Risk-adjusted outcomes in elderly endometrial cancer patients: Implications of the contrasting impact of age on progression-free and cause-specific survival

Objective To reexamine the tenet that advanced age independently impacts progression-free and cause-specific survival in patients with endometrial cancer (EC). Methods Patients undergoing surgery for stages I-IIIC EC between 1999 and 2008 were stratified by age (< 70 vs 70 years). Three propensity score (PS) methods were utilized to adjust for confounding risk factors. The PS, or conditional probability of being 70 years old, given a patient's baseline covariates, was derived using logistic regression. The Cox proportional hazards models were fit to estimate the effect of age 70 years on outcomes. Results Of 1182 eligible patients, 822 (69.5%) were < 70 and 360 (30.5%) were 70. Patients 70 were more likely to have multiple adverse risk factors. The total standardized difference of these factors was reduced by 74% and 81%, respectively, using PS-stratification and PS-matching analyses. The nonsignificant trend toward an association between progression-free survival and age 70 in an unadjusted analysis (hazard ratio [HR], 1.40; 95% CI, 0.95-2.04) was further attenuated in the 3 PS analyses. The unadjusted HR for the association between age 70 and cause-specific survival was 2.03 (95% CI, 1.32-3.13). HRs were attenuated in PS analyses but retained significance (except for PS matching), potentially reflecting differences in salvage therapies (P <.001), including a 3-fold greater use of chemotherapy in those < 70. Conclusion When risk-adjusted for the higher prevalence of adverse prognostic factors in elderly EC patients, progression-free survival after primary therapy is not age dependent but the less favorable cause-specific survival in this cohort may reflect age-related postrecurrence treatment differences.