Deltoid Anatomy

The deltoid is the largest and most powerful muscle of the shoulder girdle, a triangular multipennate mass draping the joint from the clavicle, acromion and scapular spine to the deltoid tuberosity of the humerus. Its three heads have genuinely distinct actions rather than being three portions of one uniform sheet.

Two surgical realities. The axillary nerve crosses its deep surface at a predictable distance from the acromion, and every lateral approach, portal and plate placement is designed around that number. In the cuff-deficient shoulder the deltoid is the only motor left, which is why the entire reverse arthroplasty concept is a deltoid-tensioning operation and why an axillary nerve palsy is a contraindication to it.

Why the deltoid needs the cuff. With the arm at the side the deltoid's line of pull is almost parallel to the humeral shaft, so its resultant force is predominantly superior shear, not rotation. The rotator cuff, supraspinatus superiorly, subscapularis anteriorly, infraspinatus and teres minor posteriorly, compresses the head into the glenoid and creates a fulcrum, and only against that fulcrum does deltoid shear become a rotational moment. This is the coronal force couple: deltoid pulling superiorly, the inferior cuff pulling inferiorly and medially, their resultant being rotation about a stable centre.

When the cuff fails. The fulcrum goes with it. The deltoid then translates the head superiorly against the acromion instead of elevating the arm: pseudoparalysis, then acetabularisation of the acromion and femoralisation of the head. The reverse arthroplasty answer is to move the centre of rotation medially and distally, so that the prosthesis provides the fulcrum the cuff no longer can.

The middle head. Loss of the middle head, whether from axillary nerve injury or a failed deltoid repair, causes profound abduction weakness despite an intact rotator cuff.

View larger

Torque and moment arm. The deltoid as a whole contributes approximately 50 per cent of abduction torque, rising as elevation increases. Its abduction moment arm grows from roughly a centimetre at 0 degrees to around three centimetres at 60 to 90 degrees, the mechanical mirror image of the supraspinatus, whose moment arm is largest in the low range and diminishes with elevation.

The other two heads. The anterior and posterior heads become abduction synergists above 90 degrees, when their lines of pull cross the axis of rotation. Below 90 degrees they are respectively a flexor and an extensor, and in the very low range they can even oppose abduction.

Length-tension and the reverse. The multipennate middle head operates over a short excursion, so it is sensitive to changes in resting length: both under-tensioning and over-tensioning move it away from the optimum of its length-tension curve. A reverse arthroplasty deliberately manipulates this. Medialising the centre of rotation increases the deltoid moment arm, and distalising the humerus increases deltoid resting tension and recruits anterior and posterior fibres into elevation.

Synergists and antagonists. Supraspinatus abducts and compresses alongside it, the rest of the cuff provides the fulcrum, trapezius and serratus anterior rotate the scapula upward to provide the platform, the clavicular head of pectoralis major flexes with the anterior deltoid, and latissimus dorsi extends with the posterior deltoid. Against it, latissimus dorsi and pectoralis major adduct, and the inferior cuff depresses the head, opposing deltoid shear.

Scapulohumeral rhythm. Deltoid force is useless without a stable scapular base. The overall 2:1 glenohumeral to scapulothoracic ratio across the arc of elevation means that of 180 degrees of elevation roughly 120 degrees is glenohumeral and 60 degrees scapulothoracic, and the setting phase, in the first 30 degrees, is variable and predominantly glenohumeral. Upward rotation of the scapula by trapezius and serratus anterior keeps the glenoid beneath the head and keeps the deltoid at a favourable length, so a patient with serratus anterior palsy or trapezius palsy has a normal deltoid and a normal cuff and still cannot elevate properly. Always examine the scapula from behind before attributing weakness to the deltoid.

Inspection and palpation. All three heads are subcutaneous and directly palpable. Compare contour with the opposite side from the front, side and back: the squared-off shoulder of a dislocation and the flattened contour of deltoid wasting are inspection diagnoses.

Finding the landmarks. The deltopectoral groove is found by asking the patient to adduct against resistance, and the cephalic vein is often visible within it in a thin patient. The posterior border is best seen with the arm extended against resistance, which is also how the swallow-tail sign is elicited. The anterolateral acromial corner is the landmark from which all safe-zone measurements are made; mark it before draping.

Axillary nerve injury. The surgical mechanisms are:

• A deltoid split carried beyond the safe zone
• Retraction of a split without a stay suture
• Blind plate passage in minimally invasive osteosynthesis
• Inferior capsular release or inferior anchor placement close to the inferior glenoid rim
• Over-lengthening in reverse arthroplasty

The consequence is permanent deltoid palsy, which removes both active elevation and the option of a reverse arthroplasty. This is one of the few genuinely unsalvageable complications in shoulder surgery. Avoid it by measuring the safe zone from the lateral acromion, using stay sutures, identifying the nerve where the exposure requires it, avoiding deep self-retaining retractors, and respecting lengthening limits.

Deltoid dehiscence. It follows inadequate reattachment of the deltoid origin after an anterosuperior approach, or premature active use, and produces loss of abduction power, a visible contour defect and pain. Established dehiscence is one of the most difficult problems in shoulder surgery: direct reattachment to bone can be attempted, sometimes with a local fascial or allograft augmentation, but results are unpredictable and there is no reliable salvage. Prevention is everything, and this is the strongest argument for arthroscopic and mini-open techniques wherever they will do the job.

Acromial and scapular spine stress fracture after reverse arthroplasty. Increased deltoid tension after distalisation is the mechanism, magnified by lateralised designs and by osteoporotic bone. It typically presents at 3 to 12 months with new pain over the acromion and loss of elevation, and the investigation is CT, since radiographs are often normal. Management is usually non-operative with immobilisation; fixation is technically difficult and unreliable, so prevention through correct tensioning is the meaningful intervention.

Stiffness and disuse. Prolonged immobilisation for any reason, in a patient whose deltoid is the only motor, produces rapid disuse atrophy and stiffness. Maintain passive range even while protecting a repair.

Post-injection deltoid contracture. The fibrous band in the middle deltoid tethers the arm in abduction, so that adduction to the side is impossible and the scapula wings as the child attempts it. Treatment is surgical release of the band, with careful preservation of the axillary nerve, which lies deep to the fibrotic segment.

Cosmetic and sensory sequelae. Supraclavicular nerve injury during an anterior approach produces a numb patch below the incision that patients notice more than surgeons expect; warn about it at consent. Loss of anterior deltoid bulk after harvest or dehiscence is visible and distressing.

Related pages. The nerve, and what it supplies: Axillary Nerve Anatomy is the structure every safe zone on this page is drawn around; Teres Minor Anatomy shares its posterior branch, which is why isolated teres minor wasting localises the lesion. Brachial Plexus Anatomy places it on the posterior cord, and Nerve Injury and Regeneration supplies the Seddon grades and the one-millimetre-per-day rule quoted here.

Axillary nerve palsy after dislocation is the commonest nerve injury in shoulder trauma. An anterior glenohumeral dislocation stretches the nerve, which is tethered proximally at the posterior cord and distally in the quadrangular space and on the deep surface of the deltoid; fracture-dislocations and high-energy injuries add a direct component. The incidence is reported widely, from around 5 per cent of clinically detected cases to substantially higher proportions when electromyography is used, and it rises markedly with age over 60 and with fracture-dislocation.

Presentation and natural history. Deltoid weakness, wasting appearing over weeks, and variable sensory change over the lateral deltoid; isolated teres minor involvement occurs with posterior branch injury. The great majority are neurapraxic and recover spontaneously, typically within 3 to 6 months.

Investigation and management. Electromyography at 3 to 4 weeks establishes a baseline, because fibrillation potentials take that long to appear, and is repeated at 3 months to look for reinnervation. Management is observation with a maintained passive range and prevention of stiffness. If there is no clinical or electrophysiological recovery by 3 to 6 months, exploration is considered, applying the regeneration principle of roughly one millimetre per day so that the motor end plates are reached before irreversible atrophy. The reconstructive options are:

• Neurolysis
• Direct repair
• Interpositional grafting
• Triceps branch to axillary nerve transfer, which shortens the regeneration distance and avoids a graft

Imaging. Radiographs assess the acromion and scapular spine for stress fracture after reverse arthroplasty, and the anteroposterior view is where the deltoid tuberosity index is measured before proximal humeral fixation. MRI demonstrates deltoid denervation as diffuse oedema-like T2 signal in the subacute phase followed by fatty replacement, shows deltoid detachment or dehiscence directly, and shows isolated teres minor change in posterior branch lesions. Ultrasound gives a dynamic assessment of deltoid contraction and identifies a dehiscence at the acromion. Electromyography, timed as above, is the definitive investigation for a suspected nerve lesion.

Detaching and repairing the origin. The anterosuperior approach requires deltoid origin detachment. The origin must be taken with a continuous cuff of periosteum and fascia, or with a thin flake of acromial bone, and repaired through transosseous drill holes in the acromion at closure. Fascia-to-fascia repair fails, because the deltoid origin is under continuous load from the moment the patient starts using the arm. Protect the repair for six weeks with a sling and passive motion only, avoiding active abduction and resisted elevation.

Deltoid tensioning in reverse arthroplasty. A reverse arthroplasty is a deltoid operation. Tension is set by the combination of glenoid component position, humeral component height and polyethylene thickness. Too loose, and the result is instability and dislocation, with weak elevation because the deltoid works below its optimum length. Too tight, and the result is acromial and scapular spine stress fracture, axillary nerve and brachial plexus traction neurapraxia, pain, and loss of the last degrees of elevation.

The target. Roughly 1 to 2 cm of arm lengthening compared with the contralateral side. Beyond 2 to 3 cm the neurological and acromial risks rise sharply. Lateralised designs reduce scapular notching and improve rotation but increase the load transmitted to the glenoid baseplate and to the acromion; their tensioning target differs from a classic medialised Grammont design and must follow the specific implant system.

Judging tension on the table. Intra-operative assessment rests on:

• Tension of the conjoined tendon, which should be taut but not bowstringing
• The ability to distract the components a few millimetres with a trial in place
• Stability through a full range including adduction and internal rotation
• Comparison of humeral length on an intra-operative or immediate post-operative radiograph

Anatomical variation across populations

• The distance from the acromion to the axillary nerve scales with arm length, correlating strongly in cadaveric measurement, so a fixed 5 cm rule applied without adjustment is less safe in populations of smaller average stature and in any short-armed patient. Measuring against the patient in front of you is the correct practice everywhere.
• Accessory heads of the deltoid and variations in the raphes between heads are described in cadaveric series and can make the plane of a split less obvious.
• Os acromiale, an unfused acromial epiphysis, alters the deltoid origin mechanically and has variously reported prevalence between populations. It matters both as a cause of pain after decompression and as a site of failure after reverse arthroplasty, where deltoid tension is increased.

Side-by-side guidance

Registry and outcome signals

• National joint registries (AOANJRR, NJR, AJRR) show reverse total shoulder arthroplasty as the fastest-growing category of shoulder replacement, with good medium-term survivorship and an expanding set of indications including proximal humeral fracture in the elderly.
• Registries record revision, not deltoid function, so complications such as acromial stress fracture and deltoid weakness are under-represented; their true incidence comes from institutional series and is meaningfully higher than registry revision rates suggest.
• Signal across registries and cohorts: lateralised designs reduce scapular notching and improve rotation, at the cost of greater load at the baseplate and the acromion. The tensioning target is design-specific.
• Documented neurological assessment before and after reduction of a shoulder dislocation is an established quality standard in most systems, driven as much by medicolegal experience as by clinical evidence.

High-resource versus limited-resource practice

• Well-resourced settings: routine CT for suspected acromial stress fracture, electromyography for nerve assessment, reverse arthroplasty available for the cuff-deficient shoulder, and nerve transfer surgery for failed axillary nerve recovery.
• Limited-resource settings: the assessment that matters most, palpation of deltoid contraction in all three heads, requires nothing but hands and is more informative than any imaging. Where reverse arthroplasty is unavailable, preserving deltoid function through meticulous approach selection is even more important, since there is no salvage for a deltoid palsy.
• Wherever practice occurs, the two universal rules are the same: measure the split from the acromion, and document deltoid contraction before and after any shoulder reduction.