Preoperative predictors of prolonged length of stay in radical cystectomy: a retrospective study using the American College of Surgeons-National Surgical Quality Improvement Program Dataset

Abstract

Background: Radical cystectomy (RC) is considered a complex procedure that entails significant morbidity and mortality. Objectives: We aimed to determine pre-operative patient characteristics that help predict a prolonged length of hospital stay (PLOS) following RC. Design and Methods: The American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) database was used to select patients who underwent RC between the years 2011 and 2020. Prolonged length of stay was defined as a hospital stay ⩾9 days. We compared patient demographics, pre-operative labs, surgical characteristics, and medical history between patients with or without PLOS. Multivariable logistic regression models controlling for pre-operative characteristics and propensity score matching for post-operative complications were conducted to control for possible confounders. Results: The analysis yielded details of 19,158 RC patients of which 6007 (31%) patients had a PLOS. Patients with PLOS were more likely to have post-operative complications that could serve as predictors for the PLOS rather than their pre-operative characteristics. Hence, we matched our cohort for these complications. After matching, patient pre-operative characteristics that predict PLOS included female gender (Odds Ratio (OR) = 5.91), 10-year increase in age (OR = 1.15), non-White race (OR = 1.98), partially or totally dependent functional health status (OR = 2.86), bleeding disorders (OR = 4.67), congestive heart failure (OR = 1.59), pre-operative transfusion (OR = 3.03), and a 20-min increase in operative time (OR = 1.01) ( p < 0.046). Conclusion: Patient demographics and pre-operative factors can help predict PLOS in RC patients. These predictors could serve as tools for patient counseling and risk stratification.