Predictive factors for humeral and ulnar stem loosening after total elbow arthroplasty: a retrospective cohort study.

Background: There has been a significant increase in the use of total elbow arthroplasty (TEA) in advanced osteoarthritis, post-traumatic arthritis, and complex distal humerus fractures. The purpose of this study is to determine the incidence of total elbow loosening and association with patient demographics, implant specifications, and cement mantle characteristics. Methods: A retrospective cohort study was conducted of patients who underwent primary TEA between 2013 and 2022 at a single institution. Revisions and conversions from prior open reduction and internal fixation were excluded. Pre-operative distal humeral cortical thickness and bone quality were assessed. Variables included anterior humeral flange contact with bone, graft placement behind the flange, heterotopic ossification, humeral stem thickness and length, and cement mantle length. Cement mantle quality was graded as adequate, marginal, or inadequate. Radiographic loosening of the humeral and ulnar components was assessed across 4 zones and categorized as focal (1 to 2 zones) or diffuse (3 to 4 zones). Increased translucency from initial post-operative radiograph, cement fractures, and component migration were assessed. Results: A total of 166 TEAs with an average follow-up of 24 months were reviewed in 130 females (78.3%), and 36 males (21.7%). Most common diagnoses were fracture (45.2%), rheumatoid arthritis (26.5%), and osteoarthritis (10.8%). Humeral loosening was identified in 68 elbows (41%) and ulnar loosening occurred in 55 elbows (33%). Cement mantle quality and length was significantly associated with humeral and ulnar loosening. Longer cement mantles were associated with a lower risk of ulnar component loosening. For each millimeter increase in ulnar mantle length, the odds of loosening decreased by 13%. Adequate humeral mantle was identified in only 100 patients (61%) and an adequate ulnar mantle in only 116 (70%). Cases without an adequate mantle had significantly higher rates of diffuse and focal loosening. At final follow-up, increased periprosthetic translucency was observed in 16 elbows (9.6%), cement fracture in 24 (14.5%), and component migration in 20 (12.0%). Diagnoses, implant type, flange contact, graft placement under flange, and heterotopic ossification formation were not different between cases with loosening, component migration, or cement fracture. Discussion/Conclusion: Diffuse humeral and ulnar component loosening is common after TEA at 21% and 13%, respectively, and poor cement mantle quality and length are significant predictive factors of aseptic humeral and ulnar loosening.