Aim: To assess the effect of smoking uptake on body mass index (BMI), waist circumference and height during adolescence.
Design: Five-year longitudinal study.
Setting: Thirty-six schools in South London, England, selected by a stratified random sampling procedure designed to ensure ethnic and socio-economic diversity.
Participants: A total of 5863 students took part in the HABITS (Health and Behaviour in Teenagers) Study between 1999 and 2003.
Measurements: Self-reported smoking behaviour, saliva cotinine concentrations and measured weight, waist circumference and height were obtained, along with information on gender, ethnicity, socio-economic deprivation, pubertal status, self-reported exercise and dieting. Students were examined annually from school year 7 (ages 11-12) to year 11 (ages 15-16), with response rates ranging from 74 to 84%. A total of 2665 never smokers at year 7 with complete data for years 7 and 11 were included in the analyses.
Findings: Adjusting for year 7 BMI and other potential confounders, regular smokers (more than six cigarettes a week) at year 11 had significantly lower BMI (P = 0.002) than other students. Smokers defined by a cotinine above 15 ng/ml also had lower BMI (P < 0.0001). Waist circumferences were lower in regular smokers (P = 0.014) and cotinine-defined smokers (P < 0.011). No consistent association was found between smoking and height. The adjusted difference in weight between regular smokers and other students amounted to 1.8 kg (95% CI, 0.52-3.17) for an average-height student.
Conclusion: Taking up regular smoking during adolescence may result in a lower BMI, but the effect is small and of uncertain significance.