Calculation of population attributable risk for bidi smoking and oral cancer in south Asia.

Abstract

Bidi smoking, which is widely prevalent in India and in other south Asian countries, increases the risk of oral cancer as observed in case-control studies and metaanalysis. However, population attributable risk percent (PAR%) has not been determined yet. Twelve case-control studies conducted in India, Pakistan, and Sri Lanka, which included information on bidi smoking and oral cancer, were analyzed countrywise to estimate PAR%. The cumulative cases and controls were 4778 and 6271, respectively, based on 10 case-control studies conducted in India. Among the cases, 49.1% were bidi smokers and 7.7% cigarette smokers, while they were 19.9% and 10.3%, respectively, among controls. Pooled odds ratio (OR) of bidi smoking for oral cancer was 3.3 [95% confidence interval (CI), 3.0-3.6] and 2.6 (95% CI 1.8-3.8), respectively, based on fixed- and random-effects model. Cigarette smoking, on the other hand, did not show any significant association. PAR% of bidi smoking for oral cancer ranged from 4.7% to 51.6% on individual study basis, while they were 31.4% and 24.1%, respectively, based on OR derived from fixed- and random-effects models. PAR% was 5.8% and 8.7% based on single study estimate from Pakistan and Sri Lanka, respectively. Bidi smoking is considered to account for a sizeable number of oral cancers in south Asian countries, which implies that cessation programs should be formulated and implemented vigorously.