Total Shoulder Arthroplasty

Total shoulder arthroplasty is the definitive treatment for end-stage glenohumeral arthritis, in two forms: the anatomic replacement for a shoulder with an intact rotator cuff, and the reverse for a cuff-deficient one. Rotator cuff integrity is what decides between them. Glenoid component selection and fixation remain the weak link, and glenoid loosening is the most common long-term failure mode.

Indications. The joint is replaced for:

• Glenohumeral osteoarthritis, primary or secondary
• Cuff tear arthropathy (reverse TSA)
• Inflammatory arthritis (RA) with severe joint destruction
• Osteonecrosis (AVN) of the humeral head
• Acute proximal humerus fracture (reverse TSA, over 70 years old)
• Failed prior surgery (hemiarthroplasty, fixation)

The rise of the reverse. Reverse TSA has revolutionised the treatment of cuff tear arthropathy and of proximal humerus fractures in the elderly. In the USA National Inpatient Sample, primary shoulder arthroplasty rose 103.7% between 2011 and 2017, driven by reverse TSA, which rose 191.3% to 63,845 procedures in 2017 (Wagner et al 2020). Reverse TSA is now the most commonly implanted primary shoulder arthroplasty (AOANJRR, NJR, AJRR data), exceeding anatomic in most registries, and the AOANJRR reports it as the majority of primary shoulder replacements.

The patient. Primary glenohumeral osteoarthritis has a female predominance, and the mean age is approximately 70 years.

The bones. The glenoid is pear-shaped, about 39mm high and 29mm wide, and natively sits in 2-10° of retroversion. The humeral head measures 45-55mm across with a 130° head-shaft angle. An anatomic design reproduces the offset centre of rotation, and with it normal kinematics.

The rotator cuff. Each part of the cuff has a role that the arthroplasty depends on:

• Subscapularis: the anterior dynamic stabiliser, which also resists posterior subluxation
• Supraspinatus: superior stabiliser and deltoid synergist
• Infraspinatus and teres minor: posterior stability and external rotation

A massive tear lets the head migrate superiorly and ends in cuff tear arthropathy.

Nerves at risk. The axillary nerve crosses the inferior capsule 5-10mm from the glenoid rim at the 6 o'clock position (average 7mm), which puts it at risk during inferior capsular release. The musculocutaneous nerve enters coracobrachialis 5-8cm distal to the coracoid tip, and aggressive medial retraction puts it at risk.

History. Establish what the shoulder costs the patient and what they want back from it:

• Pain: location, severity, night pain, functional limitation
• Function: activities of daily living and work affected
• Prior treatment: physiotherapy, injections, previous surgery
• Expectations: activity demands, occupation, goals
• Comorbidities: diabetes, smoking, immunosuppression

Examination. The examination is aimed at the cuff and the deltoid, because they decide the implant.

• Active against passive range of motion: pseudoparalysis suggests a cuff tear
• Cuff strength: lag signs, drop arm, external rotation weakness
• Instability: load and shift, sulcus sign
• AC joint: tenderness, cross-body adduction pain
• Neurovascular: axillary nerve (deltoid sensation), radial pulse

Radiographs are essential: a true AP, a scapular Y and an axillary lateral, with the contralateral shoulder for comparison. They show the arthritis, the glenoid version and wear pattern, the bone stock and any superior migration: an acromiohumeral distance of less than 7mm suggests a massive cuff tear.

View larger

CT with 3D reconstruction is standard. It measures glenoid version, bone stock and the Walch wear pattern, and it is the study on which component position is planned and the need for a bone graft or augmented component decided.

MRI is added when cuff status is unclear. It shows cuff integrity and fatty infiltration (Goutallier classification): stage 3-4 infiltration suggests an irreparable tear, and reverse TSA should be considered. It also assesses the subscapularis, whose integrity is critical for an anatomic replacement.

The decision. Rotator cuff integrity is the key. An anatomic TSA needs an intact cuff, especially the subscapularis and supraspinatus; a reverse TSA is for cuff deficiency, massive tears or cuff tear arthropathy, so preoperative imaging must assess the cuff.

View larger

Intact cuff to anatomic, deficient cuff to reverse is the starting rule, and it does not settle every case. The B2 and C glenoid already carry reverse TSA as an option, and a 75-year-old with a B2 glenoid should consider a reverse.

Reverse for the difficult glenoid. Reverse TSA is increasingly chosen for primary glenohumeral OA even when the cuff is intact, if the glenoid is severely retroverted or biconcave (Walch B2/B3) or has posterior bone loss that an anatomic TSA cannot reliably reconstruct. The dominant long-term failure of anatomic TSA is glenoid loosening, driven by eccentric posterior loading on a retroverted or biconcave glenoid. A reverse baseplate, fixed with a central post and peripheral screws into the scapular body, tolerates posterior bone loss and version far better than a cemented anatomic glenoid and removes that failure mode.

The trade. The reverse exchanges glenoid loosening for its own profile: scapular notching, acromial and scapular-spine stress fracture, limited external rotation, and non-anatomic, medialised kinematics.

Who gets which. In the difficult glenoid the call turns on glenoid bone stock plus age and demand, not cuff status alone. The older, lower-demand patient with a severely deformed or bone-deficient glenoid is well served by a reverse, which needs no perfect version correction. The younger patient with a correctable glenoid and an intact cuff is still better served by an anatomic TSA with version correction.

Why the reverse is not the default. Registries show reverse TSA carries a lower cumulative revision rate than anatomic TSA at comparable follow-up, in the early and mid term. A lower revision rate does not make it the better operation for every patient.

• With an intact cuff, anatomic TSA restores near-normal kinematics and external rotation. The reverse sacrifices external rotation and imposes a medialised, non-anatomic centre of rotation
• Reverse-specific problems accumulate: notching, acromial and scapular-spine stress fractures, and glenosphere and baseplate wear accrue over time. There is little long-term (15-20 year) outcome data for reverse in younger, higher-demand patients, which is why it is controversial in the young and a relative contraindication under about 65
• A failed anatomic TSA, typically from glenoid loosening or cuff failure, can usually be revised to a reverse. A failed reverse is a harder, more bone-consuming revision, so choosing anatomic in the younger intact-cuff patient keeps the reverse in reserve

Shoulder infection is its own disease. It is not hip or knee infection with a different incision. The dominant pathogen is Cutibacterium acnes (formerly Propionibacterium acnes), an anaerobic commensal of the sebaceous glands that are dense around the shoulder girdle, which is also why infection is substantially commoner in men.

How it presents. C. acnes is low-virulence and indolent. There is often no sinus, no erythema and no fever, CRP and ESR are frequently normal, and the presentation is simply a stiff or painful arthroplasty, or unexplained early loosening. It is the reason a shoulder revision that "looks aseptic" so often is not: all three deep infections in the Neer Award series were in the revision group, and its authors advise vigilance for low-grade infection when revising.

Treatment by timing.

• Acute, under 4 weeks. I&D with component retention, exchange of modular parts and long-term suppressive antibiotics, with a success rate of 50-70%
• Chronic, over 4 weeks. Two-stage revision is the gold standard: component removal, an antibiotic spacer for 6-12 weeks, then reimplantation, with a success rate of 80-90%
• Culture-negative. In the shoulder, treat "culture-negative" as a hypothesis about the culture rather than a conclusion about the joint. Biofilm and insufficient incubation are the usual explanations

Activities. The elbow and hand are unrestricted from the start.

• No lifting over 5 lbs for 12 weeks
• Driving at 6-8 weeks, if off opioids with adequate range of motion
• Return to work at 4-6 weeks for sedentary work, 3-6 months for manual labour
• Contact sports avoided permanently, for the risk of dislocation or fracture

Follow-up.

• 2 weeks: wound check, sutures or staples out, range of motion assessed
• 6 weeks: radiographs (AP, axillary, scapular Y), progress to active motion
• 12 weeks: clinical and radiographic assessment, full range expected
• 6 months, 1 year and 2 years: surveillance for loosening and wear
• Annually thereafter if asymptomatic

Scores. Functional scores improve after arthroplasty:

• ASES score from 30-40 before surgery to 75-85 after
• Constant score from 20-30 to 60-70
• Pain VAS from 7-8/10 to 1-2/10
• Forward elevation improves by 50-80°; external rotation is variable
• 70-80% return to golf or swimming

Registries. The common indications for revision are instability or dislocation, infection, and glenoid component loosening.