Background: Malignancy is the leading cause of death in Hong Kong, and lung cancer tops the list of all cancer deaths.
Methods: A cohort of clients aged ≥65 years, enrolled at 18 elderly health centres in Hong Kong from 2000 to 2003, was followed up prospectively through linkage with the territory-wide death registry for causes of death until 31 December 2008, using the identity card number as unique identifier. All subjects with suspected cancer, significant weight loss of >5% within past 6 months or obstructive lung disease at the baseline were excluded.
Results: After a total of 423 061 person-years of follow-up, 932, 690 and 1433 deaths were caused by lung cancer, other tobacco-related malignancies and non-tobacco-related malignancies, respectively. Body mass index (BMI) was independently (and negatively) associated with death from lung cancer after adjustment for other baseline variables, whereas there was only a minor or no effect for other smoking-related malignancies and non-tobacco-related malignancies. Obesity with BMI ≥30 [adjusted hazard ratio (HR), 0.55, 95% confidence interval (CI) 0.38-0.80] was associated with reduced lung cancer mortality, which was more prominent than the opposing effect of underweight (adjusted HR, 1.38, 95% CI 1.05-1.79). Consistent effects of BMI were observed after stratification into never-smokers and ever-smokers and in sensitivity analysis after excluding deaths within the first 3 years.
Conclusion: Obesity was associated with lower lung cancer mortality in this prospective cohort analysis. As the effect was rather specific for lung cancer, further studies are indicated to explore the underlying mechanism.