Between March 1994 and June 2003, 80 patients with brachial plexus palsy underwent a trapezius transfer. There were 11 women and 69 men with a mean age of 31 years (18 to 69). Before operation a full evaluation of muscle function in the affected arm was carried out. A completely flail arm was found in 37 patients (46%). Some peripheral function in the elbow and hand was seen in 43 (54%). No patient had full active movement of the elbow in combination with adequate function of the hand. Patients were followed up for a mean of 2.4 years (0.8 to 8). We performed the operations according to Saha’s technique, with a modification in the last 22 cases. We demonstrated a difference in the results according to the pre-operative status of the muscles and the operative technique.
The transfer resulted in an increase of function in all patients and in 74 (95%) a decrease in multidirectional instability of the shoulder. The mean increase in active abduction was from 6° (0 to 45) to 34° (5 to 90) at the last review. The mean forward flexion increased from 12° (0 to 85) to 30° (5 to 90).
Abduction (41°) and especially forward flexion (43°) were greater when some residual function of the pectoralis major remained (n = 32). The best results were achieved in those patients with most pre-operative power of the biceps, coracobrachialis and triceps muscles (n = 7), with a mean of 42° of abduction and 56° of forward flexion. Active abduction (28°) and forward flexion (19°) were much less in completely flail shoulders (n = 34).
Comparison of the 19 patients with the Saha technique and the 15 with the modified procedure, all with complete paralysis, showed the latter operation to be superior in improving shoulder stability. In all cases a decrease in instability was achieved and inferior subluxation was abolished.
The results after trapezius transfer depend on the pre-operative pattern of paralysis and the operative technique. Better results can be achieved in patients who have some function of the biceps, coracobrachialis, pectoralis major and triceps muscles compared with those who have a complete palsy. A simple modification of the operation ensures a decrease in joint instability and an increase in function.
Our earlier experience of trapezius transfer has shown the importance of a reduction of joint instability in patients with brachial plexus palsy. This leads to better control of the paralysed arm and a decrease in pain. A sling may no longer be necessary to control the arm. However, in spite of our considerable experience with the procedure14 some of our results have been poor. We have therefore modified the operative technique as described by Saha.15 We are not aware of studies dealing with the different patterns of paralysis in individual patients with brachial plexus palsy and its influence on the outcome after trapezius transfer. The role of inferior dislocation of the humeral head has not been well described.
We now present our results of the transposition of trapezius, taking into account the influence of the muscle function before operation and a modification of the operative technique.
Patients and Methods
In a prospective study, 80 patients with lesions of the brachial plexus (78 traumatic, one congenital, one iatrogenic), underwent a trapezius transfer between March 1994 and June 2003. The mean age of the 11 women and 69 men was 31 years (18 to 69). At the time of the accident or onset of paralysis, the mean age of the patients was 26 years (0 to 60) with a mean period of time from paralysis until the transfer operation of six years (0.8 to 37).
A neurosurgical procedure was carried out on 62 (78%) patients before muscle transposition. Before surgery the complete muscle status of the affected arm was evaluated according to the MRC grading of muscle power.17 In order to determine the degree of joint instability physical examination included the measurement of translation of the humeral head in relation to the glenoid. The assessment of the sulcus sign and anteroposterior passive movement was graded as follows: +, 25%; ++, 50%; and +++, > 50%. Pre-operative radiographs of the shoulder were taken and inferior subluxation of the head assessed on films taken with the arm dependant. The patients were followed up for a mean of 2.4 years (0.8 to 8) (Table I⇓).
Indication for trapezius transfer.
The operation is indicated for patients with multidirectional shoulder instability and insufficient or weak abduction/forward flexion because of damage to the deltoid and supraspinatus muscles, leading to poor function of the elbow and hand.14 Before considering a trapezius transfer full neurosurgical treatment including exploration of the plexus, neurolysis, repair, nerve anastomosis and nerve grafting must have been carried out and no improvement obtained after intensive conservative therapy. Paralysis of the deltoid muscle must be complete and assessed by EMG studies. The trapezius muscle must show full strength against resistance. In order to carry out the transfer it is essential that the pre-operative passive shoulder abduction is to at least 80°.
The first 58 operations were performed according to Saha’s technique.15 The patient was placed in a lateral decubitus position. After a sabrecut incision, the trapezius and deltoid muscles were exposed together with the acromion, clavicle and spine of the scapula. The deltoid muscle was mobilised from the lateral third of the clavicle, the acromion, and the lateral half of the spine of the scapula. The acromion, including the insertion of the trapezius muscle, was then separated from the spine of the scapula and the lateral third of the clavicle by an oblique osteotomy, leaving the coracoclavicular ligament. The proximal humerus was exposed by longitudinal splitting of the deltoid and its surface prepared by decortication. The rotator cuff was left untouched. With the shoulder in 80° to 90° of abduction, the acromion was transferred to the humerus just below the greater tuberosity and fixed with two 6.5 mm cancellous screws. The point of fixation to the humerus is a decisive factor in determining post-operative function. Transfer of the acromion may result in an improvement of abduction and of forward flexion and internal rotation in an anterolateral direction, and an increase of retroversion and external rotation posterolaterally.
Using Saha’s technique,15 the partially freed deltoid muscle was then sutured on top of the trapezius (Fig. 1⇓). In our latest 22 cases we modified this stage by suturing the deltoid muscle under maximum tension on top of the trapezius as far medially as possible (Fig. 2⇓). The power of the trapezius was then carried over to the humerus along the area of the insertion of deltoid.
The wound was closed over two suction drains. Radiographs taken immediately after surgery assessed the position of the screws and the transferred portion of the acromion.
Post-operatively, the arm was immobilised in a support in 80° of abduction for six weeks. Physiotherapy began on the first post-operative day with assisted and active training of the elbow, hand, and fingers. During the first week the abduction support was removed for physiotherapy, but the position of 80° of abduction was maintained during the exercises. After six weeks a radiograph was taken and progressive adduction of the arm was begun for one week as an inpatient procedure.
To investigate the influence of different patterns of paralysis on the outcome, the power of those muscles which might support post-operative function were examined. No adequate function was present in those rated MRC 0 to 2. Contraction at least against gravity was shown in those of MRC 3 to 5.
The trapezius was graded as MRC 5 in all patients. In all, the deltoid, and in 91%, the supraspinatus, showed evidence of insufficient power (MRC 0 to 2) (Table II⇓).
A completely flail arm was present in 37 patients. In 18 the flexors and extensors of the hand and fingers, as well as the abductors and extensors of the thumb, had some function (MRC 4 to 5).
Elbow flexion > 90° (elbow flexors MRC 4 to 5) was noted in 15 patients. None had complete function of the hand, fingers and thumb. Full active movement of the elbow was never combined with adequate function of the hand.
In all patients the transfer resulted in an increase of function and in 74 (92.5%) in a decrease of multidirectional instability of the shoulder.
We achieved a mean increase of active abduction from 6° (0 to 45) before to 34° (5 to 90) after surgery. The mean increase in forward flexion was from 12° (0 to 85) to 30° (5 to 90). Before operation the patients had a mean deficit of external rotation of −19° (10 to −70). There was no marked difference after the surgery with a mean deficit of external rotation of −18° (15 to −60). External rotation was slightly increased in 19 patients and decreased in 11, but no change occurred in 50 (Table III⇓).
Before surgery there was +++ multidirectional instability in 59 patients (74%) and ++ instability in 17 (21%); thus there was a sulcus sign in 76 (95%). At the time of the last review +++ instability was seen in two patients (3%) ++ in nine (11%) and + in 66 (83%). In three patients (4.0%) the instability had been completely eliminated (Table III⇑).
Radiologically, 76 patients (95%) showed inferior subluxation of the humerus before but only 14 (18%) after operation. Pseudarthrosis did not occur. Subjective assessment of the operation by the patients was excellent in 15 (19%), good in 47 (59%), uncertain in 14 (17.5%), poor in two (3%) and bad in two (3%). As regards the most important result of the operation, 61 (76%) rated improved stability, nine (11%) increase in function, seven (9%) the cosmetic effect, and three (4%) decrease in pain.
Complications included two intra-operative fractures of the head of the humerus, two cases of pressure-induced skin necrosis brought about by the abduction splint, four infections (two early and two late) and two cases of paralysis of the musculocutaneous nerve which recovered spontaneously. In two patients persistent instability and dissatisfaction necessitated arthrodesis of the shoulder.
In more than 90% of patients, the deltoid, supraspinatus and teres minor muscles showed a power of MRC 0 to 2 (Table II⇑). This resulted in reduced stability of the shoulder with subluxation of the humeral head together with loss or limitation of abduction, forward flexion and external rotation. The trapezius, levator scapulae and rhomboids had a power of MRC 4 to 5 in over 96% of patients. The biceps, coracobrachialis and triceps muscles, which arise from the scapula, span the shoulder and elbow and insert into the bones of the forearm, had insufficient function in more than 70%.
There were 48 patients (60%) with a power of MRC 0 to 2 in the pectoralis major and 32 (40%) with a power of MRC 3 to 5 (Table II⇑). After trapezius transfer, abduction and especially forward flexion were greater in those with some function of the pectoralis major. No marked difference was found for increase in joint stability (Table IV⇓).
In 38 patients (47%) the biceps, coracobrachialis and triceps had a power of MRC 0 to 2. When considering abduction and forward flexion the results obtained depended on the initial power available with those with least power achieving the least satisfactory result. Multidirectional instability of the shoulder was improved in all patients who had reasonable function of biceps, coracobrachialis and triceps. However, in flail shoulders in which these muscles and the pectoralis major all had power of MRC 0 to 2 the gain in active movement was much less (Table V⇓).
Our first 58 patients underwent trapezius transfer according to the technique of Saha.15 In 14 (24%) of these, inferior subluxation of the humeral head persisted and in four (7%) there was no improvement in multidirectional instability. Subsequently, four revision procedures were required. There were two arthrodeses of the shoulder and two shortenings of the tendon of trapezius. In our last 22 patients we modified the technique as described in order to avoid these complications. We had no further persistent subluxations and no more revision procedures were required.
We compared two groups of patients who had these different operations. Nineteen of the patients with the Saha technique and 15 with the modified procedure showed no adequate function (MRC 0 to 2) of the pectoralis, biceps, coracobrachialis and triceps muscles.
With the modified operation abduction and forward flexion were slightly better than after the Saha technique, movement to 32°/25° and 21°/17°, respectively. The modified procedure also gave more improvement in stability. In all patients a decrease in instability was achieved and inferior subluxation was abolished, whereas the latter persisted in six patients (32%) in the group with the Saha technique (Table VI⇓).
After brachial plexus palsy, trapezius transfer is indicated for patients with multidirectional shoulder instability with subluxation and insufficient or weak abduction/forward flexion caused by inadequacy of the deltoid and supraspinatus muscles and leading to poor function of the elbow and hand. The trapezius must have full strength against resistance. Because of its innervation, the trapezius is not paralysed after injury to the brachial plexus, whereas the supraspinatus and deltoid muscles are involved in 91% and 100%, respectively (Table II⇑).
Muscle transfer operations are superior to shoulder fusion in terms of passive function and because of the higher rates of complications, the longer duration of operation and the irreversibility of arthrodesis. Cofield and Briggs18 found that 25 (35%) of 71 arthrodeses required revision because of persistent pain, pseudarthrosis, and incorrect positioning. If muscle transfer fails fusion can be undertaken, but if arthrodesis fails, no further option is available. Goldner5 considered that arthrodesis should be regarded as the final option and that muscle transfer was the best option. In patients with complete lesions and those with no adequate function of the elbow and hand, trapezius transfer is the optimum procedure because it is successful in solving their particular problems. The advantage of shoulder arthrodesis is that it allows a greater increase in active function and is therefore suitable for those patients who are engaged in physically demanding work and who still have satisfactory function in the elbow and hand.14
In other, larger series about 40° of active abduction and forward flexion have been obtained after trapezius transfer (Table VII⇓).1,9,13 An additional aim is to abolish subluxation of the humeral head1,13 which with the improved stability will lead to better control of the paralysed arm.
External rotation is not markedly improved. In patients with elbow flexion and hand function the lower arm may strike the thorax if the hand is brought up to the mouth. Because 37 (46%) of our patients had a completely flail arm and full active motion of the elbow was never combined with adequate function of the hand, this did not present as a significant problem.
The outcome after trapezius transfer is influenced by the pre-operative muscle power and the technique employed. The pectoralis major, biceps and coracobrachialis can support forward flexion and decrease inferior subluxation of the humeral head. The triceps pulls it in a cranial direction. We achieved about 40° of active abduction and forward flexion only in patients with some function in the biceps, coracobrachialis and triceps or pectoralis major muscles (Tables III⇑⇑⇑ to VI⇑), and the highest values were obtained in those patients with the best residual power (Table V⇑). However, our results were not as good as those of Karev 8 and Mir-Bullo et al.12
Only a combination of transferred muscles can improve the functional outcome. Narakas19 achieved a mean abduction of 120° in four patients with a combined transfer of teres major, levator scapulae and latissimus dorsi. Chunlin and Yonghua20 described seven patients with pectoralis major transfers of whom three had an additional trapezius transfer. The combined transfers achieved abduction of 70° to 90° and forward flexion of 60° to 150°, but isolated pectoralis transfer only gave 40% of abduction.
In all muscle transpositions the maximal tension of a transferred muscle is the decisive factor influencing postoperative function. To achieve this necessary tension in trapezius transfer the fixation of the acromial fragment must be as far as distal on the humerus as possible, but will be limited by the degree of mobilisation of the muscle which is possible.13
Saha,15 describes suture of the partially freed deltoid muscle on top of the trapezius without tension after fixation of the acromion (Fig. 1⇑). We sutured the deltoid muscle under maximum tension on the trapezius muscle, as far medially as possible (Fig. 2⇑). The power of the trapezius is then carried over to the humerus via the fibres of the atrophied deltoid muscle, particularly to the area of insertion of the deltoid. This point lies some centimetres distal to the acromial fragment, so that a second further point of caudal fixation arises. The deltoid muscle also protects the transferred trapezius which is stressed because of the maximum tension. Our functional results after modification of the technique do not differ much from those of previous publications.1,8,9,11,12,14,16 In all the modified operations an increase in stability was achieved and inferior subluxation was abolished, whereas in one-third of the patients in the Saha group the subluxation persisted. Inferior subluxation is a notable problem with pain arising from constant pulling on the capsule of the joint and of the adjacent structures. Trapezius transfer is a successful procedure in overcoming this difficulty. Kotwal et al9 reported poor results in five (19%) patients who all had persistent subluxation of the head. In our entire group of 80 patients caudal subluxation of the humeral head persisted in 14 (18%) patients. Four required revision procedures. All failures occurred before we modified the operative technique.
We conclude that trapezius transfer for the flail shoulder produces a satisfactory outcome as regards function and stability, and gives subjective help to patients. The results depend on the pre-operative pattern of paralysis and the operative technique. Better results can be achieved in patients with some function of the biceps, coracobrachialis, pectoralis major and triceps compared with those who have a complete palsy. A simple modification of the operative technique using the atrophied deltoid to carry over the force of the trapezius improves the stability of the joint and increases function.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received August 28, 2003.
- Accepted March 4, 2004.
- © 2005 British Editorial Society of Bone and Joint Surgery