Chronic posterior shoulder dislocation : current evidence, treatment options, and a decision-making algorithm

Abstract

Background and aims Chronic posterior shoulder dislocation (CPSD), also known as locked or chronic posterior glenohumeral dislocation, is a rare but frequently missed injury, accounting for 2–5% of shoulder dislocations. It often results from trauma, seizures, or electrical shock, leading to reverse Hill-Sachs lesion (rHSL), posterior glenoid erosion, and soft-tissue contractures. Delayed diagnosis (50–79% initially overlooked) transforms acute cases into chronic ones (>3–6 weeks), complicating management and increasing risks of instability, osteoarthritis, and poor function. This narrative literature review aims to synthesise current evidence on the aetiology, diagnosis, imaging, treatment strategies, and outcomes of CPSD, and to propose a surgical decision-making algorithm to guide clinicians. Methods A comprehensive search was conducted (up to March 13, 2026). Inclusion criteria included peer-reviewed articles (2006–2026) on adult patients with chronic PSD, focusing on interventions, clinical outcomes (pain, ROM, function, complications), and rehabilitation. Narrative synthesis was performed. Results Aetiology is predominantly traumatic; pathoanatomy features rHSL (30–90% of cases) and posterior capsulolabral damage. Diagnosis relies on axillary radiographs, CT for bony defects, and MRI for soft tissues. Treatment depends on dislocation duration, rHSL size, glenoid integrity, and patient factors. A stepwise algorithm integrates reducibility, defect size, and associated pathologies. Conclusion CPSD requires high clinical suspicion and advanced imaging for timely diagnosis. Joint-preserving reconstructions yield superior outcomes in moderate defects with early intervention; arthroplasty is preferred for extensive damage. The proposed algorithm provides structured guidance, though high-level evidence remains limited due to rarity and study heterogeneity. Future multicenter RCTs and standardised classifications are needed.