Alcohol and Alcoholism Advance Access originally published online on May 9, 2007
Alcohol and Alcoholism 2007 42(5):442-447; doi:10.1093/alcalc/agm033
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Amount of alcohol consumption and risk of developing alcoholism in men and women
1 Institute of Preventive Medicine, Copenhagen University Hospital, Centre for Health and Society, Denmark
2 Centre of Alcohol Research, National Institute of Public health, Copenhagen, Denmark
3 Department of Psychiatry, Division of Neuroscience, Columbia University Medical Center, New York
4 Department of Health Psychology, Institute of Public Health, University of Copenhagen, Denmark
5 Alcohol Unit, Hvidovre Hospital, University of Copenhagen, Denmark
* Author to whom correspondence should be addressed at: Centre of Alcohol Research National Institute of Public health, Øster Farimagsgade 5 A, 2, 1399 København K, Denmark. Tel: (+45) 39 20 77 77; Fax: (+45) 39 20 80 10; E-mail: mg{at}niph.dk
Received 22 January 2007; first review notified 18 February 2007; in revised form 12 March 2007; accepted 20 March 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Aims: It is generally accepted, but not yet documented that the risk of future alcoholism increases with the amount of alcohol consumed. The objective of this study was to investigate this association using the Copenhagen City Heart Study. Methods: Quantity and frequency of alcohol intake was measured in 19 698 men and women randomly drawn from the Copenhagen Population Register in 1976–78. The study population was linked to three different registers in order to detect alcoholism, and average follow-up time was 25 years. Results: After adjustment for all putative confounders, the risk of alcoholism for women increased significantly at 1–7 drinks per week with a hazard ratio (HR) of 2.02 (95% confidence interval (CI): 1.16, 3.53) compared to never/almost never drinking; the HR for drinking monthly was 1.75 (95% CI: 1.08, 2.85). The risk for men did not increase significantly before 22–41 drinks per week (HR = 3.81, 95 % CI: 2.18, 6.68) or if they had a daily alcohol intake (HR = 3.55, 95 % CI: 2.11, 5.99). Smoking was independently associated with the risk of alcoholism for both men and women. Conclusion: The risk of developing alcoholism increased significantly by very low intakes of alcohol in women, while the risk is only increased significantly in men consuming more than 21 drinks per week.
|
Introduction |
|---|
|
TOPABSTRACTIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Alcohol is one of the leading risk factors and is responsible for 4% of the global disease burden (World Health Organization, 2002
). Moreover, harmful drinking is associated with increased risk of developing alcoholism, often leading to social, financial and interpersonal losses, stigmatization and social marginalization (Poznyak et al., 2005). Several studies have demonstrated that both genetic and environmental factors are involved in the development of alcoholism, but the complex interactions between genotype and environment have made it difficult to identify individual determinants of alcoholism. In 2002 a meta-analysis was performed including 50 family—twin—and adoption-studies in which problem drinking and alcohol dependence served as the primary outcome measures. The results indicated an upper limit of 20–26% for the heritability of alcohol misuse, which is a fairly lower than the rate often cited in the literature (Walters, 2002). Individual characteristics such as devotion and conservatism (Kendler et al., 1997), have been shown to be inversely associated with alcoholism while factors such as sensation seeking behaviour (Kampov-Polevoy et al., 2004; Lejoyeux and Marinescu, 2006) seasonal affective disorder (Sher, 2002), the personality trait of novelty seeking (Grucza et al., 2006) and low sense of coherence (Badura et al., 2000) is positively related to alcoholism. In addition, low education and low social status have been associated with alcoholism in cross-sectional studies (Fukuda et al., 2005; Poznyak et al., 2005; Subramanian et al., 2005) and longitudinal studies (Wray et al., 2005). While large amounts of coffee (Stevenson and Masters, 2005) has been associated with alcohol problems in one cross-sectional study, and longitudinal studies have shown prospective associations between heavy drinking and smoking (Jensen et al., 2003) as well as beverage preference (Jensen et al., 2002), surprisingly few studies have investigated associations between amount of alcohol consumption and risk of developing alcoholism. Although exposure of alcohol is a necessary link in the complex causal network leading to alcoholism, the empirical relationship between amount of consumption and the risk of developing alcoholism has not been settled. In the present study we prospectively analyzed the relationship between amount of alcohol intake and risk of developing alcoholism in a dataset that enabled us to take sex, educational level, income, smoking, physical exercise, housing status, and marital status, into account. |
Materials and Methods |
|---|
|
TOPABSTRACTIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study population
Subjects from the first examination of the Copenhagen City Heart Study (CCHS-I) were used. The CCHS is an ongoing study initially comprising 19 698 men and women over 20 years of age examined in 1976–78. Using the unique Danish personal identification number, the sample was drawn from a population of approximately 90 000 inhabitants living within ten wards surrounding Rigshospitalet, the National University Hospital of Copenhagen. The sample was selected randomly within age and sex strata, and invited by letter to participate. Detailed descriptions of the study have been published elsewhere (Appleyard et al., 1989
; Hein et al., 1993) Of the 19 698 invited persons there were 371 persons who died before the examination, and of the remaining 19 327 persons, there were 14 223 (73.6%) responders.
Alcoholism
The 19 698 persons invited to CCHS-I were linked to three different registers in order to determine alcoholism: the Danish Psychiatric Central Research Register, the Danish Hospital Discharge Register, and the Winalco-database. The Danish Psychiatric Central Research Register (Munk-Jorgensen and Mortensen, 1997) contains records of all individuals that have been admitted to a psychiatric hospital in Denmark since 1969; the Danish Hospital Discharge Register (Jurgensen et al., 1986) contains records of all individuals that have been admitted to a Danish hospital since 1976; and the Winalco-database (Becker, 2004) contains records of all individuals who have been treated for alcohol problems in the Alcohol Unit, Hvidovre Hospital—an outpatient clinic for alcoholics covering the greater Copenhagen and Frederiksberg municipalities since 1954. Individuals who were given an International Classification of Diseases (ICD) – diagnosis of alcohol abuse in either the Danish Psychiatric Central Research Register or Danish Hospital Discharge Register and individuals who were registered in the Winalco-database were considered to be alcoholic at the given time. The following ICD-8 and ICD-10 diagnoses, including secondary-diagnoses, were in this study used to define alcoholism:
ICD-8: 291.09; 291.19; 291.29; 291.39; 291.99; 303.09; 303.19; 303.20; 303.28; 303.29; 303.90; 303.91; and 303.99.
ICD-10: F10.1; F10.2; F10.3; F10.4; F10.5; F10.6; F10.7; F10.8; F10.9.
In order to test whether a more narrow definition of alcoholism would change our results, we also carried out the analyses using only two diagnoses: ICD-8; 303.20 (chronic alcoholism) and ICD-10; F102 (alcohol dependence).
Alcohol intake
Subjects were asked whether they ‘hardly ever/never’, ‘monthly’, ‘weekly’ or ‘daily’ drank alcohol, and if this intake was daily the average daily intake was recorded. Thus, an absolute amount of consumed alcohol was obtainable only for persons stating a daily alcohol intake. However, the weekly intake in CCHS-I was calculated on the basis of CCHS-II that contains additional information of the average weekly intake of each beverage type. The weekly amount of consumed alcohol in CCHS-I was obtained by means of a series of regression models estimated from CCHS-II, previously constructed byBecker et al. (Becker et al., 1995) that includes the explanatory variables age, sex, alcohol intake patterns and the daily alcohol intake. The average weekly intake of beer, wine and spirits was summed to the total alcohol intake (with one bottle of beer being approximately equivalent to the alcohol contents of one glass of wine or one glass of spirits—assuming each drink to contain 12 g of alcohol). The quantitative alcohol intake was divided into the following groups: <1, 1–7, 8–14, 15–21, 22–41, and >41 drinks per week, and the frequency of drinking was measured as hardly ever/never, monthly, weekly or daily.
Covariates
Subjects filled out a self-administered questionnaire containing questions about lifestyle and general health. The following variables were assumed to be possible confounders: Education (less than 8 years, 8–12 years, and more than 12 years); Income (monthly income in 1976–78: <4000, 4000 to 10 000, and >10 000 Danish crowns, which is approximately equivalent to <666, 666 to 1667, and >1667 US$ for exchange rates in 1977); Smoking (never smoker, previous smoker, and current smoker); Physical activity in leisure time (almost completely physically passive or light physical activity <2 h per week, light physical activity 2 to 4 h per week, exhausting physical activity >4 h per week, or regular hard training >4 h per week); Marital status (currently married, or not currently married).
Statistical analyses
The purpose of the analyses was to estimate the hazard ratios (HR) of developing alcoholism by considering the amount and frequency of alcohol consumed, while taking potential confounders into account. Data were analyzed by means of multiple Cox Regression analysis and delayed entry was implemented. To ensure maximal adjustment for confounding by age we used age as the time scale. Subjects were followed from their date of entry, when they received the questionnaire between 1976–78, to the date of their first alcoholic diagnosis, death, disappearance, or emigration or until the end of follow-up (January 2002)—whichever occurred first. Although the Winalco-database was updated until April 2005, the end of follow-up was chosen to be the date where the first register (the Danish Psychiatric Central Register) ended its update. This was done in order to avoid misclassification of alcoholism in the last years of follow-up. Individuals that were registered as alcoholics before 1976–78 were eliminated from the analyses, and in addition to the results shown in the upcoming tables, all analyses were repeated using a time-window of 3 years. Using this time-window we eliminated individuals registered as alcoholics before—and 3 years after they received the questionnaire. The method of ‘complete-subject analysis’ was used—hence only individuals with values recorded for all covariates in the given analyses were retained. All analyses were stratified according to gender.
In order to investigate the effect of each possible confounder, HRs for developing alcoholism were computed separately for each (Table 2). Secondly, HRs were computed for the quantitative weekly alcohol intake and the frequency of alcohol intake respectively, adjusted for: (i) no covariates, (ii) smoking, (iii) confounders that were significant in a final model built on backwards elimination, and (iv) adjusted for all covariates (Table 3 and 4). All statistical analyses were done by using the statistical software package SAS 9.1.
|
Results |
|---|
|
TOPABSTRACTIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Of the 19 698 individuals originally invited to participate in CCHS-I, 1566 persons (7.95%) were registered with alcohol problems at least one time in their life. Of the 14 223 individuals who answered the questionnaire, information of the amount of alcohol consumed each week was available for 14 123 participants—of these, 879 persons (6.22%) were registered as alcoholics while 11 074 persons died or emigrated during the follow-up period.
The proportion of men and the proportion of smokers increased with higher alcohol consumption. The proportion of participants who consumed more than 41 drinks per week was relatively young, did little exercise, was frequently separated or divorced, and was more frequently registered as alcoholics than participants in the other consumption categories (Table 1).
|
Investigating the HRs for each confounder showed that for women, being a current smoker and being under 40 years of age increased the risk of developing alcoholism. Low income, being a current smoker, living alone, being unmarried, and being under 40 years were risk factors for men (Table 2).
|
The risk of alcoholism varied according to the quantitative weekly alcohol-intake (Table 3), and the limit where the risk increased significantly was rather different for men and women respectively. For women, there was a strong dose-dependent increase in risk of alcoholism with increased alcohol intake, hence the crude HR was 1.91 when drinking 1–7 drinks per week (95% confidence interval (CI): 1.21, 3.02), 3.26 when drinking 8–14 drinks per week (95% CI: 1.98–5.37), and 39.84 (95% CI: 20.92–75.89) when drinking more than 41 drinks per week. For men, however, the risk of developing alcoholism only increased with consumption of more than 21 drinks per week. The HR for drinking 22–41 drinks per week was 3.99 (95% CI: 2.28–6.97) and the HR for >41 drinks per week was 8.22 (95% CI: 4.69–14.39). After adjustment for smoking, the HRs diminished slightly, but the significances of the unadjusted results were not altered. Using backwards elimination, only smoking was a significant confounder for women, while smoking, physical exercise, and marital status were significant confounders for men. Neither adjusting for factors that were significant in the final model, nor adjusting for all possible confounders chosen in the present study changed the HRs considerably (Table 3).
|
Frequency of alcohol intake was also positively associated with the risk of alcoholism (Table 4). For women, the risk increased dose-dependently with higher frequency, hence the raw HR for drinking monthly was 1.75 (95% CI: 1.08–2.85) compared to never/almost never drinking, 2.66 (95% CI: 1.62–4.35) for drinking weekly, and 6.93 (95% CI: 4.32–11.11) for drinking daily. Men only increased their risk of developing alcoholism if they had a daily alcohol intake (HR = 3.55, 95% CI: 2.11–5.99). The HRs were not adjusted for the amount of weekly alcohol intake; however the risk did not alter considerably by controlling for other confounders, and using backwards elimination, it was the same confounders that were significant as it was in Table 3.
|
Inserting a time-window of 3 years did not change the results in neither Table 3 nor 4 notably. In both tables, the HR's for women decreased slightly while the HR's for men increased a little; however, the significances of the results were the same. Using a more narrow definition of alcoholism with only two diagnoses meant that there were fewer cases of alcoholism –31 for women and 96 for men. Hence, it was not possible to carry out Cox regression analyses for women. For men however, the significances of the results did not alter, although the HR's were smaller (data not shown).
|
Discussion |
|---|
|
TOPABSTRACTIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We have demonstrated that alcohol—not very surprisingly—is a strong predictor of developing alcoholism. In addition, however, our findings suggest different thresholds of harmful drinking in relation to alcoholism among men and women, since women increase their risk by much smaller amounts and frequencies than men. Smoking had an independent effect on the risk of alcoholism for both men and women. Physical exercise and marital status confounded the risk for men, while education, income, and housing status were not significant confounders.
The present study was based on a large study population sample, and because of the prospective design, selection and recall bias was minimized. The follow-up time of 26 years meant that we were able to discover the majority of outcomes of alcoholism that expectably will ever occur for the study population—although several persons were eliminated from the analyses due to diagnosis before 1976–78. Long duration of follow-up is usually considered to be a strength in prospective cohort studies, as the number of cases, and hence the statistical power, will increase. However, in the present study, the long follow-up time probably implied individual changes in exposure, as changes in alcohol intake have occurred during the follow-up period. Since the self-reported questionnaire was not validated by biochemical markers or interviews, the assessment of alcohol consumption may be biased. Self-reports of alcohol consumption and alcohol problems are generally believed to be biased towards an underestimation among the heaviest drinkers (Poikolainen, 1985), which would diminish the statistical power of the analyses for the groups with highest alcohol intake—this misclassification is possibly sex-specific (Gronbaek and Heitmann, 1996). Misclassification of outcome is also plausible, as the concept of alcoholism in this study was defined as having being admitted to a hospital with an alcohol related diagnosis or having attended an outpatient clinic for alcoholics. With these register data only the most severe cases of alcoholism can be expected to be detected, but nevertheless, we found a rather high percentage of alcoholism, 6.22%, among the respondents.
Alcoholics diagnosed before answering the questionnaire in 1976–78 were eliminated from the analyses. It is, however, likely that people were alcoholics for a period of time before they were registered, and consequently the presented analysis would include a sub-sample of individuals that are already alcoholic, but are not yet registered. To evaluate the size of this problem we inserted a time-window of 3 years, but analyses based on this reduced sample essentially showed the same results as presented in the tables, and consequently we assume that undiagnosed alcoholics did not seriously bias our results.
It is generally assumed that non-response is associated with increased alcohol consumption (Lahaut et al., 2002), and according to our results, non-responding was also associated with alcoholism—8% of the invited persons were defined as alcoholics while only 6% of the respondents were defined as alcoholics. Consequently, selection bias may have occurred, which may limit the generalizability of our results. However, there are no strong reasons to believe that the relation between alcohol intake and alcoholism is different among responders and non-responders and that non-responding has seriously affected the obtained HRs.
While other studies have mainly used the diagnosis, ‘alcohol dependence’ (Caetano et al., 1997; Caetano and Cunradi, 2002), the definition of alcoholism was based on information from three alternative registers. Using a more strict definition of alcoholism based on only two diagnoses, too few cases were observed among women to carry out Cox regression analyses. For men, however, we found that analyses based on the strict definition showed the same patterns of significant associations—only the HR's were smaller.
Age has previously proven to be an important predictor of developing alcoholism (Hingson et al., 2006a,b), and in our study, age also seemed to affect the risk, as respondents over 40 years of age had a decreased risk of developing alcoholism (Table 2). However, in order to retain a sufficient number of cases in each sub-category, the analyses were not stratified according to age. Nevertheless, age was the underlying time-scale in our regression analyses, and therefore the modifying role of age was controlled for in all HRs.
An important aspect of the present study is the fact that the average year of birth among the respondents was 1924—suggesting possibly important differences between this study population and younger generations. Especially among women, differences in drinking patterns and alcohol-culture may make it difficult to generalize our results to younger women. The fact that even very small amounts of weekly alcohol intake implied increased risk of developing alcoholism for women in our study sample, may reflect the fact that alcohol consumption was relatively rare in women, and that the categories of drinking women included sub samples of vulnerable women, exposed to several other risk-factors and perhaps being genetically at risk.
To our knowledge, the association between the amount and frequency of alcohol intake and risk of developing alcoholism has not been documented in other studies. However, studies have shown that overall alcohol intake is a good predictor of alcohol-related harm. A Finnish study showed, that the probability of alcohol-related consequences increases with the annual intake of alcohol (Makela and Simpura, 1985), and several studies have found associations between heavy drinking patterns and increased risks for alcohol-related consequences such as drunk driving, injuries, job problems and criminality (Cherpitel et al., 1995; Midanik et al., 1996; Greenfield, 1998; Greenfield and Rogers, 1999).
Several of the covariates included in the present study such as economic status, educational level, smoking, and beverage preference have been shown to be associated with alcoholism in other studies (Jensen et al., 2002; Sher, 2002; Jensen et al., 2003; Subramanian et al., 2004; Averina et al., 2005). In this study we used CCHS-I that contained information on various lifestyle related risk factors; but many other factors may modify the association between amount of alcohol consumption and risk of developing alcoholism. These factors include individual characteristics as well as a number of social factors described in the introduction, and we believe that future research would gain much by exploring how these factors affect the relation between alcohol and alcoholism.
In conclusion, we found that both the amount and frequency of alcohol intake was positively associated with later risk of developing alcoholism, and that the risk for women increased with very low levels of consumption while the risk for men only increased with consumption of more than 21 units per week.
Danish, and several other national drinking limit recommendations for alcohol intake, are 14 and 21 drinks for women and men respectively. Seen in the light of alcoholism, the present study confirms the fact that that these limits may be relevant for men. Women, however, appear to be very susceptible to alcohol consumption, as their risk for alcoholism increases significantly by much lower intakes than 14 drinks per week. We find it important that research is conducted to clarify whether this striking sex-difference can be demonstrated in younger generations.
|
References |
|---|
|
TOPABSTRACTIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Appleyard M., Hansen A. T., Schnohr P., et al. The Copenhagen City Heart Study. Osterbroundersogelsen. A book of tables with data from the first examination (1976–78) and a five year follow-up (1981–83). The Copenhagen City Heart Study Group. Scandinavian Journal of Social Medicine Supplementum (1989) 41:1–160.[Medline]
Averina M., Nilssen O., Brenn T., et al. Social and lifestyle determinants of depression, anxiety, sleeping disorders and self-evaluated quality of life in Russia—a population-based study in Arkhangelsk. Social Psychiatry and Psychiatric Epidemiology (2005) 40:511–518.[CrossRef][Web of Science][Medline]
Badura K., Gorczyca P., Tomalczyk E., et al. Estimation of a sense of coherence in patients with alcoholic dependence syndrome—introductory report. Wiadomosci lekarskie (2000) 53:488–492.[Medline]
Becker U. Documentation of Patient Admission in the Alcohol-unit Hvidovre Hospital, Winalco—A Clinical Database [Danish] (2004) Alkoholenheden: Hvidovre Hospital.
Becker U., Deis A., Sørensen T. I. A., et al. Alcohol intake in a population study Assessment and characterization. ALCOLOGIA (1995) 7:35–42.
Caetano R., Cunradi C. Alcohol dependence: a public health perspective. Addiction (2002) 97:633–645.[CrossRef][Web of Science][Medline]
Caetano R., Tam T., Greenfield T., et al. DSM-IV alcohol dependence and drinking in the U.S. population: a risk analysis. Annals of Epidemiology (1997) 7:542–549.[CrossRef][Web of Science][Medline]
Cherpitel C. J., Tam T., Midanik L., et al. Alcohol and non-fatal injury in the U.S. general population: a risk function analysis. Accident Analysis and Prevention (1995) 27:651–661.[CrossRef][Web of Science][Medline]
Fukuda Y., Nakamura K., Takano T. Accumulation of health risk behaviours is associated with lower socioeconomic status and women's urban residence: a multilevel analysis in Japan. BMC Public Health (2005) 5:53.[CrossRef][Medline]
Greenfield T. K. Evaluating competing models of alcohol-related harm. Alcoholism-Clinical and Experimental Research (1998) 22:52S–62S.[CrossRef]
Greenfield T. K., Rogers J. D. Alcoholic beverage choice, risk perception and self-reported drunk driving: effects of measurement on risk analysis. Addiction (1999) 94:1735–1743.[CrossRef][Web of Science][Medline]
Gronbaek M., Heitmann B. L. Validity of self-reported intakes of wine, beer and spirits in population studies. European Journal of Clinical Nutrition (1996) 50:487–490.[Web of Science][Medline]
Grucza R. A., Robert C. C., Bucholz K. K., et al. Novelty seeking as a moderator of familial risk for alcohol dependence. Alcoholism-Clinical and Experimental Research (2006) 30:1176–1183.[CrossRef]
Hein H. O., Sorensen H., Suadicani P., et al. Alcohol consumption, Lewis phenotypes, and risk of ischaemic heart disease. Lancet (1993) 341:392–396.[CrossRef][Web of Science][Medline]
Hingson R. W., Heeren T., Winter M. R. Age at drinking onset and alcohol dependence: age at onset, duration, and severity. Archives of Pediatrics & Adolescent Medicine (2006a) 160:739–746.
Hingson R. W., Heeren T., Winter M. R. Age of alcohol-dependence onset: associations with severity of dependence and seeking treatment. Pediatrics (2006b) 118:e755–e763.
Jensen M. K., Andersen A. T., Sorensen T. I., et al. Alcoholic beverage preference and risk of becoming a heavy drinker. Epidemiology (2002) 13:127–132.[CrossRef][Web of Science][Medline]
Jensen M. K., Sorensen T. I., Andersen A. T., et al. A prospective study of the association between smoking and later alcohol drinking in the general population. Addiction (2003) 98:355–363.[CrossRef][Web of Science][Medline]
Jurgensen H. J., Frolund C., Gustafsen J., et al. Registration of diagnoses in the Danish National Registry of Patients. Methods of Information in Medicine (1986) 25:158–164.[Web of Science][Medline]
Kampov-Polevoy A. B., Eick C., Boland G., et al. Sweet liking, novelty seeking, and gender predict alcoholic status. Alcoholism-Clinical and Experimental Research (2004) 28:1291–1298.[CrossRef]
Kendler K. S., Gardner C. O., Prescott C. A. Religion, psychopathology, and substance use and abuse; a multimeasure, genetic-epidemiologic study. American Journal of Psychiatry (1997) 154:322–329.[Abstract]
Lahaut V. M., Jansen H. A., van de M. D., et al. Non-response bias in a sample survey on alcohol consumption. Alcohol and Alcoholism (2002) 37:256–260.
Lejoyeux M., Marinescu M. Alcohol dependence and abuse and psychiatric disorders. Revue Du Praticien (2006) 56:1081–1085.[Web of Science][Medline]
Makela K., Simpura J. Experiences related to drinking as a function of annual alcohol intake and by sex and age. Drug and Alcohol Dependence (1985) 15:389–404.[CrossRef][Web of Science][Medline]
Midanik L. T., Tam T. W., Greenfield T. K., et al. Risk functions for alcohol-related problems in a 1988 US national sample. Addiction (1996) 91:1427–1437.[CrossRef][Web of Science][Medline]
Munk-Jorgensen P., Mortensen P. B. The danish psychiatric central register. Danish Medical Bulletin (1997) 44:82–84.[Web of Science][Medline]
Poikolainen K. Underestimation of recalled alcohol intake in relation to actual consumption. British Journal of Addiction (1985) 80:215–216.[CrossRef][Web of Science][Medline]
Poznyak V., Saraceno B., Obot I. S. Breaking the vicious circle of determinants and consequences of harmful alcohol use. Bulletin of the World Health Organization (2005) 83:803.[Web of Science][Medline]
Sher L. Relationships between seasonality and alcohol use: a genetic hypothesis. Medical Hypotheses (2002) 59:85–88.[CrossRef][Web of Science][Medline]
Stevenson J. S., Masters J. A. Predictors of alcohol misuse and abuse in older women. Journal of Nursing Scholarship (2005) 37:329–335.[CrossRef][Web of Science][Medline]
Subramanian S. V., Nandy S., Kelly M., et al. Patterns and distribution of tobacco consumption in India: cross sectional multilevel evidence from the 1998–9 national family health survey. British Medical Journal (2004) 328:801–806.
Subramanian S. V., Nandy S., Irving M., et al. Role of socioeconomic markers and state prohibition policy in predicting alcohol consumption among men and women in India: a multilevel statistical analysis. Bulletin of the World Health Organization (2005) 83:829–836.[Web of Science][Medline]
Walters G. D. The heritability of alcohol abuse and dependence: a meta-analysis of behavior genetic research. American Journal of Drug and Alcohol Abuse (2002) 28:557–584.[CrossRef][Web of Science][Medline]
World Health Organization. The World Health Report 2002 - Reducing Risks, Promoting Healthy Life (2002) Geneva: World Health Organization. 65–67.
Wray L. A., Alwin D. F., McCammon R. J. Social status and risky health behaviors: results from the health and retirement study. Journals of Gerontology Series B-Psychological Sciences and Social Sciences (2005) 60:85–92. Spec No 2.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||