Predictors of length of stay after elective total shoulder arthroplasty in the United States

Abstract

Total shoulder arthroplasty (TSA) is an increasingly used treatment of glenohumeral arthritis and proximal humerus fractures. However, patient-specific characteristics affecting length of hospital stay postoperatively have not been elucidated. All patients undergoing primary unilateral TSA between 2005 and 2011 were isolated from the National Surgical Quality Improvement Program database. Patient demographics, medical comorbidities, and selected surgical variables were extracted, and length of stay was established as the primary end point of interest. Risk factors were expressed as odds ratios (ORs) with 95% confidence intervals by bivariate and multivariable analysis. A total of 2004 patients were identified; the average age was 68.8 years, and 57% were women. Mean length of stay after TSA was 2.2 days (standard deviation, 1.7), and 91% of cases received hospital discharge in <3 days. Multivariable logistic regression analysis identified renal insufficiency (OR, 11.35; P = .0002), increased age (OR, 2.13; P = .011), longer operative time (OR, 1.94; P = .0041), and American Society of Anesthesiologists class ≥3 (OR, 1.86; P = .0016) as the most significant risk factors for length of stay. Gender also influenced length of stay; women were more likely to stay ≥4 days (OR, 0.44; P < .0001). Perioperative risk stratification and preoperative counseling are paramount for patients undergoing TSA, particularly for those individuals with cardiac and renal disease or of advancing age. These variables may effectively predict prolonged hospital stay after TSA.