Fatores associados ao consumo regular de refrigerante não dietético em adultos de Pelotas , RS Factors associated with regular non-diet soft drink intake among adults in Pelotas , Southern Brazil

MÉTODOS: Estudo transversal de base populacional com 972 adultos (20 a 69 anos) do município de Pelotas, RS, realizado em 2006. A freqüência de consumo nos 12 meses anteriores à pesquisa foi medida por meio da pergunta: “em geral desde o <Mês> do ano passado, quantas vezes tu tomaste refrigerante não dietético?”. As respostas categorizadas foram dicotomizadas para fi ns de análise. Foi considerado consumo regular de refrigerante não dietético a freqüência de cinco ou mais vezes por semana. A associação entre o desfecho e variáveis demográfi cas, socioeconômicas, comportamentais e nutricionais foi analisada pelo teste qui-quadrado para heterogeneidade e tendência linear e a análise multivariável foi realizada por meio de regressão de Poisson, com variância robusta.


INTRODUCTION
Soft drink consumption has become increasingly prevalent in recent decades.Many authors 13,18,21 have associated soft drink overconsumption with increased prevalence of obesity and its consequences.Review 13 and meta-analysis studies 18 have showed that the growth in the consumption of sugar-sweetened beverages is associated with increased energy intake, weight gain, obesity and diabetes.In a systematic review in 2008, Wolff and Dansinger 22 argued that although observational studies support the hypothesis that consumption of soft drinks cause obesity, there are few clinical trials to confi rm this hypothesis.While more research studies are needed, the World Health Organization (WHO) 21 emphasizes that there is suffi cient evidence to discourage consumption of soft drinks in a healthy diet.
The WHO recommendation has been a priority of public health policies in many countries.The recommended maximum calorie intake is 10% of total dietary energy intake. 21Soft drinks have been banned from schools in England and France and their sale has been restricted or prohibited in the United States (Los Angeles, Philadelphia and Miami). 18The Dietary Guide for the Brazilian Population a recommends a reduction of at least one-third the current consumption of these foods as a strategy for healthy eating.
The identifi cation of factors associated with soft drink consumption is a fi rst step to approach high-risk groups.Previous studies 8,17 have pointed out gender, low education and age as major determinants of soft drink consumption.Recent studies 7,18,23 have also showed that dietary quality may be associated with the consumption of soft drinks.Compared to other countries such as the United States, 7,17 and European countries, 19 there are few data in Brazil about soft drink consumption and factors associated.The most recent Family Budget Survey conducted in Brazil b showed a signifi cant increase in the consumption of soft drinks (400%) compared to the 1970s.The consumption of soft drinks may be even higher since the Family Budged Survey data included only consumption at home.The Telephone-Based Surveillance of Risk and Protective Factors for Chronic Diseases (Vigitel) 2007 c showed that the rate of adults reporting regular consumption of soft drinks diet ranged from 21.0% in Aracaju, Northeastern Brazil, to 31.7% in Porto Alegre, Southern Brazil, and 38.4% in Macapá, Northern Brazil.The consumption rate was higher among men and decreased with age and education.
The present study aimed to assess factors associated with non-diet soft drink consumption in adults.

METHODS
A cross-sectional population-based study was conducted in the city of Pelotas, Southern Brazil.Pelotas is located in the extreme south of the state of Rio Grande do Sul with about 340,000 inhabitants.According to the Instituito Brasileiro de Geografi a e Estatística (IBGE -Brazilian Institute of Geography and Statistics), there are 408 urban household census tracts, which are arranged in a "spiral" from downtown to the districts.Of 404 census tracks including households, 40 were randomly selected for inclusion in the study, covering the city's entire urban area.We generated a random starting point for each track drawn for the selection of households to be included in the study.The fi rst household was selected and then the next ones were systematically selected by skipping every seven households, up to 15 households were included in each census track.There were selected 600 households for the study.All residents aged 20 and 69 years were sampled, except those with severe physical or mental disabilities that would prevent them from answering the questionnaire.
We performed two estimates for the sample size.The sample size to assess the prevalence of soft drink consumption was estimated at 35.0% for a margin of error of 4.0%, 95% confi dence interval, an addition of 10% for losses and refusals and a design effect of 1.5.It was estimated a sample of 901 respondents.For the assessment of the association between soft drink consumption and independent variables, the following parameters were used for estimating the sample size: 95% confi dence level, 80% power, exposure prevalence ranging from 10% to 90%, prevalence ratio of 1.7, addition of 10% for non-respondents and 15% for controlling for confounders and a design effect of 1.5.The estimated sample consisted of 990 respondents.
Information on demographic and socioeconomic characteristics and health and nutritional status were collected using a standard questionnaire with 100 questions.The variable skin color was reported by interviewer observation and socioeconomic status was classifi ed according to the Associação Brasileira de Empresas de Pesquisa (Brazilian Association of Research Companies) criteria.d The level of physical activity during leisure time was assessed using the long version of the International Physical Activity Questionnaire (IPAQ). 3A cutoff of 150 minutes per week was set to classify individuals as physically active or inactive. 3Nutritional status was evaluated using body mass index (BMI), which was calculated based on self-reported weight and height and classifi ed according to WHO criteria. 20e outcome, i.e., regular consumption of non-diet soft drinks, was investigated through the following question: "In general, since the <Month> in the past year, how often did you have a non-diet soft drink?."The answer options were: (a) once or less than once a month; (b) two to three times a month; (c) once or twice a week; (d) three to four times a week; (e) fi ve or more times a week.For the analysis, the outcome was dichotomized and positive respondents were those who reported consuming non-diet soft drinks fi ve or more times a week (regular use).Consumption was evaluated over a 12-month period prior to the interview.
Both male and female interviewers aged 18 years or more with a minimum requirement of complete high school education attended theory and practical sessions of a 20-hour training program.Interviewers were consistently supervised during the fi eldwork.The supervisors, who were physical education and physical therapy undergraduate students, reviewed the questionnaires and carried out quality control checks through re-visits to 25% of the sample.Data obtained from questionnaires were double entered using EpiInfo 6.0 and the analysis was conducted using Stata 9.0.
The crude analysis tested the relationship between the outcome and gender, age, skin color/ethnicity, socioeconomic status, smoking, leisure-time physical activity, nutritional status, regular consumption of fruits and vegetables (fi ve or more times a week) and weekly consumption of snack foods (burgers and hot dogs) (once or more a week).The chi-square test was used for testing heterogeneity and linear trend.Multivariate analysis was performed through Poisson regression with robust variance.The sampling effect was taken into consideration in the analysis using Stata svy commands.The level of signifi cance was set at 5%.

RESULTS
Of 514 households selected, 1,062 adult residents were eligible to participate in the study.Of them, 972 answered the questionnaire (nonresponse rate of 9.3%).The independent variable with most missing values was BMI as 87 respondents did not know their weight and/or height.
The study sample was characterized by: 57% were female; 26.1% aged between 20 and 29 years; 41.5% were categorized in class C; 82% were white; 49.4% have never smoked; 69.8% were sedentary; 51.3% were overweight or obese; 20.9% reported regular consumption of fruits and vegetables; and 36.6% reported snack food consumption on a weekly basis (Table 1).
Regular consumption of non-diet soft drink was 20.4% (95% CI 17.9;22.9).The design effect, considering the census track as the sampling unit, was 1.6.Additional analyses showed that the design effect using the household as the sampling unit was 1.4 for regular consumption of non-diet soft drinks.
Table 2 shows the frequency of consumption of regular non-diet soft drinks and the prevalence ratios for the independent variables.There were no statistically signifi cant differences in age, socioeconomic status, skin color/ethnicity, leisure-time physical inactivity, nutritional status and regular consumption of fruits and vegetables.In the adjusted analysis, males, current smokers, and those who reported snack food consumption had the highest prevalence of regular consumption of non-diet soft drinks.After adjusting for potential confounders, it was found that the variables regular consumption of fruits and vegetables (inverse association) and weekly consumption of snack foods (direct association) among females and smoking and consumption of snack foods (direct association) among males remained associated with regular consumption of non-diet soft drinks (Tables 3 and 4).

DISCUSSION
The study sample was representative of adults aged 20 to 69 years living in the city of Pelotas in view of the high proportion of respondents and low loss and refusal rate.One of the study limitations is that comparisons with results from other studies should be treated with caution due to different sampling and data collection procedures and frequency of use classifi cations.Reverse causality bias may be present in the association between frequency of non-diet soft drink consumption and nutritional status as obese overweight may modify their consumption because of their nutritional status.In addition, there may be recall error as the period of consumption studied was 12 months prior to the interview.
A fi fth (20.4%) of the adults studied in Pelotas regularly consumed non-diet soft drinks.The design effects for both the census track and the household were very close to those obtained in the sample size estimation and controlled for in the statistical analyses.
The rate of non-diet soft drink consumption found in this study is much lower than that reported in the United States.In 2007 Dhingra et al 5 found that 35% of adults in the Framingham Study consumed non-diet soft drinks at least once a day.Our fi ndings are similar to that reported in a Brazilian study on risk factors for noncommunicable diseases (Vigitel).c Corroborating other studies, 10,17 the present study found higher rates of non-diet soft drink consumption among men, and no signifi cant differences were seen regarding age, socioeconomic status and skin color/ethnicity.However, other authors reported differences in age, 10,15 socioeconomic status 4 and color skin/ethnicity. 4Among current smokers the rate of non-diet soft drink consumption was signifi cantly higher, which is consistent with other studies. 5,11,15This fi nding confi rms the trend that a health risk behavior is associated with a range of other health risk behaviors. 5,11 did not find any association between excess weight and non-diet soft drink consumption, which contrast with the results from recent reviews and metaanalyses. 13,18A recent systematic literature review 13 has pointed a positive association between higher consumption of sugar-sweetened beverages and weight gain and obesity in children and adults.In a metaanalysis including 88 studies, Vartanian et al 18 found a strong association between soft drink consumption and increased energy intake and body weight.Possibly the cross-sectional design of our study was a limiting factor in the analysis of this association.
One possible explanation for weight gain resulting from overconsumption of sugar-sweetened beverages is different physiological effects of energy intake from solid and liquid foods on satiety.Thus, carbohydrates form liquid foods would promote a higher positive energy balance.Possibly because liquid foods cause less gastric distension and have faster transit time, the body would not properly "identify" the energy contained in these foods with incomplete compensation for energy consumed in liquid form. 1,6ntrasting with previous studies, 11,15 no signifi cant association was seen between regular consumption of non-diet soft drinks and physical inactivity in our study.However, this result should be treated with caution as these studies used different instruments and criteria for classifi cation of physical activity levels.
A signifi cant association was found between weekly consumption of snack foods and regular consumption of non-diet soft drinks.Other authors 2,5,12,16 have argued that in general people who regularly consume soft drinks also eat nutritionally poor diets because these beverages would stimulate their appetite for non-nutritious foods.
A study 16 showed that individuals who consume more soft drinks have diets with a higher glycemic index, confi rming the assumption that intake of foods with high glycemic index such as soft drinks would stimulate the intake of other similar foods. 12Another study have showed that consumption of soft drinks is associated with the consumption of burgers and pizza. 2 Regular consumption of fruits and vegetables was inversely associated with that of non-diet soft drinks, especially in females.A similar association was reported by Marshall et al 14 in U.S. children, by Figueiredo et al 9 in adults in São Paulo and in young adults in Japan. 23 conclusion, the consumption of non-diet soft drinks in adults in Pelotas is high, especially among males, current smokers and those who eat snack foods every week and do not regularly consume fruits and vegetables.People should reduce their consumption of non-diet soft drinks as it is associated with increased energy intake of low nutrition foods and other unhealthy eating habits as well.
The study was approved by the Research Ethics Committee of the Escola Superior de Educação Física da Universidade Federal de Pelotas on 01/17/2008 (protocol No. 005/2008).Data were collected after respondents signed an informed consent form.

Table 2 .
Factors associated with regular consumption of non-diet soft drinks.Pelotas, Southern Brazil, 2006.(n=972) a All variables were adjusted for b Wald's test for heterogeneity c Wald´s test for linear trend

Table 4 .
Factors associated with regular consumption of non-diet soft drinks in men.Pelotas, Southern Brazil, 2006.(n=418) a All variables were adjusted for b Wald's test for linear trend c Wald's test for heterogeneity