The association between tobacco use and lung cancer risk: difference caused by smoky coal use

Background Xuanwei is one of the towns with highest lung cancer mortality in China. Residents there frequently use smoky coal for domestic cooking and heating. The smoking prevalence there is also high. Methods We conducted a population-based case-control study. 388 lung cancer cases and 388 gender-age-area-matched controls were surveyed. The exposure levels to potential lung cancer risk factors of multi-aspects 10 years ago were collected. After checking Spearman correlations between variables and referring to previous studies, we built conditional logistic regression model adjusting education, family history of malignancy, diet, alcohol consumption, occupational history, housing type and exposure to second hand smoke. We then introduced household annual consumption of smoky coal and total consumption of cigarettes from born to 10 years ago to the model in the form of ordered categorical variable (coal level 1-3: never use, less than 3 tons per year, 3 tons or more per year; tobacco level 1-3: never use, less than 130 thousands cigarettes, 130 thousands or more cigarettes). Results There were 27% and 31.2% ever-smokers in case and control group, respectively. Without interactive item, the effect of tobacco use on lung cancer was (all comparing with level 1, with 95% confidence intervals) level 2: OR=0.90(0.22, 3.67), level 3: OR=1.59(0.45, 5.70). After adding interactive item of coal and tobacco in the model, the effect of tobacco use among people never used smoky coal was level2: OR=1.09(0.09, 13.78), level3: OR=2.20(0.16, 29.44); among people using relatively small amount of coal was level2: OR=0.59(0.03, 11.37), level3: OR=0.18(0.01, 4.82); among people using large amount of coal was level2: OR=0.93(0.32, 2.72), level3: OR=1.78(0.30, 10.67). Household annual amount of smoky coal used, 10 years ago Total consumption of cigarettes from born to 10 years ago (reference group: never smoker) β SE OR(95%CI) Never