Increased peri-implant diaphyseal fractures after shoulder arthroplasty with long stem vs. short stem implants

Background: The purpose of this study was to assess the impact of stem design on periprosthetic humerus fracture subtype after shoulder arthroplasty. Methods: A single tertiary referral center database was retrospectively reviewed for all patients with periprosthetic humerus fracture after shoulder arthroplasty between January 2013 and March 2024. Index arthroplasty type, stem length, stem design, fracture type, and management were recorded. Stem length was classified as stemless, short stem (<90 mm), or long stem (≥90 mm). Fracture types were classified using the Unified Classification system (UCS)–Sanchez-Sotelo classification system on postfracture radiographs (type 1: tuberosity fracture; type 2: peri-implant; type 3: distal). Type 2 fractures were subcharacterized based on stem stability and remaining bone stock (2A: well-fixed implant; 2B: loose implant with adequate bone stock; 2C: loose implant with inadequate bone stock). Results: Ninety periprosthetic shoulder fractures were included. Fractures involved 60 reverse shoulder arthroplasties, 20 anatomic shoulder arthroplasties, 7 hemiarthroplasties, and 3 antibiotic spacers. Stem fixation data were available for 71 shoulders, of which 42 were long stems, 28 were short stems, and 1 was stemless. Seventy patients (77%) experienced a peri-implant fracture (type 2), followed by 8 fractures (9%) below the stem (type 3), and 6 (7%) tuberosity fractures (type 1). Fracture type was not assigned in 6 cases because of either lack of adequate radiographs, fracture around an articulating antibiotic spacer, or intraoperative fracture. Overall, 77% of fractures occurred around the prosthesis (type 2) and were stable (70%). The observed odds ratio of a peri-implant fracture (type 2) with long stems compared to short stems was 4.08 (95% confidence interval: 1.19, 14.07; P = .03). Short-stemmed implants underwent revision arthroplasty at a higher rate than long stem implants (32% vs. 14%), whereas long stem implants were more likely to initially be treated nonoperatively (31% vs. 11%) (P = .03). In addition, stable stems more frequently were managed with open reduction internal fixation compared to unstable stems (61% vs., 13%), while unstable stems more frequently underwent revision than stable stems (48% vs. 12%) (P < .001). Conclusion: Most periprosthetic humerus fractures after shoulder arthroplasty were diaphyseal, occur around the existing prosthesis (77%), and resulted in a stable implant (70%). Fractures around short stem implants were more likely to need revision. Humeral stem design does appear to influence fracture characteristics, as patients with long stems were more likely to sustain a type 2 (peri-implant) fracture than those with short stems.

Duke Scholars

Author Christopher Scott Klifto Orthopaedic Surgery

Author Oke Adrian Anakwenze Orthopaedic Surgery

Author Chad E. Cook Orthopaedic Surgery, Physical Therapy