Karnofsky performance status predicts overall survival, cancer-specific survival, and progression-free survival following radical cystectomy for urothelial carcinoma.

OBJECTIVE: Radical cystectomy (RC) can provide a survival advantage in patients with urothelial carcinoma of the bladder, but not without significant morbidity rates. Whether the ability of preoperative comorbidity or performance status metrics can stratify patients to overall survival (OS), cancer-specific survival (CSS), and progression-free survival (PFS) following RC is unclear. We analyze our RC experience from 2005 to 2010 to assess the prognostic power of American Society of Anesthesiologists (ASA) score, Charlson Comorbidity Index (CCI), and Karnofsky Performance Status (KPS) index as they relate to OS, CSS, and PFS.

MATERIALS AND METHODS: A retrospective analysis was performed of 234 patients who underwent RC between January 2005 and December 2010; of these, 148 patients had sufficient data for OS, CSS, and PFS analysis. Multivariate Cox proportional hazard modeling generated hazard ratios using as independent variables patient age at surgery, gender, ethnicity, preoperative KPS, CCI, and ASA values, pathologic T-staging, the presence of nodal disease, use of radiation therapy, neoadjuvant chemotherapy, and adjuvant chemotherapy. A recursive partition analysis tree divided the population into high- and low-performance groups, and 5-year survival outcomes were evaluated. OS, CSS, and PFS were employed as Kaplan-Meier dependent variables with similar populations comprising high- and low-performance subgroups.

RESULTS: Mean CSS was 46.8 months (95 % CI 43.2-50.4) with a 5-year CSS of 75 % and OS of 69 %. Patient age, pathologic T-stage, and KPS were identified as independent predictors of OS and CSS. Analysis of PFS as the continuous dependent variable identified only KPS as a statistically significant predictor of freedom from radiologic progression. No statistically significant predictive value was identified for nodal disease, neoadjuvant chemotherapy, adjuvant chemotherapy, gender, ethnicity, CCI, or ASA in terms of OS, CCS, or PFS. Patients with a KPS ≤ 80 had a shorter survival than patients with a KPS ≥ 90 in terms of OS, CSS, and PFS (log-rank Mantel-Cox: p < 0.01). For patients with a KPS ≤ 80, ~5-year CSS was 42 %, while for patients with a KPS ≥ 90 the 5-year survival was 81 %. These survival curves can be further stratified based on T-stage where patients with a KPS ≥ 90 and <T2 have a 5-year CSS of 83 %, patients with a KPS ≥ 90 and >T2 have a 5-year CSS of 80 %, whereas patients with a KPS ≤ 80 and >T2 have a ~5-year CSS of 43 % (p < 0.0001).

CONCLUSIONS: Our study suggests the use of KPS to have predictive capacity in terms of OS, CSS, and PFS. This information can be used to inform patients' survival expectations prior to proceeding with radical cystectomy.