Any use of the hand will be wasted without elbow function to maintain it in a useful position. Therefore, the first goal in the treatment of a flail arm is the restoration of elbow flexion by primary direct nerve surgery or secondary reconstructive surgery.^17,^172,¹⁷⁶ An exception is in polio-type brachial plexus paralysis, in which priority should be given to restoration of shoulder abduction rather than elbow flexion. In these cases, early nerve transfers within 1 year of paralysis achieve significantly more shoulder abduction than late nerve transfer or palliative reconstruction with muscle/tendon transfer or shoulder arthrodesis, whereas elbow flexion may spontaneously recover in up to 90% of cases.¹⁴² In obstetric brachial plexus palsy, timing of surgical intervention is highly controversial. Although most authors agree that surgical intervention is indicated when no recovery of biceps function occurs, some have recommended intervention at 3 months⁹² or as late as 9 months.⁶⁰

In all timing of reconstructive procedures for neurologically disabled patients, achieving stability of neurologic status, mental status (acceptance of the permanency of the disorder), and social status (clearly established goals of rehabilitation) is an essential prerequisite before any surgery is undertaken that is not itself concerned with stability.¹¹⁰ Conditions that interfere with progression toward the goals of rehabilitation before neurologic stability has been achieved include flexion contractures of the elbow and supination contractures of the forearm, which must be corrected before further reconstruction is undertaken.

CORRECTION OF PRE-EXISTING JOINT CONTRACTURE

Reasonably functional passive motion of the elbow must be obtained before nerve, tendon, or muscle transfers are performed. This often requires the joint and muscle contractures to be stretched, and as a result, the functioning muscles are strengthened by progressive resistance exercises. Passive flexion of at least 120 degrees is desirable before tendon transfer is carried out. Static adjustable splints (see Chapter 11) are useful in regaining passive extension.⁹⁵ Occasionally, surgery is required to release a severely damaged elbow flexor, extensor muscle, or joint capsule before the joint can be mobilized. A preliminary or concurrent V to Y release of a contracted triceps, in conjunction with flexorplasty, may be necessary to correct an extension contracture in an arthrogrypotic child. If soft tissue cover is necessary, it usually can be provided as a skin paddle on the transposed latissimus dorsi flap.

When contractures are greater than 60 degrees and interfere with or are exaggerated during ADLs, surgical release is indicated.¹²⁶ The contracture correction can consist of combinations of Z-lengthening of the biceps tendon, myofascial release of the brachialis, and release at the origin of the wrist flexor-pronators. Capsular release is necessary only in severe, long-standing contracture. The contracture that occurs in Erb’s palsy is paradoxical in nature, given the normal triceps function in the face of weak or absent biceps function. A dynamic splint¹²⁶ or serial casting, followed by biceps tenotomy with further casting,⁸⁴ may also be used. The latter method is associated with improved joint range of motion without loss of elbow flexion strength in 80% of cases. Drawbacks of this treatment are pressure sores and stiffness, and the limitations in correcting pronation as the elbow reaches full extension. When an upper motor neuron unit is involved in the contracture, botulism toxin or selective crush of the musculocutaneous nerve is used with success concurrent with casting, especially in contractures developing early after spinal cord injury. These patients are especially prone to developing a pronation contracture due to biceps weakness. Pronation and supination contractures should also be addressed before fixed bony changes occur in the radius and ulna.

CHOOSING AN APPROPRIATE MUSCLE FOR TENDON TRANSFER

The type of muscle transfer depends on the muscle groups that are available.¹⁷² The original functional properties between the donor and recipient muscles must be matched.⁸⁶ A transferred muscle must provide both adequate strength and excursion in its new role.

Clinical and electromyographic testing of various muscles should be performed before transfer. A muscle to be transferred should show an adequate electromygraphic trace and have appropriate strength. The extent of a brachial plexus lesion dictates which muscles will have retained their innervation. The most common cause of failure in tendon transfers about the elbow, as elsewhere, is overestimating the strength of the transferred muscle. The muscle to be transferred should contract against gravity resistance (grade 4, “M4”) if significant function is to be anticipated (Table 71-1). Preoperative electromyography (EMG) of the muscle proposed for transfer is a good idea, but mild denervation changes do not preclude the transfer.²²⁶

TABLE 71-1 Muscle Strength Grading System According to the British Medical Research Council ¹⁶¹

Grade	Definition	Explanation
0	Absent	No function
1	Trace	Slight contraction, no motion
2	Poor	Complete motion, gravity eliminated
3	Fair	Complete motion against gravity
4	Good	Complete motion against gravity and some resistance
5	Normal	Apparently normal strength

In addition to muscle strength, which is determined by clinical examination, it is helpful to consider the comparative work capacities of the various muscles available for transfer.²⁴⁵ The work capacity is determined by multiplying the power of the muscle (3.65 kg/cm² of physiologic cross section) by its amplitude (the distance through which the tendon moves when its muscle contracts). The work capacity of the biceps for elbow flexion is 4.8 kg/m, and for forearm supination, it is 1.1 kg/m. The sternocostal head of the pectoralis major muscle has nearly three times as much work capacity at 10.4 kg/m. By contrast, the work capacity of all forearm flexors ¹ (pronator teres, flexor carpi radialis and ulnaris, and palmaris longus) is 3.8 kg/m.¹¹⁶

The dispensability of the transferred muscle must be considered when a muscle is chosen for transfer around the elbow. Triceps transfer for elbow flexion is not often indicated because the triceps is not really expendable. The latissimus dorsi is not expendable in polio patients, who rely on it to elevate the pelvis when the other pelvic elevators are paralyzed.

Finally, cosmesis may be a consideration. The bow-stringing of the sternocleidomastoid transfer across the side of the neck makes it an undesirable transfer, even though, theoretically, it is one of the most mechanically efficient ones. Restoration of nearly normal bulk of the arm in a patient with loss of flexor or extensor muscle mass is a nice cosmetic byproduct of the bipolar transplant of the latissimus or pectoralis muscle.

RESTORATION OF ELBOW FLEXION

INTRODUCTION

Neurotization or brachial plexus reconstruction is the treatment of choice in early cases of the flail upper extremity.^* In chronic cases or failed early reconstruction, several tendon transfer procedures to restore elbow flexion are available⁴⁹: Steindler flexorplasty and its modifications,^28,^36,^40,^76,^146,^158,²²² anterior transposition of the triceps tendon,⁴³ pectoralis-to-biceps transfer, with or without the pectoralis minor,^† pectoralis minor alone,²⁰⁰ sternomastoid-to-biceps transfer with or without shoulder arthrodesis,^40,⁴⁴ transfer of the flexor carpi ulnaris,¹ and unipolar or bipolar transposition of the latissimus dorsi muscle.^11,^15,^33,^48,^117,^207,^237,²⁵⁵ Free muscle transfer^‡ is fast becoming a mainstay of the armamentarium of upper extremity reconstructive surgeons and has proved useful in both early and chronic cases of brachial plexus and traumatic injuries. Each of these treatment modalities has its advocates, and knowledge of a number of different options will allow treatment to be tailored to the needs and expectations of the patient.

The muscle groups chosen for tendon transfer depend on those that are available. Triceps-to-biceps transfer is advocated when there are biceps-triceps co-contractions, and Steindler’s procedure when the elbow flexors reach only grade 2.⁷ The latter may be contraindicated when the elbow flexors are grade 0, when the wrist flexors are weak, or when wrist and finger extensors are paralyzed.¹⁷² The Steindler flexorplasty can also be augmented by a pectoralis minor transfer. In children with obstetric brachial plexus palsy at the C5, C6 level, flexorplasties are indicated when there is a functional hand but with weak elbow flexion and inability to position the hand.¹¹³ In these patients, the Steindler procedure is reserved for elbows with excellent forearm muscles for wrist and digit flexion. When these are not all M5, the triceps-to-biceps transfer is advocated, by some authors using only a normal triceps,¹¹³ and by others,²⁰² even with a weak triceps. When neither the triceps nor the forearm flexors are judged as being of sufficient strength, the pectoralis major or minor transfer can be used, or preferably, a free-functioning muscle transfer can be performed.

NEUROTIZATION

Neurotization is the technique of nerve transfer used to restore motor or sensory function after brachial plexus injuries when the nerve damage is too far proximal for standard nerve-grafting techniques to be feasible. In 1903, Harris and Low ¹⁰⁴ inserted half of the distal fascicles of the C5 root, which was damaged at the foramen, into the healthy spinal nerve C6 or C7 in three cases of Erb’s palsy. No results were recorded. In 1913, Tuttle²⁴¹ mentioned one case of neurotization using the spinal accessory nerve. In 1963, Seddon²¹¹ first reported neurotization of the musculocutaneous nerve using intercostal nerves in two patients, both times obtaining active elbow flexion.

Subsequently, many early cases of neurotization using intercostal nerves have reported a modicum of success.^§ Early reports on neurotization of the musculocutaneous nerve using parts of the cervical plexus ³⁵ and spinal accessory nerve^5,^122,¹⁸¹ also mention some limited success.⁹⁸ Later series report satisfactory restoration of elbow flexion after intercostal nerve transfer to the musculocutaneous nerve without intercalated grafts.^68,^128,¹⁷⁴ The goal is to provide the musculocutaneous nerve with the best motor donors.²³³ This can be achieved by reconstruction using intraplexus donors such as nerve grafts from the proximal stumps of avulsed C5 roots. It must be stressed, however, that nerve transpositions are effective only when they are performed within 6 to 9 months of injury.^5,^24,^51,^128,^163,^181,^201,²¹⁷

Currently used extraplexus transpositions for elbow flexion, in general order of decreasing frequency, are the intercostal nerve,^88,¹⁷⁹ spinal accessory nerve,⁶ thoracodorsal nerve,²⁰⁴ pectoral nerve,³¹ phrenic nerve,^53,⁹⁸ hypoglossal nerve,¹⁷⁷ platysma motor branch,²² and high cervical nerve roots.²⁵² The ulnar¹⁸⁶ and median nerves^148,^150,¹⁵¹ can be used as intraplexus donors, with their vascular supply left intact. This is recommended only for global plexopathy or tetraplegia involving C7, C8 and T1 so as not to sacrifice any remaining or potential nerve function.^139,¹⁸⁶ Other intraplexus donors can be obtained by the selective contralateral C7 method^56,^97,²³⁵ or by transfer of the medial pectoral nerve.^205,²⁴⁶ Cross-chest radial nerve transfer has been proposed and is feasible anatomically but requires tendon transfer reconstruction of radial nerve innervated muscles in the contralateral arm.²¹ Another rarely used intraplexus donor is the long thoracic nerve,¹⁷⁷ which provides the sole innervation for the serratus anterior, causes scapular winging after transfer, and therefore, is not recommended.¹⁴⁴

The phrenic nerve can be transferred without pulmonary complications,⁹⁸ even with concurrent intercostal nerve transfer,¹⁰¹ although the risk of reduced vital capacity is greater when the right phrenic nerve is used.¹⁴⁹ Its full length can be harvested using video-assisted thoracic surgery techniques.²⁵¹ Transfer of intercostal nerves alone have been associated with a nonsignificant mild decline in pulmonary function, without a subjective effect on respiratory status or stamina.¹²⁸

Intraoperative somatosensory evoked potential (SEP) recording is mandatory to ascertain the level of injury—even when a root appears to have sustained a pos-ganglionic injury, nerve grafting is never indicated in a root with a negative SEP.¹⁷⁵

TECHNIQUE

The exposure of the brachial plexus as performed by Sedel is described in this section.¹⁹ The patient, under general anesthesia, is given no neuromuscular blocking agents. An incision is made parallel to the clavicle, with an extension parallel to the sternocleidomastoid muscle. The skin flaps are elevated, and the external jugular vein is ligated. The sternocleidomastoid is retracted, allowing dissection to proceed in the prescalene area. The phrenic nerve is identified on the scalenus anterior. The fifth cervical root is identified at the point where the phrenic nerve crosses the lateral aspect of the scalenus anterior proximally. The sixth cervical root is more distal and medial with a more horizontal direction. The seventh cranial nerve root is more horizontal still and more posterior; more posteriorly and medially lie the more difficult to identify eighth cervical and first thoracic nerve roots. The scalenus anterior is divided, the subclavian artery is retracted ventrally, and the pleural dome is retracted dorsally. The infraclavicular incision is extended onto the deltopectoral groove, sometimes extending over the arm. The axillary artery is dissected and the lateral, posterior and medial cords of the brachial plexus are identified. If exposure is required behind the clavicle, this is divided and repaired with an AO plate later. A supraclavicular or infraclavicular approach can be used.

In accordance with the goals of reconstruction of the flail upper limb, the first priority in neurotization is to attach a motor nerve donor to the musculocutaneous nerve. Grafts used for this purpose include the suprascapular and spinal accessory nerve, alone or in conjunction with two or three intercostal nerves (Fig. 71-1). In decreasing order of priority, the radial nerve (after dissection and removal of the triceps or posterior cord branch), median nerve, and axillary nerves are reconstructed in a similar fashion. The operative strategy depends on the type and extent of the lesion and the number, quality and length of available grafts, as well as the aspect of the proximal neuroma. If there is a neuroma in continuity, there is controversy as to whether it should be resected or treated by simple neurolysis.²⁵⁰ In any case, conduction across the neuroma is best assessed using nerve action potential rather than muscle action potential, because the former requires 40 to 50 times as many fibers of reasonable caliber to elicit a response. There is discussion as to the advantage of using interposition nerve grafting when performing intercostal nerve transfers. It has been demonstrated that intercostal nerves lose up to 10% of motor axons per 10 cm measured from the midaxillary line to the sternum.¹⁶³ Coaptation of the transfer closer to the motor endplates, avoidance of tension on the nerve repair, and earlier shoulder mobilization are other theoretical advantages. This includes coaptation directly to the dominant motor branch of the biceps muscle rather than to the musculocutaneous nerve.^31,¹²⁸ There is no consensus on the optimum number of intercostal nerves to be transferred to the musculocutaneous nerve to obtain optimal biceps function. Also, there is no consensus on using the spinal accessory nerve over intercostal nerves. A theoretical advantage of the spinal accessory nerve for transfer is the high number and ratio of motor fibers (1500-3000) compared with an intercostal nerve (500-700 motor fibers per nerve).^5,^88,²⁴⁴ However, a clear disadvantage cited by most authors is the need for an interpositional graft when using the spinal accessory nerve. If necessary, the spinal accessory nerve can be harvested through a posterior approach.^12,^99,¹⁹⁴ An average extra length of 12 cm can be gained compared with the standard anterior approach, allowing the terminal branches of the brachial plexus to be neurotized without the need for an interposition graft (Fig. 71-2).²⁴⁴ Vathana et al.²⁴⁴ measured an average of 817 myelinated axons at the most distal level; however, these authors warn that the size mismatch with terminal branches in the brachial plexus (varying between 5000 and 14000 axons) may predict poorer outcome and that this approach may be more suitable for reinnervation of the suprascapular nerve at the level of the supraclavicular notch. When interposition grafts are unavoidable, results may be improved with the use of vascularized nerve grafts.⁵³ The vascularized ulnar nerve graft has been shown to maintain its blood supply and viability even in a scarred bed.²³⁵ Pedicled on the superior ulnar collateral vessels and ulnar vessels, this yields better results than including the superior ulnar collateral vessels alone.⁹⁷ This intraplexus donor has a greater number of axons than any extraplexus donor, potentially increasing the probability for successful neurotization.²³⁵

FIGURE 71-1 Intercostal nerve neurotization of the musculocutaneous nerve.

FIGURE 71-2 The posteriorly harvested spinal accessory nerve can be used to reach distal anterior targets in the brachial plexus.

Distal nerve-to-nerve transfers are possible, however, only when the lower plexus is intact. Possible transfers include ulnar nerve to the biceps branch of the musculocutaneous nerve or to biceps and brachialis, medial pectoral, and thoracodorsal nerves (Fig. 71-3).^147,^150,^151,¹⁸⁷ A shorter distance and time to reinnervation is provided by these transfers.

FIGURE 71-3 Oberlin ¹⁸⁷ transfer of the ulnar nerve to the musculocutaneous nerve. When the ulnar nerve is intact, a fascicle can be transferred to the biceps motor branch to restore elbow flexion.

RESULTS

Intraplexus donors provide better results than extraplexus donors, although direct neurotization of the musculocutaneous nerve from ipsilateral intercostal nerves produces results comparable to intraplexus neurotizations,²³³ as does direct neurotization of the suprascapular nerve with the spinal accessory nerve for shoulder function.²³⁵ A meta-analysis of the English literature (26 studies, 965 nerve transfers) showed M3 biceps strength in 72% of direct intercostal-to-musculocutaneous transfers versus 47% with interposition grafts (P < 0.001), with a trend observed at ≥M4 biceps strength (41% vs 32%).¹⁶³ Spinal accessory nerve–to–musculocutaneous nerve transfers for elbow flexion were comparable to intercostal nerves without interposition grafts at ≥M3 strength (77% vs 72%). Intercostal nerves, however, were significantly better at providing ≥M4 biceps strength relative to spinal accessory nerve (41% vs 29%; P < 0.001). The same authors were unable to demonstrate a difference when two versus three or four intercostal nerves were used for restoration of elbow flexion to ≥M3 strength.

This contrasts with a prospective study involving 205 patients performed by Waikakul et al.,²⁴⁷ comparing spinal accessory to intercostal nerve transfer. These authors found that spinal accessory nerve transfers produced significantly better motor outcomes, but patients with intercostal nerve transfers had significantly better protective sensation and pain relief. Also, the shoulder subluxation in the spinal accessory group required more frequent shoulder fusion. Spinal accessory nerve transfer is preferred by these authors in isolated biceps paralysis, whereas intercostal transfer is recommended for global plexus palsy, particularly those with deafferentation pain.²⁴⁷ A beneficial effect of neurotization that frequently precedes functional return by months is the alleviation of deafferentation pain and an improvement in protective sensation.¹⁸¹ A lack of pain allows patients to concentrate on rehabilitation and improve their function and dexterity.²³⁵

The contrasting results between different studies may be due to the use of an interposition graft with the spinal accessory nerve that may offset the advantage of containing more motor fibers.¹⁶³

Oberlin’s method of transfer of two ulnar nerve fascicles show promising results for restoration of elbow flexion,^139,^186,²³² as does the double fascicular nerve transfer from the ulnar and median nerves to biceps and brachialis branches of the musculocutaneous nerves.^147,¹⁵¹

Several nerve transfers may be combined to decrease the time to functional return and provide better results. Leechavengvongs and colleagues ¹⁴⁰ reported the results of 15 combined spinal accessory–to–suprascapular nerve, ulnar nerve–to–biceps motor branch, and long head of the triceps branch–to–the axillary nerve transfers for C5 and C6 avulsion injuries. Elbow flexion strength was M4 in 13 patients and M3 in two patients at an average follow-up of 24 months.

TENDON TRANSFERS

Steindler’s Flexorplasty

Steindler first described his simple but ingenious concept of proximally shifting the origin of the flexor-pronator muscle group to increase its lever arm in flexing the elbow in 1918.^220,²²¹ His technique consisted of the subperiosteal dissection of the origin of the flexor-pronator muscles of the forearm from the medial epicondyle. The muscle flap was then transposed proximally between the brachialis and the triceps, and was sutured to the medial epicondylar ridge through two drill holes 2 inches above the epicondyle. Steindler recommended the procedure only if the wrist flexors were normal or only slightly weakened.

Subsequent authors differed on the prerequisites for Steindler’s flexorplasty. Mayer and Green ¹⁵⁸ believed that the epicondylar muscle group strength should be grade 3 or better, and they re-emphasized the need for strong wrist flexors if the operation is to succeed. They described this simple test:

The arm is abducted to 90 degrees to eliminate gravity. Any patient who can flex the elbow in this position (using the epicondylar muscle group) is a candidate for the operation. Nyholm believed that the criteria for forearm muscle strength should be liberalized because all of his patients achieved 90 degrees of elbow flexion, even though a relatively large number had forearm muscles that were “primarily paretic.”¹⁸³ Dutton and Dawson⁷⁶ performed the operation if the strength of the forearm flexor-pronator group was rated fair or better. Alnot⁷ believed that the shoulder flexorplasty was most indicated to reinforce a biceps that had recovered to M2 strength. It seems that there is considerable variability in the preoperative muscle strength needed for active flexion, and the axiom that the muscle loses a grade in transferring it does not necessarily apply in this particular situation.

Others have modified and refined Steindler’s original concept to increase flexor strength and decrease the tendency toward development of a pronation deformity. Bunnell ⁴⁰ described extending the common tendon of origin with a fascia lata graft that would reach 2 inches up the lateral border of the humerus. This resulted in moderate but not complete correction of the pronation tendency. Mayer and Green ¹⁵⁸ detached the flexor-pronator origin with a portion of the medial epicondyle and attached it through a window cut in the anterior cortex of the humerus 5 to 7.5 cm proximal to the joint. Their description of the technical details of the Steindler flexorplasty, which includes careful dissection of the ulnar and median motor branches, is the best in the literature. Lindholm and Einola¹⁴⁵ used a screw to fix the epicondylar fragment to the humerus in two cases. Eyler⁷⁹ advocated omitting the flexor carpi ulnaris to allow the surgeon to work in the “internervous plane” between the flexor sublimis, anteriorly, and the flexor profundus and flexor carpi ulnaris, dorsally.

Efforts have been made to augment the strength of elbow flexion, especially in patients who have weakness of the flexor-pronator muscle group. Initially, Steindler recommended proximal transfer of the radial wrist extensor muscle origins off the lateral epicondyle in conjunction with the medial transfer, but he did not comment on it in his later communications.²²² Mayer and Green reported having difficulty mobilizing the lateral epicondylar muscle group without damaging the nerve supply¹⁵⁸ and gave up the procedure after two unsatisfactory results. Lindholm and Einola¹⁴⁶ were unable to detect any increase in flexion strength in the six patients who had the additional lateral transfer.

Anatomy

The pronator teres arises from two heads. The humeral head originates from the medial supracondylar ridge, the medial intermuscular septum, and the common flexor tendon. The smaller ulnar head arises from the coronoid process of the ulna. The median nerve enters the forearm between the two heads of the pronator teres. The flexor carpi radialis, palmaris longus, and humeral head of the flexor superficialis all originate from the common flexor tendon. The flexor carpi ulnaris arises from two heads: the humeral head originates from the common flexor tendon, and the ulnar head originates along the medial border of the olecranon and posterior border of the upper three fifths of the ulna. Branches originating from the medial surface of the median nerve supply the pronator teres (C6, C7), the flexor carpi radialis (C6, C7), the palmaris longus (C7, C8), and the humeral head of the flexor superficialis (C8 and T1). The flexor carpi ulnaris (C8 and T1) is innervated by two or three branches of the ulnar nerve, the first of which usually leaves the nerve just as it passes between the two heads of the muscle.

Technique (modified from Mayer and Green)

A sandbag is placed under the opposite hip (Fig. 71-4). A tourniquet usually is not used. The incision begins on the anterior aspect of the arm about 7.5 cm above the elbow and swings gently in a medial direction. At the elbow, it runs just posterior to the epicondyle. It then curves anteriorly, following the direction of the pronator teres, ending about 10 cm below the elbow. The ulnar nerve is isolated and freed distally to the branches of the flexor carpi ulnaris. Preserving these motor branches, the surgeon continues the dissection distally for 5 cm. The lacertus fibrosus is divided, and the median nerve is exposed above the elbow and dissected distally, exposing the motor branches (all of which leave the medial aspect of the nerve) to the common flexor-pronator muscle group. The common flexor-pronator muscle origin is then detached with a flake of epicondyle (cartilage in children and bone in adults). The flake of bone or cartilage is then grasped with a clamp, and while traction is exerted in distal and anterior directions, the muscles are stripped from the anterior surface of the joint and from the coronoid process of the ulna. As the ulnar head of the flexor carpi ulnaris is detached from the ulna with an elevator, the assistant puts gentle traction on the median and ulnar nerves, demonstrating the motor twigs that must be carefully avoided. Dissection is continued distally as far as the anatomic distribution of the nerves permits. The common tendon is then transfixed with a modified pull-out suture of No. 1 prolene. The elbow is flexed to 120 degrees, and traction is exerted on the transfer to determine how far above the elbow the transfer will reach. This is usually between 5 and 7.5 cm. The ulnar nerve often seems to have less tension on it if it is transferred anterior to the epicondyle. The atrophic fibers of the brachialis are slit longitudinally, the periosteum is incised, and the anterior humerus is exposed subperiosteally. An opening in the anterior cortex of the humerus nearer to the lateral than to the medial border is made at the point to which the transplant reaches when the elbow is flexed. Two small drill holes are made from anterior to posterior through this cortical window. The prolene suture ends are then threaded through the holes and out through the triceps muscle and skin. To be sure that they are not subjected to undue tension on twisting, the nerves are inspected as the transfer is pulled into the cortical window. The distal portion of the wound is closed, up to the bend in the elbow. The sutured ends are drawn tight with the elbow in maximal flexion and tied over a button under which a thick piece of felt has been placed. Several auxiliary sutures between the periosteum and the transplanted epicondylar tissues are placed, and the wound is closed with a drain.

FIGURE 71-4 Steindler’s flexorplasty.²²¹ A, The incision. B, The ulnar nerve is mobilized proximally and distally, and its motor branches to the flexor carpi ulnaris are identified and protected. C, The common flexor-pronator origin is detached with a flake of medial epicondyle. The motor branches of the median nerve to the flexor-pronator group are identified and protected. D, The detached flexor-pronator group is mobilized distally as far as the motor branches of the medial and ulnar nerves will permit. The brachialis muscle is divided. E, The distal humerus is prepared, and a prolene pull-out suture is used to anchor the transferred muscles to the anterolateral surface of the humerus 5 to 7.5 cm above the elbow.

(After Mayer, L., and Green, W.: Experiences with Steindler flexorplasty of the elbow. J. Bone Joint Surg. 36A:775, 1954.)

A posterior splint is then applied with the elbow in about 120 degrees’ flexion and the forearm in full supination. Four weeks after surgery, the pull-out suture is removed. A removable Orthoplast splint is applied, and active flexion and supination and extension exercises are begun. No special retraining is required because the muscles transferred functioned as an accessory elbow flexor before transfer.⁵⁸ Splinting is discontinued 6 to 8 weeks after surgery, the longer time being necessary for patients with normal triceps function. A dynamic extension splint is often useful if the patient has no triceps function (Fig. 71-5).

FIGURE 71-5 A and B, A dynamic flexion and extension elbow splint. The dynamic extension is particularly useful in patients who have Steindler’s flexorplasty but lack triceps function.

In patients with C5-7 palsy in which the wrist and finger extensors are weakened or paralyzed, the forearm flexor-pronators are strong but the triceps is not available or co-contracts with the biceps, the Steindler procedure should be complemented with wrist arthrodesis.¹⁷² Furthermore, success of the procedure relies on sufficient power of the flexor-pronator muscles.^76,^124,^145,^146,^155,¹⁵⁸

Results

Published results of the Steindler flexorplasty reflect a high degree of success in achieving a functional range of elbow flexion against gravity. Most transfers reported have been performed for poliomyelitis. Steindler achieved 79.5% good results (flexion against gravity of not less than 90 degrees) in 39 cases.²²⁴ Fifteen good results (useful range of flexion with good to fair power) in 27 flexorplasties were reported by Carroll and Gartland.⁴² Using strict criteria for success, including subtracting from the total score for flexion contracture of more than 15 degrees and for supination of less than 45 degrees, Mayer and Green recorded 11 excellent results, five good, four fair, and two poor results among the 22 flexorplasties they followed up (Fig. 71-6).¹⁵⁸ Segal and associates compared the results of 13 Steindler flexorplasties (transplantation of both flexor and extensor origins) and 17 Clark pectoralis major transfers.²¹³ The flexorplasty results were better than the pectoralis transfers, but the average flexion contracture in the flexorplasty group was 60 degrees, whereas it rarely exceeded 15 degrees in the Clark transfer group. Kettelkamp and Larson, evaluating 15 flexorplasties using Mayer and Green’s scoring system, noted eight excellent or good results, six fair results, and one poor one.¹²⁴ They also measured the carry-lift strength (flexion against gravity, plus weight in 1-pound increments) and found that nine of the 14 patients who were able to flex the elbow 110 degrees or more against gravity were able to flex through this range with 1 pound or more. The maximum weight lifted was 6 pounds by one patient.

FIGURE 71-6 A, A 21-year-old man sustained a brachial plexus injury in a motorcycle accident. Steindler’s flexorplasty performed 16 months after injury restored excellent elbow flexion. Subsequent transfer of the flexor carpi ulnaris into the extensor carpi radialis brevis restored wrist extension, and finger and thumb extension were restored using the superficialis muscle of the middle and ring fingers. Shoulder arthrodesis was also performed. B, Fixed flexion deformity of 35 degrees was present after Steindler’s flexorplasty.

Nyholm ¹⁸³ reported on 24 patients, six of whom achieved a normal degree of elbow flexion after flexorplasty. Five patients (three who had brachial plexus trauma and two polio) recovered some biceps function, which did not seem to be compromised by the previous flexorplasty. Nyholm noted that all of the patients except two preferred to flex the elbow with the forearm pronated and hand clenched.

Hoffer and Phipps ¹¹³ reported on three Steindler flexorplasties in children with obstetric brachial plexus palsies (C5, C6). They achieved flexion strength M4 with 20 to 30 degrees of extension deficit.

One of the largest series of flexorplasties was reported by Lindholm and Einola,¹⁴⁵ who found that 50 of their 61 patients achieved a range of flexion of at least 90 degrees. Eleven were able to lift a weight heavier than 1 kg at right angles. Dutton and Dawson⁷⁶ noted that 20 patients had excellent or good function of the transferred muscles. Eighteen could lift at least 1 to 2 kg through an arc of full elbow motion.

Finally, several investigators have noted the beneficial effect of arthrodesis of the flail shoulder in patients who have had flexorplasties.^76,^124,^145,²²³ Shoulder arthrodesis permits abduction at the scapulothoracic joint that decreases the gravitational forces that must be overcome by the transferred flexor-pronator muscle group. Stabilizing the humerus also maximizes the power of the transfer in preventing backward movement of the elbow, which diminishes the advantage of the flexion contracture by increasing the length of the arc through which the reconstructed flexor muscle must move the weight against gravity.¹²⁴

A retrospective review of 12 patients with C5-7 brachial plexus palsy and extremely weak elbow flexion (M2 or less) were treated by Steindler flexorplasty and wrist arthrodesis. Eleven patients were found to have good or very good outcomes based on the criteria established by Alnot and Abols,⁸ and one patient had mild active flexion of the elbow.¹⁷² These results emphasize that despite recent advances in neurolysis nerve grafting and neurotization procedures, a direct approach to the neurologic lesion can give incomplete results, and for this reason, tendon transfers still play an important role. Chen⁴⁹ reported results in eight patients, in four of whom one or several flexor-pronator tendons had been used in tendon transfer procedures to restore wrist and finger extension. The results after modified Steindler flexorplasty were not negatively affected by these previous procedures.

Liu and associates ¹⁴⁶ reported on the largest cohort of 71 patients with a follow-up of 4 to 15 years. The mean range of motion during active flexion against gravity was 114 degrees (range: 80 to 140 degrees). A mean extension lag of 28 degrees (range: 0 to 50 degrees) was created. Fourteen patients (20%) had significant loss of supination, for 11 of whom flexor carpi ulnaris–to–extensor carpi radialis brevis tendon transfers were performed to regain supination. Using Mayer and Green’s scoring system,¹⁵⁸ results were excellent in 32%, good in 47%, fair in 13%, and poor in 8%. A clear improvement was seen over time; the results in operations performed between 1981 and 1987 were significantly better than those performed between 1970 and 1980 (57% versus 16% excellent results). This difference was attributed by the authors to the addition of shoulder arthrodesis as a stabilizing supplementary procedure in 45 cases, and the aforementioned tendon transfers in patients without active supination.

Andrisano ⁹ reported on the long-term follow up (3-13 years, average 8.6 years) in 22 patients. These authors found good results in 50%, fair in 19% and poor results in 38%. These results were independent of the etiology of the elbow paralysis and the patient’s age, but were strongly dependent on surgical factors, such as the correct tension and placement of the epitrochlear mass.

Brunelli and coworkers ³⁶ have modified the procedure such that during transfer of the flexor carpi ulnaris, flexor carpi radialis, and palmaris longus, they are carefully separated from the flexor digitorum superficialis, which is left in place. Together with placement of the transfer on the anterior aspect of the humerus, the so-called Steindler effect (pronation and simultaneous elbow and finger flexion) could be avoided in 28 of 32 cases.

Complications

The main complications of Steindler’s transfer are pronation and flexion contractures. Pronation contractures are related to (1) preoperative absence of active supination in most of these patients and (2) tightening of the pronator teres as a result of its proximal shift. Inserting the transfer more anteriorly and laterally on the humerus diminishes this tendency, but if no active supination is present, pronation deformity still may be a problem. Segal and associates ²¹³ and Nyholm¹⁸³ did not notice any great change in passive supination after Steindler flexorplasty in their patients. Carroll and Gartland⁴² noted that approximately half of the 23 patients who were improved by the flexorplasty had pronation defects. Lindholm and Einola¹⁴⁵ encountered pronation contracture in 17 of 61 patients who had had a flexorplasty, but 14 of them had had no active supination preoperatively. Dutton and Dawson noted an average loss of supination of 39 degrees in the 25 patients on whom they reported.⁷⁶

The strength of the transferred muscles, the duration of immobilization, and the strength of the opposing triceps muscle affect the extent of the flexion contracture. A flexion contracture of the elbow does increase the mechanical advantage of the transferred muscle, and this is especially important in patients who have significant weakness of the transferred muscle group. The degree of fixed flexion contracture of the elbow that is considered acceptable varies a good deal among surgeons. Steindler believed that a 60-degree flexion contracture was acceptable,²²³ and Carroll and Gartland accepted 40 degrees.⁴² Kettlekamp and Larson¹²⁴ observed that, when strength rather than appearance was important (typically for men and patients with severe paralysis of the contralateral arm), flexorplasties are more satisfactory with a flexion contracture between 30 and 60 degrees. Mayer and Green¹⁵⁸ were not satisfied if the flexion contracture exceeded 15 degrees, and they used turnbuckle splints postoperatively to minimize the flexion deformity. Dutton and Dawson⁷⁶ preferred a flexion contracture of 15 to 30 degrees, which allows the arm to “hang at the side” more normally in patients concerned with appearance, and a 30- to 45-degree contracture in patients who were more concerned with function.

Dutton and Dawson ⁷⁶ reported transient ulnar nerve paresthesias in two patients. Lindholm and Einola¹⁴⁵ reported on two patients with postoperative ulnar nerve paresthesias, one of whom was still having symptoms at follow-up. Description of the transferred muscle occasionally occurs with variable impact. Steindler reported an excellent result after reinsertion,²²³ whereas Kettelkamp and Larson recorded a poor result in a patient with an untreated description.¹²⁴

Inherent in the Steindler flexorplasty is the increased pronatory effect on the forearm and the associated risk of pronation deformity (Steindler effect). The incidence in the literature varies from 28% to 50%.^42,^76,¹⁴⁵ Theoretically, inserting the flexor mass more anteriorly and laterally on the humerus will decrease this effect.¹⁴⁶ The flexor carpi ulnaris can be transferred to the flexor carpi radialis⁴⁰ or extensor carpi radialis brevis¹⁴⁶ primarily or secondarily, or the flexor digitorum superficialis origin can be left in place.³⁶

Latissimus Dorsi Transfer

History

Schottstaedt and colleagues ²⁰⁷ were the first investigators to report successful restoration of elbow flexion by rotating the entire latissimus dorsi muscle on its neurovascular pedicle, reattaching its insertion to the coracoid and its origin to the distal biceps muscle and tendon. One patient could lift 4 pounds. They noted that the procedure was contraindicated when the latissimus dorsi was the only muscle capable of elevating the pelvis for a forward step. Zancolli and Mitre refined the technique described by Schottstaedt and associates and coined the term bipolar transfer.²⁵⁴ Brones and colleagues³³ reported that the bipolar myocutaneous latissimus dorsi flap restored contour and function in a post-traumatic condition.

Restoration of elbow flexion by transferring only the origin of the latissimus dorsi to the biceps insertion (unipolar transfer) was reported by Hovnanian in 1956.¹¹⁷ He noted that the cross-sectional area, fiber length, and excursion of the latissimus dorsi muscle compared favorably with those of the biceps and brachialis. Axer ¹¹ and others modified the unipolar technique by using only the upper third of the latissimus dorsi, transferring its origin off several spinous processes to the biceps tendon. Excursion of 7 cm was obtained by stimulating this part of the muscle. According to these authors, transferring only a portion of the muscle avoids the sometimes bulky enlargement of the arm that results when the whole muscle is used. Preservation of some function in the undisturbed portion of the latissimus dorsi was noted in one of the patients. Botte and Wood²⁸ decided whether to do bipolar or unipolar transfer at the time of surgery. If the line of pull stabilized in the plane of elbow motion and there was no tension on its neurovascular pedicle, the transfer was completed as a unipolar transfer. If the line of pull deviated outside the plane of elbow flexion or there was tension on the neurovascular pedicle, the transfer was converted to a bipolar transfer.

The latissimus muscle flap is widely used as a pedicled muscle flap for restoration of elbow flexion.^53,^55,^155,¹⁸²