[Selection of population-based controls in pair-matched case-control study analyzing risk factors for lung cancer]

To explore the methods to select population-based controls in a pair-matched case-control study of risk factors for lung cancer and to lay foundation for development of their monitoring in urban areas.

Data of cases of lung cancer and their three different groups of pair-matched control based on population from the "Study on surveillance for incidence and mortality of common malignant tumor and their risk factors" in Beijing, Tianjin, Shanghai and Chongqing were personally-interviewed and analyzed with conditional multivariate logistic regression using STATA software involving 1 312 pairs aged 35 - 74.

Odds ratios for main risk factors were consistent with the three groups of controls, and there was certain difference in ORs for dietary factor, etc. between the studies with three difference control groups. There was little differences in frequencies of exposure to risk factors between studies with neighbor-controls and normal-controls. Cigarette smoking, lower body mass index, long-time depressed mood, lower capability of emotion regulation, previous respiratory diseases, family history of cancer in the first-degree relatives significantly increased the risk for lung cancer after adjusted for age, education level, family average income and other risk factors.

It was feasible to use population-based controls instead of neighbor controls in surveillance for risk factors of lung cancer.