A fifty-two-year-old right-hand-dominant woman presented with left shoulder pain and deformity. The patient had cerebral palsy characterized by spastic left hemiparesis and athetosis. She noted that the shoulder and elbow deformity became more prominent while she was walking, as is typical of upper motor neuron disorders with spasticity, making balance and walking difficult. She had been diagnosed with severe degenerative osteoarthritis and an inferior glenohumeral subluxation, which had been unresponsive to nonoperative treatment including botulinum toxin injections, intra-articular steroid injections, and a suspension orthosis to reduce the glenohumeral subluxation.
On physical examination, the patient had marked spasticity with only minimal volitional movement of the left shoulder. There was no active flexion or extension of the elbow, wrist, or fingers. There were, however, spontaneous athetoid movements of the left upper extremity as well as a positive sulcus sign. Passive shoulder motion was limited by pain caused by underlying arthritis and by increased tone. The resting shoulder posture was 15° of extension, 60° of abduction, and 30° of internal rotation. The left elbow was held in 90° of flexion, although it could be moved passively from 0° to 150°. Radiographs (Fig. 1-A) and magnetic resonance imaging (MRI) (Fig. 1-B) demonstrated advanced degenerative arthritis of the glenohumeral joint and inferior glenohumeral subluxation but no evidence of a rotator cuff tear or other associated intra-articular shoulder pathology.
In creating a treatment plan, we were uncertain whether there was any substantial volitional movement of the left upper extremity since this can often be masked by increased upper-extremity tone. The clinical question was whether the limited motion was a result of absent or weak activity of the shoulder muscles, or the result of dyssynergy or cocontraction of the antagonist muscles during attempted shoulder movement. There was also concern that the spasticity could cause instability of the shoulder following surgery. The patient underwent electrodiagnostic evaluation of the left shoulder and elbow. Dynamic polyelectromyography (poly-EMG) was used to identify dyssynergy or cocontraction. Dynamic poly-EMG recordings from specific muscles (the long head of the triceps, pectoralis major, teres major, and latissimus dorsi) in conjunction with the movement tracings provided detailed information regarding the activity of each muscle. Dynamic poly-EMG demonstrated that volitional muscle activity was present but limited. Additionally, it was noted that dyssynergic activity increased in the long head of the triceps and also in the entire deltoid muscle (Fig. 2-A) during active attempts at walking and shoulder flexion/extension. The elbow flexor muscles were moderately to highly active during walking and elbow flexion/extension (Fig. 2-B). It was believed that these dyssynergic muscles contributed to the increased tone and limitation of motion, as well as creating the resting shoulder posture.
In cases in which there is a question regarding whether contractures are too rigid or spasticity is too great for fractional lengthening to result in a substantial gain in active motor function or decreased tone, selective nerve blocks of the dyssynergic muscles can be used to demonstrate temporary improvement of active motion and improved tone. On the basis of the dynamic poly-EMG findings, the patient underwent diagnostic nerve blocks of the posterior and middle portions of the deltoid muscle and the long head of the triceps muscle to determine if tone and athetosis in the shoulder could be successfully controlled so that insertion of a stable humeral head replacement would be possible. With use of a long-acting local anesthetic agent, it was possible to control the shoulder position, improve the resting posture, and alleviate some of the pain in the shoulder. This served for diagnostic purposes only and was not considered a reasonable long-term solution since the glenohumeral joint remained subluxated and painful despite the nerve blocks.
On the basis of a comprehensive review of the patient's physical examination, imaging, and dynamic poly-EMG findings, hemiarthroplasty of the shoulder to address the joint arthrosis was recommended. A glenoid implant was not considered because of concern about loosening and excessive wear caused by spasticity and athetosis. To address the inferior glenohumeral subluxation, a biceps suspension procedure4 was included in the surgical plan. Importantly, addressing the spasticity and contractures was required to maintain concentric shoulder reduction, reduce pain, and allow for postoperative rehabilitation. As a result, a fractional lengthening of the long head of the triceps muscle was planned. Since it is not possible to fractionally lengthen the deltoid muscle surgically, the patient had chemodenervation of the posterior and middle portions of the deltoid muscle with botulinum toxin injections the week prior to surgery. It was understood that botulinum injections in the deltoid might need to be repeated in the postoperative setting. Botulinum toxin was selected, instead of alcohol or phenol injections, to avoid the destructive effects on local tissue, chronic painful dysesthesia, local muscle transformations, and vascular reactions that have been reported with these other injectables6. In addition, a myotendinous lengthening of the short head of the biceps muscle proximally as well as the brachialis and brachioradialis muscles distally was recommended to address the spastic elbow flexion. Although botulinum injections into the long head of the triceps, biceps, brachialis, and brachioradialis muscles would also control spasticity and cocontraction in these muscles, fractional lengthenings were considered a more long-term solution. After a long discussion with the patient and family, full informed consent was obtained regarding treatment choices as well as the expected outcome and possible complications of surgery.
Operative Technique
The patient was positioned supine with the arm on a hand table and a small bolster under the scapula. A standard deltopectoral surgical approach was used. The cephalic vein was identified. The interval between the deltoid and pectoralis major muscles was developed, and the clavipectoral fascia was incised to expose the subscapularis muscle and tendon. The subscapularis was tenotomized to expose the glenohumeral joint. Fractional lengthening of the long head of the triceps was done through an anterior approach. Access to the triceps was possible because substantial pectoralis major, deltoid, and biceps muscle atrophy allowed for easy retraction of these anterior structures. The conjoint tendon was retracted medially, and the brachial plexus was identified and retracted laterally. The long head of the triceps was identified posterior to the latissimus dorsi and teres major tendons; it extends distally and somewhat laterally from the infraglenoid tubercle to eventually meet with the lateral head of the triceps. The long head of the triceps has a musculotendinous junction located on its anterior surface, which was fractionally lengthened by transecting the tendon over the muscle belly. Subsequently, a fractional lengthening of the short head of the biceps was undertaken. The humeral head was resected and prepared for the hemiarthroplasty prosthesis.
Before the humeral implant was placed, the biceps suspension was undertaken4. The long head of the biceps tendon and the proximal part of the humerus were exposed subperiosteally over the bicipital groove. Two unicortical 5-mm drill holes were made in the humerus proximally and within the bicipital groove distally. The tendon of the long head of the biceps was dissected free from the muscle while the proximal origin of the tendon was left intact. A Krackow suture of Ticron 2-0 (Tyco, Waltham, Massachusetts) was placed in the distal end of the tendon, and the biceps tendon was passed proximally to distally through the tunnel in the humerus. Before tensioning of the biceps suspension, the humeral prosthesis was inserted. Cephalad traction was then placed on the biceps tendon to reduce the inferior subluxation. The distal end of the biceps tendon was sutured to the proximal portion of the biceps tendon with multiple sutures of Ticron 2-0, creating a suspension sling. Finally, the subscapularis tendon was repaired.
Attention was then turned toward the elbow. A curvilinear incision was made on the anterolateral aspect of the antecubital space and carried down through the subcutaneous tissues and fascia. The plane between the brachialis and brachioradialis muscles was developed to expose the brachialis muscle. The radial nerve was identified and carefully retracted on the lateral side of the arm. The brachialis muscle was carefully exposed to its medial border. The brachial artery and median nerves were identified and protected. The brachialis muscle was then lengthened by transecting the tendon over the muscle belly. Next, an incision was made on the dorsal-radial aspect of the forearm at the junction of the proximal and middle thirds of the forearm. The fascia was opened in line with the tendon of the extensor carpi radialis longus muscle. The undersurface of the brachioradialis muscle was then fully exposed. The radial nerve was identified and protected. The brachioradialis muscle was lengthened by transecting the tendon over the muscle belly. A layered closure of all incisions was undertaken. No intraoperative complications were noted.
Postoperative Course
Postoperatively, pain management was supplemented with oral diazepam therapy for the first two weeks to control the increased postoperative muscle spasticity. The repair was protected in a sling for three months to allow bone-to-tendon healing of the biceps suspension and subscapularis repair. The patient performed passive motion exercises of the shoulder and elbow during this three-month period. Importantly, the arm was supported at all times during therapy to prevent strain on the biceps suspension. The postoperative course was uncomplicated and, at two months, the patient had a pain-free shoulder both at rest and with passive motion. At the 4.5-year follow-up examination, she continued to be pain-free and satisfied with the outcome of surgery. Radiographs demonstrated a well-positioned hemiarthroplasty implant and reduced glenohumeral joint (Fig. 3). Passive motion of the elbow ranged from 0° to 150° with markedly decreased flexion tone. Additionally, with the decreased tone, the patient had gained 40° of active forward shoulder flexion and 20° of active external rotation compared with no active shoulder motion preoperatively.
Cerebral palsy is a nonprogressive neurological disorder of movement and posture. Shoulder spasticity is common in patients with cerebral palsy and can create considerable problems in the positioning of the upper extremity. Additionally, shoulder spasticity can lead to inferior glenohumeral subluxation and subsequent arthrosis. The combination of increased upper-extremity tone, contractures, glenohumeral subluxation, and arthrosis can be very challenging to manage. The pain of the arthrosis often produces increased muscle tone, thus creating a vicious cycle of pain and spasticity. While nonoperative management with bracing, steroid injections, and botulinum toxin injections is an option for the prevention of spasticity after upper motor neuron injury, in our experience, these modalities have not been a reliable long-term solution for fixed contractures, joint subluxation, or chronic arthrosis. Our patient had previously undergone botulinum injections in the muscles of the shoulder and elbow. These injections had improved resting muscle tone and had decreased athetoid movements but were a temporary treatment and did not address the pain resulting from the glenohumeral subluxation and arthrosis.
With joint contractures or long-standing spasticity, muscles can appear absent on physical examination, yet they have the potential for satisfactory function when positioned appropriately, after lengthening, or through denervation of contracted or severely spastic antagonist muscles. As a result, the preoperative evaluation of patients with shoulder spasticity must be individualized since the treatment goal for each patient is based on specific and realistic potential functional levels.
As demonstrated in our case, dynamic poly-EMG can be extremely useful in identifying the dyssynergic muscles that are apt to result in muscle imbalance and potential instability of the shoulder after surgery and in assessing whether shoulder surgery can improve a patient's function. With dynamic poly-EMG, not only can a patient's volitional control of a selected muscle be determined, but any dyssynergic muscle contraction can be identified7. As a result, preoperative botulinum toxin injections or intraoperative tendon lengthenings can be planned to reduce tone. Additionally, dynamic poly-EMG prevented the release or lengthening of muscles that may not have been involved in our patient's deformity. These adequately functioning muscles are potentially critical in restoring active function once shoulder spasticity and deformity are improved.
In our case, preoperative injection of the posterior and middle parts of the deltoid with botulinum toxin as well as selective lengthening of the dyssynergic long head of the triceps reduced muscle tone, thereby maximizing conditions to allow for both the healing of the biceps suspension in the immediate postoperative setting and the maintenance of long-term joint congruity. Additionally, controlling the elbow tone and position was considered important for control of the postoperative shoulder position. We believed that the elbow spasticity and persistent flexion posture, if not treated, could have changed the shoulder position in the postoperative setting or increased the overall tone in the upper extremity.
Our patient experienced pain relief, improved resting shoulder posture, decreased spasticity, and some restoration of active motion after surgery. Glenohumeral fusion was also considered as an option but was not selected because of the possibility of improved active motion with shoulder arthroplasty as determined with poly-EMG. Total shoulder arthroplasty was considered for this patient but was not selected because of concern about early glenoid loosening and wear given the presence of spasticity and athetosis. Hattrup et al.5 described three cases of total shoulder arthroplasty in patients with cerebral palsy and glenohumeral arthrosis. None of their patients had preoperative glenohumeral subluxation, and they did not report on the use of preoperative poly-EMG. Although arthroplasty was effective in relieving pain in all three patients, two patients experienced postoperative subluxation thought to be the result of a rotator cuff tear. The patient without postoperative subluxation of the glenohumeral joint underwent botulinum toxin injections in muscles selectively chosen after assessment of the patient's dystonic pattern on physical examination. We believe that using preoperative botulinum toxin injections in patients with upper motor neuron syndrome to be paramount in optimizing the soft-tissue environment necessary for successful shoulder arthroplasty.
In this report, we present a case in which selective tendon lengthenings and botulinum injections of the shoulder and elbow, in combination with a biceps suspension procedure and hemiarthroplasty, were effective in achieving a concentrically reduced glenohumeral joint, relieving pain, and improving spasticity. Although this is a rare and complex problem, we believe that the preoperative use of dynamic poly-EMG was important in guiding the surgical plan in our patient.