Chapter 83: Nerve Transfers

In this scenario, there is no realistic prospect of meaningful distal motor reinnervation from revision of the neuroma given the distance and the time since injury. The time elapsed is of less relevance for sensory function. Nerve transfers could be used to shorten the reinnervation distance, provided a suitable expendable donor is available which is distal enough to allow reinnervation before motor end plate degeneration, is close enough to coapt directly without grafting, and is ideally synergistic in function to aid rehabilitation. The ideal donor should be predominantly motor or sensory to match the desired outcome, it should have a large axon count (particularly in a motor transfer), it should be well matched in size to the recipient to facilitate an accurate coaptation, and it should ideally have normal function with no previous injury. Example donors in this case would be branches of the radial nerve (extensor carpi radialis brevis [ECRB]/supinator) or the brachialis branch of the musculocutaneous nerve onto the anterior interosseous nerve (AIN).1–3

A key principle for selection of a donor nerve for nerve transfer is that it is uninjured and has a full MRC grade 5 function. In this case, this does apply to the phrenic nerve. However, it is still generally reserved for last resort reconstructive reinnervation procedures, as it has both a significant donor-site deficit (albeit usually transient) and is not synergistic to upper limb motor function, and therefore can be more difficult to rehabilitate. Its use would usually reflect there being very few suitable donor nerves available, such as a situation with complete root avulsion C5–T1 and where the spinal accessory nerve has also been injured. Many surgeons would prefer to use intercostal nerves rather than the phrenic nerve to reinnervate the elbow flexors because of the limitations described. A key element of any donor nerve is that it is expendable and while the phrenic nerve is considered to be expendable when required in scenarios of last resort, which does rely on the patient having sufficient respiratory reserve to withstand 1 to 2 years of measurably reduced respiratory function and a permanent reduction in inspiratory force generated.1 The negative effects on respiratory function seem to be much greater when the right phrenic nerve is used rather than the left.2 Its use is therefore not recommended in the presence of multiple rib fractures, significant underlying lung disease, or where there are problems with the contralateral hemidiaphragm. In patients with no underlying respiratory problems, the measurable reduction in lung function does not necessarily correspond to a clinical problem, even if both phrenic and some intercostal nerves are harvested.3 Because the phrenic nerve in the neck is proximal/medial and short in length, an intervening nerve graft is usually required to transfer its function to that of elbow flexion.4 It therefore does not facilitate shortening of the reinnervation distance that would have been achieved if direct repair or nerve grafting of the plexus injury had been possible to restore elbow flexion. There are, however, thoracoscopic techniques reported to harvest the phrenic nerve distally at its insertion to the diaphragm, greatly increasing its length and facilitating direct coaptation.5,6

3. Tung TH, Mackinnon SE. Nerve transfers: indications, techniques, and outcomes. J Hand Surg Am 2010;35(2):332–341

In general, there are some key technical principles in nerve transfer surgery which need to be adhered to in order to optimize functional outcomes. In this example, when considering reinnervation of the suprascapular nerve territory to restore glenohumeral stability, initiation of shoulder abduction and external rotation of the upper limb, the maximum number of motor fascicles available should be directed into the recipient nerve. This is generally achieved by having the shortest distance for reinnervation to occur, with the least coaptation possible and by performing the most direct repair which will normally be an end-to-end coaptation. Any interposed nerve graft requires an additional coaptation, increases length of the reinnervation pathway, and reduces the number of axons reaching the target muscles, as there will always be losses crossing even the most technically excellent coaptation site.1 In this scenario, as the recipient nerve is completely denervated with no prospect of recovery, an end-to-end coaptation will be preferred especially as this will not be sacrificing any residual function that might have been spared by a supercharged reverse end-to-side technique, and will target a greater number of regenerating axons to the distal target than an end-to-side technique. A supercharged reverse end-to-side technique would involve dividing the spinal accessory nerve and coapting it end to side onto an epineurial window in the suprascapular nerve. Conversely, an end-to-side coaptation would involve suturing the distal divided stump of the nonfunctioning suprascapular nerve onto the side of the intact spinal accessory nerve, and has not been demonstrated to be reliable, particularly if no donor axonal injury is induced (Fig. 83.1).2 End-to-end nerve transfer End-to-side nerve transfer Supercharged end-to-side nerve transfer Donor Distal Proximal Recipient Fig. 83.1: Three types of nerve coaptations. End-to-end is the gold standard for all transfers and the only functional anastomosis for motor nerve transfers. End-to-side is preferred for sensory nerve transfers. Reverse end-to-side, or supercharged, coaptations are utilized to provide small but quick reinnervation of motor endplates while primary or transferred nerve axons regenerate.

morbidity? Isolated long thoracic nerve palsy resulting in serratus anterior paralysis and scapula winging is usually a closed traction injury related to exercise. Initial management is most commonly nonoperative with physiotherapy to reduce abnormal adaptive movements and reduce stretching out of the weak or paralyzed serratus anterior muscle. In those patients who are not progressing with a recovery of function, a nerve transfer using the thoracodorsal nerve as a donor coapted end-to-end onto the long thoracic nerve to serratus anterior can restore function over a short reinnervation distance. It is technically fairly straightforward as the two nerves are in close proximity either side of a relatively bloodless dissection plane on an accessible area of the chest wall.1 Transfer of the medial pectoral nerve has also been described for long thoracic nerve palsy, but is much less common, involves more dissection, and may require interpositional grafting to achieve reach.2 There has been a cadaveric study supporting intercostal nerve transfer. This study showed that the size match was reasonable between the donor and recipient nerves and that the required donor site nerve length was adequate to perform direct end-to-end coaptation. This would not be a synergistic transfer.3 A thoracodorsal nerve to axillary nerve transfer can be used to improve shoulder function by reinnervating the deltoid and teres minor but is not relevant in this scenario.4 A spinal accessory nerve to suprascapular nerve transfer would be indicated to reinnervate the infraspinatus and supraspinatus muscles in order to improve shoulder abduction following plexus injury but again is not relevant in this scenario.5

6. Xu WD, Gu YD, Xu JG, Tan LJ. Full-length phrenic nerve transfer by means of video-assisted thoracic surgery in treating brachial plexus avulsion injury. Plast Reconstr Surg 2002;110(1):104–109, discusion 110–111 INDICATIONS FOR NERVE TRANSFER 3. Which one of the following represents a key indication for nerve transfer following an upper limb motor nerve injury? There are a number of key indications for nerve transfer following motor nerve injury in the upper limb, for example, in cases where a primary nerve repair is not possible, a nerve graft is unlikely to be successful, and where there is a long distance between the level of injury and the target muscle. In cases where nerve grafts greater than 7 cm are required, the outcomes tend to be poor. The benefit of using a nerve transfer in cases where there is a long distance between the injury and the target muscle is that it may prevent irreversible changes to the motor endplates which will occur if they remain denervated for more than about 14 months. By shortening the distance required for nerve recovery, this can sometimes be avoided. This is particularly important in cases where there is a delayed presentation following nerve injury. Nerve transfers are less often used in more distal nerve injuries, as the distance to nerve recovery is shorter and hence there is less of a time constraint in relation to the motor end plates. In these scenarios, nerve repair or grafting is generally preferred. Most nerve gaps of 2 cm can be repaired effectively with nerve grafts. A nerve injury with early signs of recovery does not usually warrant nerve or tendon transfer, as this suggests a neuropraxia and that recovery will continue. In some cases, a tendon transfer may be considered (e.g. pronator teres to ECRB end to side in a high radial nerve injury) in order to provide wrist extension as an alternative to an external splint in the interim, while recovery occurs. If there is no evidence of recovery at 20 months post–motor nerve injury, then the motor endplates are unlikely to remain functional; so, the window of opportunity has been missed in terms of benefits of nerve transfer. Contraindications to nerve transfer include an extended time interval between injury and transfer (usually more than 14–18 months is considered too long), where there is no expendable donor, and where the donor nerve has a motor strength below MRC grade M4.1,2

function? The clinical priority for restoration of upper limb function following a brachial plexus injury is elbow flexion in most situations. The second target is usually achieving shoulder stability combined with abduction and external rotation of the shoulder. This combination allows a hand to mouth action that can be a useful adjunct to the contralateral uninjured limb even if no further functional recovery is achieved. Lack of shoulder stability results in poor upper limb function, and a lot of the power generated for elbow flexion can be wasted in relocating a subluxated glenohumeral joint. There is also usually long-term discomfort associated with an unstable scapulothoracic interface and glenohumeral joint. Wrist extension is important for passive tenodesis hand grip or to increase the power of active finger flexion. Hand and wrist function are of limited use if shoulder and elbow paresis leave the patient unable to position the hand in space. Elbow extension is generally assisted by gravity and so is less important to correct with surgery, provided at least one arm has active elbow extension for activities such as pushing up to standing from a seated position, reaching overhead objects, and chair to bed transfers in spinal cord injury patients. In cases where nerve transfer is performed for elbow extension, it is usually coapted to the nerve branch to long head of triceps and the “gold standard” donor is to use the teres major motor branch.1–5

Elbow flexion is key to upper limb function and is lost in a C5/6 plexus injury. The Oberlin procedure remains the “gold standard” for restoration of elbow flexion by transfer of an expendable ulnar motor fascicle of the flexor carpi ulnaris onto the biceps branch of the musculocutaneous nerve (Fig. 83.2).1 It is possible to further augment the Oberlin procedure with additional transfer of an expendable median nerve motor fascicle, such as to flexor carpi radialis (FCR) or flexor digitorum superficialis (FDS), onto the brachialis branch of the musculocutaneous nerve.2 There is, however, debate as to whether this double fascicular modification to the Oberlin procedure makes a significant difference to the functional outcomes and indeed it may not do so (Fig. 83.3).3,4 Other options for restoration of elbow flexion in the brachial plexus injury patient include intercostal nerve transfer, thoracodorsal nerve transfer, and medial pectoral nerve transfer. The intercostal nerve transfer shows inferior results to the Oberlin transfer but is indicated in five root plexus avulsions. Two to three nerves must be transferred. The thoracodorsal nerve is recommended if the Oberlin transfer is not possible but is contraindicated if shoulder adduction is weak. The medial pectoral nerve transfer is used for C5–C6 nerve root injuries as pectoralis function is preserved. Unfortunately, nerve grafting is also required in many cases which impairs overall outcomes.5,6 Biceps branch (proximal) Biceps branch (distal) Musculocutanous nerve Biceps brachii muscle Ulnar nerve Median nerve Flexor carpi fascicles Triceps brachii muscle Brachial artery Brachial artery Brachialis muscle Fig. 83.2: Oberlin transfer involves coapting flexor carpi ulnaris motor fascicles from the ulnar nerve to the biceps branch of musculocutaneous. The double fascicular transfer combines this with transfer of flexor carpi radialis motor fascicles from the median nerve to brachialis branch.

posterior interosseous nerve. There are a number of different transfers that can be used to restore wrist and digit extension following proximal radial nerve injury. A dual nerve transfer approach is preferred to address both wrist and digital extension. Ideally, two nerve branches to flexor digitorum superficialis (FDS) are transferred to extensor carpi radialis brevis (ECRB) for wrist extension while flexor carpi radialis (FCR) and palmaris longus (PL) nerve branches are transferred to the posterior interosseous nerve (PIN) for finger extension. With this approach, most patients are able to recover from individual finger extension.1 The median nerve is accessed through an incision in the volar forearm just distal to the antecubital fossa. All median nerve branches are identified via electrical stimulation. Nerve branches to the FDS and FCR will cause finger flexion and wrist flexion, respectively. The radial sensory nerve can be accessed through the same incision and followed proximally to the PIN and ECRB branch. These branches are divided as proximally as possible. The FDS and FCR branches are then divided as distally as possible and nerve coaptation is performed. An alternative is to use pronator quadratus motor branch of the AIN transfer to PIN, but this is technically more difficult than dual nerve transfer.2 Transfer of the supinator motor branch onto the PIN is indicated in cases Spinal accessory nerve (distal) Spinal accessory nerve (proximal) Long head of triceps branch (distal) Axillary nerve (distal) Axillary nerve (proximal) Suprascapular nerve Long head of triceps branch (proximal) Fig. 83.3: Double nerve transfer for shoulder abduction showing distal branches of spinal accessory nerve to trapezius transferring to suprascapular nerve and motor branch to the long head of the triceps (radial nerve) transferring to the axillary nerve. of C7–T1 nerve root injuries and can provide excellent results, as supinator innervation (C6) is preserved.3,4 It involves the transfer of one or both nerve branches to supinator and helps recover function of extensor pollicis longus and extensor digitorum communis muscles. Nerve transfers for wrist extension tend to target the ECRB to give a more balanced centralized extension of the wrist, rather than ECU or ECRL.

Intrinsic muscle innervation is critical to balancing the extrinsic flexor and extensor tendons when opening and closing the hand, and to most types of grip. The majority of the intrinsic muscles are innervated by the ulnar nerve, and by the T1 nerve root. The median innervated intrinsics can be remembered by “LOAF” (lateral 2 lumbricals, opponens, abductor, and flexor pollicis brevis). A hand without good intrinsic function tends to adopt a claw position. In a proximal ulnar nerve lesion, the reinnervation distance and therefore time may exceed viability of the motor endplates in the hand, leading to little or no functional recovery. The “gold standard” nerve transfer for this scenario is to use the AIN terminal branch to pronator quadratus transferred either end to end or supercharged reverse end to side onto the motor branch of the ulnar nerve, according to whether a late motor recovery is anticipated or not. The posterior interosseous nerve branch to extensor carpi radialis brevis, musculocutaneous nerve branch to brachialis, median nerve motor branch to flexor carpi radialis, and radial nerve branch to medial triceps are all nerves that can be sacrificed for nerve transfers in the arm, but are all a more significant distance from the intended recipient muscles than the distal pronator quadratus anterior interosseous branch.1–3