Comprehensive Evaluation of the Ability of Comorbidity and Health Status Indices to Improve the Prediction of Perioperative Morbidity and Long-Term Survival Outcomes After Radical Cystectomy.

BACKGROUND

In the era of standardized outcome reporting, it remains unclear if widely used comorbidity and health status indices can enhance predictive accuracy for morbidity and long-term survival outcomes after radical cystectomy (RC).

PATIENTS AND METHODS

In this monocentric study, we included 468 patients undergoing open RC with pelvic lymph node dissection for bladder cancer between January 2009 and December 2017. Postoperative complications were meticulously assessed according to the EAU guideline criteria for standardized outcome reporting. Multivariable regression models were fitted to evaluate the ability of ASA physical status (ASA PS), Charlson comorbidity index (± age-adjustment) and the combination of both to improve prediction of (A) 30-day morbidity key estimates (major complications, readmission, and cumulative morbidity as measured by the Comprehensive Complication index [CCI]) and (B) secondary mortality endpoints (overall [OM], cancer-specific [CSM], and other-cause mortality [OCM]).

RESULTS

Overall, 465 (99%) and 52 (11%) patients experienced 30-day complications and major complications (Clavien-Dindo grade ≥IIIb), respectively. Thirty-seven (7.9%) were readmitted within 30 days after discharge. Comorbidity and health status indices did not improve the predictive accuracy for 30-day major complications and 30-day readmission of a reference model but were associated with 30-day CCI (all P < .05). When ASA PS and age-adjusted Charlson index were combined, ASA PS was no longer associated with 30-day CCI (P = .1). At a median follow-up of 56 months (IQR 37-86), OM, CSM, and 90-day mortality were 37%, 24%, and 2.9%, respectively. Both Charlson and age-adjusted Charlson index accurately predicted OCM (all P < .001) and OM (all P ≤ .002) but not CSM (all P ≥ .4) and 90-day mortality (all P > .05). ASA PS was not associated with oncologic outcomes (all P ≥ .05).

CONCLUSION

While comorbidity and health status indices have a role in predicting OCM and OM after RC, their importance in predicting postoperative morbidity is limited. Especially ASA PS performed poorly. This highlights the need for procedure-specific comorbidity assessment rather than generic indices.