Presented by: Arun Rai MD, MBA
Northwell

Introduction

While frailty has emerged as a predictor of morbidity following urological surgery, it remains difficult to measure. A new frailty index (modified Frailty Index-5(mFI-5)) has emerged for endoscopic treatment of benign prostatic hyperplasia (BPH); however, its discriminatory abilities for postoperative complications has not been investigated. Therefore, we compared mFI-5 to two most commonly used indexes, the American Society of Anesthesiologists physical classification (ASA) and modified Charlson Comorbidity Index (mCCI) using the National Surgical Quality Improvement Program (NSQIP) database.

Materials

We retrospectively queried the 2015-2020 NSQIP datasets for patients who underwent endoscopic treatment for BPH using CPT and ICD codes. Patients were stratified by procedure type (transurethral resection of the prostate (TURP), laser vaporization (LVP) and enucleation (LEP) of the prostate). Risk indexes were calculated and compared as predictors of postoperative outcomes using C-statistics (AUC).

Results

,

38,128 patients were included with a mean age of 71[RA1] . In aggregate, ASA Class, mFI-5, and mCCI were fair models of mortality (AUC>0.7) and not superior to the other (p>0.05). ASA Class (AUC=0.605) and mCCI (AUC=0.635) were poor models for any readmissions and surgical complications; [RA2] however, all other indexes failed to model any other postoperative outcomes (AUC<0.6).

Upon stratifying by endoscopic procedure, all three indexes were fair models of mortality (AUC>0.7) and not superior to the other (p>0.05). When stratifying for patients undergoing TURP, mFI-5 and ASA class were poor models of surgical complications (AUC=0.638) and readmission (AUC=0.600), respectively, but were superior to other indexes (p<0.05). For LVP, mCCI (AUC=0.636) and ASA (AUC=0.600) were poor models of length of stay ≥ 2 days, but mCCI was superior (p<0.05). Additionally, all indexes were poor models of readmission, but mCCI was superior (AUC=0.626, p<0.05) for Greenlight. Lastly, for LEP, ASA and mCCI were poor models of any complication and surgical complication (AUC>0.6), respectively, but not significantly different (p<0.05). All three indexes were poor models of readmission (AUC>0.6) and not significantly different for LEP.

Conclusion

All three indexes are fair models of mortality for endoscopic treatment of BPH. However, these indexes were inadequate models for all other postoperative outcomes, despite previous studies demonstrating an association. Endoscopic treatment of BPH needs better risk indexes to predict post-operative complications.

Funding

None

Co-Authors

Justin James, BS
Sophie Davis School of Medicine

Zeph Okeke, MD
Northwell Health

Greg Mullen, MD
Northwell Health