Objective: To investigate the capability of a modified self-administrable comorbidity index recommended in the standard sets for neoplastic diseases published by the International Consortium for Health Outcomes Measurement (ICHOM) to predict 90-day and long-term mortality after radical cystectomy.
Methods: A single-center series of 1,337 consecutive patients who underwent radical cystectomy for muscle-invasive or high-risk non-muscle-invasive urothelial or undifferentiated bladder cancer were stratified by the modified self-administrable comorbidity index and Charlson score, respectively. Multivariate logit models (for 90-day mortality) and proportional-hazards models (for overall and non-bladder cancer mortality) were used for statistical workup.
Results: Considering 90-day mortality, both comorbidity indexes contributed independent information when analyzed together with age (p < 0.0001). The Charlson score performed slightly better (area under the curve [AUC] 0.74 vs. 0.72 for the ICHOM-recommended comorbidity index). Considering 5-year overall mortality in 727 patients with complete observation, the performance of both measures was similar (AUC 0.63 vs. 0.62, including age AUC 0.66 for both indexes). With 6-sided stratifications, the modified self-administrable comorbidity index separated the risk groups slightly better (p values for directly neighboring curves: 0.0068-0.1043 vs. 0.0001-0.8100).
Conclusion: The ICHOM-recommended modified self-administrable comorbidity index is capable of predicting 90-day mortality and long-term non-bladder cancer mortality after radical cystectomy similarly to the commonly used Charlson score.
Keywords: Bladder; Comorbidity; Cystectomy; International Consortium for Health Outcomes Measurement; Logit model; Proportional hazards model; Standard set.
© 2020 S. Karger AG, Basel.