Obesity and influenza associated mortality: Evidence from an elderly cohort in Hong Kong

Objective: Obesity was not identified as a risk factor for influenza until the recent 2009 H1N1 pandemic. Based on a cohort of 66,820 subjects aged 65. years and over with the follow-up period from July 1998 to December 2010 in Hong Kong, we assessed the modifying effect of obesity on mortality risks specifically attributable to influenza infections (termed as "influenza associated mortality risks"). Methods: A Cox proportional model with time dependent covariates was adopted to assess the hazard ratio of mortality in each obesity group when influenza activity increased 10% in the community. Results: Hazard ratio of influenza-associated all-cause mortality was 1.081 (95% confidence interval 1.013, 1.154), 1.047 (1.012, 1.084), 0.981 (0.936, 1.028), 1.018 (0.980, 1.058) and 1.062 (0.972, 1.162) in the underweight, normal, overweight, moderate obesity and severe obesity groups, respectively. A similar U shape pattern across the obesity groups was also observed in influenza associated mortality risks of respiratory diseases, pneumonia and influenza. This pattern was more evident among ever smokers, although the influenza effect estimates in each obesity group had overlapping confidence intervals. Conclusion: There is some but limited evidence to suggest that underweight and obesity were associated with higher mortality risks of influenza in old population.