Association between radiographic soft-tissue thickness and increased length of stay, operative time, and infection rate after reverse shoulder arthroplasty

Abstract

Reverse shoulder arthroplasty (RSA) is a widely performed surgical procedure to address various shoulder pathologies. Several studies have suggested that radiographic soft-tissue thickness may play a role in predicting complications after orthopedic surgery, but there have been limited studies determining the use of radiographic soft-tissue thickness in RSA. The purpose of this study was to evaluate whether radiographic soft-tissue thickness could predict clinical outcomes after RSA and compare the predictive capabilities against body mass index (BMI). We hypothesized that increased radiographic shoulder soft-tissue thickness would be a strong predictor of operative time, length of stay (LOS), and infection in elective RSA. A retrospective review of patients undergoing RSA at an academic institution was conducted. Preoperative radiographic images were evaluated including measurements of the radius from the humeral head center to the skin (HS), deltoid radius-to-humeral head radius ratio (DHR), deltoid size, and subcutaneous tissue size. Different correlation coefficients were used to analyze various types of relationships, and the strength of these associations was classified based on predefined boundaries. Subsequently, multivariable linear and logistic regressions were performed to determine whether HS, DHR, deltoid size, and subcutaneous tissue size could predict LOS, operative time, or infection while controlling for patient factors. HS was the most influential factor in predicting both operative time and LOS after RSA, with strong associations indicated by standardized β coefficients of 0.234 for operative time and 0.432 for LOS. Subcutaneous tissue size, deltoid size, and DHR also showed stronger predictive values than BMI for both outcomes. In terms of prosthetic joint infection, HS, deltoid size, and DHR were significant predictors, with HS demonstrating the highest predictive power (Nagelkerke R2 = 0.44), whereas BMI did not show a statistically significant association with infection. Low event counts resulted in wide confidence intervals for odds ratios in the infection analysis. Greater shoulder soft-tissue thickness as measured with concentric circles on radiographs is a strong predictor of operative time, LOS, and postoperative infection in elective primary RSA patients.