Artigos

Conteúdo (article):

ALCOHOL CONSUMPTION AND OVERWEIGHT IN BRAZILIAN ADULTS – CUME PROJECT**

Luís Paulo Souza e Souza1,2
Helen Hermana Miranda Hermsdorff3
Aline Elizabeth da Silva Miranda4
Josefina Bressan5
Adriano Marçal Pimenta1,6

1 Universidade Federal de Minas Gerais. Faculdade de Medicina. Programa de Pós-Graduação em Saúde Pública. Belo Horizonte, MG, Brasil.
2 Universidade Federal do Amazonas. Instituto de Saúde e Biotecnologia. Departamento de Medicina. Coari, AM, Brasil. E-mail: luis.pauloss@hotmail.com; ORCID: 0000-0002-9801-4157.
3 Universidade Federal de Viçosa. Departamento de Nutrição e Saúde. Viçosa, MG, Brasil. E-mail: helenhermana@gmail.com; ORCID: 0000-0002-4441-6572.
4 Universidade Federal de Minas Gerais. Departamento de Nutrição. Programa de Pós-Graduação em Nutrição e Saúde. Belo Horizonte, MG, Brasil. E-mail: alineesnutri@hotmail.com; ORCID: 0000-0002-4993-9436.
5 Universidade Federal de Viçosa. Departamento de Nutrição e Saúde. Viçosa, MG, Brasil. E-mail: josefinabressan@yahoo.com.br; ORCID: 0000-0002-4441-6572.
6 Universidade Federal de Minas Gerais. Escola de Enfermagem. Departamento de Enfermagem Materno-Infantil e Saúde Pública. Belo Horizonte, MG, Brasil. E-mail: adrianomp@ufmg.br; ORCID: 0000-0001-7049-7575.


**Article extracted from the Thesis “Consumption of alcoholic beverages and overweight in adults of the baseline of the Cohort of Universities of Minas Gerais (CUME), Brazil”, defended by Luís Paulo Souza e Souza in the Public Health Graduate Program of the School of Medicine of Federal University of Minas Gerais.

Acknowledgements: To participants in The Cohort of Universities of Minas Gerais (CUME), as well as the Coordination of Support to Higher Level Personnel (CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) (Ministry of Education – Ministério da Educação, funding code 001) and the Research Support Foundation of Minas Gerais (FAPEMIG) for financial support. Hermsdorff HHM e Bressan J are Research Productivity Fellows from the National Council for Scientific and Technological Development (CNPq).

ABSTRACT
This is a cross-sectional study with 2909 participants (aged ≥18 years) from the baseline of the Cohort of Universities of Minas Gerais [CUME - Coorte de Universidades Mineiras] which verified the association between alcohol consumption and overweight. Data on sociodemographic factors, lifestyle, eating habits, anthropometric factors and clinical conditions were collected through an online questionnaire. Body mass index ≥ 25 kg/m² was considered an indicator of overweight. The daily consumption of alcohol was evaluated in grams (alcohol) and according to type in milliliters (beer, wine, hard liquor). The prevalence of alcohol consumption and overweight was 73.6% and 40.8%, respectively. There was a significant tendency of an increase in overweight with higher beer consumption (tendency p value of 0.038), which was not observed for the other types of alcohol. After sensitivity analyses, alcohol consumption was associated to overweight, with a tendency of increase in prevalence with higher daily consumption. There is a crucial need to curb the widely accepted idea that a low or moderate alcohol consumption is not harmful to one’s health, and to be cautious of such a proposition. The influence of alcohol consumption regarding weight gain must be considered in public health policies and policies of alcohol consumption control.
Keywords: Alcohol Drinking; Overweight; Obesity.

RESUMO
Estudo transversal com 2.909 participantes (≥ 18 anos) da linha de base da Coorte de Universidades Mineiras (CUME), Brasil, que verificou a associação entre o consumo de bebidas alcoólicas e o excesso de peso. Por meio de questionário virtual, coletaram-se dados sociodemográficos, estilo de vida, hábitos alimentares, antropométricos e condições clínicas. Considerou-se Índice de Massa Corporal ≥ 25 kg/m2 como excesso de peso. Avaliou-se o consumo diário de bebidas alcoólicas no total em gramas (álcool) e segundo tipo em mililitros (cerveja, vinhos e destiladas). As prevalências do consumo de álcool e do excesso de peso foram 73,6% e 40,8%, respectivamente. Houve uma tendência significativa de aumento da prevalência de excesso de peso quanto maior o consumo de cerveja (p de tendência = 0,038), fato não observado para os demais tipos de bebidas. Após análises de sensibilidade, a ingestão de álcool associou-se com excesso de peso, com tendência de aumento da prevalência quanto maior o consumo diário. Ressalta-se a necessidade de reduzir a visão amplamente aceita de que o consumo leve a moderado de álcool não é nocivo à saúde, adotando cautela nesta proposição. Deve-se considerar a influência da ingestão de bebidas alcoólicas no ganho de peso nas políticas públicas de saúde e de controle do consumo do álcool.
Palavras-chave: Consumo de Bebidas Alcoólicas; Sobrepeso; Obesidade.


INTRODUCTION
The overweight (or excess weight) is currently the major nutritional disorder and is considered one of the main public health problems. In 2016, the global prevalence of this disorder in adults was 39%1, and in the countries with a high prevalence of excess weight, it accounted for 1 to 3% of total cost of healthcare2. In Brazil, 55.7% of adults were considered to have excess weight in 20183, and in 2017, the costs related to the disorder amounted to 2.1 billion dollars4.
The causes of excess weight include genetic, metabolic, psychosocial and cultural aspects, food intake and sedentary lifestyle. In general, studies attempting to explain such factors have mainly focused on decreased energy expenditure due to sedentary lifestyle and increased intake of energy-dense foods and sugary and alcoholic beverages5,6,7.
Regarding the relation between alcohol consumption and excess weight, it is noteworthy that alcoholic beverages have an elevated energy density – one gram of alcohol has 7.1 kilocalories (kcal), whereas one gram of carbohydrates has 4 kcal, one gram of protein has 4 kcal, and one gram of lipids has 9 kcal6,8. Due to its energy value, it is assumed that alcohol could potentially provide one’s daily energy requirements and lead to excess weight, depending on the frequency and type of consumption, quantity and type of alcohol9,10,11,12.
Alcohol abuse, in addition to causing dependence, increases the risk of excess weight in both sexes7,8,12,13,14,15,16, regardless of the type of alcohol (wine, beer or hard liquor)17,18,19,20. On the other hand, moderate alcohol use has been shown to be a protective factor, especially the intake of wine and beer5,7,12.
Excess weight and alcohol consumption are important public health issues that need to be investigated due to the relation between them and the fact that both represent risk factors for non-communicable diseases and injuries (NCDI)6,21,22, which are the main causes of morbidity and mortality in Brazil and represent very high costs for the Brazilian Unified Health System4.
Even though the Brazilian population presents a high mean daily alcohol consumption, there are few studies analyzing its association to excess weight, leading to little consistency in the scientific findings on this topic in Brazil23,24,25,26.
This study is also relevant because it evaluates a cohort of the Brazilian population which is underexplored in epidemiological investigations. These individuals are distributed throughout Brazil, have a high schooling level, hold high positions of employment that are crucial to the country’s economy, present a high prevalence of unhealthy habits27; therefore, when their contributions are cut short by getting ill or death, these could result in significant social and economic burdens for the country.
Thus, the objective of our study was to verify the association between alcohol consumption and overweight in adults from the baseline of the Coorte de Universidades Mineiras (CUME – Cohort of Universities of Minas Gerais), Brazil.

METHOD
This is a cross-sectional study that analyzes the baseline of Cohort of Universities of Minas Gerais (CUME - Coorte de Universidades Mineiras), an open cohort that included a restricted population group, to evaluate the impact of the Brazilian dietary pattern and of the nutritional transition experienced by former students from the Instituições Federais de Ensino Superior (IFES) [Federal Institutions of Higher Education] in the State of Minas Gerais, Southeastern region of Brazil. The study design, communication strategies and baseline profile of participants were detailed in a previous article27.
Data was collected between March and August 2016, in a virtual environment owned by CUME Project, of former students of Universidade Federal de Viçosa (UFV) and Universidade Federal de Minas Gerais (UFMG) who were invited via e-mail. Those who signed the informed consent form were granted access to the data collection instrument.
The invitation to participate in the study was sent to all 64,202 former students of UFMG and the 16,945 former students of UFV, who graduated between 1994 and 2014 (undergraduates and graduates), whose e-mail address were listed on the database of the alumni associations or the IT directories of IFES. A total of 4,949 participants answered the baseline questionnaire (Q_0), which represented a rate of 6.1%.
The study analysis excluded individuals who did not complete the Food Frequency Questionnaire (FFQ) (n = 1,679); pregnant women and women who had given birth within one year prior to the study (n = 123); those who reported inconsistent values for total calorie intake [ < 500 kcal/day (n = 1) or > 6,000 kcal/day (n = 92)];28 non-Brazilian former students (n = 11); and Brazilians students residing abroad (n = 134). With that, the final sample was made up of 2,909 participants.
The data was collected through an online self-managed baseline questionnaire divided into two parts due to its length (access at: http://www.projetocume.com.br/questionario). The first part was made up of questions related to sociodemographic and economic characteristics; lifestyle; referred individual and family morbidity; use of medication; personal history of clinical and biochemistry tests from the previous two years; and anthropometric data. The second part was sent one week after completion of the first part and included the FFQ, composed by 144 food items based on the original version previously validated in Brazil about food intake in the last year29.
The outcome variable chosen was “excess weight”, defined according to self-referred data on weight (in kilos) and height (in centimeters), which were used to calculate the body mass index (BMI). The individuals were classified as follows: BMI < 25.0 kg/m² = no excess weight and BMI ≥ 25.0 kg/m² = excess weight.
The self-referred data on weight, height and BMI were previously validated in a study conducted with a subsample from CUME, which generated intraclass correlation coefficients that indicated excellent agreement; i.e., 0.989 (weight); 0.995 (height); and 0.983 (BMI)30.
To evaluate alcohol consumption, the study used information collected from the FFQ. In the part of the FFQ regarding alcoholic beverages, for each beverage selected (cachaça; hard liquor - rum, vodka and whisky; beer; wine), each participant indicated the portion size expressed in measurements commonly used in Brazil (glasses, cans) and the usual consumption frequency (day/week/month/year).
The consumption frequency of each beverage was transformed into daily frequencies which, in turn, were multiplied by the portion size to calculate the daily consumption of each beverage (grams – g, or milliliters – mL). To calculate alcohol intake and the intake of each alcoholic beverage (beer, wine and hard liquor), Brazilian tables of the nutritional composition of the foods31 and, whenever necessary, the table from the United States Department of Agriculture, were used32.
At the end, the exposure variables considered were: a) alcohol consumption (no, yes); b) weekly frequency of alcohol consumption (< 1 day/week; 1–4 days/week; ≥ 5 days/week); c) daily alcohol consumption (g) (in quartiles); d) daily consumption of the types of alcoholic beverages (mL) (in quartiles) = d1) beer; d2) wine; d3) hard liquor, including cachaça, rum, vodka and whisky.
The daily intake of each alcoholic beverage and alcohol in general, in addition to the covariables related to food intake, were adjusted by calorie intake through the residual method33, before the statistical analyses.
The covariables considered were: a) sociodemographic characteristics: sex (male, female); age (years – continuous); skin color (white, brown/black, yellow/indigenous); marital status (single, married/common law marriage, separated or divorced/widow(ed)); professional status (retired/homemaker/unemployed, student, full-time work, part-time work, informal work); and household income (continuous); b) life habits: smoking (never smoked, former smoker, current smoker) and physical activity, evaluated through a list of 24 activities and expressed in minutes per week (active = individuals with ≥ 150 minutes/week of moderate activity or ≥ 75 minutes/week of vigorous activity; insufficiently active = individuals with < 150 minutes/week of moderate activity or < 75 minutes/week of vigorous activity; inactive = individuals with no leisure physical activity)34; c) clinical characteristics: self-reported medical diagnosis for chronic diseases: stroke, myocardial infarction, celiac disease, inflammatory bowel diseases (Crohn’s disease, ulcerative colitis), nonalcoholic liver steatosis, alcoholic cirrhosis, gastritis, gastric or duodenal ulcer, acute renal failure, chronic renal failure, lung cancer, skin cancer, colon cancer, breast cancer, cervical cancer, prostate cancer, type 2 diabetes mellitus, hypertension, high triglyceride levels, high cholesterol levels, depression (no, yes); d) food intake: these variables were also obtained through the FFQ, with a methodology similar to the one used for alcohol consumption estimates, and were calculated in daily grams of each food. To calculate calorie (kcal) and nutrient intake, Brazilian Tables of the Nutritional Composition of the Foods31 and the Table from the United States Department of Agriculture were used32. Food items were classified according to the NOVA classification as minimally processed, processed and ultra-processed foods35.
The characterization of the participants was done with the distribution of absolute and relative frequencies, means and standard deviations of the sociodemographic, economic, lifestyle and health conditions variables, according to the presence or absence of excess weight. Statistical differences were evaluated through Pearson’s chi-square test or the Student´s t test.
Prevalence ratios (PR) and their respective 95% confidence intervals (CI) of the associations between each exposure variable and excess weight were estimated through Poisson’s regression models with robust variances, initially adjusted for sex and age, and later for skin color, marital status, professional status, smoking, physical activity, medical diagnosis of depression, household income, total alcohol consumption and total energy intake.
Sensitivity analyses were conducted to evaluate the robustness of results, and excluded: (a) individuals with previous medical diagnosis of obesity; (b) individuals who gained 10 kilograms or more before the study; (c) all individuals considered in the analyses (a) and (b), besides those who were on weight control medication.
The data was analyzed through the software Stata, version 13.0, and statistical significance was set at 5%.
The study was approved by the Human Research Ethics Committees of UFV and UFMG (protocol 596.741-0/2013).

RESULTS
Of all the participants, 40.8% (n = 1,188) had excess weight; in that, 27.6% (n = 804) were overweight and 13.2% (n = 384) were obese. Moreover, 3.0% (n = 87) were underweight and 56.2% (n = 1,634) were eutrophic.
In comparison to the participants without excess weight, individuals with excess weight were, in their majority, male, married, working full time and with a higher mean age and household income (p < 0.05) (Table 1).
Furthermore, individuals with excess weight presented higher frequencies of smoking or previous smoking, no physical activity and medical diagnoses of depression and chronic diseases, in addition to presenting a higher mean of calorie intake (kcal/day) and intake of macronutrients (g/day) [carbohydrates, proteins, lipids (polyunsaturated fatty acid – PUFA; monounsaturated fatty acids – MUFA; saturated fat; trans fat)] and a lower mean of intake of minimally processed food (p < 0.05) (Table 2).
Regarding alcohol intake, 73.6% reported having consumed alcohol, and the mean daily alcohol consumption was equal to 5.9 g (standard deviation – SD = 9.1 g). Regarding the type of alcohol, the mean daily consumption of beer was 76.2 mL (SD = 22.6 mL); of wine was 16.1 mL (SD = 5.0 mL); and of hard liquor was 2.9 mL (SD = 1.6 mL).
Participants with excess weight, in comparison to those without, presented higher frequencies of alcohol consumption between 1 to 4 days/week, and a higher mean of daily consumption of alcohol, beer, wine and liquor (p < 0.05) (Table 3).
In the multivariate analysis, the participants in the intermediate quartile of daily alcohol consumption (Q2 – PR: 0.87; 95% CI: 0.77-0.99) had a lower prevalence of excess weight in comparison to the lowest quartile. In the analysis by type of alcohol, participants in the intermediate quartiles of beer consumption (Q2 – PR: 0.84; 95% CI: 0.74-0.96); hard liquor consumption (Q3 – PR: 0.84; 95% CI: 0.74-0.94) and wine consumption (Q3 – PR: 0.85; 95% CI: 0.75-0.96) had a lower prevalence of excess weight in comparison to those in the lower quartiles. However, there was a significant tendency of increase in the prevalence of excess weight with the increase in beer consumption (tendency p = 0.038), which was not observed with the other types of alcohol (p ≥ 0.05) (Table 4).
The sensitivity analyses showed that when there was exclusion of: 1) participants with a previous medical diagnosis of obesity (n = 256); 2) those who gained ≥10 kilograms before the study (n = 203); or 3) a combination of both criteria plus participants who were on weight loss medication (n = 420), the results were as follows: a) an increase tendency of prevalence of excess weight with a higher intake of alcohol in exclusion criteria 1 and 3, and with a higher intake of beer in all exclusion criteria (tendency p < 0.05); b) the elimination of the negative association between intermediate wine consumption (Q2) and excess weight, and the appearance of a negative association between the intermediate consumption of hard liquor (Q3) and excess weight in all exclusion criteria (Table 5).



DISCUSSION
In this study, the high prevalence of alcohol consumption and excess weight among the participants was demonstrated. On the other hand, the means of total daily alcohol intake and of the types of alcohol were low. It is also noteworthy that the consumption of alcohol, beer, wine, and hard liquor in intermediate quartiles was negatively associated to excess weight. However, after the sensitivity analyses, such finding was maintained only for liquor, while the consumption of alcohol and beer showed a positive association to excess weight.
The portion of participants who consumed some type of alcoholic beverage (73.6%) was higher than those found for the global population (43%) and the Brazilian population in general (40%)36. On the other hand, the mean daily alcohol consumption (5.9 g) of our participants was below that of the world population, which was 33 g (equivalent to 300 mL of wine, 750 mL of beer, or 80 mL of liquor)36, and it was within the limits of intake recommended by the World Health Organization (10 to 12 g – equal to one 100-mL glass of wine, a 330-mL can of beer or a 30-mL dose of hard liquor)37. It is important to highlight that the quantity of daily alcohol consumption that is considered safe, and the recommendations on maximum daily intake of alcohol, vary within and between different countries, and there is no consensus regarding an intake limit37,38, which ranges from 10 to 14 g. A study recently published by the Global Burden of Disease (GBD)36 discourages the consumption of any quantity (“zero tolerance”).
In the analysis by type of alcohol, there was a higher mean of daily beer intake (76.2 mL), followed by wine (16.1 mL) and hard liquor (2.9 mL). Globally, hard liquor is the most frequently consumed type (44.8%), followed by beer (34.3%) and wine (11.7%). In the Americas, beer is the most consumed type of alcohol (53.8%), followed by hard liquor (31.7%) and wine (13.5%). In Brazil, the results are similar (beer: 62%; hard liquor: 34%; wine: 3%), which means Brazil ranks third in Latin American and fifth in the Americas regarding alcohol consumption36.
The fact that the participants of our study reported a high consumption of alcohol, but at a low daily quantity, can be considered a positive aspect in the participants’ behavior since a high daily alcohol intake can be harmful to one’s health36. Such results are likely to be explained by the high schooling level of participants because, in general, individuals with higher levels of education tend to have healthier habits39.
The intermediate consumption of alcohol (Q2; 1.18 to 3.3 g/day), beer (Q2; 8.1 to 31 mL/day), wine (Q3; 6.4 to 17.6 mL/day) and hard liquor (Q3; 1.1 to 2.5 mL/day) is negatively associated to excess weight, which suggests an inverse association, even though the tendency p is significant only in the analysis of daily beer consumption – therefore, one can infer a higher consumption of beer increases the prevalence of excess weight. Other studies present similar findings10,14,38,40.
Among the moderate and heavy consumers of alcohol, a larger fraction of the energy of alcohol may not be an available source of energy due to the induction of the microsomal ethanol oxidizing system (MEOS)40, which means that the calories contained in alcohol are less related to weight gain and more related to alcohol toxicity, such as hepatotoxicity or carcinogenesis38,40. Moreover, individuals with normal weight who consume a low or moderate quantity of alcohol can maintain their habits of consuming alcohol without gaining weight, which may have conferred this protective effect to the intermediate quartiles in our study38.
It is important to highlight that many cross-sectional, longitudinal and experimental studies evaluated by systematic reviews and meta-analyses demonstrated findings that differ from ours, which indicated a positive association between alcohol consumption and excess weight5,10,12,14. Alcohol has a high potential to interact with different aspects of body weight regulation. Therefore, other causes for controversial results might be the different confounding factors included in the data analysis38,40. Then, it is important to try to control these biases in epidemiological studies and consider, in addition to the known and widely used aspects (sex, age, physical activity, smoking, psychosocial factors, income and professional status), those aspects related to the pattern of consumption and intake of alcohol; previous body weight; and factors that may have previously influenced weight gain or loss, such as past illnesses, medications, caffeine consumption and individual genetic factors38,40.
Some of the potential biases mentioned are difficult to control in cross-sectional studies; hence, our study opted for sensitivity analyses in an attempt to mitigate such biases. In this way, participants with a previous medical diagnosis of obesity, those who gained 10 or more kilograms before the study, and those who were on weight control medications were excluded. After these procedures, only the intermediate consumption of hard liquor (Q3; 1.1 to 2.5 mL/day) kept its negative relation to excess weight.
Regarding wine and beer, after the sensitivity analysis, our findings (beer: 8.1 to 31 mL/day; wine: 6.4 to 17.6 mL/day) were different from the findings of other studies, which showed that a moderate consumption of beer (11 to 22 mL/day)41 and wine (5 to 13 mL/day)5,12 provided protection against weight gain. Among the possible explanations for this effect is the presence of polyphenols in wine and beer, which provide antioxidant and anti-inflammatory action, with the reduction of leukocyte adhesion molecules and inflammatory biomarkers, which improves lipid profile42, and establishes another favorable characteristic of these beverages.
It is important to highlight that this difference between our results and those from other investigations can be explained by the fact that in our analysis we did not differentiate between red and white wine - the former presents better protective effects due to the presence of more than 10 times the amount of phenolic compounds17,42. Regarding beer, our questionnaire did not consider different types of beer separately. A study conducted in Brazil analyzed the phenolic compounds in the different types of Brazilian beers and found that the physical-chemical attributes (density, refractive index, bitterness and ethanol content) influence the quantity of phenolics43. Moreover, this same study identified that the phenolic profile of Brazilian beers was different from that of European beers, with a low content of ferulic acid43, which could be a hypothesis to explain the lower protection offered by the moderate consumption of Brazilian beer in comparison to European beer5,10,12,14. However, it is worth mentioning that in our tendency analyses it was found that the more the beer consumption increased, the higher the prevalence of excess weight, which corroborates other studies12,14,15.
Hard liquor presents higher concentrations of alcohol and lower concentrations of polyphenols, which could mean a higher risk of excess weight43. However, in this study, a protective effect was observed in the intermediate quartile (Q3; 1.1 to 2.5 mL/day), which also is corroborated by other studies10,12. Such finding may be explained by the fact that alcohol is positively related to HDL-c42. In a meta-analysis of prospective studies, only a very low alcohol intake (between 0.1 and 5 g/day) was proven a protective factor for metabolic syndrome44. The consumers in the intermediate quartile (Q3) of hard liquor in this study had a median alcohol consumption of 3.2 g/day.
The Seguimiento Universidad de Navarra (SUN), in Spain, found that individuals who consume at least seven drinks per week of beer and hard liquor present a higher risk of overweight and obesity (OR: 1.32 – 95% CI; 1.00-1.74) in comparison to non-drinkers after multivariate data adjustment45. It is worth mentioning that the aforementioned study was a longitudinal analysis, which guarantees the causality of associations and thus strengthens our findings.
In this study, we found an increase tendency of the prevalence of excess weight with a higher daily alcohol consumption, which was also found by other studies1,9,10,14,16,18,46, even though there are studies who found a null or inverse association10,38,40.
Some of the studies showed a J-shaped curve in the analyses of alcohol consumption and its effects on health1,9,10,14,16,18,46. For some authors, the J-shaped pattern must be carefully evaluated, considering that, at times, this relation may occur due to an erroneous classification of consumption patterns. The lower risk among those with zero and moderate alcohol consumption may be due to the inclusion of individuals who already presented alterations and were advised to lower or eliminate alcohol use46. It is worth mentioning that our analyses were adjusted for diagnosis of chronic diseases that could potentially affect the consumption of alcohol. Recently, despite finding the J-shaped curve in the relation between alcohol consumption and ischemic diseases, in particular, the GBD36 argues that when mortality in general is evaluated, the curve loses is J shape, since the risk of death from other causes surpasses the protection offered by lower or intermediate alcohol use. As a recommendation, the GBD emphasizes that, regardless of the amount, alcohol consumption is harmful to human health among populations36.
Of all the participants, 40.8% presented excess weight. Although this is a high prevalence, it is lower than the one found in the general Brazilian population (55.7%)3 and in other national and international studies28,47,48,49,50, and it is similar to the prevalence found in the SUN cohort (38%)45, which shows samples with similar characteristics to our cohort.
These differences can be explained by the characteristics of the population from this study, which was mostly composed by young adults with high levels of income and education, which are protective factors against excess weight9. Individuals in a more favorable socioeconomic position are more likely to adopt healthy behaviors and have more access to healthcare39, even if they do present a higher consumption of alcohol15,25,27,51.
Although this study presents a privileged sample profile with individuals in a good socioeconomic position, such characteristic allowed detailing and depth in the questions of the questionnaire. A study with a similar sample has shown reliable and valid results in addition to a high retention rate45,52.
Regarding limitations in this study, we highlight the fact that the participants’ level of alcohol consumption was self-reported. There are, however, some authors who claim that in samples of communities where alcohol is legal, socially accepted, and seen as part of the traditional diet standard, the issue of incorrect self-reporting is minimized53. In the Spanish cohort SUN, it was shown that individuals with a higher level of education are less likely to inaccurately report alcohol consumption52,54. However, we acknowledge that these issues do not eliminate possible measurement errors inherent to the instrument used and to the specific characteristics of our population.
Some studies with adults and using the FFQ showed good reproducibility and validity in the measurement of alcohol consumption52,55,56, and the Spanish cohort SUN52 indicated that, even if there might be some inaccuracy in the information, such inaccuracy would not constitute a differential due to the higher level of education of the participants. Nevertheless, it is worth mentioning the estimation of alcohol intake in our study might also be a limitation, since the FFQ has not been specifically validated for alcohol consumption.
Another study limitation was the number of participants who were excluded for not completing the FFQ (n = 1,679), which reduced our sample. This may have increased the probability of type II statistical error; however, even after the multivariate analysis, significant associations were found between alcohol consumption and excess weight. Moreover, when comparing the participants and the excluded individuals in relation to sex, age, frequency of alcohol consumption, and excess weight, only one statistically significant difference was found. Among the excluded individuals, there were more males (39% vs. 31.3%; p < 0.001 as per Pearson\'s chi-square test), which means that this finding must be considered in the interpretation of our results, considering that men consume more alcohol and present a higher prevalence of excess weight than women in Brazil3.
As strengths of this study, we highlight the fact that there is still little data that effectively characterizes the dietary pattern of the Brazilian population, so the CUME Project can contribute in that sense. This study also has the advantage of better accuracy, obtained by describing the portions of alcohol beverages, which could minimize the memory bias and the expected underestimation of consumption55. Finally, with the sensitivity analyses, there was an attempt to mitigate the possibility of reverse causality biases, which cross-sectional studies are highly susceptible to, as previously mentioned, based on other studies recommending sensitivity analyses38,40.
It was concluded there was a high prevalence of alcohol consumption among the participants of CUME, and an inverse association of the intermediate quartiles of the intake of alcohol, beer, wine, and hard liquor with excess weight was observed. On the other hand, after sensitivity analyses, the high consumption of alcohol and beer was positively associated to excess weight, even if the intermediate hard liquor intake was negatively associated to excess weight. However, regarding this last finding, it is worth highlighting the need to reduce the widely accepted view that a low or moderate alcohol consumption is not harmful to one’s health5,7,12 and approach this proposition cautiously. Thus, those individuals with a low use or no use at all of alcohol cannot be encouraged to expect a substantial benefit from alcohol intake. On the other hand, for those who already consume alcohol, moderation is recommended to avoid binge drinking.
Additional longitudinal analyses (which consider the temporal sequence) are required to support these associations, especially in populations with a high variability of alcohol consumption, and evaluate the specific role of different types of alcoholic beverages.

REFERENCES
1. World Health Organization (WHO). Obesity and overweigh. Geneva: WHO; 2018.
2. Organisation for Economic Co-operation and Development (OECD) [serial on the Internet]. Obesity Update; 2014 [cited 2018 Sep 01]. Available from: http://www.oecd.org/health/Obesity-Update-2014.pdf
3. Brasil. Ministério da Saúde (MS). Vigitel Brasil 2018: vigilância de fatores de risco e proteção para doenças crônicas por inquérito telefônico: estimativas sobre frequência e distribuição sociodemográfica de fatores de risco e proteção para doenças crônicas nas capitais dos 26 estados brasileiros e no Distrito Federal em 2018. Brasília: MS; 2019.
4. Bahia L, Coutinho ES, Barufaldi LA, Abreu GA, Malhão TA, Souza CP, Araujo DV. The costs of overweight and obesity-related diseases in the Brazilian public health system: Cross-sectional study. BMC Public Health 2012; 12:440-447.
5. Poppitt SD. Beverage Consumption: Are Alcoholic and Sugary Drinks Tipping the Balance towards Overweight and Obesity? Nutrients 2015; 7(8):6700–6718.
6. Fazzino TL, Fleming K, Sher KJ, Sullivan DK, Befort C. Heavy Drinking in Young Adulthood Increases Risk of Transitioning to Obesity. Am J Prev Med 2017; 53(2):169-175.
7. Bezerra IN, Alencar ES. Association between excess weight and beverage portion size consumed in Brazil. Rev Saude Publica 2018; 52:21.
8. Reis NT, Rodrigues CSC. Nutrição clínica no alcoolismo. Rio de Janeiro: Rubio; 2003.
9. Kachani AT, Brasiliano S, Hochgraf PB. O impacto do consumo alcoólico no ganho de peso. Rev Psiquiatr Clín 2008; 35(Supl. 1):21-24.
10. Orea CS, González MAM, Rastrollo MB. Alcohol consumption and body weight: a systematic review. Nutr Rev 2011; 69(8):419-431.
11. Brandhagen M, Forslund HB, Lissner L, Winkvist A, Lindroos AK, Carlsson LM, Sjöström L, Larsson I. Alcohol and macronutrient intake patterns are related to general and central adiposity. Eur J Clin Nutr 2012; 66(3):305-313.
12. Traversy G, Chaput JP. Alcohol Consumption and Obesity: An Update. Curr Obes Rep 2015; 4(1):122–130.
13. Farinelli MA, Partridge SR, Roy R. Weight-related dietary behaviors in young adults. Curr Obes Rep 2016; 5:23–29.
14. Yeomans MR. Alcohol, appetite and energy balance: is alcohol intake a risk factor for obesity? Physiol Behav 2010; 100(1):82-89.
15. Costanzo S, Castelnuovo AD, Donati MB, Iacoviello L, Gaetano G. Wine, beer or spirit drinking in relation to fatal and non-fatal cardiovascular events: a meta-analysis. Eur J Epidemiol 2011; 26:833–850.
16. Bendsen NT, Christensen R, Bartels EM, Kok FJ, Sierksma A, Raben A, Astrup A. Is beer consumption related to measures of abdominal and general obesity? A systematic review and meta-analysis. Nutr Rev 2013; 71(2):67-87.
17. Lopez MTB, Rastrollo MB, Orea CS, Lopez MS, Montero AF, Gea A, González MAM. Different types of alcoholic beverages and incidence of metabolic syndrome and its components in a Mediterranean cohort. Clin Nutr 2013; 32(5):797-804.
18. Artero A, Artero A, Tarín JJ, Cano A. The impact of moderate wine consumption on health. Maturitas 2015; 80(1):3-13.
19. Chakraborty S. Analysis of NHANES 1999-2002 data reveals noteworthy association of alcohol consumption with obesity. Ann Gastroenterol 2014; 27(3):250-257.
20. Bernal RTI, Malta DC, Iser BPM, Monteiro RA. Método de projeção de indicadores das metas do Plano de Ações Estratégicas para o Enfrentamento das Doenças Crônicas não Transmissíveis no Brasil segundo capitais dos estados e Distrito Federal. Epidemiol Serv Saude 2016; 25(3):455-466.
21. Sabóia RF, Araújo AP, Barbosa JMA, Galvão CEP, Cruvel JMS, Ferreira SCN. Obesidade abdominal e fatores associados em adultos atendidos em uma clínica escola. Rev Bras Promoc Saude 2016; 29:259-267.
22. Cruz MF, Ramires VV, Wendt A, Mielke GI, Mesa JM, Wehrmeister FC. Simultaneity of risk factors for chronic non-communicable diseases in the elderly in Pelotas, Rio Grande do Sul State, Brazil. Cad Saude Publica 2017; 33(2):e00021916.
23. Domingos JBC, Jora NP, Carvalho AMP, Pillon SC. Consumo de álcool, sobrepeso e obesidade entre caminhoneiros. Rev Enferm UERJ 2010; 18:377-382.
24. Silva ABJ, Oliveira AVK, Silva JD, Quintaes KD, Fonseca VAS, Nemer ASA. Relação entre consumo de bebidas alcoólicas por universitárias e adiposidade corporal. J Bras Psiquiatr 2011; 60(3):210-215.
25. Cibeira GH, Muller C, Lazzaretti R, Nader GA, Caleffi M. Consumo de bebida alcoólica, fatores socioeconômicos e excesso de peso: um estudo transversal no sul do Brasil. Cien Saude Colet 2013; 18(12):3577-3584.
26. Carvalho CA, Fonseca PCA, Barbosa JB, Machado SP, Santos AM, Silva AAM. Associação entre fatores de risco cardiovascular e indicadores antropométricos de obesidade em universitários de São Luís, Maranhão, Brasil. Cien Saude Colet 2015; 20(2):479-490.
27. Domingos ALG, Miranda AES, Pimenta AM, Hermsdorff HHM, Oliveira FLP, Santos LC, Lopes ACS, González MAM, Bressan J. Cohort Profile: The Cohort of Universities of Minas Gerais (CUME). Int J Epidemiol 2018; 47(6):1-10.
28. Schmidt MI, Duncan BB, Mill JG, Lotufo PA, Chor D, Barreto SM, Aquino EM, Passos VM, Matos SM, Molina MC, Carvalho MS, Bensenor IM. Cohort profile: Longitudinal Study of Adult Health (ELSA-Brasil). Int J Epidemiol 2015; 44(1):68-75.
29. Henn RL, Fuchs SC, Moreira LB, Fuchs FD. Development and validation of a food frequency questionnaire (FFQ-Porto Alegre) for adolescent, adult and elderly populations from Southern Brazil. Cad Saude Publica 2010; 26(11):2068–2079.
30. Miranda AES, Ferreira AVM, Oliveira FLP, Hermsdorff HHM, Bressan J, Pimenta AM. Validação da síndrome metabólica e de seus componentes autodeclarados no estudo CUME. Rev Min Enferm 2017; 21:e1069.
31. Núcleo de Estudos e Pesquisas em Alimentação. Tabela Brasileira de Composição de Alimentos. Campinas: Universidade Estadual de Campinas; 2011.
32. United States Department of Agriculture (USDA) [serial on the Internet]. National Nutrient Database for Standard Reference Release. USDA Food Composition Databases; 2018 [cited 2019 Mar 20]. Available from: https://ndb.nal.usda.gov/ndb/foods
33. Willett WC, Colditz GA. Approaches for conducting large cohort studies. Epidemiol Rev 1998; 20:91-99.
34. World Health Organization (WHO). Global recommendations on physical activity for health. Geneva: WHO; 2010.
35. Monteiro CA, Cannon G, Moubarac JC, Levy RB, Louzada MLC, Jaime PC. The UN Decade of Nutrition, the NOVA food classification and the trouble with ultra-processing. Public Health Nutr 2018; 21(1):5-17.
36. Global Burden of Disease. Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2018; 392(22):1015-1035.
37. Kalinowski A, Humphreys K. Governmental standard drink definitions and low-risk alcohol consumption guidelines in 37 countries. Addiction 2016; 111(7):1293-1298.
38. Wang L, Lee MI, Manson JAE, Buring JE, Sesso HD. Alcohol consumption, weight gain, and risk of becoming overweight in middle-aged and older women. Arch Intern Med 2010; 170:453-461.
39. Faleiro JS, Giatti L, Barreto SM, Camelo LV, Griep RH, Guimarães JMN, Fonseca MJM, Chor D, Chagas MCA. Lifetime socioeconomic status and health-related risk behaviors: the ELSA-Brazil study. Cad Saude Publica 2017; 33(3):e00017916.
40. Suter PM. Is alcohol consumption a risk factor for weight gain and obesity? Crit Rev Clin Lab Sci 2005; 42(3):197-227.
41. Romeo J, Gross MG, Wärnberg J, Díaz LE, Marcos A. Does beer have an impact on weight gain? Effects of moderate beer consumption on body composition. Nutr Hosp 2007; 22(2):223-228.
42. Huang J, Wang X, Zhang Y. Specific types of alcoholic beverage consumption and risk of type 2 diabetes: A systematic review and meta-analysis. J Diabetes Investig 2017; 8(1):56-68.
43. Nunes NM, Brito TC, Fonseca ND, Aguiar PF, Monteiro M, Perrone D, Torres AG. Phenolic compounds of Brazilian beers from different types and styles and application of chemometrics for modeling antioxidant capacity. Food Chem 2016; 15(199):105-113.
44. Sun K, Ren M, Liu D, Wang C, Yang C, Yan L. Alcohol consumption and risk of metabolic syndrome: a meta-analysis of prospective studies. Clin Nutr 2014; 33(4):596-602.
45. Orea CS, Rastrollo MB, Cordoba JMN, Gortari FJB, Beunza JJ, González MAM. Type of alcoholic beverage and incidence of overweight/obesity in a Mediterranean cohort: the SUN project. Nutrition 2011; 27(7-8):802-808.
46. Gaetano G, Costanzo S. Alcohol and Health. Praise of the J Curves. JACC 2017; 70(8):923-925.
47. Garcia LP, Freitas LRS. Consumo abusivo de álcool no Brasil: resultados da Pesquisa Nacional de Saúde 2013. Epidemiol Serv Saude 2015; 24(2):227-237.
48. O\'Donovan G, Stamatakis E, Hamer M. Associations between alcohol and obesity in more than 100 000 adults in England and Scotland. Br J Nutr 2018; 119(2):222-227.
49. Rezende FAC, Rosado LEPL, Ribeiro RCL, Vidigal FC, Vasques ACJ, Bonard IS, Carvalho CR. Body Mass Index and Waist Circumference: Association with Cardiovascular Risk Factors. Arq Bras Cardiol 2006; 87(6):728-734.
50. Lima NP, Horta BL, Motta JVDS, Valença MS, Oliveira V, Santos TV, Gigante DP, Barros FC. Evolution of overweight and obesity into adulthood, Pelotas, Rio Grande do Sul State, Brazil, 1982-2012. Cad Saude Publica 2015; 31(9):2017–2025.
51. Pham CV, Tran HTD, Tran NT. Alcohol Consumption and Binge Drinking Among Adult Population: Evidence from the CHILILAB Health and Demographic Surveillance System in Vietnam. J Public Health Manag Pract 2018; 24(Sup. 2):S67-S73.
52. Gea A, Rastrollo MB, Toledo E, Lopez MG, Beunza JJ, Estruch R, González MAM. Mediterranean alcohol-drinking pattern and mortality in the SUN (Seguimiento Universidad de Navarra) Project: a prospective cohort study. Br J Nutr 2014; 28;111(10):1871-1880.
53. Lourenço S, Oliveira A, Lopes C. The effect of current and lifetime alcohol consumption on overall and central obesity. Eur J Clin Nutr 2012; 66(7):813-818.
54. Gómez MS, Fuente C, Vázquez Z, Irala J, González MAM. Cohort profile: the \'Seguimiento Universidad de Navarra\' (SUN) study. Int J Epidemiol 2006; 35(6):1417-1422.
55. Lopes ACS, Caiaffa WT, Mingoti SA, Costa MFFL. Food intake in epidemiological studies. Rev Bras Epidemiol 2003; 6(3):209-219.
56. Nakahata NT, Takada AN, Imaeda N, Goto C, Kuwabara KH, Niimura H, Arai Y, Yoshita K, Takezaki T. Validity of a food frequency questionnaire in a population with high alcohol consumption in Japan. Asia Pac J Clin Nutr 2016; 25:195-201.