Predictors of competing mortality in advanced head and neck cancer.

PURPOSE Death from noncancer causes (competing mortality) is an important event in head and neck cancer, but studies identifying predictors of this event are lacking. We sought to identify predictors of competing mortality and develop a risk stratification model for competing events. PATIENTS AND METHODS Cohort study of 479 patients with stage III to IV carcinoma of the head and neck diagnosed between August 1993 and November 2004. Patients were treated on consecutive prospective clinical trials involving organ-preserving chemoradiotherapy and surgery. We used multivariable competing risks regression models to analyze factors associated with the cumulative incidence of competing mortality, locoregional and distant failure, and second malignancies as first events. Results Median follow-up was 52 months median for survivors. The 5-year cumulative incidence of competing mortality was 19.6% (95% CI, 15.8 to 23.4). On multivariable analysis, competing mortality was associated with female sex (hazard ratio [HR], 1.72; 95% CI, 1.13 to 2.63), increasing age (HR, 1.30; 95% CI, 1.04 to 1.62), increasing Charlson Comorbidity Index (HR, 1.24; 95% CI, 1.05 to 1.47), decreasing body mass index (HR, 0.33; 95% CI, 0.13 to 0.84), and decreasing distance traveled to the treating center (HR, 0.65; 95% CI, 0.44 to 0.98). Patients with zero, one, two, and > or = three risk factors had 5-year competing mortality of 8.9% (95% CI, 3.0% to 14.8%), 12.4% (95% CI, 7.0% to 17.8%), 22.1% (95% CI, 14.5% to 29.7%), and 39.3% (95% CI, 28.6% to 50.1%), respectively. CONCLUSION Competing mortality in advanced head and neck cancer is associated with several demographic and health status characteristics. Analyses of risk factors for competing mortality may be useful in outcomes reporting and designing clinical trials.