Prediction of major surgical complications by comprehensive geriatric assessment in elderly patients with gynecologic cancers: A prospective cohort study.

Abstract

e20530 Background: To evaluate the ability of a preoperative comprehensive geriatric assessment (CGA) to predict the risk of postoperative complications in elderly patients who underwent primary surgery for gynecologic cancers. Methods: Sixty consecutive patients (26 ovarian, 16 cervical, 7 endometrial, 5 uterine sarcoma, 4 vulvar, and 2 metastatic cancers) older than 70 years scheduled to take surgery electively for gynecologic cancer were preoperatively assessed by CGA. Every category of CGA was evaluated for in-hospital postoperative complications and mortality within 30 days of surgery. Major complication included any complication of 2 or higher grade according to Memorial Sloan-Kettering Surgical Complication Criteria 12/99. Results: Mean age was 73.6±3.5 years (range 70-85). Overall postoperative complication rate was 30.0% (18/60) including 9 (15.0%) major and 8 (13.3%) multiple complications. Mean hospital stay was 18.4±13.3 days (range 3-81), during which in-hospital mortality was 2 (3.3%). Univariate analysis revealed that dependent instrumental activity of daily living (IADL) was associated with any postoperative complication (p=0.023), especially for multiple complications (p=0.019). However, poor performance status (PS) was associated only with major and multiple complications (p=0.021 and p=0.014, respectively). Multiple regression analysis revealed that dependent IADL was the independent risk factor for any postoperative complication (relative risk [RR] 8.0, 95% confidence interval [CI] 1.1-59.8, p = 0.042). Notably, poor PS was significantly associated with postoperative mortality (p = 0.006), however, dependent IADL with a marginal significance (p = 0.059). Conclusions: IADL and PS were useful to predict postoperative complications and related mortality in elderly patients with gynecologic cancers. Postoperative complications and related mortality can be prevented by identifying the elderly patients with gynecologic cancers who showed dependent IADL and poor PS in CGA.