Smoking at diagnosis significantly decreases 5-year cancer-specific survival in a population-based cohort of 18 166 colon cancer patients.

SummaryBackground Accumulating evidence suggests smoking may adversely affect cancer patients’ outcomes. Previous studies of smoking and survival in colon cancer have been limited by size and/or lack of a population basis and results have been inconsistent. Aim To investigate in a large population-based cohort whether smoking status at diagnosis is an independent prognostic factor for cancer-specific survival in colon cancer and whether treatment modifies any impact of smoking. Methods Colon adenocarcinomas diagnosed between 1994 and 2012 were abstracted from the National Cancer Registry Ireland, and classified by smoking status at diagnosis. Cancer-specific death rates over 5 years were compared in current, ex- and never smokers using multivariable Cox proportional hazards models, and subgroup analyses by treatment (combinations of cancer-directed surgery and chemotherapy) were conducted. Results Of 18 166 colon cancers, 20% of patients were current smokers, 23% ex-smokers and 57% never smokers. Compared to never smokers, current smokers had a significantly raised cancer death rate [multivariable hazard ratio (HR) = 1.14, 95% CI: 1.07–1.12]. There was a significant interaction between treatment and smoking (P = 0.03). In those who had cancer-directed surgery only, but not other groups, current smokers had a significantly increased cancer death rate compared to never smokers (HR = 1.21, 95% CI: 1.09–1.34). Conclusions Smoking at diagnosis is an independent prognostic factor for colon cancer. The limitation of the association to surgically-treated patients suggests that the underlying mechanism(s) may be related to surgery. While further research is needed to elucidate mechanisms, continued efforts to encourage smoking prevention and cessation may yield benefits in terms of improved survival from colon cancer.