Identifying recovery trajectories following primary total shoulder arthroplasty: a cohort study of 3,358 patients from the Dutch Arthroplasty Register

Abstract

Background and purpose: Some patients do not improve after total shoulder arthroplasty (TSA), indicating different recovery trajectories. We aimed to identify recovery trajectories after TSA based on the Oxford Shoulder Score (OSS). Second, we investigated whether recovery trajectories were associated with patient or procedure characteristics.
Methods: We included primary anatomical and reversed TSAs (ATSAs/RTSAs) for osteoarthritis (OA) or cuff arthropathy/rupture with preoperative, 3-month, and/or 12-month postoperative OSS, registered between 2016 and 2022 in the Dutch Arthroplasty Register (n = 3,358). We used latent class growth modeling (LCGM) to identify recovery patterns, and multinomial logistic regression analyses to investigate associations between potential risk factors and class membership (odds ratio [OR], 95% confidence interval [CI]).
Results: We identified 3 recovery patterns: “Fast responders” (59%), “Steady responders” (27%), and “Poor responders” (14%). Factors associated with “Steady responders” vs “Fast responders” were female vs male sex (OR 2.0, CI 1.5–2.7), ASA III–IV vs ASA I (OR 1.9, CI 1.2–3.1), Walch A1 vs B2 (OR 1.6, CI 1.1–2.5), and most vs medium socioeconomic deprivation (OR 1.4, CI 1.1–1.9). Factors associated with “Poor responders” vs “Fast responders” were ASA II vs ASA I (OR 2.0, CI 1.1–3.6), ASA III–IV vs ASA I (OR 3.0, CI 1.6–5.5), Walch A1 vs B2 (OR 2.1, CI 1.3–3.3), previous shoulder surgeries (OR 1.8, CI 1.3–2.4), most vs medium socioeconomic deprivation (OR 1.5, CI 1.2–2.0), RTSA for OA vs ATSA for OA (OR 1.8, CI 1.2–2.7), and RTSA for cuff arthropathy or rupture vs ATSA for OA (OR 2.3, CI 1.5–3.4).
Conclusion: 3 recovery trajectories were identified following TSA, which we labelled as “Fast responders,” “Steady responders,” and “Poor responders.” “Steady responders” and “Poor responders” were more likely to have higher ASA scores, a Walch A1 vs B2 classification, and greater vs medium socioeconomic deprivation than “Fast responders.” Moreover, “Steady responders” were more likely to be female, while “Poor responders” were more likely to have previous shoulder surgeries and RTSA for OA or for cuff arthropathy or rupture than “Fast responders.”