Risk factors and reasons for revision after reverse total shoulder arthroplasty

Abstract

BackgroundAs the indications for reverse shoulder arthroplasty (RSA) continue to expand, the need for revision surgery after RSA will become more frequent. The objective of this study was to characterize patient-related risk factors for revision RSA and to compare reasons for early vs. late revision after RSA. MethodsPatients who underwent primary and revision RSA from 2015 to 2019 were identified in a national insurance database. Subgroups of early revision (defined as revision within 1 year postoperatively) and late revision (more than 1 year postoperatively) were also identified. The primary outcome of interest was patient-related risk factors for revision RSA. Secondary outcomes of interest were patient-related risk factors for early vs. late revision RSA and to compare surgical diagnoses for early vs. late revision RSA. Univariate analysis using chi-square tests was performed to analyze any differences in reasons for revision. Multivariate regression was subsequently utilized to control for any confounding variables when identifying risk factors for revision. ResultsA total of 28,880 patients were identified who underwent RSA, with 553 (1.9%) patients undergoing revision RSA. Three hundred eighty-five patients (69.6%) were classified as early revision (within one year), while 141 (30.4%) underwent late revision more than a year postoperatively. Risk factors for overall revisions included age <65 years (odds ratio [OR] = 1.23, P = .032), male sex (OR = 2.21, P < .001), type I diabetes mellitus (OR = 1.44, P = .039), congestive heart failure (CHF) (OR = 1.79, P < .001), and depression (OR = 1.33, P = .002) in addition to RSAs performed for fracture (OR = 1.63, P < .001) and glenohumeral instability (OR = 2.25, P < .001) compared to RSA performed for arthritis. Risk factors for early revision RSA included male sex (OR = 2.54, P < .001) and CHF (OR = 1.81, P < .001) in addition to RSAs performed for fracture (OR = 1.84, P < .001) and glenohumeral instability (OR = 2.44, P < .001). Risk factors for late revision RSA included male sex (OR = 1.62, P = .004), CHF (OR = 1.83, P = .005), steroid use (OR = 1.79, P = .036), human immunodeficiency virus (OR = 3.50, P = .038), and RSA performed for glenohumeral instability (OR = 1.92, P = .004). Early revision RSA was more commonly performed for instability (63.1% vs. 25.0%, P < .001) and stiffness (5.5% vs. 1.2%, P = .021) than late revisions. ConclusionRevision RSA is uncommon at early follow-up. Overall patient-related risk factors for revision include male sex, age <65 years, type I diabetes mellitus, CHF, and depression in addition to RSAs performed for fracture and glenohumeral instability. Instability and stiffness were more common indications for early compared to late revision. Instability remained the most common reason for overall revision followed by periprosthetic infection.