Associations of smoking with the risk of second smoking-associated primary cancer among adults with a history of cancer.

Associations between smoking and smoking-associated second primary cancers (SSPC) have been documented predominantly among survivors of smoking-associated first primary cancers (FPC), with limited evidence among survivors of non-smoking-associated FPCs. We examined survivors of FPCs diagnosed from 1992 to 2015 in the Cancer Prevention Study II Nutrition Cohort, followed through June 30, 2017. Self-reported smoking status, intensity, and years since quitting were collected a mean of 1.6 years before FPC diagnosis. Among 28,272 survivors (58.8%, male; mean age at FPC diagnosis, 72.2 years [SD=7.08]), the most common FPCs were prostate (32.8%), breast (18.5%), colorectum (9.0%), and non-Hodgkin lymphoma (8.3%). Over a median follow-up of 7.0 years, 1926 survivors developed SSPCs. Based on Cox proportional hazards models, adjusted for key confounders, SSPC risk was higher among current (HR=3.78; 95%CI=3.22-4.44) and former (HR=1.63, 95%CI=1.46-1.81) smokers compared with never smokers, with similar associations for survivors of smoking-associated (HR=4.08; 95%CI=2.90-5.74 for current smokers) and non-smoking-associated (HR=3.74; 95%CI=3.11-4.51 for current smokers) FPCs. Among former smokers, risk decreased with increased years since quitting, approaching never-smoker levels after > 30 years of cessation. Post-diagnosis cessation (599 of 1518 current smokers) was associated with a 25% lower SSPC incidence versus continued smoking (HR=0.75; 95%CI=0.54-1.03). In conclusion, pre-diagnostic smoking was strongly associated with higher risk of SSPCs among survivors of both smoking- and non-smoking-associated FPCs. Longer duration since quitting and cessation after diagnosis reduced the elevated risk, highlighting smoking cessation at any time can help reduce the burden of SPCs among the growing population of cancer survivors.