Content area
Full Text
A cohort of 42,655 clients that were first registered with the Elderly Health Service in 2000 were followed prospectively through the tuberculosis (TB) notification registry until the end of 2002. A total of 286 active TB cases (186 culture confirmed) were identified. The annual TB notification rates were 735, 427, and 174 per 100,000 among current smokers, ex-smokers, and never-smokers, respectively (p < 0.001). The trend in TB risk persisted after the control of background characteristics using Cox proportional hazards analysis (adjusted hazard ratios [HRs]: 2.63,1.41, and 1, p < 0.001). In comparison with never-smokers, current smokers had an excess risk of pulmonary TB (adjusted HR, 2.87; 95% confidence interval [Cl], 2.00-4.11; p < 0.001), but not extrapulmonary TB (adjusted HR, 1.04; 95% Cl, 0.33-3.30; p = 0.95). Among the current smokers, those who developed TB smoked more cigarettes per day than those who did not (13.43, SD 8.76 vs. 10.96, SD 7.87, p = 0.01 ). A statistically significant dose-response relationship was observed with respect to active TB and culture-confirmed TB (both p < 0.05). Smoking accounted for 32.8% (95%Cl, 14.9-48.0%), 8.6% (95%Cl, 3.3-15.1%), and 18.7% (95% Cl, 7.7-30.4%) of the TB risk among males, females, and the entire cohort, respectively. Approximately 44.9% (95% Cl, 20.7-64.6%) of the sex difference was attributable to smoking.
Keywords: Chinese; elderly; notification; smoking; tuberculosis
Although several studies have shown a relationship between smoking and tuberculosis (TB), many of them were based on infection (1-3) or mortality rate (4-7). Because of the diagnostic uncertainty related to latent TB infection and the rapidly reduced TB mortality after the introduction of effective treatment, it is not clear whether the association is causal. Although a few studies examined the relationship between smoking and the development of TB (8-13), they had various limitations such as the case-control or cross-sectional design, small sample size, and inadequacy in controlling confounders. A large longitudinal study that adequately controls for the main confounders and other background socioeconomic variables is required to clarify the ongoing controversy.
In Hong Kong, smoking and TB are both common conditions. The prevalence of smoking is much higher in males than females (14). Over 20% of adult males are daily smokers (14). The annual TB notification rate is approximately 100 per 100,000...