Shoulder Hemiarthroplasty for Acute Proximal Humerus Fracture

Indication. A 3-part or 4-part proximal humerus fracture in an elderly patient (greater than 70 to 75 years), a head-splitting fracture (greater than 40 percent of the articular surface), a fracture-dislocation with a high predicted risk of avascular necrosis, or a failed ORIF with head collapse or AVN — where the humeral head cannot be salvaged. A 1-part fracture (all fragments displaced less than 1 cm or angulated less than 45 degrees) is managed non-operatively. The formal Neer classification that drives this decision is laid out under Background and Evidence. The modern shift to reverse shoulder arthroplasty. This is an evolving area. Reverse shoulder arthroplasty (RSA) is increasingly preferred over hemiarthroplasty for acute complex fractures in the elderly because its function is independent of tuberosity healing. The one decision to make before scrubbing is which implant to commit to:

Preoperative planning. Review an AP in the scapular plane, a scapular-Y and an axillary lateral, plus a CT with 3D reconstruction — the CT is essential to define fragment size, displacement, comminution and any head split. Measure the contralateral normal shoulder for head size and height landmarks. Assess bone quality (DEXA and cortical thickness on CT), rotator-cuff integrity (a chronic tear favours RSA), and optimise the patient medically (diabetes control, smoking cessation, nutrition). Plan the implant: head size from CT (typically 43 to 49 mm), a standard versus long stem (long if metaphyseal comminution extends greater than 5 cm), cemented fixation in the fracture setting, and have RSA components available as a backup. Consent specifically for the dependence of the result on tuberosity healing (a 20 to 40 percent nonunion rate with poor function and possible revision to RSA), axillary nerve injury (5 to 10 percent already injured by the fracture, 1 to 2 percent iatrogenic), stiffness, instability, infection (2 to 4 percent, higher than elective arthroplasty), and the option of primary RSA. Setup. Beach chair at 60 to 80 degrees (typically 70), head in a horseshoe headrest, a bump under the medial scapular border to stand the glenoid vertical, the whole arm draped free including the hand. C-arm fluoroscopy from the contralateral side; a fracture table is not needed. General anaesthesia with an interscalene block for postoperative analgesia — avoid indwelling catheters because of the phrenic-nerve-palsy risk. Document the preoperative axillary nerve examination before draping.

The goal is to replace the unsalvageable head while restoring anatomic prosthetic height and version, then to reconstruct the tuberosities around the stem with bone graft and a heavy suture construct that will heal and restore rotator-cuff function. The exposure is the extensile deltopectoral approach laid out in full below (and in depth on the deltopectoral approach to the shoulder page).

View larger

Rehabilitation. Tuberosity protection is the priority for the first 6 weeks — abduction pillow at all times, passive motion only. | Phase | Timing | Immobilisation | Therapy | |-------|--------|----------------|---------| | 1 | 0 to 6 weeks | Sling with abduction pillow at 30 degrees at all times | Pendulums from day 1 to 2; passive elevation to 90 degrees by week 4 and 120 by week 6; passive external rotation to 20 degrees by week 4 and 30 by week 6; no active motion, no lifting | | 2 | 6 to 8 weeks | Wean the abduction pillow on radiographic evidence of healing | Active-assisted elevation 90 to 140 degrees, external rotation to 40 to 45 degrees, begin gentle internal rotation | | 3 | 8 to 12 weeks | Sling for comfort only | Active elevation to 120 to 140 degrees, begin gentle isometrics | | 4 | 12 to 16 weeks | None | Theraband and light weights 0.5 to 1 kg once there is cortical bridging on 3 of 4 cortices | | 5 | 4 to 6 months | None | Progressive functional strengthening; plateau by 6 to 12 months | Expected outcomes depend almost entirely on tuberosity healing. If the tuberosities heal (60 to 80 percent with bone graft and good fixation), expect 100 to 130 degrees of forward elevation, 30 to 45 degrees of external rotation, minimal pain and 70 to 80 percent satisfaction. If they do not heal (20 to 40 percent), expect pseudoparalysis with elevation less than 90 degrees, moderate to severe pain, satisfaction less than 40 percent, and often conversion to RSA. For comparison, primary RSA for acute fractures achieves 120 to 140 degrees of elevation and 85 to 90 percent satisfaction regardless of tuberosity healing. Follow-up. Wound check and drain removal at 2 weeks; radiograph (AP, scapular-Y, axillary) at 6 weeks to assess tuberosity healing and advance to active-assisted motion, again at 12 weeks to confirm progressive union and begin strengthening, at 6 months for final tuberosity position, at 1 year for function, then 2-yearly to monitor for late loosening, glenoid erosion and heterotopic ossification. Complications

Epidemiology. Proximal humerus fractures are among the commonest fragility fractures of the elderly, classically grouped with fractures of the hip, distal radius and spine as the osteoporotic fracture sites. Incidence rises sharply with age, with a female predominance and a peak in patients over 70 — the population in whom arthroplasty is considered. Most result from a low-energy fall onto the outstretched hand in osteoporotic bone. Pathoanatomy — why 3- and 4-part fractures need arthroplasty. The humeral head blood supply is dominated by the ascending branch of the anterior circumflex humeral artery, which travels in the lateral bicipital groove and enters the head as the arcuate artery (greater than 70 percent of supply); the posterior circumflex humeral artery contributes the remainder. In a 3-part fracture the arcuate artery is usually disrupted (AVN risk 15 to 25 percent); in a 4-part fracture the head is completely free of soft-tissue attachment and the AVN risk reaches 75 to 90 percent. This is the rationale for replacing rather than fixing the head in displaced 4-part fractures in the elderly. The Neer classification still drives management. Displacement is defined as 1 cm of translation or 45 degrees of angulation.

Key evidence. Boileau (2002) showed that tuberosity malposition occurred in half of 66 hemiarthroplasties for fracture and was the dominant reason for poor outcomes — mean active elevation only 101 degrees — with excessive prosthetic height and retroversion the correctable surgeon errors. Two cadaveric CT studies established the pectoralis major insertion as the height landmark: Torrens (2008) found the upper edge a mean of 5.64 cm distal to the head top (and 24.65 degrees to the posterior fin for version), and Ponce (2013) refined this to about 58.9 mm in men and 55.2 mm in women, scaling with patient height. On the implant-choice question, Sebastiá-Forcada (2014) randomised patients over 70 to RSA versus hemiarthroplasty and found RSA superior (forward elevation 120 versus 80 degrees), with only 56.6 percent of hemiarthroplasty tuberosities healing and function independent of tuberosity healing in the RSA group. Gallinet (2009) confirmed RSA gives better elevation and abduction at the cost of rotation and scapular notching. This consistent message — tuberosity healing is the determinant of hemiarthroplasty outcome, and RSA removes that dependence — is why practice is shifting toward RSA for elderly complex fractures.