Operative Management of Proximal Humerus Nonunions in Adults: A Systematic Review.

Objectives: To characterize the literature on operative interventions for proximal humerus nonunions in adults. Second, to identify prognostic factors associated with outcomes for locked plate open reduction and internal fixation (ORIF).

Data Sources: PubMed, EMBASE, MEDLINE, Scopus, and Cochrane databases were searched for articles from 1990 to 2020.

Study Selection: Studies reporting outcomes of proximal humerus nonunions managed with ORIF, hemiarthroplasty (HA), total shoulder arthroplasty (TSA), or reverse TSA (RTSA) were included. Studies failing to stratify outcomes by treatment or fracture sequelae were excluded.

Data Extraction: Two authors independently extracted data and appraised study quality using MINORS score.

Data Synthesis: Descriptive statistics were reported. Outcomes for ORIF and arthroplasty groups were not compared due to differing patient populations.

Conclusions: Thirty-seven articles were included, representing 508 patients (246 ORIF, 137 HA/TSA, and 125 RTSA). Patients managed by ORIF were younger with simpler fracture patterns than those managed by arthroplasty. Regarding ORIF, locked plates achieved highest union rates (97.0%), but clinical outcomes were comparable with all plate fixation constructs [forward flexion (FF): 123-144°; external rotation: 42-46°; Constant score: 75-84]. Complication and reoperation rates for ORIF were 26.0% and 14.6%, respectively. Furthermore, subgroup analysis of locked plate ORIF demonstrated shorter consolidation time with initial conservative fracture management (4.3 vs. 6.0 months) and autograft use (3.9 vs. 5.5 months). With arthroplasty, RTSA demonstrated greater forward flexion (109.4° vs. 97.2°) but less external rotation (16.5° vs. 36.8°) than HA/TSA. Complication and reoperation rates were 18.2% and 10.9% for HA/TSA and 21.6% and 14.4% for RTSA, respectively.

Level Of Evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.