Population-based study of competing mortality in head and neck cancer.

PURPOSE Patients with head and neck cancer (HNC) are at high risk of death resulting from noncancer causes and second malignancies (ie, competing mortality). Variation in competing mortality risk complicates individual treatment choices and design and interpretation of clinical studies. METHODS Using the Surveillance, Epidemiology, and End Results registry, we identified 34,568 patients with nonmetastatic squamous cell carcinoma of the head and neck diagnosed between 1994 and 2003. We developed a multivariable competing-risk regression model to stratify patients according to competing mortality risk and evaluate the impact of this risk on power loss in clinical studies. RESULTS The 5-year cumulative incidences of all-cause mortality, HNC-specific mortality, and competing mortality were 51.3% (95% CI, 50.8% to 51.9%), 23.8% (95% CI, 23.3% to 24.2%), and 27.6% (95% CI, 26.8% to 28.3%), respectively. Factors associated with increased competing mortality were increasing age, male sex, black race, unmarried status, localized disease, higher socioeconomic status, nonsurgical treatment, and hypopharyngeal, nasopharyngeal, and oral cavity subsites. The 5-year cumulative incidences of competing mortality for patients in low-, medium-, and high-risk score tertiles were 20.0% (95% CI, 18.8% to 21.3%), 27.7% (95% CI, 26.3% to 29.1%), and 33.7% (95% CI, 32.2% to 35.2%), respectively. Compared with patients with low competing mortality risk, relative sample sizes required to show benefit of a treatment regarding all-cause mortality were 12% and 42% higher in the medium- and high-risk groups, respectively. CONCLUSION Multiple factors affect risk of competing mortality among patients with HNC. Risk stratification would be useful to identify patients most likely to benefit from treatment intensification.