Smoking history before surgery and prognosis in patients with stage IA non-small-cell lung cancer--a multicenter study.

The prognosis of lung cancer patients with surgically resected non-small-cell lung cancer (NSCLC) can be predicted generally from age, sex, histologic type, stage at diagnosis, and additional treatment. Nine studies have reported that a history of smoking before diagnosis influences the prognosis of the disease in lung cancer patients. In this study, a total of 3082 patients who underwent surgery and were diagnosed with primary pathological stage IA NSCLC at 36 national hospitals from 1982 to 1997 were analyzed for the effect of smoking on survival. Smoking history and other factors influencing either the overall survival or the disease-specific survival rates of patients were estimated with the Cox proportional hazards model. Multivariate analysis demonstrated significant associations between overall survival and age (P < 0.0001), sex (P = 0.0002), and performance status (PS) (P < 0.0001). Disease-specific survival was associated with age (P = 0.0063), sex (0.00161), and PS (P = 0.0029). In males, disease-specific survival was associated with age (P = 0.0120), PS (P = 0.0022), and pack-years (number of cigarette packs per day, and years of smoking) (P = 0.0463). These results indicate that smoking history (pack-years) is important clinical prognostic factor in estimating disease-specific survival, in male patients with stage IA primary NSCLC that has been surgically resected.