of 363 men and 330 women between the ages of 29 and 34 years, selected from the Copenhagen Perinatal Cohort on the basis of perinatal records.
Main Outcome Measures
Socioeconomic status, education, IQ, personality, psychiatric symptoms, and health-related behaviors, including alcohol consumption, were analyzed. The outcome variables were subjected to linear and logistic regression analyses with 2 factors (beer and wine), each with 2 levels (drinking or not drinking a certain beverage type).
Wine drinking was significantly associated with higher IQ, higher parental educational level, and higher socioeconomic status. Beer drinking was significantly associated with lower scores on the same variables. On scales concerning personality, psychiatric symptoms, and health-related behaviors, wine drinking was associated with optimal functioning and beer drinking with suboptimal functioning.
Our data demonstrate that wine drinking is a general indicator of optimal social, cognitive, and personality development in Denmark. Similar social, cognitive, and personality factors have also been associated with better health in many populations. Consequently, the association between drinking habits and social and psychological characteristics, in large part, may explain the apparent health benefits of wine.
DANISH epidemiological studies of beverage choice and health have revealed many apparent health benefits related to drinking wine. In this series of studies light to moderate wine consumption was related to good self-perceived health1; lower risks for stroke,2 upper digestive tract cancer,3 lung cancer,4 and hip fracture5; and, perhaps more remarkably, to a lower rate of all-cause mortality as compared with abstaining, or beer or liquor drinking.6
However, all the Danish studies based their analysis of the beneficial health effects of wine on observational data collected between 1976 and 1982. At that time, only a small proportion of the Danish population regularly drank wine. This raises the question as to whether the apparent health benefits of wine reflected the direct physiologic effects of this beverage or other health-related individual differences between wine drinkers and nonwine drinkers. In this study, we examined the association between beverage choice and a broad spectrum of social, cognitive, and personality characteristics in a large sample of Danish adults evaluated between November 26, 1990, and October 5, 1994.
The Copenhagen Perinatal Cohort consists of 9125 individuals born at the Copenhagen University Hospital from September 22, 1959, to December 20, 1961. As part of a study addressing the effects of prenatal exposure to prescribed maternal medications, a sample was selected on the basis of perinatal records for detailed evaluation of physical, psychological, and social development.7 (Preliminary analyses revealed very few significant interactions between beverage choice and prenatal medication exposure.) The sample consisted of 693 subjects (363 men and 330 women) between the ages of 29 and 34 years (mean age, 32.0 years). The mean number of school years attended for the total sample was about 10 months above the average reported for this Danish generation (ie, 10 years).8
The assessment procedures included the following standardized tests and questionnaires: Wechsler Adult Intelligence Scale,9 Eysenck Personality Questionnaire,10 Zuckerman's Sensation Seeking Scale,11 and Millon Clinical Multiaxial Inventory12 (data will only be presented for the clinical symptom scales). Detailed information was also collected on health and health behaviors and on social status and educational level, including parental educational level and social status (recorded at the subjects' 1-year birthday). An Alcohol Usage Questionnaire included detailed information about alcohol consumption during the last week (or the last typical week, if consumption in the most recent week was atypical).
Cross-tabulation of beer and wine drinking resulted in the following 4 groups based on beverage preference during the last week: no beer/no wine drinkers, beer drinkers, wine drinkers, and beer and wine drinkers. The distribution of men and women within these groups is listed in Table 1. Consumption of liquor was rare, the median consumption being 0 to 1 drinks per week, irrespective of beer- and wine-drinking pattern.
The number of drinks during the last week is given in Table 2. For the total sample, the mean number of drinks per week and the percentage of abstainers was close to results for the same age group in a 1995 Copenhagen Health Survey.13
Beer and wine are conceptualized as 2 factors with 2 levels (drinking or not drinking). A linear regression program (SPSS Inc, Chicago, Ill) was used to test the main effects of these 2 factors and their possible interactions. Statistically significant interactions between beer and wine were disclosed for only women's parental educational level (Table 3). Thus, the characteristics associated with the 2 beverage types may be interpreted independently. The differences between wine drinkers and no wine drinkers were not signficantly different in the beer-drinking and no beer-drinking parts of the sample. Similarly the characteristics associated with beer drinking did not depend on wine-drinking status. Preliminary analysis revealed many interactions between sex and alcohol type, and consequently, data for male and female subjects were analyzed separately.
Table 3, Table 4, and Table 5 present data for the 4 subgroups resulting from cross-tabulation of beer and wine drinking. As there was no interaction between beer and wine drinking, the tables also give the estimates of the main effects associated with the 2 beverage types. In general, across all tables, the differences between the pure beer and pure wine drinkers were close to the differences expected from the main effects of beer and wine drinking, respectively.
The data in Table 3 show that for men and women, wine drinking was consistently associated with higher scores in parental social status, parental education, and subjects' years in school and social status.
Table 4 presents Verbal, Performance, and Full-Scale Wechsler Adult Intelligence Scale IQs. Significant main effects of beer and wine were obtained for all IQs except for Performance IQ in women. For both sexes, beer drinking was consistently related to lower scores on the IQ scales; wine drinking consistently related to higher scores on the IQ scales. The IQ differences between male pure beer and pure wine drinkers were dramatic (Full-Scale IQs, 95.2 and 113.2, respectively).
For Sensation Seeking, Table 4 shows a significant effect of wine drinking for males only. With regard to the Eysenck Personality Questionnaire, beer-drinking men showed higher Neuroticism scores while for women, wine drinking was related to lower Neuroticism scores and to higher Extroversion scores. Analysis of Millon Clinical Multiaxial Inventory clinical symptoms revealed that wine drinking was consistently related to lower scores and beer drinking to higher, more pathological scores for both men and women (Table 5). Moreover, pure wine drinkers obtained the lowest (healthiest) mean scores; pure beer drinkers obtained the highest mean scores on 8 of 9 symptom scales, with the largest differences in Anxiety and Alcohol Abuse.
Beverage choice was unrelated to body mass index, height, and weight. Beer drinking was statistically significantly associated with a higher prevalence of risk drinking, smoking, and illicit drug use, while wine drinking was associated with lower prevalence of smoking (data not shown).
Data in Table 2 show that total alcohol consumption is larger for beer drinkers (particularly in male subjects). Since studies investigating health effects of beverage choice usually attempt to control total alcohol consumption, all analyses were repeated with total alcohol consumption as a covariate. As may be expected from the correlation between beer drinking and total consumption, most effects associated with beer drinking became smaller or insignificant. In contrast, the effects associated with wine drinking remained significant and for some Wechsler Adult Intelligence Scale IQs and Millon Clinical Multiaxial Inventory symptom scales became statistically stronger.
The Copenhagen City Heart Study6 demonstrated lower mortality associated with wine drinking based on measurement of drinking habits 3 to 5 years after Denmark joined the European Community in 1973. Before joining the European Community, Denmark was traditionally a beer-drinking country. In the 1990s it is much more common for Danes to drink wine, and consequently intellectual, social, and personality differences between wine drinkers and no wine drinkers may be far less pronounced than in the 1970 population study by Grønbæk et al.6 Nevertheless, 20 years later, we found 2 sets of characteristics significantly associated with beer and wine drinking in a Danish population. In support of our conclusions, a 1999 study of the introduction of coffee into the traditionally tea-drinking society of Scotland shows similar relationships to social status and health risk factors.14
Social status gradients in health are one of the most consistent findings of public health epidemiology.15,16 The present study demonstrates strong relationships among beverage choice and intelligence and socioeconomic status (these results were corroborated by a parallel analysis of a 20- to 26-year-old sample from the same cohort). A substantial part of the variation in intelligence (and personality) is independent of social status, and consequently, residual confounding is likely when a single index of social status is used as a covariate in studies of the beneficial health effects of wine.
Our results suggest that wine drinking is associated with optimal social, intellectual, and personality functioning, while beer drinking is associated with suboptimal characteristics. According to Bouchard,17 virtually any dimension of behavior scaled from the less valued to the more valued correlates positively with IQ. The dramatic IQ differences and differences in personality measures between wine drinkers and beer drinkers strongly suggest that wine and beer drinking in Denmark is associated with many known and unknown factors that may affect health, morbidity, and mortality.
Accepted for publication January 11, 2001.
This study has been supported by US Health Service (USPHS) grant DA05056 from the National Institute on Drug Abuse (Drs Reinisch and Sanders), USPHS grants HD17655 and HD20263 from the National Institue of Child Health and Human Development (Dr Reinisch), Bethesda, Md, and grant 1400/2-4-1997 from the Danish National Board of Health, Copenhagen (Dr Mortensen).
We thank Vibeke Munk, BA, for help with the manuscript and for critical comments.
Corresponding author: Erik L. Mortensen, PhD, Institute of Preventive Medicine, Kommunehospitalet, DK-1399 Copenhagen K, Denmark (e-mail: firstname.lastname@example.org).