Tobacco smoking-associated risk for tuberculosis: a case-control study.

Tobacco smoking as a risk factor for tuberculosis (TB) is often ignored. This study was carried out to study tobacco smoking-associated risk for tuberculosis. Ninety-five patients with TB, admitted in the Himalayan Institute Hospital, Dehradun, India and 190 controls were randomly selected and questioned regarding their smoking status and type, quantity and duration of tobacco used. Pulmonary tuberculosis (PTB) was present in 57.8%, extra-pulmonary tuberculosis (EPTB) in 26.3% and disseminated TB in 15.7% cases. The odds ratios (ORs) for association of smoking tobacco (cigarette and beedi [a thin, Indian cigarette made of flaked tobacco wrapped in a rectangular piece of dried Diospyros melanoxylon leaf] collectively), cigarette and beedi with TB were 3.53 (P < 0.0001), 1.65 (P = 0.337) and 4.49 (P < 0.0001) respectively. The ORs for PTB, EPTB and disseminated TB among smokers were 3.44 (P = 0.0001), 3.56 (P = 0.0067) and 2.51 (P = 0.145) respectively. The OR for sputum positivity for acid fast bacilli (AFB) among smokers was 4.65 (P = 0.0001). The ORs for a positive association with TB increased with duration of smoking (P < 0.0001) and number of cigarettes/beedis smoked (P < 0.0001). Univariate analysis showed a significant association between TB and other confounding risk factors viz., previous TB contact (OR 6.93, P = 0.0001), previous TB infection (OR 38.27, P < 0.0001) and malnutrition (OR 3.77, P = 0.0032). In multivariate analysis using a logistic regression model, factors independently associated with TB were smoking (OR 3.05, P = 0.000), previous TB contact (OR 6.52, P = 0.001), previous TB infection (OR 37.72, P = 0.001) and malnutrition (OR 3.73, P = 0.009). Therefore, smoking has a strong association with PTB and EPTB but not with disseminated TB. Significant association exists between beedi smoking and TB but not between cigarette smoking and TB. The risk for TB depends upon dose, duration and type of tobacco smoked.