Smoking and mortality from tuberculosis and other diseases in India: retrospective study of 43000 adult male deaths and 35000 controls.

BACKGROUND In India most adult deaths involve vascular disease, pulmonary tuberculosis, or other respiratory disease, and men have smoked cigarettes or bidis (which resemble small cigarettes) for several decades. The study objective was to assess age-specific mortality from smoking among men (since few women smoke) in urban and in rural India. METHODS We did a case-control study of the smoking habits of 27000 urban and 16000 rural men who had died in the state of Tamil Nadu, southern India, from medical causes (ie, any cause other than accident, homicide, or suicide), and of 20000 urban and 15000 rural male controls. The main analyses are of mortality at ages 25-69 years. FINDINGS In the urban study area, the death rates from medical causes of ever smokers were double those of never smokers (standardised risk ratio at ages 25-69 years 2.1 [95% CI 2.0-2.2]). The risks were substantial both for cigarette smoking (the main urban habit) and for bidi smoking. Of this excess mortality among smokers, a third involved respiratory disease, chiefly tuberculosis (4.5 [4.0-5.0], smoking-attributed fraction 61%), a third involved vascular disease (1.8 [1.7-1.9], smoking-attributed fraction 24%), 11% involved cancer (2.1 [1.9-2.4], smoking-attributed fraction 32%), chiefly of the respiratory or upper digestive tracts, and 14% involved alcoholism or cirrhosis (3.3 [2.9-3.8], not attributed to smoking). Among ever smokers, the absolute excess mortality from tuberculosis was substantial throughout the age range 25-69 years. (A separate survey of 250000 men living in the urban study area found that ever smokers are three times as likely as never smokers to report a history of tuberculosis, corresponding to a higher rate of progression of chronic subclinical infection to clinical disease.) The proportional excesses of respiratory, vascular, and neoplastic mortality at ages 25-69 years among ever smokers in the urban study area were replicated, each with similarly narrow CI for the risk ratio, in the rural study area (where bidi smoking predominated), and are taken to be largely or wholly causal. For urban and for rural death from medical causes at older ages (> or =70 years), the standardised risk ratio was 1.3. INTERPRETATION Smoking, which increases the incidence of clinical tuberculosis, is a cause of half the male tuberculosis deaths in India, and of a quarter of all male deaths in middle age (plus smaller fractions of the deaths at other ages). At current death rates, about a quarter of cigarette or bidi smokers would be killed by tobacco at ages 25-69 years, those killed at these ages losing about 20 years of life expectancy. Overall, smoking currently causes about 700000 deaths per year in India, chiefly from respiratory or vascular disease: about 550000 men aged 25-69 years, about 110000 older men, and much smaller numbers of women (since few women smoke).