|Home | About | Journals | Submit | Contact Us | Français|
Nerve transfers are increasingly utilized for repair of servere brachial plexus injuries and, indeed, are the only option when the proximal spine nerve roots have been avulsed from the spinal cord. The procedure essentially involves the coaption of a proximal foreign nerve to the distal denervated nerve, so that the latter will be reinnervated by the donated axons. The primary goal of surgery in the severe obstetrical brachial plexus palsy case is to return proximal arm function, particularly elbow flexion and a stable, dynamic shoulder that can abduct and externally rotate. Neural input should thus be directed first to the biceps via the musculocutaneous nerve or its branches, and next to reconstructing the suprascapular nerve. Unlike an adult with a complete palsy, where the return of distal hand function is virtually impossible, the infant has better odds of successful reinnervation of the hand. If donor nerve sources are available following repair of the musculocutaneous and suprascapular nerves, grafts can be directed to the radial (for wrist and finger extention) and median (for elbow and finger flexion and critical hand sensibility) nerves. Recovery of intrinsic hand muscle function from plexus reconstruction in the global severe palsy case is not a realistic possibility at present. Cortical plasticity, which is likely more prevelent in the baby than the adult, appears to play an important physiological role in the functional recovery of the reinnervated muscles following nerve transfers.
Traction on the brachial plexus during maneuvers to clear the shoulder in a difficult delivery is felt to be the predominant mechanism by which plexus damage occurs. Experimental studies done at the turn of the century by Clark and Sever1,2 demonstrated that the severity and extent of obstetrical brachial plexus palsy (OBPP) were related to the amount of traction placed on the arm against the neck. Over 90% of injuries occur in overweight infants (3807 to 5000 g) with cephalic presentation, resulting cephalopelvic disproportion, and shoulder dystocia with neck hyperextension.3,4,5,6 In contrast, babies with plexus injury delivered as a breech presentation are usually small (<3000 g).5,6 Here, manipulation of the arm and hyperextension of the neck with the head extracted imposes considerable traction on the brachial plexus.3 This mechanism predisposes toward lower root injuries, bilateral involvement,3 and greater likelihood of severe avulsions.5
Shortly after delivery, the diagnosis of OBPP is entertained by the observation of a flail upper extremity, with the entire plexus affected in most infants at birth. By 2–6 weeks of age, however, the predominant level of the lesion should be well established. Three typical syndromes of brachial plexus injuries are seen. The majority of obstetrically injured infants will have sustained damage to the upper (C5 and C6) roots, with a resulting Erb palsy.7 The second most frequent type of injury is the pan-plexus lesion, in which there is some or complete damage to all the roots.8 The lower plexus palsy (Klumpke paralysis) involves the C8 and T1 nerve roots, but in practice, isolated lower plexus injuries are actually rare (2–3%). The remaining cases are combined total lesions.
In the complete severe palsy, the arm is totally flail and insensate and appears pale and sometimes swollen because of loss of vasomotor control. The complete lesion is often associated with a Horner's sign, which should be sought for in dark room conditions in subtle cases. The lower extremities should be checked for muscle tone and hyperreflexia, which are seen with the most severe plexus injuries in which multiple nerve roots have been avulsed and the spinal cord injured. Phrenic nerve palsy also may rarely accompany a severe lesion,9 and a chest x-ray to assess for an elevated hemidiaphragm is invaluable. A complete injury frequently evolves over time into an injury that primarily reflects an upper root injury. In practice, at the 4-week visit, the injury has usually declared itself as an unchanged complete paralysis or an upper root paralysis. The 3-month visit is critical. For those patients exhibiting spontaneous recovery, especially in biceps and deltoid, further watchful waiting is recommended.10 If no substantial recovery has occurred by the 3-month visit, an electromyogram is performed to confirm the degree of brachial plexus injury and further define the particular nerve roots affected.6,11,12 Unfortunately, the utility of electrodiagnostic studies beyond this in the evaluation and management of OBPP remains unclear, as evidence of electromyographic reinnervation over time may have poor correlation with clinical recovery.13 Our approach is to use the 3-month evaluation as an important initial flag. If the patient has a complete palsy (especially with a coexisting Horner pupil)14 and no recovery, surgery is almost certainly to be required, and preparations are undertaken to perform it. With Erb palsy, no recovery in biceps and deltoid at 3 months also identifies a likely surgical candidate. Surgery is typically undertaken at 4–6 months of age in these patients. The patients demonstrating early (3-month) recovery in deltoid and biceps are followed, and most gain further recovery and are destined for a good outcome. For those few who plateau with poor recovery, a delayed surgical exploration is recommended. We do not offer surgery to isolated Klumpke palsy patients, as the prognosis for hand intrinsic functional recovery following plexal nerve reconstruction is poor.
We reserve a computed tomography myelogram to cases that are felt to need surgical exploration as a preoperative study. Thin-section computed tomography postmyelography is the most sensitive test to date for revealing complete nerve root avulsions, with or without a pseudomeningocele (Fig. 1),15 where the absence of rootlets within a pseudomeningocele best predicts a root avulsion.16 At the current time, computed tomography myelography provides a better determination for nerve root avulsion than magnetic resonance imaging.17
The surgical approach to the severe OBPP patient is to first ascertain whether the preoperative impression that all the nerve roots are truly avulsed is in fact the case. The second aspect of surgery is to perform nerve repair, which in the severe cases with avulsion frequently incorporates nerve transfer operations to reanimate the extremity. The surgical exploration therefore warrants exposure of the entire supraclavicular and infraclavicular plexus, with appropriately placed incision, as well as marking of incisions that will allow exposure of donor nerves that may need to be transferred (Fig. 2). In the infant, a transverse skin incision above the clavicle, followed by well-developed subplatysmal flaps, allows for complete dissection in the posterior triangle, including C3–T1 spinal nerve and trunks, as well as exposing the entire trajectory of the accessory nerve (Fig. 2A). Identification and protection of the phrenic nerve over the scalenus anterior muscle (the identification of which is assisted by using a disposable nerve stimulator) should be made early, as it may be extremely small or tethered by neuroma or fibrosis in the young infant. For the infraclavicular plexus, the cords of the plexus can be exposed by an incision along the deltopectoral groove (Fig. 2). Access to the main nerve branches from the cords requires division of the pectoralis minor and is aided by release of the upper part of the tendinous attachment of the pectoralis major to the humerus.
After complete and thorough exposure of the plexus, including intraforaminal dissection and external neurolysis of the nerve-in-continuity, intraoperative electrodiagnostic studies follow.18 First, motor-evoked stimulation is used to determine which roots are conducting by observing the distal muscles contraction, occasionally augmented by needle electromyography. Conduction of nerve action potentials across the neuroma or from clinically nonfunctioning roots are next recorded, with large, fast-conducting (preganglionic) nerve action potentials seen in the situation in which the nerve root is avulsed.18,19,20,21 Somatosensory evoked potentials from the opposite cortex help determine the integrity of sensory roots but appear to not be as useful in infants as in adults and, moreover, do not assay the motor (ventral) roots.22 The intraoperative electrical tests and operative findings are used in concert with the preoperative clinical exam, electromyograms, and imaging to determine the extent of injury and presence of root avulsion to guide operative decisions about the type of nerve reconstruction, as detailed below.
Because the mandate of this chapter is to discuss management of the severe OBPP (which implies multiple nerve root avulsions) with nerve transfers, we will deliberately not discuss the circumstances of neuroma neurolysis and plexoplexal repairs for extraforaminal spinal nerve rupture, which is the most typical scenario. The management of upper root paralysis is guided with a goal of complete repair. In the situation in which the upper two roots are avulsed from the spinal cord, the only repair option is a nerve transfer. The exact detail of the repair is therefore dictated by the number of root avulsions, including the consideration if C7 is also avulsed. However, even in the complete severe palsy, the C5 spinal may be singularly spared, thus allowing it to be used as the source of axons for a plexoplexal repair to distal elements (Fig. 3).8,23 To deny a plexoplexal repair from a useable C5 spinal nerve to its distal outflow would be a disservice, given the relatively poor number of extraplexal transfer possibilities. There are a few cases in which it is unclear whether the proximal root stump has sufficient integrity to form the basis of a graft. We have found the use of very proximal intraforaminal dissection of the nerve roots to be invaluable in these cases. Intraforaminal dissection, along with a frozen section of the very proximal stump to assess fascicular pattern and absence of ganglion cells, has been useful in decision making; other researchers have assessed the degree of myelin staining to predict the quality of the stump.24 However, in uncertain circumstances, we prefer to use a nerve transfer rather than a questionable proximal stump. The possible permutations and combinations for repair therefore include intraplexal grafts alone from a single functioning root (Fig. 3), and intraplexal grafts and selective transfers or transfers alone for the devastating but decidedly rare cases in which all nerve roots are avulsed (Fig. 4). We focus next on the nerve transfer options available.
Nerve transfers (or “neurotization”) involve the repair of a distal denervated nerve element using a proximal foreign nerve as the donor of neurons and their axons, which will reinnervate the distal targets. The concept is to sacrifice the function of a (lesser valued) donor muscle to revive function in the recipient nerve and muscle that will undergo reinnervation.25 In adult plexus surgery, where avulsions are much more common,26 nerve transfers have become increasingly popular and used for the repair of brachial plexus injuries, with many different ingenious transfers associated with improving results, as reported and reviewed recently.27,28,29,30
The anatomical and physiological principles that underlie nerve transfers are relatively straightforward. Because motor recovery has been the main goal, the choice of a nerve input that has a reasonable aliquot of motor fibers is required.25 The loss of the donor muscle must not represent loss of important or critical function.31 Obviously, the value of the element to be reinnervated must greatly exceed the utility of the sacrificed element. There are several important principles to adopt to maximize outcome in nerve transfers, the first of which is to reinnervate the recipient nerve as close to the target muscle as possible,32 which is exemplified by the transfer of an ulnar nerve fascicle directly to the biceps branch of the musculocutaneous, in close proximity to its entry into the muscle.33 The second principle is to perform a direct repair, without intervening grafts—a tactic that seems to be associated with improved outcomes.34,35,36,37,38 The third principle is to use combinations of similarly behaving neuromuscular units, maximized when agonistic donors and recipients are chosen, as cortical readaptation is the physiological basis for functional recovery.39,40 This may also be the physiological underpinning that explains why intraplexal (e.g., medial pectoral-musculocutaneous) nerve donors may garner superior results as compared with extraplexal (e.g., intercostal-musculocutaneous) nerves.41 The last principle is to perform the surgery as early as possible after irreversible injury to maximize outcomes.42,43,44
The earliest report of a nerve transfer (by Tuttle in 1913) mentions the use of the accessory nerve as the donor.45 An improved appreciation of the extracranial accessory nerve anatomy based on cadaveric studies46 has led to the adoption of a more distal dissection of the nerve close to the trapezius muscle insertional points, where the very distal nerve is divided before its transfer (Fig. 4).47 For restoration of dynamic shoulder function, both the suprascapular and axillary nerves have been chosen as targets. Although the former can be repaired directly by end–end suture by the distal accessory nerve, the latter requires an interposed nerve graft. Recent series have confirmed Alnot's bias,48 that the suprascapular nerve is a superior target (compared with axillary nerve) for accessory nerve transfer, with generally good results reported in the majority of patients.49,50,51,52 In a recent meta-analysis of the literature related to nerve transfers, Merrell concluded that for shoulder abduction, best results were achieved with an accessory-suprascapular transfer, as compared with other transfers.28 Other researchers have noted that shoulder function will be further optimized if both suprascapular and axillary nerves and their muscles are reinnervated.41
The concept for intercostal nerves transfers to repair brachial plexus injuries can be credited to Yeoman, working with Seddon,45 with a report of this early experience reviewed by Seddon in his classical textbook.53 Standardized techniques, consisting of the use of three intercostal nerves (third to fifth) to the distal musculocutaneous, without interposed grafts, led to more consistent results, approaching Medical Research Council grade 3 (M3) or better function in ~50% of the patients.54 More recent series using intercostal nerve transfers to musculocutaneous report significantly improved results, ranging from 64 to 88% for at least M3 elbow flexion.30,34,35,36,37,38,55 Intercostal nerve transfers to musculocutaneous in infants with OBPP produced reliably good outcomes, with close to 85% of patients obtaining M4 or better elbow flexion.56 These authors stress the importance of well dissecting the intercostal nerves distally along the rib to allow their transfer easily to the axilla and direct repair without graft, as has been nicely demonstrated in anatomic studies (see incision in Fig. Fig.22B).57 In adults, a large series using gracilis muscle grafts for elbow flexion reported that intercostal neurotization of nerve to free muscle was a better choice than using the accessory nerve.58 In severe OBPP, a gracilis free muscle transfer (innervated by accessory nerve) can obtain elbow flexion,59 whereas double free muscle transfers may restore prehensile function successfully in very severe cases.60 Unilateral intercostal nerve transfers do not embarrass respiratory function61 but should not be performed if phrenic nerve is dysfunctional or used on the ipsilateral side as a donor for transfer.31
In some cases of pure Erb palsy, where both C5 and C6 are avulsed, the C7 spinal nerve is intact and available as a (intraplexal) donor for reinnervating the distal upper truncal or its divisional outflow.23,62,63 Such a transfer can be associated with very good outcomes, related to the recipient elements,23,63,64 with little risk of loss of function from taking the C7 spinal nerve, as long as there is an absence of coexistent lower plexus lesions.65 The redundancy of the C7 spinal nerve allowing for its safe sacrifice has been verified by authors who have used the contralateral C7 as a donor for transfer.66,67 Other than clinical mild loss of triceps function and clinically inconsequential loss of the triceps reflex, the procedure appears to be safe as far as motor loss is concerned.66,67 Sensory abnormalities are common following C7 sacrifice and may be permanent in 5% of cases.68 Chuang and colleagues first reported the use of the contralateral (from the normal, noninjured side) C7 spinal nerve, reporting modest results of nonindependent movement in the paralyzed limb.69 Subsequent series in adolescent and adult patients reported the results of contralateral C7 transfer via cross chest sural or vascularized ulnar nerve grafts directly to recipient infraclavicular plexal nerve elements.66,67,70,71,72,73,74 The long-term functional outcome was claimed to be very good in the largest reported series, but data were only reported in a minority of the patients operated on.70,71 The best results in the use of contralateral C7 for total avulsion noted that when the median nerve was the recipient, good sensory function was achieved in about half of the adolescents, and some also experienced forearm muscle recovery.72 In a most carefully reported 3-year follow up, Songchareon reported median nerve motor recovery to an M3 or M4 in ~20% of patients, whereas another 20% had an M2 outcome in wrist flexion.75 Outcome in the sensory domain was somewhat better, especially in adolescents, with half of these patients having useful sensory restoration in the median nerve distribution.75 We conclude that this technique remains limited given the rather modest results in motor recovery,64 the fact that synchronous movement of the healthy side is required,70,71 and the small but real risk of some functional deficit incurred in the donor side.68,75 Perhaps a targeted approach to obtain median nerve distribution sensory recovery is warranted,55 although similar outcomes can be obtained with a safer and less cumbersome transfer from the lower intercostals to the sensory (lateral cord) head of the median nerve.76
The use of the selective motor branches of the cervical plexus (C3 and C4), for transfer to the distal elements usually supplied by (the avulsed) upper spinal nerves, was first advocated by Brunelli, with modest results.77 In a more recent series, Yamada transferred the anterior primary rami of C3 and C4, via sural nerve grafts, to the upper trunk (in cases of C5 and C6 avulsion and a few flail arm cases) and claimed fantastic results.78,79 Others have been unable to validate these remarkable results, demonstrating more modest success with using the motor components of cervical plexus to branches of the upper trunk when both C5 and C6 are avulsed.23,62,80 Unlike the cervical plexus, which contains a variable number of motor fibers,77 the phrenic nerve contains a large number of pure motor axons that allow the possibility of entire or partial transfer with success.81 In particular, the transfer to musculocutaneous has been an excellent tactic, with 11 of 12 patients achieving better than antigravity and 58% achieving M4 function.31 An important issue when the phrenic nerve is sacrificed is the resulting respiratory function compromise, which has been measured to average about a 10% decrease in vital capacity.31 Although not clinically important in the majority of situations, this degree of loss in respiratory reserve will produce symptoms in higher-demand situations and may be severely detrimental to infants and children who develop respiratory infections. In our opinion, this essentially precludes the use of the phrenic nerve as a donor in infants who are undergoing nerve reconstruction for OBPP.
The pectoralis major muscle has dual input from both the medial and lateral pectoral nerves, arising from the medial and lateral cords, respectively. Because C5 and C6 avulsion interrupts the lateral cord supply, the muscle remains innervated (and strong) as long as a significant injury is not incurred to the C7 and C8 elements. Although popularized recently for upper plexus injuries,82 the medial pectoral nerve as a donor for transfer has previously been considered and infrequently used, as reviewed by Narakas for adults and Gilbert for obstetric palsy.5,45 A resurgence of interest in this transfer has been associated with reports of useful outcomes (defined as M3 or better) in elbow flexion in ~84% of patients.83 Excellent results in OBPP too have been claimed, with success in 68% of cases.84 Others have criticized the use of this transfer in OBPP because of the loss of arm adduction, which can be useful for the infant/toddler/child to hold objects against his or her trunk,39 especially as the intercostal transfer appears to be so favorable in this clinical setting.56 However, the results of various series vary, and Samardzic noted that medial pectoral transfers were associated with significantly improved outcomes in elbow flexion as compared with intercostal and accessory nerve transfers.41 The latter group has also been one of the few to demonstrate consistently good results with transfers to the axillary nerve, reporting useful results in over 80% of patients.83 Only one or, at times, two of the terminal branches to the pectoralis major need be taken (and quite distally) for the transfer, with the practical implication that some pectoralis major supply can be preserved and a direct repair without intervening graft can be performed to the musculocutaneous nerve in the distal axilla.32 Significant caution needs to be exercised if there is substantial injury involving C7 and C8; in this case, the pectoralis major will be quite weak preoperatively, which is a contraindication to considering a medial pectoral transfer.
One of the most exciting recent developments in the neurotization field has been the transfer of portions of functioning distal plexal elements to directly reinnervate nerve branches going to critical muscles that are parlayed.32 This era really began with Oberlin's anatomical studies of the fascicular pattern and then the application in several patients, where a single redundant ulnar nerve fascicle was transferred to biceps branches in the medial arm to restore elbow flexion.85 A key aspect of the procedure is to reinnervate the biceps branch close to its motor entry into the muscle.86 The initial report of excellent results has subsequently been validated by several other authors,29,85,87,88,89,90 including their use in birth palsies.90 Most impressive have been the results reported by Sungpet, who used a single ulnar nerve fascicle directed to biceps and obtained a grade of M3 or better in 34 of 36 patients.89 A recent report indicates that elbow flexion function will be further augmented (especially in delayed surgery cases) by also concomitantly reinnervating brachialis muscle via a graft from the medial pectoral nerve.91 Another alternative to using the ulnar fascicle is to use a fascicle of the adjacent median nerve to transfer to biceps muscle nerve, with good results ranging from 6429 to 80% of patients.92 An emerging transfer is the direct repair of the anterior branch of the axillary nerve by the nerve that goes to the long head of triceps in the posterior arm.93,94
The level of evidence concerning which technique is best for a given lesion is rather poor. However, based on some recent reviews of the literature, some general guidelines can be provided. A meta-analysis conducted on the nerve transfer literature noted that for restoration of shoulder abduction, it is best to use an accessory nerve transfer to the suprascapular nerve, whereas for elbow flexion, intercostals without graft should be performed.28 This set of transfers is certainly appropriate for the situation in which the patient has a complete flail arm with all five spinal nerve roots avulsed. In the grim scenario of complete avulsions, the addition of a contralateral C7 transfer with an interposed vascularized ulnar nerve graft directed to the entire or perhaps the lateral root of median nerve in the axilla could be considered.64 Alternatively, and preferred by us, would be to transfer some lower intercostals or sensory cervical plexus elements to the sensory aspect of the median nerve. Although the above strategy is appropriate for the pan plexus injury, the tactics are very different for the isolated Erb palsy. If C5 and C6 are avulsed but C7 is clearly intact, intraplexal graft repairs from the C7 may be considered to reinnervate the shoulder abductors and elbow flexors.23 Alternatively, directed discrete intraplexal transfers should be performed, and based on the most recent literature, this seems to be a favored approach. A combination of distal accessory to suprascapular, ulnar nerve fascicle to biceps nerve (perhaps augmented by a portion of medial pectoral nerve via graft to brachialis nerve), and long head of triceps nerve (or thoracodorsal nerve)41 to the anterior portion of the axillary nerve should be performed.
The primary goal of surgery in the severe OBPP case is to return proximal arm function, particularly elbow flexion, and a stable, dynamic shoulder. Restoration of elbow flexion and arm abduction allows the patient to place the arm into a position of function, which is functionally meaningful (Fig. 5). Neural input should thus be directed first to the biceps via the musculocutaneous or to its branches to restore elbow flexion.95 The next order of priority is to achieve a stable shoulder that can abduct, which best obtained by reconstructing the suprascapular nerve to reanimate the supraspinatus rather than the axillary nerve.49 Unlike the adult with a complete palsy, in whom the return of distal hand function is virtually impossible, the infant has better odds of successful reinnervation of the hand, and this may be attempted if donor nerve sources are available following repair of the musculocutaneous and suprascapular nerves.5,96 In this case, grafts can be directed to the radial (for wrist and finger extension) and median (for elbow and finger flexion and critical hand sensibility) nerves. Recovery of intrinsic hand muscle function from plexus reconstruction is not a realistic possibility at present.