Alcohol Use among Adolescents: a Population-based Study

OBJECTIVE: To estimate the prevalence and factors associated with alcohol use among adolescents. METHODS: Population-based, cross-sectional study performed from 2005 to 2006, in the city of Pelotas, Southern Brazil, with 1,056 adolescents aged between 11 and 15 years. An anonymous, self-applied questionnaire was used, based on the World Health Organization model for drug use. Poisson regression was employed to analyze data. RESULTS: Prevalence of adolescents who reported alcoholic beverage consumption in the last month was 23.0% (95% CI: 20.4;25.4), prevalence was 21.7% among females and 24.2% among males. Prevalence of alcohol consumption at 11 years of age was 11.9%. In the multiple regression analysis, alcoholic beverage use by male adolescents was higher among those who reported tobacco use in the last month, those who were older, and those who had already had sexual intercourse. Among female adolescents, age was the only variable associated with alcoholic beverage use. CONCLUSIONS: Alcoholic beverage use was prevalent in both sexes and began at a very early age. Smoking and previous sexual intercourse were also associated with this use. Preventive measures need to begin earlier, aiming to control alcohol use in the 11-to-15-year age group.


INTRODUCTION
Alcohol use early in life is one of the most relevant predictive factors of future health, socio-cultural and economic problems.Consumption before the age of 16 years signifi cantly increases the risk of excessive drinking in adulthood, in both sexes. 14olescents comprise the population group that has the greatest alcohol consumption problems.Studies show that even low consumption is associated with greater risk of accidents, a use of psychotropic drugs, 8,11,17,19,22 and highrisk behavior, including unprotected sex (without condom). 7,18 the long term, alcoholic beverage consumption can lead to suicide and chronic diseases, including mental disorders, cancer, 7 systemic arterial hypertension, 6 obesity, cerebrovascular accident, polyneuropathies, dementia, convulsions and digestive tract neoplasias. 18Laranjeira R, Pinski I, Zaleski M, Caetano R. I Levantamento nacional sobre os padrões de consumo de álcool na população brasileira.Brasília: Ministério da Saúde; 2007.
Studies with adolescents have identifi ed the following as factors associated with alcohol use: age, 2,22 male sex, 2,12,17,20,22 economic status, 11,19 studying in public schools, 22 not living with parents, family history of alcohol, entering the job market, not having a religion, 5 tobacco 16 and illicit drug use, 2 little support or lack of understanding of family, association with family behavior towards alcohol use and abuse 15,21 and depression. 3,16e present study aimed to estimate the prevalence of and factors associated with alcohol use among adolescents.The following parameters were used to calculate the sample size of the extensive research project: estimated prevalence of 62.3% 22 for alcohol use in the last month, 95% confi dence interval, and 80% statistical power, thus requiring 948 adolescents.A total of 10% for losses and 10% to control confounding factors were added, totaling 1,137 adolescents.However, innumerable outcomes were investigated and calculations were made, according to the appropriate prevalences and exposures.The one requiring the greatest size, i.e. depression in the beginning of adolescence, established the sample number of 1,145 adolescents aged between 11 and 15 years, living in the urban area, for the analysis of this study.

METHODS
Sampling was performed in multiple stages.A total of 6,794 households were estimated as necessary, based on the estimate that there were 3.4 residents per household and about 16% of residents in the age group assessed, in Pelotas urban area.Of all the 448 census tracts of the urban area, 79 were systematically selected.One block from each tract was randomly selected.In addition, one street corner was randomly selected to be the starting point, and 86 households were subsequently visited, following a predetermined order.Of all the 1,145 adolescents located, 89 (7.8%) were classifi ed as losses.Losses varied from adolescents who did not answer the question referring to the outcome to those few who were not found in their homes after three visits.
A total of 1,056 adolescents answered an anonymous self-applied questionnaire with 81 questions about health behavior.All adolescents in the household participated in the study.After being fi lled out, questionnaires were put into a sealed box.
The dependent variable was alcoholic beverage use in the last 30 days.
Associations with demographic (sex and age) and socioeconomic variables (economic status, level of education in completed years, and school failure), life habits and depression symptoms were tested.
Economic status categories were established from ownership of household assets, payment of home employees and head of household's level of education.a The following life habits were analyzed: presence of intimidation (victimization or bullying), physical activity practice, tobacco use in the last month, illicit drug use in the last month, participation in religious activities, sexual behavior (previous sexual intercourse) and presence of depressive symptoms.
The intimidation variable was constructed from the following set of questions: "Has anyone given you nicknames?","Have you been threatened by anyone?", "Has anyone stolen or damaged any of your possessions?","Have you been left out of games or groups?","Has anyone punched or kicked you?", "Has anyone spread gossips or rumors about you?", and "Have you been forced to do something you did not want do to?". 13 Children's Depression Inventory was used to characterize the presence of depressive symptoms. 10ude analysis was performed using the SPSS software (version 11.0), considering effect measures, with their respective 95% confi dence intervals and association tests. 1 Analysis adjusted by Poisson regression, according to the hierarchical model, 23 was performed with the Stata software (version 7.0).
The theoretical model presupposes hierarchy among levels in relation to the outcome, i.e. more distal variables determine alcohol consumption.In the present study, the model was divided into two levels.
On the fi rst level, the following variables were included: sex, age, economic status, level of education and school failure; on the second, the following: intimidation, physical activity, tobacco use, illicit drug use, religious activity, sexual behavior, and depressive symptoms.Variables that entered the model had a p-value <0.20 in the crude analysis and variables whose p-value continued to be <0.05 after adjustment were maintained in the analysis.This project was approved by the Comitê de Ética em Pesquisa da Universidade Federal do Rio Grande do Sul (Rio Grande do Sul Federal University Research Ethics Committee).Adolescents answered the questionnaire after written consent was given by their parents or legally responsible adults.

RESULTS
The sample of adolescents interviewed was divided similarly according to sex, with a predominance of low-class individuals (36.5%), aged 11 years (20.6%), with six to seven years of schooling (43.6%), without intimidation (68.2%), who did not practice a physical activity (55.0%), did not smoke in the last month (93.1%), did not use illicit drugs in the last month (96.5%), did not participate in religious activities (54.7%), had not had sexual intercourse (87.8%) and did not have depressive symptoms (97.9%) (Table 1).
As regards age, prevalences of alcohol consumption increased according to age group, for each sex and for all individuals in the study (Table 2).
In the analysis adjusted for both sexes, an increase in the prevalence of alcohol consumption compared to the increase in age group, beginning at 13 years, was observed.In addition, prevalences of alcohol consumption were higher among adolescents who had failed school, reported tobacco use in the last month and had already had sexual intercourse (Table 3).
Among female adolescents, in the crude analysis, alcoholic beverage consumption in the last month was greater among those aged 14 and 15 years, who had eight or more years of schooling, had failed school, smoked or used illicit drugs in the last month, had already had sexual intercourse, and had depressive symptoms.Those who did not practice a physical activity showed lower prevalence of alcohol use.Socioeconomic level and presence of intimidation were not associated with alcoholic beverage consumption in the last month (Table 4).In the adjusted analysis, there was association with age: the older the age group, the higher the prevalence of alcohol use.However,  confi dence intervals showed association from the age of 14 years (Table 4).
Among males, in the crude analysis, the following were variables associated with higher prevalence of alcohol use: 14 years of age or older, eight or more years of schooling, smoking in the last month, use of illicit drugs in the last month, and previous sexual intercourse.Social status, history of school failure, intimidation and practice of physical activity were not associated with alcohol in the last month (Table 5).
In the adjusted analysis, the proportion of adolescents who reported alcoholic beverage use in the last month was found to be directly associated with age.Smoking and previous sexual intercourse were also associated with alcoholic beverage use (Table 5).Even though they were maintained in the adjusted analysis, the "participation in religious activities" and "presence of depressive symptoms" variables were not signifi cantly associated with the outcome.

DISCUSSION
Cross-sectional studies are not indicated to analyze etiological factors, once they are infl uenced by reverse causality.However, the present study is populationbased and representative of the city of Pelotas.
Studies on adolescent health are usually performed in schools or in the relationship between workers and non-workers, not on a population basis.The present study sought to include a representative sample of adolescents of this city to answer questions about early alcohol consumption.
As questions are of a personal nature, reliability of responses was guaranteed with the use of an anonymous, self-applied questionnaire, which was put into a sealed box by the respondent, assuring them secrecy of information and encouraging them not to omit alcoholic beverage consumption.However, prevalence of consumption may have been underestimated as this is a stigmatizing habit.
Previous studies indicated that individuals with lower economic conditions showed greater alcohol consumption in adolescence and also adulthood. 16,20 However, alcohol use among females has been changing, which may be harmful due to women's physiological peculiarities.They have lower alcohol dehydrogenase serum levels, higher proportion of fat compared to body water and alcohol metabolism variations in different menstrual cycle phases. 4Such characteristics may result in chemical dependence, even with lower amounts, and greater harmful effects caused by alcohol, when compared to men. 18 regards age, in both sexes, results showed that 12% of interviewees aged 11 years had already consumed alcoholic beverages in the last month, and these values rose to 20.2% at 13 years, thus indicating beginning of consumption at an early age.
Addictive types of behavior are associated with each other and preventive campaigns could not only approach one issue.Studies point to an association with tobacco use 12,22 in those aged between 11 and 13 years, confi rmed by the present investigation among male adolescents.Psychoactive substance use at such ages increases risk of chemical dependence in adult life. 11Even experimental use is harmful and may lead to behavioral changes.Accidents, aggressions, and undesired and unprotected sexual intercourse occur more frequently among adolescents after consumption of any psychoactive substance. 2,9,24 the present study, level of education, school failure, intimidation, depressive symptoms, physical activity and religious activity were not associated with the outcome.Another study with a similar design, such as the one performed by Dalgalarrondo et al 5

A
population-based, cross-sectional study was performed in the city of Pelotas, Southern Brazil, between September 2005 and April 2006.This study was part of a more extensive research project on several health outcomes in adolescence, performed by the Programa de Pós-Graduação em Saúde e Comportamento da Universidade Católica de Pelotas (Post Graduation Program in Health and Behavior, Pelotas Catholic University).

Table 2 .
Prevalence of alcohol use in the last 30 days, according to sex and age.Pelotas, Southern Brazil, 2005-2006.
a Linear trend test b Adjusted for age and among themselves (fi rst level) c Adjusted for age, school failure and among themselves (second level)

Table 4 .
Crude and adjusted analyses of alcohol use in the last 30 days for females.Pelotas, SouthernBrazil, 2005-2006.
a Linear trend test b Adjusted forage and among themselves (fi rst level) c Adjusted for age and among themselves (second level)

Table 5 .
Crude and adjusted analyses of alcohol use in the last 30 days for males.Pelotas, SouthernBrazil, 2005-2006.
in 2004, found out that adopting a religion consists in a set of values, including acceptance or refusal of alcohol and drug use.In the present study, almost half of adolescents (45.3%) reported participating in religious activities, but differences in alcohol use were not found.It is possible that the type of religion may interfere with adolescent behavior towards alcohol use.Alcoholic beverage use showed high prevalence in both sexes and began at an early age, revealing changes in adolescent profi le.Thus, this set of fi ndings may point to alcohol use among adolescents as a public health problem.Preventive measures to control alcohol consumption must include the 11-to-15-year age group, aiming to reduce the social cost of this abuse.