Studies on the effects of alcohol consumption on cognitive function have conflicting results. Several longitudinal studies show an association of light-to-moderate alcohol consumption with a reduced risk of cognitive impairment and dementia,1–10 whereas others show no association.13–17 However, an association between increasing alcohol consumption and cognitive deterioration or dementia has been found in subjects with a history of heavy drinking and in those with an APOE-ε4 allele.2,9,11,12 In most studies, the follow-up period has ranged up to 6 years,2,4–6,8,9,11–16 whereas 4 studies have a follow-up of 10 to 25 years.1,3,10,17
The outcome measures of the longitudinal studies have ranged from cognitive deterioration1–3,10,12,13,16 to Alzheimer disease,4–6,8,9,14,17 vascular dementia,4,9,15 or dementia in general.4,6,9,11,17 The most common definition of light-to- moderate alcohol consumption is 3–4 drinks per day,5,6,10 although this varies from drinking alcoholic beverages at least once a week8,14,15 to more than 2 drinks per day.2 This variation in definition of light-to-moderate alcohol consumption makes it difficult to compare the different studies. Apart from quantity of alcohol consumed, specific drinking patterns are rarely reported. Furthermore, many studies have assessed alcohol use at only 1 point in time. For example, the long-term effects of 2 drinks everyday may differ from the effects of 7 drinks twice a week. Light-to-moderate drinkers who engage in binge drinking have been reported to have higher mortality risks than those who do not.18
In the present study, we examined a population-based sample of 554 subjects who belong to the prospective Finnish Twin Cohort Study.19 During the initial years of the follow-up period of 25 years, these subjects had received 2 structured questionnaires 6 years apart; these questionnaires included detailed questions about alcohol use. We investigated the long-term effects of midlife alcohol consumption on cognitive performance later in life. We were particularly interested in investigating whether various drinking patterns have different effects on cognition.
The subjects of this study were recruited through the Finnish Twin Cohort, which was compiled from the Central Population Registry of Finland.19 The population registry is a computerized databank of personal information on all Finnish citizens from 1967 onward. All sets of persons with the same birth date, sex, surname at birth, and local community of birth were identified to yield a nationwide cohort of 13,888 pairs of twins of known zygosity20 born before 1958.
Of these pairs, 1089 were monozygotic pairs born 1908–1933. From this group, all 826 monozygotic twin individuals (including subjects living in institutions) from 413 pairs with both alive in 1999, and age 65 years and over at that time, were asked to participate in a telephone interview using the TELE instrument21 to assess cognitive status, as described subsequently. The interviews were conducted by trained nurses from the Department of Public Health, University of Turku, Finland. A week before the telephone interview, each twin was sent a letter containing information about the study and its purpose. The nurses called the subjects and asked permission to conduct an interview. If the caregivers answered, they were asked their opinion about the subjects’ ability to participate in the interview. If the subject also gave consent, the interview was started.
The study protocol was approved by the Joint Ethical Committee of the University of Turku and the Turku University Central Hospital. The telephone interviews took place between 6 September 1999 and 5 November 2001.
Alcohol Intake and Other Baseline Measurements
The twins received structured questionnaires in 1975 and 1981 assessing their demographic, social, environmental, medical, and lifestyle characteristics. Among other questions, the subjects’ drinking habits were queried in detail. Total alcohol consumption was computed7 on the basis of beverage-specific responses to the average amount consumed weekly or monthly. Individuals were categorized as abstainers, light drinkers (3 or fewer drinks per week), moderate drinkers (more than 3 but no more than 7 [for women] or 14 [for men] drinks per week), and heavy drinkers (on average, more than a drink a day for women and more than 2 for men). The alcohol frequencies were asked separately for beer, wine, and spirits (never, ≤2 days, 3–8 days, 9–15, or 16 or more days per month). Binge drinking was defined by the question “Do you drink at least once a month more than 1 bottle of wine, half a bottle of spirits, or the equivalent amount of other alcoholic beverages on the 1 and same occasion?” A question on passing out as a result of excessive alcohol intake was in the 1981 questionnaire only with 5 response categories. This variable was analyzed as never, once, or twice or more during the past year. In the 1975 questionnaire, the twins reported whether their alcohol consumption had been greater at any time prior. Of 135 subjects who were then abstainers, 2 reported more alcohol use earlier and both had been light drinkers. At follow-up, one of them was cognitively healthy and the other cognitively declined, as defined subsequently.
Smoking habits were assessed in detail in the 1975 questionnaire22 and classified for the present analyses as a dichotomous variable (current smoker in 1975: yes or no). Education was assessed using a 9-category question on attained educational level, and this was categorized into high (≥12 years), intermediate (7–11 years), and low (≤6 years), according to the 1975 questionnaire response.
Assessment of Dementia
The subjects of this study were categorized into demented, cognitively declined, and healthy on the basis of the TELE interview. The TELE is a sensitive and specific instrument differentiating patients with Alzheimer disease from healthy controls.23,24 In addition to the cognitive domain questions in the TELE interview, the subjects were asked if they had noticed any changes in their memory or had attended medical examinations concerning memory problems. Questions were also asked about independent living, assistance with practical tasks in everyday life, general health, current diseases, and medications. These additional questions were asked after the cognitive domain questions, and thus the information obtained from them did not reveal the subjects’ possible cognitive difficulties before the screening test. For subjects with dementia, the cutoff score of <16 was established for the TELE (maximum score being 20 points; low scores indicating dementia and high scores normal memory). On the basis of our previous study,23 this cutoff score yields a sensitivity of 77% and a specificity of 100%, which ensured that the false-positive rate was reduced in the group of twins with dementia. For the healthy subjects, the cutoff score of >17.5 was established. This score yields a sensitivity of 97% and a specificity of 69%, ensuring that the false-negative rate was reduced in the group of the healthy subjects. The category of milder cognitive decline was established as TELE scores of 16.0 to 17.5.
The TELE interview was completed for 594 subjects (response rate 72%). Among those who were not interviewed, 121 declined to participate in the interview, 48 were not reached by phone, 32 were ill, 20 died before being contacted, and 11 were not contacted (eg, the cotwin who was first contacted had declined to participate). Those who participated were more often men (78% participation) than women (67%), but after adjustment for sex in a logistic regression model, the participants and nonparticipants did not differ in age, education, or reported alcohol consumption. Of the 594 subjects who completed the TELE interview, 554 had complete data on alcohol consumption in the 1975 questionnaire and 503 had complete alcohol data in the 1981 questionnaire. The mean age (± standard deviation [SD]) of the subjects at the study entry was 49.4 (± 5.9) years.
Among the 594 twins with TELE scores, there were 276 pairs with scores on both twins, whereas in 42 pairs, only 1 twin had TELE scores. The mean TELE score of 16.1 (95% confidence interval [CI] 15.1–17.0) for this latter group was lower than the mean of 17.4 (17.1–17.7) for twins from full pairs, indicating possible selection of the participants based on degree of dementia.
Finally, we examined the history of hospitalizations for alcohol-related diseases and disorders based on record-linkage of the twin cohort to the national hospital discharge register, a nationwide register of all public and private hospital discharge records in Finland from 1972 to 1995. The recorded diagnoses have been shown to have a very good accuracy.25 The International Classification of Diseases, 8th Revision (ICD-8) and ICD-9 rubrics for alcoholism, alcohol intoxication, alcohol cirrhosis, and alcohol pancreatitis were used to identify those subjects with a major alcohol-related hospitalization during this period.
For all reported analyses, subjects were treated as individuals. However, because the twins had been sampled as pairs, the individual observations may not be fully independent. This clustering within pairs was adjusted for in all analyses to obtain correct confidence intervals using the cluster option in Stata (version 7; Stata Corp., College Station, TX) to obtain robust variance estimates. Log-linear models were used to estimate simultaneously the odds ratios (ORs) of dementia and milder cognitive decline as outcomes, compared with normal cognitive function, in relation to the predictors assessed 25 years earlier.
At the end of the 25 years of follow-up, 103 participants were classified as having dementia. Table 1 shows the characteristics of the study sample. The mean age of the study sample at the time of the interview was 74.6 (± SD 5.97) years, and 55% were women.
Because age, sex, and educational level were associated with dementia risk, the risk estimates are adjusted for these variables. Table 2 shows that a heavy drinker had an OR of 2.4 (0.7–8.7) for dementia and 2.5 (0.8–8.1) for milder cognitive decline. When modeled together with binge drinking, the point estimates for risks of dementia decreased to less than 1 for moderate and heavy drinkers, although with wide confidence intervals. The 1975 questionnaire assessed the frequency of consumption of specific types of alcohol beverages (beer, wine, spirits). There was no evidence for an increased risk of dementia with increased frequency of beer, wine, or spirit intake based on trend tests (data not shown).
Binge drinking in 1975 was associated with an OR of 3.9 for dementia (1.6–9.5) and 2.4 for milder cognitive decline (1.1–5.2) (Table 2). Binge drinking in 1975 was associated with an increased dementia risk even when alcohol consumption was in the same model (OR = 5.6; CI = 2.0–16). If those who were heavy drinkers in 1975 were excluded from the analyses, binge drinking in 1975 was associated with a higher risk of dementia (3.7; 1.3–10) and milder cognitive decline (3.0; 1.3–7.0).
Table 3 shows that if the person reported binge drinking in both 1975 and 1981 questionnaires, the risks of dementia (4.2; 1.2–14) and of milder cognitive decline (2.4; 0.8–7.4) were similar to those found when considering binge drinking only in 1975 (Table 2). Passing out as a result of alcohol use in 1981 was also associated with increased risk of dementia. Passing out twice or more during the previous year increased the OR to 12 (3.3–43) (Table 3). When alcohol consumption was in the same model (data not shown), the OR of dementia for passing out twice or more was 21 (4–110). After excluding those who were heavy drinkers in 1981 (n = 30), the OR of dementia associated with passing out at least twice during the previous year was 15 (2.4–93).
Smoking in 1975 was associated with an OR of 2.8 (1.3–6) for dementia and 1.5 (0.8–3) for milder cognitive decline, with no evidence for moderation of risk by alcohol use, binge drinking, or passing out (data not shown).
Finally, to examine the possibility that the increased dementia risk resulting from binge drinking patterns could be accounted for by those with alcoholism, we examined hospital records of major alcohol-related disorders. We found 7 of 554 persons with such alcohol-related hospital treatment periods. The increased dementia risk resulting from binge drinking and passouts prevailed when these 7 persons were either excluded or modeled as dichotomous variables (data not shown).
Our results show that midlife reports of binge drinking at least once per month or passing out as a result of excessive alcohol intake at least twice during the previous year were risk factors for dementia later in life. Binge drinking and passing out seemed to be risk factors even after controlling for alcohol consumption or after excluding from the analyses those who were heavy drinkers. These results are consistent for both measures of excessive drinking and at 2 separate time points.
The follow-up period of 25 years in this study was considerably longer than that in most earlier studies.2,4–6,8,9,11–16 Because most of the subjects in this study were middle-aged at the time of the first data collection in 1975, we presumed that they were cognitively intact at the study entry. However, if there were any persons with dementia who had replied to the 1975 questionnaire, it is unlikely that they would have survived 25 years to the time of the dementia assessment. Thus, it is unlikely that the persons who provided the dementia assessment had misreported their drinking habits 25 years earlier as a result of memory problems. Likewise, memory problems at baseline would have been unlikely to affect alcohol use. A relatively large proportion (24% of the study population) was abstainers. Life-abstention was much more common in women earlier in Finland than at present, reflecting the social and cultural norms that prevailed earlier last century. For example, in the age group of 50–69-year-old women, 3 independent surveys in Finland indicate that 56%, 39%, and 49% of these women were abstainers in the years 1968, 1976, and 1984, respectively.26 Thus, the abstainers in our study are less likely to reflect persons who do abstain for health reasons or who do not belong to mainstream society.
The effect of various alcohol consumption patterns on cognition has rarely been reported. However, Rehm et al18 found that light-to-moderate drinkers with occasional binge drinking have a higher mortality than those light-to-moderate drinkers without such occasions. In another study,27 the risk of death from cardiovascular disease was increased in frequent heavy or binge drinkers. In the present study, we had data not only on total alcohol consumption, but also on patterns of alcohol consumption as reported at 2 different occasions 6 years apart. We were thus able to demonstrate an association of occasional excessive alcohol consumption in light-to-moderate drinkers with an increased risk of dementia. One mechanism for this pattern might be acute alcohol intoxication, which has been shown to interfere with memory and learning processes in animal studies.28,29 However, excluding the few subjects who had been hospitalized for alcohol intoxication had no effect on the results. Another possible pathway is that people with binge drinking are at high risk of head injuries. Such injuries early in adulthood have shown to be a risk factor for dementia.30
The dementia diagnosis in the present study was based on scores received from a telephone interview assessing cognitive status.21 Telephone screening for mental status has some limitations; for example, elderly persons may have impaired hearing. Examiners inquired about hearing problems at the beginning of the telephone interview and made sure that the questions were heard. For those with severe hearing problems, the interview was discontinued. Some cognitive items such as reading, praxis, and visuospatial tasks are difficult to assess by telephone. It is impossible to ensure that persons with dementia do not use newspapers or calendars as sources of orientation during the interview. In addition, external distractions are not easily controlled by the examiner. However, the TELE interview has shown a strong correlation with mental state assessment administered face-to-face.21,23,24 Symptoms of depression, impaired hearing, and low education do not materially affect the ability of TELE to find individuals with dementia in a population.24
With the TELE method, it was impossible to differentiate the subtypes of dementia. However, in another study of alcohol consumption and risk of dementia,9 generally similar relationships of alcohol use with Alzheimer disease and vascular dementia were found. Because alcoholism is a possible cause of dementia as a result of thiamine malnutrition or alcoholic neurotoxicity,31 some subjects with alcoholic dementia might have been included in the dementia group. However, retrospective autopsy studies have shown the prevalence of Wernicke-Korsakoff syndrome (a cause of dementia associated with excessive alcohol consumption) to be low in the general population (0.4–2.8%).32,33 In 1 population-based study, only 6% of all patients with dementia were classified as having alcoholic dementia.34 In addition, an earlier study in a sample that was selected to represent the Finnish population (so-called Mini-Finland survey)35 showed primary degenerative dementia constitutes half of all dementia cases, vascular and combined dementia 39%, and secondary dementia 11% (with severe dementia resulting from chronic alcoholism 2.8%). It is unlikely that the most severe alcoholics participated in our telephone screening as a result of the high mortality and disability among them. We conclude that binge drinking in midlife substantially increases the risk of dementia and cognitive decline later in life.
We thank the nurses of the Department of Public Health, University of Turku, for carrying out the telephone interviews, and Kauko Heikkilä, for database management at the Department of Public Health, University of Helsinki.
1. DeCarli C, Miller BL, Swan GE, et al. Cerebrovascular and brain morphologic correlates of mild cognitive impairment in the National Heart, Lung, and Blood Institute Twin Study. Arch Neurol
2. Dufouil C, Tzourio C, Brayne C, et al. Influence of apolipoprotein E genotype on the risk of cognitive deterioration in moderate drinkers and smokers. Epidemiology
3. Elias PK, Elias MF, D'Agostino RB, et al. Alcohol consumption and cognitive performance in the Framingham Heart Study. Am J Epidemiol
4. Ruitenberg A, van Swieten JC, Witteman JC, et al. Alcohol consumption and risk of dementia: the Rotterdam Study. Lancet
5. Leibovici D, Ritchie K, Ledesert B, et al. The effects of wine and tobacco consumption on cognitive performance in the elderly: a longitudinal study of relative risk. Int J Epidemiol
6. Orgogozo JM, Dartigues JF, Lafont S, et al. Wine consumption and dementia in the elderly: a prospective community study in the Bordeaux area. Rev Neurol (Paris)
7. Romanov K, Rose RJ, Kaprio J, et al. Self-reported alcohol use: a longitudinal study of 12,994 adults. Alcohol Alcohol.
8. Lindsay J, Laurin D, Verreault R, et al. Risk factors for Alzheimer's disease: a prospective analysis from the Canadian Study of Health and Aging. Am J Epidemiol
9. Mukamal KJ, Kuller LH, Fitzpatrick AL, et al. Prospective study of alcohol consumption and risk of dementia in older adults. JAMA
10. Leroi I, Sheppard JM, Lyketsos CG. Cognitive function after 11.5 years of alcohol use: relation to alcohol use. Am J Epidemiol
11. 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
12. Colsher PL, Wallace RB. Elderly men with histories of heavy drinking: correlates and consequences. J Stud Alcohol
13. Cervilla JA, Prince M, Mann A. Smoking, drinking, and incident cognitive impairment: a cohort community based study included in the Gospel Oak project. J Neurol Neurosurg Psychiatry
14. Tyas SL, Manfreda J, Strain LA, et al. Risk factors for Alzheimer's disease: a population-based, longitudinal study in Manitoba, Canada. Int J Epidemiol
15. Hébert R, Lindsay J, Verreault R, et al. Vascular dementia: incidence and risk factors in the Canadian study of health and aging. Stroke
16. Launer LJ, Feskens EJ, Kalmijn S, et al. Smoking, drinking, and thinking. The Zutphen Elderly Study. Am J Epidemiol
17. Truelsen T, Thudium D, Gronbaek M. Amount and type of alcohol and risk of dementia: the Copenhagen City Heart Study. Neurology
18. Rehm J, Greenfield TK, Rogers JD. Average volume of alcohol consumption, patterns of drinking, and all-cause mortality: results from the US National Alcohol Survey. Am J Epidemiol
19. Kaprio J, Sarna S, Koskenvuo M, et al. The Finnish Twin Registry: formation and compilation, questionnaire study, zygosity determination procedures, and research program. Prog Clin Biol Res
20. Sarna S, Kaprio J, Sistonen P, et al. Diagnosis of twin zygosity by mailed questionnaire. Hum Hered
21. Gatz M, Reynolds C, Nikolic J, et al. An empirical test of telephone screening to identify potential dementia cases. Int Psychogeriatr
22. Kaprio J, Koskenvuo M. A prospective study of psychological and socioeconomic characteristics, health behaviour and morbidity in cigarette smokers prior to quitting compared to persistent smokers and non-smokers. J Clin Epidemiol
23. Järvenpää T, Rinne JO, Räihä I, et al. Characteristics of two telephone screens for cognitive impairment. Dement Geriatr Cogn Disord
24. Gatz M, Reynolds CA, John R, et al. Telephone screening to identify potential dementia cases in a population-based sample of older adults. Int Psychogeriatr
25. Keskimäki I, Aro S. Accuracy of data on diagnoses, procedures and accidents in the Finnish Hospital Discharge Register. Int J Health Sci
26. Simpura J, ed. Finnish Drinking Habits. Results From Interview Surveys Held in 1968, 1976 and 1984
. Helsinki: The Finnish Foundation for Alcohol Studies; 1987.
27. Malyutina S, Bobak M, Kurilovitch S, et al. Relation between heavy and binge drinking and all-cause and cardiovascular mortality in Novosibirsk, Russia: a prospective cohort study. Lancet
28. Ryabinin AE. Role of hippocampus in alcohol-induced memory impairment: implications from behavioral and immediate early gene studies. Psychopharmacology (Berl)
29. Dahchour A, De Witte P. Ethanol and amino acids in the central nervous system: assessment of the pharmacological actions of acamprosate. Prog Neurobiol
30. Plassman BL, Havlik RJ, Steffens DC, et al. Documented head injury in early adulthood and risk of Alzheimer's disease and other dementias. Neurology
31. Joyce EM. Aetiology of alcoholic brain damage: alcoholic neurotoxicity or thiamine malnutrition? Br Med Bull
32. Torvik A. Wernicke's encephalopathy—prevalence and clinical spectrum. Alcohol Alcohol
33. Harper C, Fornes P, Duyckaerts C, et al. An international perspective on the prevalence of the Wernicke-Korsakoff syndrome. Metab Brain Dis
34. Fratiglioni L, Grut M, Forsell Y, et al. Prevalence of Alzheimer's disease and other dementias in an elderly urban population: relationship with age, sex, and education. Neurology
35. Sulkava R, Wikstrom J, Aromaa A, et al. Prevalence of severe dementia in Finland. Neurology