A fifty-five-year-old, right-hand-dominant man was seen with residual paralysis in the right shoulder eighteen months after an initially complete brachial plexus injury. He had been in a motor vehicle collision that resulted in a fracture-dislocation of the right shoulder. The fracture involved the lower portion of the glenoid and the inferior-lateral aspect of the scapula. He had a complete brachial plexus palsy at the time of the injury. He underwent open reduction and internal fixation of the scapular fracture within two weeks after the injury at an outside hospital. His original surgeon elected to manage the brachial plexus injury by observation. Within the first month after the injury, he started to recover hand function and then progressively, over a period of a year, he had functional recovery of the hand, wrist, forearm, and elbow. However, he had severe weakness of the shoulder girdle muscles secondary to partial palsy of the posterior cord of the brachial plexus and complete palsy of the axillary nerve and suprascapular nerve. This weakness had not improved at eighteen months after the injury, and he was referred to our institution for evaluation and treatment.
At the time of presentation, the patient had an examination of the right upper extremity and demonstrated a full range of motion of the elbow, forearm, wrist, and fingers. The muscle strength in these areas was equivalent to or slightly weaker than that in the contralateral upper extremity (i.e., grade M4 to M5, according to the British Medical Research Council [BMRC]) scoring system16). According to this scoring system, M0 indicates no muscle contraction; M1, muscle contraction but no motion; M2, muscle contraction with motion but gravity eliminated; M3, muscle contraction with motion against gravity; M4, muscle contraction against resistance; and M5, normal muscle strength. Inspection of the right shoulder showed evidence of severe atrophy of the deltoid and infraspinatus muscles. The shoulder range of motion was limited. The patient had active forward flexion to 50° and abduction to 40°, presumably from contraction of the clavicular head of the pectoralis major, a weak supraspinatus muscle, and a normal upper portion of the trapezius. He had no external rotation with the arm at the side or with the arm abducted (i.e., the infraspinatus and teres minor were graded M0). We believe that this difference in the manifestation of the injury between the supraspinatus and infraspinatus was related to the fracture of the lower body of the scapula, which could have caused additional local injuries to the suprascapular nerve branch to the infraspinatus as well as the axillary nerve. Active internal rotation was to the iliac crest, and he had mildly positive results on the belly-press and lift-off tests, which indicated a weak subscapularis. He had no evidence of winging of the scapula, but clinical examination demonstrated mild scapular dyskinesia. The importance of the dyskinesia could not be well demonstrated because of the very limited active range of motion of the shoulder. The passive range of motion of the shoulder was 150° of flexion, 100° of abduction, 40° of external rotation with the arm at the side, and internal rotation to T12. Shoulder flexion strength was graded M2 and the abduction strength was graded M2, while the external rotation strength was graded M0. The latissimus dorsi and teres major muscles were also weak (M2), and the deltoid was paralyzed (M0). The patient had normal strength (M5) of the pectoralis major and trapezius muscles.
Radiographic evaluation demonstrated the plate and screws in the inferior-lateral aspect of the scapula from the previous surgery. The glenoid articular surface appeared to be well reconstructed although there was some mild narrowing of the glenohumeral joint. The humeral head was centered on the glenoid with no evidence of inferior subluxation. We believe that the lack of inferior subluxation of the glenohumeral joint was due to the presence of two partially functioning rotator cuff muscles (the subscapularis and the supraspinatus). The electromyogram and nerve conduction study showed the infraspinatus, teres minor, and deltoid to be denervated with fibrillations. In addition, there was evidence of denervation with some reinnervation of the supraspinatus, subscapularis, latissimus dorsi, and teres major. The trapezius was normal.
The major complaint of the patient was the lack of shoulder function, specifically, the lack of external rotation. He reported that his hand was "stuck on [his] belly" when he flexed the elbow, requiring passive external rotation of the shoulder with use of the contralateral hand to place the hand in a position for function. Preoperatively, the Constant and Murley score17 was 24 points, the Disabilities of the Arm, Shoulder and Hand (DASH) score18 was 69 points, and the subjective shoulder value19 (which was determined by asking the patient to try to estimate subjectively the percentage of function of the affected shoulder compared with the normal, contralateral one) was 5%.
On the basis of this information, we elected to use the middle and lower segments of the trapezius to restore external rotation of the shoulder.
Surgical Technique
The patient was placed in the left lateral decubitus position. The position of the hand and arm was controlled with use of an articulated, hydraulic arm-holder (SPIDER arm holder; TENET Medical Engineering, Calgary, Alberta, Canada). A curvilinear incision was made, starting 2 cm medial and distal to the palpable medial edge of the spine of the scapula and extending proximally and horizontally over the spine of the scapula until the lateral edge of the acromion was reached. The posterior aspect of the deltoid appeared severely atrophic and did not respond to electrical stimulation. This part of the deltoid was detached from its insertion on the spine of the scapula, and sutures (number-2 FiberWire; Arthrex, Naples, Florida) were placed in the tendinous portion for later repair. The infraspinatus and teres minor muscles appeared to be severely atrophic and did not respond to electrical stimulation. The supraspinatus appeared mildly atrophic and responded weakly to electrical stimulation, without any resulting shoulder motion. On the basis of these findings, the lower and middle segments of the trapezius were dissected and detached with use of sharp dissection from the medial 3 cm of the spine of the scapula. Further medial dissection was performed to elevate the harvested muscle from its bed, taking care to identify and protect the spinal accessory nerve, which lies close to the vertebral attachment of the trapezius. A small segment (15 mm) of the medial spine of the scapula was excised in order to facilitate the passage of the transferred tendon (Fig. 1). We smoothed this surface with a rasp and covered it with skin allograft (GRAFTJACKET; Wright Medical, Arlington, Tennessee) to protect the transferred tendon from abrasion. It was necessary to extend the trapezius tendon with an Achilles tendon allograft sufficiently long to reach the insertion site of the infraspinatus on the greater tuberosity of the humeral head. The thin broad part of the tendon was wrapped around the musculotendinous unit of the lower portion of the trapezius and was sutured in place with multiple sutures (number-2 FiberWire) (Fig. 2).
The shoulder was then placed in 40° of external rotation with the arm at the side, avoiding extension or flexion of the shoulder. This allows the scapula to be lying in a neutral position on the chest wall midway between the retracted and protracted positions. The insertion site of the infraspinatus was exposed and split in its middle part along the length of its fibers. The central 70% of the insertion of the infraspinatus was elevated from the humerus. The remaining attached portion of the infraspinatus tendon was used later to reinforce the repair of the Achilles tendon to bone. The insertion site of the infraspinatus was then gently decorticated with a burr. Four absorbable 5.5-mm suture anchors (Bio-Corkscrew; Arthrex) were placed in the footprint of the infraspinatus, with two in the medial aspect and two in the lateral aspect. As an assistant applied tension to the end of the Achilles tendon, the tendon was laid on the prepared infraspinatus footprint and the sutures from the anchors were placed approximately 10 mm medial to the portion of the tendon that was touching the bone in order to add extra tension to the repair. Then, the allograft was repaired to the footprint with use of the anchored sutures, additionally weaving the sutures through the upper and lower tendinous portions of the elevated infraspinatus in order to reinforce the repair (Fig. 3). Further folding of the muscular portion of the infraspinatus was performed to envelop the Achilles allograft and provide it with additional vascularity (Fig. 4). Then, the deltoid was repaired back to its insertion site using number-2 FiberWire sutures that were passed through drill-holes in the spine of the scapula. The skin was closed over a suction drain. An external rotation brace (DonJoy, Vista, California) was applied before the patient was awakened from anesthesia.
Postoperative Protocol
The drain was removed after forty-eight hours. The patient was managed with the shoulder external rotation brace, allowing limited passive range of motion through an arc of 0° to 40° of external rotation, for six weeks. This motion was performed for fifteen minutes three times per day either by a therapist or a home assistant. At six weeks, the brace was discarded and the patient began active and active-assisted range-of-motion exercises. We found the transfer easy to train, making use of the coupling of external rotation to scapular retraction with the arm at the side. That is, the patient reeducated the transfer by attempting to retract the scapula, observing the resulting external rotation. Alternatively, with the arm at the side, the therapist placed the shoulder in external rotation and asked the patient to maintain this position while retracting the scapula.
After three months, a computed tomographic arthrogram was performed and the images were suggestive of graft integrity (Fig. 5). This postoperative study was not optimal to ensure integrity of the allograft muscle transfer construct. A magnetic resonance imaging study would have been better, but the patient had hardware in the scapula that would interfere with the quality of the study. At that point (i.e., three months after the transfer was performed and confirmed to be healed), the patient was started on gentle strengthening exercises for another six weeks, followed by a return to unrestricted activities.
Outcome
Outcome measures include shoulder range of motion, subjective shoulder value, the Constant and Murley score, the DASH score, and patient satisfaction. At the time of the last follow-up, nine months after surgery, the shoulder range of motion, specifically external rotation, had improved substantially. The patient had no pain, and the shoulder flexion had improved from 50° to 90°; abduction, from 40° to 80°; and external rotation with the arm at the side, from an inability to initiate any external rotation to 30° of active external rotation. Shoulder flexion and abduction strength had improved from M2 to M4, and external rotation strength, from M0 to M5. There was no evidence of scapular winging or dyskinesia. The subjective shoulder value had improved from 5% to 40%; the Constant and Murley score, from 24 to 69 points; and the DASH score, from 69 to 37 points. The patient was very satisfied with the result of the surgery, and he stated that he would undergo the procedure again. He was able to perform many activities of daily living without problem and was swimming on a daily basis with no complaints.