Impact of Smoking on Outcomes in HPV+ Oropharynx Cancer: It′s More Than Pack-Years

Abstract

We sought to examine the impact of smoking status on clinical outcomes in human papillomavirus positive (HPV+) squamous-cell carcinoma (SCC) of the oropharynx. Smoking status and outcomes for 484 pts with HPV+ SCC of the oropharynx from 2001-2015 were determined from an institutional review board approved database. Stage I-IVB pts were included. Smoking status was coded on each patient as never, former (quit > 3 months before diagnosis), or active smoking (quit within 3 months of diagnosis or continued smoking during and/or after treatment). Pack-years (pk-yrs) were also captured for smokers. Co-morbidities were scored according to the Charlson comorbidity index. Kaplan-Meier plots were generated for overall survival (OS), progression-free survival (PFS), locoregional failure free survival (LRFFS), and distant metastasis-free survival (DMFS). Cox proportional hazard regressions were used for univariate (UVA) and multivariate analysis (MVA). Separate MVA were performed to assess the relative impact of pk-yrs and smoking status. Active smokers included 94 pts (19%), 226 pts (47%) were former smokers, and 164 (34%) were never smokers. Of smokers, 247 pts (51%) had >10 pk-yrs and 179 (37%) had >20 pk-yrs. Median follow-up for the entire cohort was 36.3 months (range = 1-162.6 months). Active smokers had significantly worse outcomes compared to never/former smokers for 5-year OS (56.8% vs. 89.6%, P<0.0001), PFS (60.1% vs. 80.4%, P<0.0001), and DMFS (78.1% vs. 90.9%, P = 0.002). There was no difference in the rate of 5-year LRFFS for active vs. never/former smokers (88.4% vs. 92.8%, P = 0.12). On UVA, all smoking parameters (any smoker, >10 pk-yrs, >20 pk-yrs, active smokers) were significantly associated with inferior OS. On MVA, active smoking was the smoking related variable most strongly associated with inferior OS and PFS respectively (HR = 3.4, P<0.0001; HR 2.2, P = 0.0004). Other parameters significantly associated with inferior OS and PFS respectively included higher Charlson scores (HR = 4.2, P<0.0001; HR = 4.0, P<0.0001), and T4 stage (HR = 2.9, P = 0.006; HR = 3.0, P = 0.0008). When smoking was analyzed by pk-yrs in a separate MVA, ≥20 pk-yrs was significantly associated with worse OS and PFS (HR = 3.1, P<0.0001; HR = 1.6, P = 0.03), in addition to higher Charlson scores and T4 disease. Active smoking at diagnosis or continuing to smoke during and/or after treatment is the most powerful smoking related predictor of PFS and OS. While less strongly correlated, >20 pk-yrs was also associated with inferior outcomes, perhaps driven by the active smoking subpopulation within that group. These findings should be incorporated into future risk stratification schemes and clinical trial design.