Prevalence of Alcohol and Tobacco Use among Brazilian Adolescents: a Systematic Review

Prevalência de consumo de álcool e tabaco entre adolescentes brasileiros: revisão sistemática ABSTRACT OBJECTIVE: To analyze alcohol and tobacco use among Brazilian adolescents and identify higher-risk subgroups. METHODS: A systematic review of the literature was conducted. Searches were performed using four databases (LILACS, MEDLINE /PubMed, Web of Science, and Google Scholar), specialized websites and the references cited in retrieved articles. The search was done in English and Portuguese and there was no limit on the year of publication (up to June 2011). From the search, 59 studies met all the inclusion criteria: to involve Brazilian adolescents aged 10-19 years; to assess the prevalence of alcohol and/or tobacco use; to use questionnaires or structured interviews to measure the variables of interest; and to be a school or population-based study that used methodological procedures to ensure representativeness of the target population (i.e. random sampling). RESULTS: The prevalence of current alcohol use (at the time of the investigation or in the previous month) ranged from 23.0% to 67.7%. The mean prevalence was 34.9% (refl ecting the central trend of the estimates found in the studies). The prevalence of current tobacco use ranged from 2.4% to 22.0%, and the mean prevalence was 9.3%. A large proportion of the studies estimated prevalences of frequent alcohol use (66.7%) and heavy alcohol use (36.8%) of more than 10%. However, most studies found prevalences of frequent and heavy tobacco use of less than 10%. The Brazilian literature has highlighted that environmental factors (religiosity, working conditions, and substance use among family and friends) and psychosocial factors (such as confl icts with parents and feelings of negativeness and loneliness) are associated with the tobacco and alcohol use among adolescents. CONCLUSIONS: The results suggest that consumption of alcohol and tobacco among adolescents has reached alarming prevalences in various localities in Brazil. Since unhealthy behavior tends to continue from adolescence into adulthood, public policies aimed towards reducing alcohol and tobacco use among Brazilians over the medium and long terms may direct young people and the subgroups at higher risk towards such behavior. 902 Use of alcohol and tobacco among adolescents Barbosa Filho VC et al Tobacco use is an important modifi able risk factor for the development of non-communicable diseases (NCD). Globally, 60% of deaths are caused by NCD, which is the leading cause of mortality. Tobacco use is responsible for 5.1 million deaths per year, second only to high blood pressure, …

Tobacco use is an important modifi able risk factor for the development of non-communicable diseases (NCD).Globally, 60% of deaths are caused by NCD, which is the leading cause of mortality.Tobacco use is responsible for 5.1 million deaths per year, second only to high blood pressure, which kills 7.5 million people per year. 65If serious action is not taken, annual tobacco-related deaths are projected to increase to 8 million by 2030 (10% of all deaths). 67e of alcohol is responsible for 2.3 million deaths per year (4% of all deaths).Alcohol use is associated with more than 60 types of disease and injury.It can also decrease the risk of coronary heart disease, stroke and RESUMO OBJETIVO: Analisar o uso de álcool e tabaco em adolescentes brasileiros e identifi car os subgrupos de maior risco.

INTRODUCTION
diabetes. 65Harmful use of alcohol is a major risk factor for premature death and disability worldwide. 26cohol and tobacco are used in different countries across all income groups. 65Therefore, monitoring the exposure of individuals and populations to these substances can be considered to provide an important assessment of the health status of the world's population. 64Since alcohol and tobacco use frequently begin in childhood and adolescence and continue into adult life, use of alcohol and tobacco among young people is a predictor of the future burden of alcohol-and tobacco-related diseases. 21,45Estimating the consumption of alcohol and tobacco among adolescents is an a The following institutional websites were accessed: Brazilian Department for Alcohol and Other Drugs Policies (www.obid.senad.gov.br); Brazilian Center for Information on Psychotropic Drugs (www.cebrid.epm.br);National Institute of Cancer (www2.inca.gov.br),Health Ministry (portalsaude.saude.gov.br); and the Brazilian Institute for Geography and Statistics (www.ibge.gov.br).
important tool for monitoring the population's health status. 66udies have monitored the consumption of alcohol and tobacco among young people.The Youth Risk Behavior Survey (YRBS) 17 has been conducted biennially in the United States since 1991.These surveys provide estimates and temporal trends of alcohol and tobacco use among American adolescents.The Global Schoolbased Student Health Survey (GSHS) is an international survey that monitors several health-related risk factors among young people. 21,46The GSHS has contributed to knowledge of alcohol and tobacco use, especially among adolescents in developing countries.The fi rst studies on alcohol and tobacco use among adolescents in Brazil were conducted in the mid-1980s. 22,25Since then, more comprehensive epidemiological studies on these habits among Brazilian youth have been conducted. 24,32,36e present study aimed to systematically review the literature on the prevalence of alcohol and tobacco use among Brazilian adolescents, in order to identify high-risk subgroups and to contribute towards developing national policies focused on combating alcohol and tobacco use among adolescents.This review may contribute towards better awareness of the main cities and locations where the rates of alcohol and tobacco use are higher, especially among adolescents who are usually not evaluated in national studies (i.e.teenagers in small towns).

METHODS
The literature search was performed in June 2011 on the LILACS, MEDLINE/PubMed and Web of Science electronic databases and the Google Scholar portal.A combination of Medical Subject Headings (MeSH), Descritores de Ciências da Saúde (DeCS; for terms in Portuguese) and text words were used to generate the list of citations.The search process was constructed specifi cally for each database, and no limits were used in these searches.The key terms were used to search in MEDLINE/PubMed, LILACS and Google Scholar and by topic in the Web of Science.The search was conducted in English and Portuguese.
Our search strategy was based on a combination of three search parameters: risky behavior, sample type and nationality.The keywords for behavioral risk factors were stratifi ed into three subgroups, featuring the behavioral outcomes included: (i) general terms for behavioral risk factors (adolescent behavio* OR adolescent health OR cardiovascular health OR cardiovascular diseas* OR health behavio* OR lifestyle OR risky behavio* OR risk factor*); (ii) alcohol use (alcohol use OR illicit drugs OR alcohol drinking OR alcoholism OR alcoholic beverages); and (iii) tobacco use (smoking OR illicit drug OR tobacco use OR tobacco).Each subgroup of the behavioral risk factor descriptors was used in combination with keywords for sample type (youth OR teen* OR adolesc* OR child*) and nationality (Brazil*) to locate potentially relevant studies.The Boolean operator "AND" was used to combine the three groups in the search.The truncation symbols for each specifi c database (e.g., *, $, or #) were used to capture all suffi x variations of a root word.
Articles were selected in accordance with a systematic method.First the article titles and abstracts were read, then the inclusion criteria were applied to the analysis.All of the full-text articles were obtained and included if they met the inclusion criteria.The references for all selected papers were examined to identify other publications that should be reviewed.Searches were conducted on institutional websites to reach national survey data not published in scientific journals.a The inclusion criteria for studies were as follows: (i) original research study; (ii) sample including Brazilian adolescents aged ten to 19 years (or with mean age within this range); (iii) observational study that assessed the prevalence of alcohol or tobacco use, regardless of whether the study considered this variable to be an exposure or an outcome; (iv) questionnaires or structured interviews to measure alcohol and tobacco use; and (v) school-or populationalbased study that used methodological procedures to ensure representativeness of the target population (i.e., random sampling).Studies were not included if participants were selected based on a convenience sample or another non-probabilistic method, for example, if studies selected participants by evaluating adolescents from the largest school in town or a single class.Studies that included only adolescents from specifi c areas within a city (for example, slums or poor neighborhoods) were not included.These criteria were selected in order to increase comparability between the studies.
Articles were excluded if they averaged the measurements of smoking or alcohol consumption (for example, average number of cigarettes or alcoholic beverages per day), because these measurements would not assess the prevalence of the behavior in the target population.Studies that evaluated only lifetime alcohol and/or tobacco use and that did not measure the prevalence of alcohol or tobacco use in adolescents were also excluded.Theses, dissertations and monographs were not included, as it would be impractical to perform a systematic search of these papers.Likewise, case reports and reviews were not included.In the case of duplication, the most recent or the most complete publication was used, and the other publications were excluded.
The following data were extracted: location and year of survey, sample size, age of participants and defi nition of alcohol and/or tobacco use.The prevalence of current alcohol and tobacco use and the respective 95% confi dence intervals (95%CI) are presented for the total sample sizes.Prevalence rates and 95%CI are presented for the total sample and stratifi ed according to gender for frequent and heavy use of alcohol and tobacco.The 95%CI were obtained directly from original articles whenever possible, 1,5,7,11,20,23,28,32,33,36,39,40,[51][52][53]55,56,59,68 or were calculated using the statistics program Stata 10.0 (Stata Corp., College Station, United States) the 'cii' command (95%CI exact for binomial distribution).
Overall mean prevalences were calculated to demonstrate the central trend for current use of alcohol and tobacco in Brazilian studies.These (weighted) mean prevalence rates were calculated by summing the numbers of cases in all the studies considered, dividing by the sum of the number of participants in all the studies considered, and multiplying by 100.The studies were organized according to the criteria used to identify the behavior of interest (i.e.current, frequent or heavy use of alcohol or tobacco).The studies were further organized according to geographical region of Brazil, in alphabetical order by study site and in chronological order by survey year.

RESULTS
The literature search yielded 3,195 potentially relevant articles.After reading the titles, 647 articles were selected on the basis of the inclusion criteria.The 647 abstracts were reviewed and 160 articles were selected for full text review.Of these, 65.4% were excluded for the following reasons: 32 used a non-probabilistic method for selecting participants, 28 studies did not present the variables of interest, 39 were duplicated studies and fi ve did not include adolescents.Thus, 56 studies met our inclusion criteria.One additional study was obtained from the reference search, and two national surveys b,c were obtained from the website search.Therefore, 59 studies were reviewed (Figure 1).

DISCUSSION
The Brazilian literature includes a large number of studies on alcohol and tobacco use during adolescence.Alcohol and tobacco use among adolescents has been treated fairly in the national literature, through development of several comprehensive high-quality epidemiological studies. 2,22,24,32,36,47The high number of Brazilian studies indexed in international databases (e.g.MEDLINE/PubMed and Web of Science; Figure 1), indicate increasing interest in alcohol and tobacco-related research relating to young Brazilians and the impact of risky behavior on the individual's health.
However, there are some gaps in the national literature on alcohol and tobacco use among young Brazilians.
There was a high concentration of studies focusing on the southern and southeastern regions of Brazil, whereas few of the studies included were conducted in the northern, northeastern and central-western regions.The development of scientifi c research relating to these risky behaviors in Brazil has yet to cover all regions of the country.Efforts have been made by regional and national institutions to conduct surveys involving adolescents in all regions of Brazil.These institutions have invested fi nancially and academically in this research area, and have developed policies for diagnosing and preventing behavioral risk factors among young people in different regions of Brazil.
Although the sale of alcoholic beverages to children under 18 years has been prohibited in Brazil, d many Brazilian adolescents have consumed alcohol.Several studies observed prevalence rates of current alcohol use of more than 50%. 12,29,31,44,54,56,60,62The most worrisome fi ndings, however, were the high prevalence rates (above 10%) of frequent 6,14,15,19,22,37,55,60,61 and heavy 10,18,20,43,52,58,62 alcohol use in several studies.The prevalence rates of tobacco among Brazilian adolescents were also high, with several studies fi nding rates of more than 15%. 28,30,41,44,60,680][41] These behaviors are exhibited by large portions of the young Brazilian population at several sites and are contributing negatively to the health status of these adolescents.
Several studies included in this review sought to identify risk factors associated with current, frequent and heavy use of alcohol and tobacco among young Brazilians (Table 2).A large number of studies highlighted a positive association between age and consumption of these substances, which may explain the different prevalence estimates, since these studies examined different age groups.
Some of the highest estimates for current alcohol 12,31,56 and tobacco 30,51,68 use were obtained in studies with older adolescents (14 years or older).However, studies that included young adolescents (those younger than 14 years) also revealed high prevalence rates.For example, the Pesquisa Nacional de Saúde do Escolar (National Adolescent School-based Health Survey), 36 a national survey among eighth-grader students (over 90% were 13-15 years) in all 27 Brazilian state capitals, found high prevalences of current alcohol (27.3%) and tobacco (6.3%) use.Another study, which evaluated young adolescents (13-15 years) in three state capitals in Southern Brazil, found high estimates for current tobacco use (10.7%, 12.6%, and 17.7% in Florianópolis, Curitiba, and Porto Alegre, respectively). 28One study on adolescents from a birth cohort study in Pelotas, Southern Brazil, demonstrated the early onset of smoking among the participants. 40Of these, 7.5% consumed cigarettes on a weekly basis at 14-15 years of age, and this prevalence nearly tripled in late adolescence to 21.5% at 17-19 years of age. 40Cross-sectional data from this cohort also indicate that more than half the adolescents with any lifetime tobacco use smoked their fi rst cigarette before reaching ten years of age. 38hese data suggest that the prevalence of alcohol and tobacco use is already high in early adolescence and that these estimates tend to increase with age.
Gender was frequently analyzed in studies to identify subgroups at higher risk of alcohol and tobacco use.However, the direction of association between gender and alcohol and tobacco use has important variations according to the particular behavior presenting risk.Male adolescents are more likely to drink alcohol, and this result was obtained for current, frequent and heavy alcohol use (Table 2).However, for tobacco use, some  studies showed that female adolescents had higher rates of smoking habits (Table 2), while some studies indicated that male adolescents were at higher risk of smoking.These confl icts may be related to historical and social conditions or concepts of male and female identity. 30However, the current trend indicates that the consumption of alcohol and tobacco may be increasing among female adolescents, thus indicating an at-risk population in future generations.
The Brazilian literature has also indicated that sociocultural and environmental factors (such as religiosity, working condition and substance use among family and friends) are associated with tobacco and alcohol use during adolescence.Adolescents who are not religious are more likely to use alcohol and tobacco (Table 2), a fi nding that may be related to the doctrines followed by some religious denominations, which prohibit unhealthy behavior. 7Some studies have suggested that secondary exposure to smoking and alcohol consumption by parents, older brothers and friends may be an important risk factor for developing these behaviors (Table 2).Data from a study conducted in three state capitals in Southern Brazil indicated that adolescents with friends who smoke are nine times more likely to smoke than adolescents without friends who smoke. 28ychosocial factors (such as confl ict with parents and negative feelings such as sadness and loneliness) also contributed towards alcohol and tobacco use among Brazilian adolescents.One study on adolescents in all 27 state capitals found that those in poor relationships with their parents were 50% more likely to be heavy alcohol drinkers. 24Another study on students in Gravataí, Southern Brazil, showed that adolescents who experienced feelings of sadness over the previous 12 months were 2.6 and 1.7 times more likely to report current tobacco and alcohol use, respectively. 63The national literature has indicated the importance of environmental and psychosocial factors on alcohol and tobacco consumption among adolescents.Reducing alcohol and tobacco use in adolescents may involve reducing exposure to these risk factors.
The national literature has highlighted the association between alcohol consumption and smoking during adolescence. 1,19,29,49Students in João Pessoa, Northeastern Brazil, who drank alcohol on a weekly basis were 20 times more likely to consume tobacco on a weekly basis. 19In Brasilia, in Central-Western Brazil, nearly eight in ten adolescent smokers also drank alcohol. 49These two behaviors are associated, but the association does not necessarily imply causality.However, interventions aiming to reduce alcohol consumption among adolescents could also reduce tobacco use, and vice-versa.
Some studies have examined temporal trends in the consumption of alcohol and tobacco consumption among adolescents.Data from a series of surveys conducted between 1987 and 1997 by the Centro Brasileiro de Informações sobre Drogas (Brazilian Center for Information on Psychotropic Drugs -CEBRID) have shown important estimates for temporal trends in these behaviors among adolescents in ten Brazilian state capitals. 22The prevalence of frequent alcohol use (six or more drinks per month) increased in six of the cities studied.However, the prevalences of frequent alcohol use (13.2% in 1987 and 15.0% in 1997) and frequent tobacco use (7.0% in 1987 and 6.2% in 1997) were generally stable. 22Another study from CEBRID presented estimates for tobacco use in 2004, which were compared with data from 1997. 35With the exception of Recife and Rio de Janeiro, there was a signifi cant decrease in heavy use of tobacco (defi ned as 20 or more cigarettes in the previous month).The decrease in tobacco use was related to changes in public policy in Brazil during the interval between the surveys (cigarette advertising was banned from the Brazilian media in 2000). 35blic policy can be effective in reducing unhealthy behavior.Given the high prevalence of current, frequent and heavy alcohol use among Brazilian adolescents (Figure 2 and Table 1), there is an urgent need for policies that are effective for reducing alcohol use among adolescents.
Some limitations of this review should be highlighted.The fi rst limitation is the variability between the studies reviewed.Although several studies have used questionnaires based on instruments from the GSHS, 21,46 there are important variations in the defi nitions of alcohol and tobacco use patterns, recall periods and sample selection methods.These inter-study methodological differences may be related to the objectives established for each study.Future studies should aim to minimize these variations.Local and regional studies with a standardized methodological procedure (i.e.instruments, cutoff points, etc.) would be required to contribute towards diagnosing and monitoring alcohol and tobacco use among Brazilian adolescents.
The second limitation relates not only to this review but also to most studies included in it: the use of questionnaires to estimate alcohol and tobacco use among adolescents.Many studies did not test the validity of the questionnaires, and when they were tested, they found that adolescents tended to underestimate their consumption. 35Thus, the real prevalence rates may be greater than those found in these studies. 34e last limitation is that some important studies may not have been included in this review.A large number of studies on alcohol and tobacco use in Brazilian adolescents are theses and dissertations that have not been published in journals.However, most of the main research concerning alcohol and tobacco use among Brazilian adolescents is presented in this review.

CONCLUSIONS
There are important inter-study variations in the prevalence rates of current, frequent and heavy alcohol and tobacco use among Brazilian adolescents.However, a large proportion of the studies included reached worrying prevalence rates of alcohol and tobacco consumption.
This review pointed out the relationship between some environmental factors (e.g.religiosity, working condition and substance use among family and friends) and psychosocial factors (e.g.confl ict with parents and negative feelings such as sadness and loneliness) and alcohol and tobacco use among young Brazilians.However, we recommended that further studies should be conducted to identify risk factors, especially in relation to alcohol use, because the national literature does not comprehensively assess these variables.Since unhealthy behaviors continue through adolescence and adulthood, medium and long-term public policies created to reduce the use of alcohol and tobacco among Brazilians should focus on interventions among young people and subgroups that are at higher risk, so as to promote adoption of healthier behaviors.

Table 1 .
Prevalence of alcohol and tobacco use among Brazilian adolescents in studies included in this review, by gender, according to region, survey site, year, sample, participant's ages, and defi nition of alcohol or tobacco use.Location (survey year) a Figure 3. Prevalence of current tobacco use among Brazilian adolescents according to region, year of survey, sample size and participants' ages.

Table 1 .
Continuation Location (survey year) a

Table 1 .
Continuation Location (survey year) a

Table 2 .
Summary of risk factors associated with alcohol and tobacco use among Brazilian adolescents (the numbers are the study reference number).