Latarjet / Coracoid Transfer for Anterior Shoulder Instability

Indication. The Latarjet coracoid transfer is performed for recurrent anterior shoulder instability with significant bone loss or a high-risk soft-tissue failure pattern — the setting in which an arthroscopic Bankart repair alone has an unacceptably high recurrence rate. It is an open, bony augmentation procedure. Absolute indications (strong evidence): - Glenoid bone loss greater than 20–25% on 3D CT (best-fit circle method) — the inverted-pear morphology on a plain radiograph is the clinical proxy

• Off-track Hill-Sachs lesion: the Hill-Sachs interval is wider than the glenoid track (Di Giacomo glenoid-track concept) even with glenoid bone loss less than 20%
• Failed arthroscopic Bankart with recurrent instability — especially with progressive bone loss
• ISIS score of 6 or greater in a contact athlete — a 70% predicted failure rate from arthroscopic Bankart justifies a primary Latarjet Strong relative indications: - Bone loss of 13.5–20% in a high-demand contact athlete (rugby, football) or an ISIS score of 4–5
• Epilepsy with recurrent seizure-related dislocations — repeated re-dislocation mandates a durable bony reconstruction
• Poor labral tissue quality (attenuation, absence, prior failed repair) — there is no tissue to repair arthroscopically
• A significant Hill-Sachs lesion (greater than 25% of humeral head width) where remplissage would excessively limit external rotation The decision — when to abandon arthroscopic Bankart. Bone loss, not soft-tissue technique, is the dominant driver of Bankart failure (Burkhart and De Beer, 2000). The Instability Severity Index Score (ISIS, Balg and Boileau 2007) quantifies that risk pre-operatively. It is a 10-point score (range 0–10 — it caps at 10, not 11): a score greater than 6 predicts a 70% recurrence risk after arthroscopic Bankart, supporting a primary Latarjet instead.

Consent. Counsel specifically on the true early complication burden: transient or permanent nerve injury (most often the musculocutaneous nerve, 0.8–8%), recurrent instability (5–8%), graft lysis on CT (15–30% at 5 years, which does not equal failure), hardware problems (3–5%), and glenohumeral arthritis progression (around 30% radiographic arthritis at 10 years). Document elbow flexion and lateral forearm sensation on day one. Setup. Beach-chair position, 30–45 degrees head-up, arm free-draped over the edge of the table. Mark the coracoid tip, anterior acromion and deltopectoral groove before prep — the coracoid is the key landmark, and all dissection references from it.

The goal is to expose the anterior glenoid through the deltopectoral approach, harvest the coracoid, pass it through a subscapularis split, and fix it flush to the anteroinferior glenoid rim — restoring stability by three mechanisms: a bone block that extends the articular arc, the sling effect of the conjoint tendon, and capsular reconstruction using the preserved coracoacromial ligament stump. The exposure and coracoid harvest are laid out in full as the first steps below.

View larger

View larger

View larger

Rehabilitation | Phase | Timing | Immobilisation | Therapy | |-------|--------|----------------|---------| | 1 — Protection | Weeks 0–4 | Sling in 15 degrees forward flexion, neutral rotation | Pendulum exercises from day 1; hand, wrist, elbow ROM unrestricted; no active shoulder motion against resistance | | 2 — Early motion | Weeks 4–8 | Sling discontinued at 4 weeks | Active-assisted ROM: elevation in the scapular plane to 90 degrees, external rotation to neutral then 20 degrees; scapular stabilisation; no subscapularis strengthening until week 6 | | 3 — Strengthening | Weeks 8–16 | — | Full active ROM; progressive rotator-cuff strengthening; closed-chain proprioception; sport-specific (non-contact) drills from 12 weeks; optional CT at 12 weeks to confirm graft incorporation | | 4 — Return to sport | Months 4–6 | — | Contact training from 4 months when full strength and ROM achieved; full contact sport return at 6 months minimum; criterion-based discharge (greater than 90% limb symmetry index, full pain-free ROM) |

Expected outcomes. Recurrence rate 5–8% overall (2–4% in appropriately selected patients at 10 years); return to sport 74–97% across series (contact athletes 80–90%); mean external rotation loss 5–10 degrees (clinically acceptable); CT-confirmed graft union in 70–85% at 6 months; 85–95% satisfied or very satisfied at 5 years. Complications

Graft lysis does NOT equal failure. CT evidence of graft resorption appears in 15–30% at 5 years, but the sling mechanism preserves stability provided the conjoint tendon remains correctly positioned. Revise only for symptomatic recurrent instability when CT confirms the sling is compromised (graft medially displaced or conjoint tendon detached). Factors associated with lysis include screw divergence, a thin graft, excessive bone-bed preparation and infection. Long-term arthritis. Arthritis progression is the most significant long-term concern — around 30% radiographic arthritis at 10 years, comparable to post-instability arthritis without surgery. Causes include an over-lateral graft (proud screw heads), hardware impingement and loss of normal articular geometry. When severe, reverse shoulder arthroplasty is preferred over anatomic TSA because the subscapularis is unreliable after Latarjet.

Critical exam pitfalls

Why bone loss drives failure. Anterior dislocation is the most common direction of shoulder dislocation, and recurrence after a first dislocation is frequent — especially in the young, contact-sport population. The dominant determinant of failure after a soft-tissue (Bankart) repair is bony deficiency: an inverted-pear glenoid (loss of the normal inferior glenoid width) and an engaging Hill-Sachs lesion. Burkhart and De Beer (2000) showed that shoulders with significant bone defects had a 67% recurrence rate after arthroscopic Bankart versus 4% without, and 89% in contact athletes with defects — the foundational rationale for choosing Latarjet. Surgical anatomy that governs the operation. - Subscapularis and the split. Subscapularis arises from the subscapular fossa and inserts on the lesser tuberosity, with distinct upper (upper subscapular nerve, C5) and lower (lower subscapular nerve, C5–C6) neurovascular territories. The split is made at the junction of the lower two-thirds and upper one-third — the intact lower two-thirds becomes the sling, tensioning in abduction-external rotation. A split that is too superior damages the upper half (responsible for internal-rotation strength) and weakens the lift-off and belly-press.

• Coracoid and conjoint tendon. The coracoid projects anterolaterally from the supraglenoid scapular base; its tip lies 1–2 cm medial and 1–2 cm inferior to the AC joint. Pectoralis minor inserts on the medial coracoid (released during harvest); coracobrachialis and the short head of biceps arise at the tip (preserved as the conjoint tendon). The flat inferomedial coracoid surface is the graft contact face; the deep coracoid surface bears a hyaline-like fibrocartilage. Only the coracoacromial ligament is divided — the coracoclavicular ligaments (conoid and trapezoid) are preserved.
• Axillary nerve. Exits the quadrangular space with the posterior circumflex humeral artery and courses anteroinferiorly along the inferior subscapularis, lying approximately 12–15 mm inferior to the glenoid rim at the 6 o'clock position. At risk during inferior capsulotomy and inferior screw placement; injury (less than 0.5%) causes deltoid paralysis and lateral shoulder numbness.
• Anterior circumflex humeral vessels. Arise from the third part of the axillary artery and run along the lower border of subscapularis (contributing the arcuate artery to humeral head supply). Coagulate with bipolar before division during the inferior capsulotomy. The decision tools — ISIS and the glenoid track. Balg and Boileau (2007) derived the 10-point ISIS in 131 patients; a score greater than 6 carried a 70% recurrence risk after arthroscopic Bankart, recommending a Bristow-Latarjet instead (the score was validated multicentre by Rouleau, ICC 0.933). Yamamoto and Itoi (2007) defined the glenoid track: the medial margin of glenohumeral contact lies a mean 84% of the glenoid width medial to the rotator-cuff footprint, so a Hill-Sachs lesion extending medial to that margin is off-track and at risk of engagement — justifying bony augmentation in bipolar bone loss even when glenoid loss alone is subcritical.