Sociodemographic and Clinical Predictors of Utilization of Prostatic Artery Embolization versus Transurethral Resection of the Prostate

Monday, April 13, 2026

3:45 PM - 3:54 PM ET

Location: 717

Presenting Author(s)

Elan Meltzer, MD, MS

Resident
RWJ Community Medical Center, New Jersey, United States

Author/Co-author(s)

Abhishek Kumar, MD

Division Chief of Vascular and Interventional Radiology
Rutgers New Jersey Medical School, New York, United States
PS

Pratik A. Shukla, MD

Associate Professor of Interventional Radiology
Rutgers - New Jersey Medical School, New Jersey, United States

Purpose: To evaluate sociodemographic and clinical predictors of receiving prostatic artery embolization (PAE) versus transurethral resection of the prostate (TURP) among elective hospitalizations in the 2017 National Inpatient Sample (NIS).

Materials and Methods: Using the 2017 NIS we filtered for male elective admissions with benign prostatic hyperplasia or lower urinary tract symptoms. PAE and TURP procedures were identified. A 1:3 age-matched sample was generated, 87 patients underwent PAE and 258 underwent TURP. Covariates included race, household income quartile, insurance status, hospital type, and APDRG surgical risk status. Multivariate logistic regression was used to evaluate predictors of undergoing PAE versus TURP.

Results: High surgical risk status (APRDRG >2) was strongly associated with undergoing PAE (OR 4.44, 95% CI 2.35–8.38, p < .001). Odds of PAE were lower for Black (OR 0.42, 95% CI 0.13–1.31, p = .142) and Hispanic patients (OR 0.46, 95% CI 0.17–1.25, p = .118) compared with White patients, though not statistically significant. AAPI patients showed no difference (OR 1.07, 95% CI 0.26–4.39, p = .924). Income quartiles 2 (OR 1.15, 95% CI 0.49–2.70, p = .754), 3 (OR 1.48, 95% CI 0.62–3.50, p = .377), and 4 (OR 1.60, 95% CI 0.69–3.72, p = .266) were not significant compared with quartile 1. Insurance type, including Medicare (OR 1.32, 95% CI 0.11–15.66, p = .827) and Private (OR 2.20, 95% CI 0.16–29.5, p = .551), was also not significant. Compared with urban teaching hospitals, rural hospitals trended toward lower PAE utilization (OR 0.14, 95% CI 0.02–1.03, p = .066), while urban non-teaching hospitals had significantly lower odds (OR 0.41, 95% CI 0.19–0.88, p = .021).

Conclusion: High surgical risk status was the strongest predictor of undergoing PAE, indicating that poor surgical candidates may be more likely to be referred. Despite the historically higher disease burden among Black patients, the wide confidence interval and low lower bound suggest that Black patients may be underrepresented among those receiving PAE and that despite higher comorbidity burden, they are not being referred for PAE as often as their clinical risk might warrant {1}. Further studies with larger datasets are needed to reveal potential significance. Furthermore, patients at urban non-teaching centers were less likely to receive a PAE, which may be attributed to limited availability of interventional radiology services, lower familiarity with the procedure among referrers, or differences in adoption of newer interventions at non-academic centers.