The association of smoking with different eating and dietary behaviours: a cross-sectional analysis of 80,296 UK adults

Background and Aims Smokers typically have a lower body mass index (BMI) than non-smokers, while smoking cessation is associated with weight gain. In pre-clinical research, nicotine in tobacco smoking suppresses appetite and influences subsequent eating behaviour; however, this relationship is unclear in humans. This study measured associations of smoking with different eating and dietary behaviours. 

Design A cross-sectional analysis of data from health assessments conducted between 2004-2022. Setting An independent healthcare-based charity within the UK. Participants 80,296 men and women (mean±standard deviation [SD]: age, 43.0±10.4 years; BMI, 25.7±4.2 kg/m2; 62.5% male) stratified into two groups based on their status as a smoker (n=6,042; 7.5%) or non-smoker (n= 74,254; 92.5%). 

Measurements Smoking status (self-report) was the main exposure, while the primary outcomes were selected eating and dietary behaviours. Age, sex and socioeconomic status (index of multiple deprivation [IMD]) were included as covariates and interaction terms, while moderate-to-vigorous exercise and sleep quality were included as covariates only. 

Findings Smokers had lower odds of snacking between meals and eating food as a reward or out of boredom versus non-smokers (all odds ratio [OR]≤0.82; P<0.001). Furthermore, smokers had higher odds of skipping meals, going more than 3 hours without food, adding salt and sugar to their food, overeating, and finding it hard to leave something on their plate versus non-smokers (all OR≥1.06; P≤0.030). Additionally, compared with non-smokers, smoking was associated with eating fried food more times per week (rate ratio [RR]=1.08; P<0.001) and eating fewer meals per day, eating sweet foods between meals, and eating dessert on fewer days per week (all RR≤0.93; P<0.001). Several of these relationships were modified by age, sex, and IMD. 

Conclusions Smoking appears to be associated with eating and dietary behaviours consistent with inhibited food intake, low diet quality, and altered food preference. Several of these relationships are moderated by age, sex, and socioeconomic status.