We read with great interest the recent meta-analysis by Bates et al1 on the risk of tuberculosis from exposure to tobacco smoke.1 Because both smoking and tuberculosis are still major killers, especially in developing countries, such a meta-analysis has been long awaited. However, a few points might warrant further discussion.
In the analysis by Bates et al,1 smoking increases the relative risk (RR) of both tuberculosis infection (RR, 1.73, as measured by a positive tuberculin skin test [TST] result) and disease (RR, 2.33-2.66), and the authors therefore estimated the net RR for disease development in an infected population to be 1.4 to 1.6. However, the only cohort study included by Bates et al1 for the analysis of disease development refers to elderly subjects in Hong Kong.1- 2 With the huge past burden of tuberculosis, a very high background prevalence of tuberculosis infection is expected in that cohort.2- 3 The additional effect of smoking on infection prevalence should be minimal, but the RR for tuberculosis disease remained high, especially for culture-confirmed tuberculosis among the male subjects (unadjusted RR, 2.24 and 3.49 for ex-smokers and current smokers, respectively).2