ABC | Volume 113, Nº4, October 2019

Original Article Francisco et al. Smoking and unhealthy diet Arq Bras Cardiol. 2019; 113(4):699-709 Results Mean age of the sample was 36.4 years (95%CI: 36.1–36.6); the majority was women (53.0%) and young adults (59.4%). The prevalence of the co-occurrence of the risk factors was 8.6% (95% CI: 7.9–9.3). Table 2 shows the prevalence of smoking and an inadequate diet as well as associations with the other variables. No associations were found between smoking and BMI, physical inactivity, hypertension, diabetes or dyslipidemia (p > 0.05). All variables were associated with unhealthy eating, except for having a private health insurance plan (p = 0.102) and BMI (p = 0.196). In the simple analysis, differences were found with regard to the region of the country, sex, skin color/ethnicity, education and health insurance (p < 0.05). Associations were found between the co-occurrence of risk factors and both binge drinking (p < 0.001) and self-rated health (p < 0.001). Higher frequencies of the co-occurrence of risk factors were found among individuals residing in the southern and southeastern regions of the country compared to those residing in the central western region. Higher frequencies were also found among individuals who did not have private health insurance, binge drinkers and individuals who reported their health to be fair, poor or very poor at the time of the interview. After controlling for sex and age, the prevalence of the co-occurrence of risk factors was lower among individuals who resided in the northern and northeastern regions of the country, women, individuals with brown skin color, those with a spouse and those with excess weight. The prevalence of the two risk factors also reduced significantly with higher education levels (p < 0.001) (Table 3). Table 4 shows the hierarchical Poisson regression model of factors associated with the co-occurrence of smoking and unhealthy eating. The prevalence of both factors was lower in the northeasternregionofthecountry(PR=0.67;95%CI:0.54–0.83) andhigher in the southern region (PR=1.40; 95%CI: 1.11–1.77) compared to the central western region. The prevalence was approximately 40% lower among women and was also lower among individuals with brown skin and those who lived with a spouse. The prevalence reduced significantly with higher education levels and was approximately 16% higher among individuals without private health insurance (PR = 1.16; 95% CI: 1.05–1.27) after controlling for region of residence and other socio-demographic factors. Regarding behaviors and health conditions, excess weight was inversely associated with the co-occurrence of both risk factors. The prevalence was higher among those who considered their health to be fair or poor/very poor. Moreover, a strong, independent, statistically significant association was found between binge drinking and the co-occurrence of the risk factors considered (PR = 3.22; 95% CI: 2.70–3.85). Discussion The prevalence of the co-occurrence of smoking and an unhealthy diet was 8.6%. In a study conducted in England with a population aged 16 to 64 years, the prevalence of the co-occurrence of smoking and inadequate diet (measured by the consumption of beans, vegetables and fruits) was 25.5% among men and 23.6% among women. 21 In this study, which considered four lifestyle risk factors, the prevalence of smoking was 28.0% and insufficient intake of beans, fruits and vegetables was represented by the failure to have five portions of these foods, as recommended. The divergence in frequencies may be explained by cultural diversity and the diversity of eating habits in different populations. In a study conducted in the city of Botucatu (state of São Paulo, Brazil), Berto et al. 22 investigated the association between smoking and other behavioral risk factors in adults and found the co-occurrence of smoking and low intake of fresh fruits and vegetables (12.9% and 12.3% among men and 5.8% and 5.1% among women, respectively). In a study conducted with adults in the city of Florianópolis (state of Santa Catarina), the prevalence of the co-occurrence of smoking and an inadequate diet was 3.5%; inadequate diet was considered the reported intake of fruits and vegetables ≤5 days per week. 13 In an American study with an adult population (≥20 years), smokers had a diet of poor quality, with less intake of fruits, vegetables, dairy products and whole grains as well as a greater percentage of energy from solid fats, alcohol and added sugar. 23 A study conducted with adult smokers found that fruits and vegetables, noncaffeinated beverages, sweets and dairy products worsened the sensory attributes of cigarettes, whereas meats, alcoholic beverages and caffeinated drinks improved the sensory attributes. 24 Haibach et al. 25 found that smokers who ingested more fruits and vegetables had lower levels of nicotine dependence and greater occurrence of smoking cessation. To make the dietary indicator employed in the present study, we considered whole milk a marker of healthy diet. The results of epidemiological studies evaluating dietary sources of saturated fat have revealed the lack of an association or beneficial effects of dairy products on cardiovascular diseases. 26,27 Considering the gaps in the scientific literature on the association between milk consumption and health, Lamarche et al. 27 report the need to investigate whether whole milk and skim milk have different effects on health. Data from the 2008-2009 Brazilian Family Budget Survey revealed that the adult population has low consumption of dairy products (100 g/ml per day), 28 but high consumption of red and processed meats (90 g per day), with more than 80% of the participants exceeding the limit recommended by the World Cancer Research Fund (300 g per week). 29 It should be stressed that the Vigitel survey did not address the consumption of processed meats or quantify the consumption of red meat. Therefore, only the frequency of weekly consumption was considered in the present study, regardless of the presence of visible fat. A previous study evaluating the attributable fraction of cancer in the adult population due to different exposures found that red meat is a risk factor for colon and rectal cancer when ingested at a rate of more than 70 g/day. 30 In the present investigation, the co-occurrence of risk factors was higher in the southern region of Brazil and lower in the northeastern region than in the central western region. Reduced prevalence of smoking was found in all regions of the country between 2006 and 2013, but with the highest rates found in the South and the lowest in the 702