Interaction between current smoking, leanness, and physical inactivity in the prediction of hip fracture.

To study the association between smoking habits and the incidence of hip fracture, adjusted for leanness and physical inactivity, a cohort study with 3 years follow-up was conducted. Subjects were 34,856 adults aged 50 years or older who attended a health screening in Nord-Trøndelag County in Norway in 1984-1986 (91% of eligible subjects in 1986, n = 38,356). Of these, 421 suffered a hip fracture during the years 1986-1989. Using Cox regression models, the relative risk (with 95% confidence interval) of suffering a hip fracture for female smokers versus nonsmokers was 1.5 (1.0-2.4). These results refer to females when the female body mass index (BMI) was set at 25 kg/m2 in the female model (the mean BMI for the smoking female population in this study). Among thinner females, however, smoking had a much stronger effect. For instance, if the female BMI was set at 20 kg/m2, the relative risk was 3.0 (1.8-5.0). The relative risk of hip fracture for male smokers versus nonsmokers was 1.8 (1.2-2.9) irrespective of BMI. Smoking is associated with incidence of hip fracture in both sexes and also after adjusting for body mass index and physical inactivity (the effect of physical inactivity was adjusted for self-reported ill health because ill health was included in the model). For lean females, the association with current smoking was large, as large as if they added 10 years to their age.