Television viewing and abdominal obesity in women according to smoking status: results from a large cross-sectional population-based study in Brazil Hábito de assistir televis o e obesidade abdominal entre mulheres fumantes: resultados de um estudo transversal de base populacional no Brasil

OBJECTIVE: To investigate the associations between television viewing and abdominal obesity (AO) in Brazilian women, according to smoking status. METHODS: Data of 13,262 adult women (18-49 years) from the 2006's Demographic Health Survey, a cross-sectional household study with complex probabilistic sample and national representativeness, were analyzed. AO, defined as waist circumference ≥ 80.0 cm, was the outcome. Television viewing frequency (≥ 5 times/week, 1-4 times/week, < 1 time/week) was the main exposure variable, and smoking status (yes or no) the main co-variable. Prevalence ratios were estimated using Poisson regression models separately for smokers and non-smokers. RESULTS: A statistically significant interaction term was observed between smoking status and television viewing (p < 0.05). Prevalence of AO among smokers who reported television viewing ≥ 5 times/week amounted to 59.0%, higher than the 35.0% for those with < 1 time/week television viewing (p-value = 0.020). The values for non-smokers were 55.2% and 55.7%, respectively. Smokers with television viewing ≥ 5 times/week were 1.7 times (95% CI: 1.1 - 2.5) more likely to pre-sent AO, compared to those who reported a frequency < 1 time/week. There was no significant association among non-smokers. CONCLUSIONS: Television viewing ≥ 5 times/week may increase the prevalence of AO among women who smoke. More detailed information on media use, as hours per day, may offer better estimates. OBJETIVO: Investigar a associa o entre a frequência assistindo televis o e obesidade abdominal (OA) entre mulheres brasileiras, segundo o hábito de fumar. MéTODOS: Foram analisados os dados de 13.262 mulheres adultas (18-49 anos) estudadas na Pesquisa Nacional de Demografia e Saúde (PNDS-2006), um estudo transversal, com amostragem probabilística complexa, de representatividade nacional. OA, definida como circunferência da cintura ≥ 80.0 cm, foi considerada como desfecho. A frequência assistindo televis o (≥ 5 vezes/semana, 1-4 vezes/semana, < 1 vez/semana) foi a principal variável de exposi o e o hábito de fumar (sim ou n o) a principal covariável. Foram estimadas raz es de prevalência por meio de modelos de regress o de Poisson, para fumantes e n o fumantes separadamente. RESULTADOS: Observou-se intera o estatisticamente significante entre frequência assistindo televis o e hábito de fumar (p < 0,05). A prevalência de OA entre mulheres fumantes que assistiam televis o ≥ 5 vezes/semana foi de 59,0%, e maior do que 35,0% entre aquelas que assistiam televis o < 1 vez/semana (p-valor = 0,02). Os valores de OA