Abstract

A population based cohort study including 167 343 adults aged 34 and above living in a rural area were recruited and followed-up during January 1996–December 2004 as part of an ongoing oral cancer screening programme. Blood pressure and body mass index (BMI) were measured at baseline and at 3 year intervals. Demographic details and lifestyle factors were also collected through a questionnaire. These subjects were actively followed-up for vital status by field workers through municipal death registers, records collected from social and religious organisations, and during repeated house visits and cause of deaths were determined by verbal autopsy.

The mortality risks associated with chewing (and 95% CIs), after adjusting for age, sex, socio-economic dietary variables, and for other habits, were 0.90 (0.86 to 0.94) for all-cause, 1.07 (0.94 to 1.22) for cancer and 1.22 (1.04 to 1.44) for tobacco-related cancer; with smoking the respective mortality risks were 1.31 (1.24 to 1.39), 1.63 (1.37 to 1.94) and 1.68 (1.36 to 2.08); and with alcohol use the risks were 1.13 (1.06 to 1.20), 1.32 (1.11 to 1.57) and 1.47 (1.19 to 1.80). Reduced risk of all-cause mortality by chewing was observed only in the 60–84 years old group (0.90 (0.85 to 0.94)), and detrimental effects of chewing on cancer mortality were shown in the young and middle-age groups: 34–39 years old (1.33 (0.67 to 2.65)), and 40–59 years old (1.26 (1.03 to 1.55)). An analysis of the association between BMI and change in weight in relation to mortality risk showed that low BMI was found to be a predictor of mortality, while high BMI was not.