Heavy alcohol consumption and associated factors: a population-based study



Juvenal S Dias da CostaI; Mariângela F SilveiraI; Fernando K GazalleI; Sandro S OliveiraI; Pedro C HallalI; Ana Maria B MenezesI; Denise P GiganteI; Maria T A OlintoIII; Silvia MacedoI

IFaculdade de Medicina, Universidade Federal de Pelotas. Pelotas, RS, Brasil
IIEscola de Medicina, Universidade Católica de Pelotas. Pelotas, RS, Brasil
IIICentro de Ciências da Saúde, Universidade do Vale do Rio dos Sinos. São Leopoldo, RS, Brasil





OBJECTIVE: To determine the prevalence of heavy alcohol consumption and factors associated with it in a Brazilian adult population.
METHODS: Cross-sectional population-based study including 2,177 adults (aged 20 to 69), living in the urban area of the municipality of Pelotas, Rio Grande do Sul State, Brazil. The sample was selected in multiple stages. Heavy alcohol consumption was defined as above 30g/day. The adjusted analysis was conducted by logistic regression.
RESULTS: The prevalence of heavy alcohol consumption was 14.3% (29.2% among men and 3.7% among women). The following groups presented higher prevalences of heavy alcohol consumption after adjusted analysis: men, elderly people, blacks or mulattoes, heavy smokers, and people who present some kind of chronic disease. Men with minor psychiatric disorders showed higher prevalences of heavy alcohol consumption than other men. Among women, association between age and heavy alcohol consumption was inversely related. Furthermore, the study indicates that among hypertensive subjects, those with heavy alcohol consumption presented worse disease management.
CONCLUSIONS: Heavy alcohol consumption is high and results in countless negative consequences for the individual's health and quality of life. Our results highlight the high prevalence of heavy alcohol consumption and indicate subsections of the whole population more susceptible to alcoholism.

Keywords: Alcoholism. Prevalence. Cross-sectional studies. Risk factors. Socioeconomic factors. Social problems.




Moderate consumrtion of alcoholic beverages, according to the literature, is a protective factor for mortality due to all causes, particularly due to its effect in reducing the frequency of cardiovascular diseases.4,9 However, heavy alcohol consumption results in innumerous negative consequences for the individual's health and quality of life, increasing the rate of morbidities that result in death or functional limitations such as cirrhosis, some types of cancer, cerebral vascular accidents, violence, mental diseases, among others.10,14,16 The World Health Organization (WHO) indicates that mortality and functional limitations resulting from heavy alcohol consumption are even greater than those caused by cigarette smoking.16

Alcoholism is, of itself, also considered an illness and a multicentric Brazilian study1 indicates prevalences ranging from 7.6% to 9.2%. Another study undertaken in Brazil5 found a prevalence of alcoholism of 12.4% among hospital patients. A study undertaken in South Korea7 indicates a prevalence of heavy alcohol consumption of 16% among men and 2% among women, whereas a study undertaken in New Zealand6 indicated a prevalence of 9.9% among elderly subjects.

Heavy alcohol consumption also entails in higher costs for the Health Care System, both directly and indirectly, because the morbidities which ensue from consumption are expensive and difficult to manage. Furthermore, alcoholism also provokes family disturbances with greater frequency.

The objective of this study was to determine the prevalence and the factors associated with heavy alcohol consumption in an adult population in southern Brazil. Detecting the population groups more exposed to heavy alcohol consumption makes it possible to plan more efficacious health policies designed to decrease this morbidity and other diseases unleashed by alcoholism.



A cross-sectional, population-based study was undertaken during the period between the 3rd of December, 1999 and the 3rd of April, 2000, involving several aspects related to the health of the adult population, aged 20 to 69, living in the urban area of Pelotas, Rio Grande do Sul State, Brazil.

The sample was selected in multiple stages. The primary sample units were the 281 census tracts of the urban area of the municipality, 40 of which were selected according to a systematic random sampling technique. Within the selected tracts, systematic sampling of households was undertaken. All inhabitants of selected households within the age range of interest to this study were interviewed.

Calculation of the sample size estimated a 15% prevalence of heavy alcohol consumption, with a 95% confidence level, 80% testing power, exposures between 5 and 50%, 2.0 relative risk, as well as a 10% excess in the sample size to account for eventual losses and refusals and 15% for adjusted analysis. The estimated sample size resulting from the above mentioned calculations was composed of 1,595 individuals. Since this study was conducted in conjunction with others, the sample was increased in order to be able to account for other less common outcomes.

Alcohol consumption was measured by means of a questionnaire that evaluated the types, quantities and frequency of individual consumption of alcoholic beverages. The cut-off point for defining hazardous drinking was 30g/day or more of ethanol.9,11

The independent variables collected were: sex, age, skin color, (white or black/mulatto according to the interviewer's observation), formal education (measured in years of schooling in which the student passed on to the next grade) social level (as it is classified by the Associação Nacional de Empresas de Pesquisa – ANEP2 [National Association of Research Enterprises], conjugal status (single, living with a companion, widowed/separated), minor psychiatric disorders (Self Reported Questionnaire - SRQ-20, with a cut-off point of 6 for men and 7 for women),8 hypertension (the mean of two measurements equal to or greater than 160/95 or the consumption of anti-hypertensive medication), body mass index (normal: <25 kg/m2; overweight: 25 a 29,9 kg/m2; obesity: 30 kg/m2 or more),12 physical activity (insufficiently active: less than 1000kcal/week spent in leisure activities13), smoking (never smoked cigarettes; ex-cigarette smoker; light to moderate smoker – <20 cigarettes per day; heavy smoker – 20 cigarettes or more per day) and chronic disease (presence of at least one of the following diseases: self-referred diabetes, arterial hypertension, minor psychiatric disorders, chronic bronchitis or obesity).

The consequences of heavy consumption of alcohol on health care were indicated by analyzing adherence to treatment for systemic arterial hypertension (measured by the adequate control of blood pressure), rates of hospitalization and frequency with which the individual recurred to the health services (measured by the number of visits made to the services in the last year).

The interviewers were medical students from the Universidade Federal de Pelotas [The Federal University of Pelotas] and were submitted to a training program, including exercises on standardizing the procedures for measuring weight and height as well as participation in a pilot study, and were unaware of the objectives and hypothesis of the study.

Individuals who refused to answer the questionnaire or who weren't home at the time of the interview were visited two more times and at different periods of the day. Quality control was carried out by applying a simplified version of the questionnaire to 10% of the subjects included in the sample.

Data was codified and registered twice in a database using the Epi Info 6 Program and automatic verification of consistency as well as amplitude was undertaken. Stata 7.0 and SPSS 8.0 programs were used for analyses, which consisted of a description of the sample, crude analysis (of the entire sample and stratified by sex) and adjusted analysis. Adjusted analysis was undergone through non-conditional logistic regression, based on a conceptual model consisting of three hierarchical levels;15 in which the first level was composed of demographic variables (sex, age and skin color), the second level contained the socio-economic variables (social status – ANEP, formal education and civil status) and the third level consisted of variables related to morbidity and lifestyle (smoking, physical activity, minor psychiatric problems, BMI, systemic arterial hypertension and the presence of chronic diseases). The variables were maintained within the model when they attained a 20% level of significance and, if it reached 5%, the association was considered significant.

This project was approved by the Ethics and Research Committee of the Faculdade de Medicina of the Universidade Federal de Pelotas [School of Medicine of the Federal University of Pelotas] and was considered to be of minimum risk. Consequentially, only verbal consent was requested of people submitted to interviews.



One thousand nine hundred and sixty eight individuals of the 2,177 considered eligible for the study were interviewed, bringing the rate of losses and refusals to a total of 9.6%.

Twenty one percent of the study sample stated they never drank an alcoholic beverage; moderate consumption (up to 30 g/day of ethanol) was reported by 65.1% of the sample and the prevalence of heavy alcohol was 14.3% (95% CI 9.7%, 17.8%) being that these represented 3.7% of the women and 29.2% of the men.

The sample obtained was composed of 1,122 (57%) women, the mean age was 41.6 years, being that almost half of the individuals were less than 40 years old. The majority of the individuals (83%) were defined as being white. As to socioeconomic status, the A and E classes of the ANEP each contained approximately 5% of the individuals in the sample and the great majority of the individuals (37%) were concentrated in group C. A similar distribution was found with respect to years of schooling. The majority of the sample was composed of individuals who were married or lived in stable relationships (61%). Data related to co-morbidity and lifestyles are better described in Table 1.



The results of crude analysis for the occurrence of the outcome and of the several factors studied are also presented in Table 1. Significant statistical association was detected between heavy alcohol consumption and male sex, non-white skin color, physical activity and the presence of chronic disease, being that the greatest risks for alcohol abuse were found among the male sex (OR=9.53) and non-white skin color (OR=1.76). The variables formal education or schooling, socio-economic level and smoking presented a linear tendency to increase in prevalence according to the decrease in schooling and socioeconomic level and according to the increase in cigarette consumption. Tables 2 and 3 present the crude analysis stratified by sex, in which important differences may be noticed. As to the age groups, there was a tendency to increase according to the increase in age, however, this is a low intensity association. Among the women, sex stratification resulted in an inverse tendency. The younger age group among the women presented an increased association with abusive alcohol consumption. Among the men, the tendency remained the same as was observed when the sample was analyzed as a whole; however the magnitude of the effect was more readily demonstrated.





No association was found between heavy alcohol consumption and minor psychiatric disorders or hypertension when the entire sample was analyzed, however, when subjects of the male sex were analyzed separately, the presence of psychiatric disorders was directly associated with the outcome.

Some of the variables (socio-economic level and schooling) which presented statistical significance in the sample when analyzed as a whole, when stratified according to sex, particularly the female sex, lost their significance. The reverse holds true for civil status, wherein single women presented a higher risk of being heavy alcohol consumers.

Multivariate analysis yielded significant associations for the following variables: male sex, old age, skin color, low socio-economic level, the presence of a chronic disease and cigarette smoking (see Table 4). The above variables, with the exception of cigarette smoking and the presence of a chronic disease, had increased effects in this analysis when compared to the crude analysis.



An analysis of the relation between heavy alcohol consumption and the control of systemic arterial hypertension was undertaken and an important association between heavy alcohol consumption and poor control of hypertension was found both within the global analysis (OR=1.73, value p<0.001) and analysis stratified by sex. A lower number of medical consultations were also associated to the outcome. Hospitalization was also studied as a variable, but no significant associations were found.



A study with a cross-sectional design, despite its advantages in terms of rapid results and low costs, presents some limitations as to inferences with respect to causal effects, due to the fact that both exposures and outcomes are collected simultaneously. Particularly with respect to this study, the relations between heavy alcohol consumption and physical activity, body mass index, minor psychiatric disorders and cigarette smoking may have been affected by this bias.

Furthermore, the subject's account of alcohol consumption may have been affected by memory bias. However, if this bias occurred, it was probably conservative, that is, the frequency of heavy alcohol consumption may be higher than declared.

Alcohol consumption was evaluated, as in other analyses,6,7,9 by means of an account of the weekly consumption of alcoholic beverages according to frequency, quantity and types of beverages. However, it is known that quantification of alcoholism is a controversial issue, and that the strategy utilized may sub estimate the prevalence of alcohol abuse.14

Considering the methodological strategies undertaken in this investigation it is noteworthy that there was a low percentage of losses and refusals (less than 10%) and that there is a great similarity between the study sample and census data for the municipality (Instituto Brasileiro de Geografia e Estatistica, Censo Demográfico, 2000 [Brazilian Institute of Geography and Statistics, Demographic Census, 2000]).

The prevalence of heavy alcohol consumption (14.3%) was greater than indicated in other studies that utilized the same definition of the outcome, in which prevalences ranged from 7 to 10%.1,6 On the other hand, studies7,14 which utilized the DSM III or DSM IV criteria found higher prevalences (15 to 30%).

The greater frequency of heavy alcohol consumption among men is consistent with the literature.6 A Chinese study6 indicated 16% prevalence of alcoholism among men and 2% among women, while, in the present study, prevalences were, respectively, 27.2% and 3.8%. Another study among Brazilian hospital patients5 detected a 22% prevalence among men, whereas the prevalence among women was only 3%.

As to age, differences were also observed between the sexes. Heavy alcohol consumption was directly associated to aging among men, whereas the opposite tendency was observed among women. It is interesting to note the contrast between this and another Brazilian study9 in which a similar association between alcohol consumption and age was found among both men and women.

Blacks and mulattoes presented higher rates of heavy alcohol consumption, even after analysis was adjusted, indicating the existence of ethnic and/or cultural differences which cannot be explained by age and sex variables. Since skin color is one of the determinants of socio-economic status, its effect cannot be controlled by formal schooling or socio-economic level (mediating factors). Nonetheless, this result is consistent with that encountered in other studies undertaken in Brazil,3,7 which established controls for socio-economic indicators as well.

Alcoholism and socio-economic indicators were inversely associated, as occurs in the majority of the literature reviewed.3,7 However, the only significant difference was noted among the poorest individuals (class E). Although few years of schooling was associated with heavy alcohol consumption in the crude analysis, in the adjusted analysis this association was not significant.

Cigarette smokers presented greater prevalence of heavy alcohol consumption, particularly heavy smokers. Although there is no causal relation between these variables, this association is recurrent in the literature3. This result seems to indicate that alcoholics tend to be more careless about their health.

Finally, despite the limitations of the cross-sectional study, it was noted that people who presented some form of chronic disease consumed more alcohol excessively than other people. This is a cause of concern for it was found that these patients had worse control of systemic arterial hypertension. This finding may be a consequence of the deleterious effect of heavy alcohol consumption, as well as the lower degree of personal health care presented by these individuals, since it was also found that the number of visits to the health center was lower among this population.

Studies similar to this one should be undertaken periodically in order to monitor the tendencies of heavy alcohol consumption. However, ideally, they should be designed in such a way as to make stratified analyses according to sex viable, given that the findings suggest that there are different profiles of heavy alcohol consumption between men and women.

The results indicate that black men, individuals within lower social and economic levels, heavy cigarette smokers and people who present chronic diseases are the groups which present higher frequencies of heavy alcohol consumption of alcohol and who are therefore more susceptible to morbidity and mortality related to alcoholism. Health activities, geared towards these groups, but preferably focused on the entire population, should be planned in order to diminish heavy alcohol consumption and, consequently, its harmful effects on health.



1. Almeida Filho LM, Mari JJ, Coutinho E, França JF, Fernandes JG, Andreoli SB, et al. Estudo multicêntrico de morbidade psiquiátrica em áreas urbanas brasileiras (Brasília, São Paulo, Porto Alegre). Rev ABP-APAL 1992;14:38-42.        

2. Associação Nacional de Empresas de Pesquisa. Critério de classificação econômica Brasil. São Paulo; 1996.        

3. Chaieb JA, Castellarin C. Associação tabagismo x alcoolismo: introdução às grandes dependências humanas. Rev Saúde Pública 1998;32:246-54.        

4. Farchi G, Fidanza F, Giampaoli S, Mariotti S, Menotti A. Alcohol and survival in the Italian rural cohort of the Seven Countries Study. Int J Epidemiol 2000;29:667-71.        

5. Figlie NB, Pillon SC, Dunn J, Larangeira R. The frequency of smoking and problem drinking among general hospital inpatients in Brazil - using the AUDIT and Fagerstrom questionnaires. São Paulo Med J 2000;118:139-43.        

6. Khan N, Davis P, Wilkinson TJ, Sellman JD, Graham P. Drinking patterns among older people in the community: hidden from medical attention? N Z Med J 2002;115:72-5.        

7. Kim JM, Shin IS, Stewart R, Yoon JS. Alcoholism in older Korean men: prevalence, aetiology, and comorbidity with cognitive impairment and dementia in urban and rural communities. Int J Geriatr Psychiatry 2002;17:821-7.        

8. Mari JJ, Williams P. A validity study of a psychiatric screening questionnaire (SRQ 20) in primary care in the city of São Paulo. Br J Psychiatric 1986;148:23-6.        

9. Moreira LB, Fuchs FD, Moraes RS, Bredemeier M, Cardozo S, Fuchs SC, et al. Alcoholic beverage consumption and associated factors in Porto Alegre, a Southern Brazilian City: a population-based survey. J Studies Alcohol 1996;57:253-9.        

10. Nanchahal K, Ashton WD, Wood DA. Alcohol consumption, metabolic cardiovascular risk factores and hypertension in women. Int J Epidemiol 2000;29:57-64.        

11. National Institutes of Health. The sixth report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure. Washington (DC): NIH; 1997.        

12. Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, et al. Physical activity and public health. A reccomendation from the centers for disease control and prevention and the american college of sports medicine. JAMA 1995;273:402-7.        

13. Rehm J, Greenfield TK, Walsh G, Xie X, Robson L, Single E. Assessment methods for alcohol consumption, prevalence of high risk drinking and harm: a sensitivity analysis. Int J Epidemiol 1999;28:219-24.        

14. Victora CG, Huttly SR, Fuchs SC, Olinto MTA. The role of conceptual frameworks in epidemiological analysis: a hierarchical approach. Int J Epidemiol 1997;26:224-47.        

15. World Health Organization. Obesity: preventing and managing the global epidemic: report of a WHO consultation on obesity. Geneva; 1998. (WHO/NUT/NCD/98.1).        

16. World Health Organization. About Global Alcohol Database [on line]. 2002. Available on <URL: http://www3.who.int/whosis/alcohol/alcohol_about_us.cfm? path=whosis,alcohol,alcohol_about&language=english [2003 Set 25]        



Correspondence to
Juvenal Soares Dias da Costa
Departamento de Medicina Social – Faculdade de Medicina (UFPel)
Duque de Caxias, 250, 3º piso
96030-002 Pelotas, RS, Brasil
E-mail: jcosta@epidemio-ufpel.org.br

Received on 13/1/2003
Reviewed on 1/9/2003
Approved on 19/9/2003

Faculdade de Saúde Pública da Universidade de São Paulo São Paulo - SP - Brazil
E-mail: revsp@org.usp.br