Surgical repair of brachial plexus injury: a multinational survey of experienced peripheral nerve surgeons

Neurosurg Focus 16 (5):Preview Article 1, 2004, Click here to return to Table of Contents Surgical repair of brachial plexus injury: a multinational survey of experienced peripheral nerve surgeons ALLAN J. BELZBERG, M.D., F.R.C.S.(C), MICHAEL J. DORSI, B.A., PHILLIP B. STORM, M.D., AND JOHN L. MORIARITY, M.D. Department of Neurological Surgery, Johns Hopkins School of Medicine, Baltimore, Maryland Background. Brachial plexus injuries (BPIs) are often devastating events that lead to upper-extremity paralysis, rendering it a painful extraneous appendage. Fortunately, there are several nerve repair techniques that provide restoration of some function. Although there is general agreement in the medical community concerning which patients may benefit from surgical intervention, the actual repair technique for a given lesion is less clear. Object. The authors sought to identify and better define areas of agreement and disagreement among experienced peripheral nerve surgeons regarding the management of BPIs. Methods. The authors developed a detailed survey in two parts: one part addressing general issues related to BPI and the other presenting four clinical cases. The survey was mailed to 126 experienced peripheral nerve physicians of whom 49 (39%) participated in the study. The respondents represented 22 countries and multiple surgical subspecialties. They performed a mean of 34 brachial plexus reconstructions annually. Areas of significant disagreement included the timing and indications for surgical intervention in birth-related palsy, management of neuroma-in-continuity, the best transfers to achieve elbow flexion and shoulder abduction, the use of intra- or extraplexal donors for motor neurotization, and the use of distal compared with proximal coaptation during nerve transfer. Conclusions. Experienced peripheral nerve surgeons disagreed in important respects as to the management of BPI. The decisions made by the various treating physicians underscored the many areas of disagreement regarding the treatment of BPI including the diagnostic approach to defining the injury, timing of and indications for surgical intervention in birth-related palsy, management of neuroma-in-continuity, choice of nerve transfers to achieve elbow flexion and shoulder abduction, use of intra- or extraplexal donors for neurotization, and the use of distal or proximal coaptation during nerve transfer. KEY WORDS brachial plexus peripheral nerve surgical repair nerve injury Abbreviations used in this paper: BPI = brachial plexus injury; CT = computerized tomography; EMG = electromyography; MR = magnetic resonance; SEM = standard error of the mean; SNAP = sensory nerve action potential; SSEP = somatosensory evoked potential. The management of BPI represents one of the most complex challenges facing the peripheral nerve surgeon. The intricate anatomy and variety of potential injuries combine to yield a wide array of pathological entities, clinical deficits, potential treatments, and prognoses. Additionally, standardized measures are lacking and it is difficult to quantify and compare pre-, intra-, and postoperative findings. Therefore, surgeons often disagree on the best approach to repairing BPIs. Controversial areas include priorities for motor recovery 3,45,49 and how best to diagnose nerve root avulsion pre- and intraoperatively. 9,23, 24,38,52 In addition, treatment of particular forms of BPI, including birth-related palsies 6,26,32,46,56 and adult traction injuries causing complete, upper, or lower plexus palsies, 3,4, 34,37,44,49,53,54 vary widely among surgeons. Disagreement concerning the management of BPIs will likely persist until prospective randomized studies are conducted to determine the relative merits of various treatments. Before undertaking such a study, the areas of agreement and disagreement must be clearly defined. The present study was designed to define variability among experienced peripheral nerve surgeons as how best to treat a particular brachial plexus lesion. Our hypothesis is that when peripheral nerve surgeons are provided with the clinical details of a patient with a BPI, significant treatment-related disagreement will be observed. Additionally, it is hoped that the results of this survey will provide the less experienced practitioner with an overview of various brachial plexus lesions and the management approaches most often used by more experienced surgeons. Finally, although we recognize the great benefit to patients provided by procedures such as muscle and tendon transfers, in the present study we focused on peripheral nerve reconstruction in BPI. CLINICAL MATERIAL AND METHODS Study Participants We reviewed the literature and membership of various professional organizations and generated a list of 126 peripheral nerve experts (physicians) practicing on four conti- Neurosurg. Focus / Volume 16 / May, 2004 1

A. J. Belzberg, et al. nents, in 23 countries, and in five surgical subspecialties. A questionnaire was prepared in which participants were asked how they would manage four clinical cases and they were also requested to answer questions concerning related surgical techniques in nerve repair. Recipients were given the option to complete the questionnaire in full or simply to indicate that they did not wish to participate. After 6 weeks, a second copy of the questionnaire was sent to all physicians who had failed to respond. In total, 49 surgeons (39%) returned completed questionnaires. The Questionnaire A detailed questionnaire was developed in two parts. The first section addressed general issues related to BPIs in which participants opinions were sought on how to diagnose nerve root avulsion (both pre- and intraoperatively) and their priorities for motor recovery in brachial plexus reconstruction. The anatomy of the brachial plexus is shown in Fig. 1. The focus of the second section was the management of four hypothetical clinical cases of BPI. These cases were designed to reflect common presentations and possible incongruities in surgical management. In each case, the participants were provided with a brief clinical vignette and a diagram of the brachial plexus indicating the area of injury. They were to assume that the patient s recovery had reached a plateau and no further change in function would develop. They were then asked whether they would offer surgery and, if so, at what time after injury, and what procedure(s) they would perform. The clinical vignettes and related diagrams were as follows. Case 1. This 5-month-old infant suffered a traumatic birth injury with panplexus involvement and resultant upper-extremity flail. The hand-related deficit recovered within 48 hours of delivery. Elbow flexion began at 2 months of age. At present (5 months of age), the patient exhibits 20 of shoulder abduction and 45 of elbow flexion against gravity. Supraspinatus and infraspinatus muscle strength was Grade 3/5 (Figs. 2 and 3). Case 2. This 14-year-old patient was involved in a motor vehicle accident. There is a panplexus loss of function with a flail extremity and a Horner syndrome. Imaging studies revealed pseudomeningoceles at the C5 T1 nerve roots (Fig. 4). Case 3. This 45-year-old patient suffered a motorcycle accident. Now 4 months postinjury there is no evidence of recovery in supraspinatus, infraspinatus, deltoid, or biceps muscle function. There is Grade 3/5 power in the triceps. Imaging of the spine demonstrated normal anatomy. Sensory nerve action potential is absent in the median nerve (Fig. 5 upper). Case 4. This 22-year-old patient suffered a motorcycle accident. Now 3 months postinjury there is good function in upper trunk distribution but no motor function distal to the elbow. A Horner syndrome is present. Imaging studies revealed pseudomeningoceles at C-7, C-8, and T-1. There is a normal SNAP in the ulnar nerve (Fig. 6 left). Definition of Terms The term donor nerve refers to a nerve in continuity with the central nervous system that is used in a nerve transfer to reinnervate the distal portion of another nerve (the target nerve). Proximal nerve transfers or graft repairs terminate in the nerve roots, trunks, divisions, or cords of the brachial plexus, whereas distal nerve transfers or graft repairs terminate in the peripheral branches of the brachial plexus (for example, the median nerve). RESULTS General Information Forty-nine surgeons participated in this study, yielding response rate of 39%. The surgical subspecialties, geo- Fig. 1. Drawing depicting the anatomy of the right brachial plexus. n = nerve. Fig. 2. Case 1. Drawing of the BPI demonstrating upper trunk neuroma-in-continuity. Insert demonstrates surgical repair with nerve grafts. 2 Neurosurg. Focus / Volume 16 / May, 2004

Surgical repair of brachial plexus injury Fig. 3. Flow chart indicating how respondents would manage the patient in Case 1. Other refers to partial resection of the neuroma in which certain elements are left intact. XI = accessory nerve. Neurosurg. Focus / Volume 16 / May, 2004 graphic regions, and annual experience in treating BPIs of the participating surgeons are summarized in Table 1. Most participants were trained in neurosurgery, orthopedics, or plastic surgery (39, 35, and 18%, respectively), and they performed a mean of 33 brachial plexus reconstructions annually. The participants approaches to the preoperative diagnosis of root avulsion are delineated in Table 2. The vast majority (94%) conducted some type of imaging study, either CT myelography or MR imaging, in addition to electrodiagnostic studies and physical examinations. The most frequently sought physical finding indicative of nerve root avulsion was the presence of a Horner syndrome. The techniques used for intraoperative diagnosis of nerve root avulsion are shown in Table 3. Respondents seemed to rely most heavily on both the absence of intraoperative SSEPs and direct inspection. The surgeons priorities for upper-extremity motor recovery focused mostly on elbow flexion and shoulder abduction, although many surgeons (Table 4) also targeted reinnervation of the wrist and hand (largely wrist extension and finger flexion). Hypothetical Cases Case 1. This case depicted a 5-month-old infant with a traumatic neuroma-in-continuity of the upper trunk secondary to a birth injury (Fig. 2). A summary of the respondents approaches is shown in Fig. 3. Of note, nearly one third (29%) of the surgeons would not undertake surgery in this case. Those choosing to operate were roughly evenly divided between those who would resect the neuroma and perform nerve graft repair, and those limiting the surgery to an external neurolysis of the neuroma. In both instances, any additional use of nerve transfers focused on reinnervation of the musculocutaneous nerve and/or the suprascapular nerve. The mean recommended age for surgical correction was 8.9 0.7 months (mean SEM). Case 2. In this case, a motor vehicle accident provided nerve root avulsions at C-5 through T-1 (Fig. 4). The donor nerves suggested for surgical interventions are sum- 3

A. J. Belzberg, et al. Fig. 4. Case 2. Drawing of the BPI demonstrating avulsion of the C5 T1 nerve roots. This illustrates one of the common repair strategies involving nerve transfers: spinal accessory nerve suprascapular nerve and intercostal nerves musculocutaneous nerve. marized in Table 5. The two most frequently suggested donor nerves for nerve transfer were the spinal accessory (32 surgeons) and the intercostal nerves (25 surgeons). Other frequently used donor nerves included the cervical plexus (12 surgeons), the phrenic nerve (10 surgeons), and the contralateral C-7 spinal nerve (eight surgeons). The suggested nerve transfer procedures are summarized in Table 6. Distal nerve transfers were the most common, including spinal accessory-to-suprascapular nerve, intercostal-to-musculocutaneous nerve, and intercostal-tomedian nerve (Fig. 4). Because of the stable condition noted during physical examination, the mean recommended time between injury and surgery was 2.4 0.3 months ( SEM). Case 3. The hypothetical patient in this case suffered complete loss of C-5 and C-6 function (supraspinatus, infraspinatus, deltoid, and biceps muscle function all Grade 0/5) and partial loss of C-7 (triceps muscle Grade 3/5). The management schemes are summarized in Fig. 5. The most common donor nerve for nerve transfer was the spinal accessory nerve (31 surgeons), and this was followed by the intercostal nerves (18 surgeons). Other less common donor nerves included the ipsilateral C-7, medial pectoral, phrenic, and ulnar nerves as well as elements of the cervical plexus (Table 7). As in Case 2, the majority of nerve transfers were to distal rather than proximal targets (86 and 14, respectively), with spinal accessory-to-suprascapular, intercostal-to-musculocutaneous and medial pectoral-to-musculocutaneous nerves being the most common (Table 8, Fig. 5 lower). Because of the patient s stable status noted on physical examination, the mean recommended time between injury and operation was 4.9 0.3 months ( SEM). Case 4. This hypothetical patient suffered a lower BPI with avulsion of the C-7, C-8, and T-1 nerve roots. The Fig. 5. Case 3. Upper: Drawing of the BPI demonstrating C- 5 and C-6 avulsion with a fibrotic upper trunk. This illustrates one possible repair strategy. Lower: A second repair strategy illustrating the most common nerve transfers: spinal accessory nerve suprascapular nerve, medial pectoral nerve musculocutaneous nerve, and phrenic nerve axillary nerve. surgeons responses are summarized in Fig. 6. As in Cases 2 and 3, the most common donor nerves for transfer were the spinal accessory and the intercostal nerves (13 surgeons each). In contrast to the previous cases, however, a less diverse repertoire of nerve transfers was recommended, and the majority of transfers were to proximal rather than distal targets (29 and 22, respectively), the most common being spinal accessory to-c-7/middle trunk (Tables 9 and 10). Because physical examination demonstrated a stable condition the mean time between injury and operation was 4 Neurosurg. Focus / Volume 16 / May, 2004

Surgical repair of brachial plexus injury Fig. 6. Left: Drawing of the BPI demonstrating C7 T1 avulsions with a fibrotic lower trunk. The most common nerve transfer procedure is shown: spinal accessory nerve C-7/middle trunk. Some would add intercostal sensory and motor nerve transfers for the hand. Right: A contralateral C-7 transfer repair. 5 0.9 months (mean SEM). Of note, only 30 of 49 respondents indicated that they would operate. Of those who would operate, only the responses of the 24 who indicated nerve repair as a primary procedure were included in the study. DISCUSSION The management of BPIs remains an enigma for most medical practitioners. Some physicians believe that there is rarely a need for surgery, assuming that those that do not eventually regenerate with functional recovery were not amenable to surgical repair. Unfortunately, because of this pessimistic attitude, often long after the injury patients TABLE 1 Summary of characteristics of peripheral nerve surgeons responding to the questionnaire Characteristic No. of Respondents/Procedures (%) total 49 op subspecialties neurosurgery 19 (39) orthopedic surgery 17 (35) plastic surgery 9 (18) other (hand, trauma, pediatric) 4 (8) geographic region North America 26 (53) Europe 15 (31) Asia 8 (16) BPI reconstructions/yr* mean SEM 33 5 median 20 range 2 150 * One respondent reported conducting 800 BPI reconstructions annually. Because this skewed the mean, this value was not included in the data analysis. present to peripheral nerve surgeons when they are no longer good candidates for primary nerve repair. Most unfortunate are those patients with severe BPIs who have been advised to undergo amputation, when, with rare exceptions, some form of nerve repair in conjunction with muscle transfer would potentially provide enough function to justify maintaining the integrity of the limb. In most patients, surgical nerve repair can provide a functional, useful limb, albeit one acting as a helper to the uninjured arm. Optimal repair of a traumatized brachial plexus differs substantially among even experienced surgeons, indicating lack of a definitive protocol. The purpose of this study was to determine the extent of variability among surgeons. Forty-nine surgeons participated in this study. It became clear during the study period that several physicians included in the original mailing were not surgeons and unlikely to respond to the questionnaire; however, 39% of physicians surveyed chose to respond. Participants per- Diagnostic Modality TABLE 2 Preoperative diagnostic choices in the treatment of an avulsion injury % of Surgeons Reporting the Modality imaging study 94 CT myelogram 80 MRI 55 both 41 electrodiagnostics 71 EMG 55 SNAP 43 SSEP 39 all 3 20 clinical diagnostics 59 Horner syndrome 29 absent Tinel sign 20 winged scapula 18 Neurosurg. Focus / Volume 16 / May, 2004 5

A. J. Belzberg, et al. TABLE 3 Modalities used for intraoperative diagnosis of a nerve root avulsion injury* Diagnostic Modality % of Surgeons Using Modality intraoperative electrodiagnostics 82 SSEP 76 NAP/NCNVs 27 EMR (direct stimulation) 27 MEP 6 SCEP 4 direct visualization 67 dissection of foramina 57 other 16 histological studies 16 * EMR = evoked muscle response; NAP/NCVs = nerve action potential/ nerve conduction velocities; SCEP = spinal cord evoked potential. Laminectomy, cerebral spinal fluid leak. Frozen section examination of dorsal root ganglion. formed a mean of 33 BPI repairs per year, suggesting experience and proficiency in this procedure. Review of the demographic data indicated both multispecialty and international participation. No attempt was made to differentiate clinical approach by specialty or geographic location; this will be the subject of a subsequent data analysis. Avulsion of nerve roots from the spinal cord is a particularly devastating injury. When brachial plexus trauma results in one or more nerve root avulsions, the prognosis for recovery is limited because spontaneous recovery of nerve function is exceedingly unlikely. Success in reimplantation of avulsed nerves has been limited; 7,8 when faced with root avulsion, the vast majority opted for nerve transfer repair. 7 Nerve transfer surgery has become the mainstay of treatment for this type of lesion. The ability to diagnose an avulsion, or to suspect it, preoperatively allows better surgical planning. For preoperative diagnosis of nerve root avulsion, 94% of the respondents would have performed either CT myelography or MR imaging, with 41% using both. Consistent with published accounts that CT myelography shows equal or greater sensitivity than plain myelography in detecting pseudomeningoceles, 38 no respondents reported the use of plain myelography alone. Some authors have reported that plain myelography can be more sensitive than CT myelography for detecting C8 T1 root avulsions where skeletal artifact can reduce the clarity of CT Motor Function TABLE 4 Priorities for motor recovery No. of Surgeons shoulder abduction 36 rotation 13 adduction 4 flexion 4 elbow flexion 44 extension 6 wrist extension 20 finger flexion 28 TABLE 5 Donor nerves suggested by 47 surgeons for operative intervention No. of Donor Nerve Surgeons (%) Target (frequency) spinal accessory 32 (68) suprascapular (29), musculocutaneous (2), axillary (1) intercostal 25* (53) musculocutaneous (27), median (11), axillary (3), radial (2), medial pectoral (2), medial antebrachial cutaneous (2), medial cord (2), lat cord (2), posterior cord (1), ulnar (1), long thoracic (1) cervical plexus 12 (26) axillary (3), median (2), C-7/middle trunk (2), musculocutaneous (1), suprascapular (1), radial (1), medial pectoral (1), upper trunk (1) phrenic 10 (21) musculocutaneous (4), upper trunk (3), suprascapular (2), axillary (1) contralat C-7 8 (17) median (5), musculocutaneous (1), lat cord (1), unspecified (1) 12th hypoglossal 2 (4) musculocutaneous (1), lat cord (1) * Intercostal nerves can be coapted to more than one target in a single reconstruction procedure. scans. 30,52 Although not pathognomonic of avulsion, the presence of a pseudomeningocele is indicative that an injury mechanism has occurred that might result in nerve root avulsion. In a recent study the authors noted that nerve root avulsions were better predicted by identifying the absence of rootlets in a pseudomeningocele. When this is observed on CT myelography it may indicate an extraforaminal root avulsion because of its high specificity and high likelihood ratio. 10 Although less popular than CT myelography preoperative MR imaging was also thought to be useful in diagnosing root avulsion (80 and 55%, respectively), consistent with reports 23 in which MR imaging was found to be useful in detecting both pre- and postganglionic nerve root injuries. 40 Several authors have demonstrated that CT myelography is more sensitive than MR imaging in detecting complete root avulsions, 9,37,55 including one study in which investigators prospectively compared CT myelography and MR imaging findings and confirmed these findings when undertaking hemilaminectomy. 9 Also of concern is a report that the ability of MR imaging to detect pseudomeningocele depends on the timing after injury. 52 In addition to diagnostic imaging, 71% of the respondents would use some form of electrodiagnostic study for preoperative diagnosis of root avulsion. Electromyography, SNAP, or SSEP were used with equal frequency; support for each practice can be found in the literature. Specifically, EMG of posterior cervical musculature (especially the deeper layers) has been used to determine the presence of a very proximal lesion consistent with root avulsion. 5 Both SSEP and SNAP have been studied, and it has generally been thought that the presence of an SNAP is a more reliable indication of preganglionic injury than is the absence of an SSEP. 24 Because a preganglionic lesion (including avulsion injury) leaves the sensory axons in continuity with the neuronal cell body in the dorsal root ganglion, the sensory axon in the peripheral nerve survives, and an action potential along the nerve is demonstrated on electrical testing. The presence of an SNAP in 6 Neurosurg. Focus / Volume 16 / May, 2004

Surgical repair of brachial plexus injury TABLE 6 Summary of the most commonly recommended nerve transfers for Case 2* TABLE 7 Summary of the most commonly recommended donor nerves (and their targets) for Case 3* Donor Nerve Grouped by Target (no.) Transfer Target No. of Donor Nerve Surgeons (%) Targets (no.) proximal contralat C-7 (1) hypoglossal (1) phrenic (3) cervical plexus (1) cervical plexus (2) intercostal (1) distal intercostal (27) phrenic (4) spinal accessory (2) cervical plexus (1) contralat C-7 (1) hypoglossal (1) spinal accessory (29) phrenic (2) cervical plexus (1) intercostal (11) contralat C-7 (5) cervical plexus (2) intercostal (3) cervical plexus (3) spinal accessory (1) cervical plexus (1) other (6) lat cord upper trunk medial trunk middle trunk posterior cord musculocutaneous suprascapular median axillary radial * There were 13 proximal and 105 distal recommended nerve transfers (one respondent did not specify the target for a contralateral C-7 transfer and thus this transfer was not recommended). Other transfers include: intercostal to medial pectoral (two cases); intercostal to long thoracic (one case); cervical plexus to medial pectoral (one case). One respondent did not specify the target for a contralateral C-7 transfer; hence that transfer is not included in this figure. spinal accessory 31 (70) suprascapular (24), musculocutaneous (4), upper trunk (2), lat cord (1) intercostal* 18 (41) musculocutaneous (15), median (4), axillary (4), medial cord (2), long thoracic (1) C-7 8 (18) C-7/middle trunk (5), musculocutaneous (3), suprascapular (1), axillary (1), upper trunk (1) medial pectoral 9 (20) musculocutaneous (8), axillary (1) phrenic 7 (16) musculocutaneous (4), suprascapular (1), axillary (1), upper trunk (1) cervical plexus 6 (14) axillary (3), long thoracic (2), suprascapular (1) ulnar 4 (9) musculocutaneous (4) contralat C-7 1 (2) axillary (1) other 4 (9) axillary (2), C-7/middle trunk (1), upper trunk (1), musculocutaneous (1) * The intercostal nerves can be coapted to more than one target in a single reconstruction. Other nerve transfers included: teres major motor to axillary (1); triceps branch of radial to axillary (1); C-8 to C-7/middle trunk (1); C-4 to upper trunk (1); medial radix of median to musculocutaneous (1). the presence of extensive muscle denervation demonstrated on EMG is strong evidence of a preganglionic lesion. In contrast, postganglionic lesions (intraplexal) separate the distal axon from the cell body, leading to Wallerian degeneration in the distal sensory axon and the absence of an SNAP. Our respondents generally agreed that elbow flexion and shoulder abduction are the two most important limb functions to restore (Table 4). Based on multiple sources in the literature it appears that others concur with this belief. 17,49 After elbow flexion and shoulder abduction, respondents ranked return of both wrist extension and finger flexion as goals. Although these more distal functions are harder to restore, many thought that prehension is important for the upper extremity to be functional and, in a small number of patients, can be achieved. 13,14 Notably, extremity-related protective sensation is crucial for its survival. An extremity that lacks sensation is likely to be subjected to repeated trauma and eventually succumb to infection. Interestingly, there was little discussion concerning external rotation of the shoulder. Many respondents may have inadvertently included external rotation with shoulder abduction. Alternatively, they may have been assuming that secondary surgery, including muscle and tendon transfers, would be used to address this. Based on these results, it may be possible to simulate a prospective randomized study by allowing each surgeon to perform his/her preferred technique and prospectively collect data for later comparisons. For valid comparisons, this approach would require a uniform method of pre- and postoperative evaluation of function in nerve-injured patients as well as standardized procedures to delineate the actual pathological entity. The development and implementation of valid instruments to be used in such a study is a formidable task, but is clearly necessary. Case 1 The survey results for Case 1 highlight the lack of consensus among peripheral nerve surgeons regarding how best to manage obstetrical plexus injuries. Given the incidence of spontaneous recovery (and assuming a plateau), nearly one third (29%) of the surgeons would not perform surgery in the patient. This is in agreement with authors who reserve surgical intervention for patients in whom biceps function has not been recovered by 3, 19,25 4 to 6, 26,46 5, 56 or 6 31 months of life. In addition, a few authors have suggested that infants in whom some neurological recovery occurs in all muscle groups within the 1st or 2nd month of life will recover without operative care. 20,32,35,48 Despite a biceps muscle grade of 3/5 strength, 71% of the respondents would perform surgery. The high rate of surgical intervention is likely due to the following: although the patient did exhibit recovery of some elbow flexion and shoulder abduction, both remained very limited even at 5 months of age. The case depicted the not un- Neurosurg. Focus / Volume 16 / May, 2004 7

A. J. Belzberg, et al. TABLE 8 Summary of the most common nerve transfers for Case 3* Donor Nerve Grouped by Target (no.) Transfer Target proximal spinal accessory (2) C-7 (1) upper trunk C-4 (1) C-7 (5) C-7/middle trunk C-8 (1) medial cord spinal accessory (1) lat cord distal intercostal (15) medial pectoral (8) phrenic (4) spinal accessory (4) musculocutaneous ulnar (4) C-7 (3) radix median (1) spinal accessory (24) suprascapular cervical plexus (1) C-7 (1) intercostal (4) cervical plexus (3) medial pectoral (1) axillary contralat C-7 (1) C-7 (1) teres major motor (1) triceps branch of radial (1) intercostal (4) median cervical plexus (2) long thoracic intercostal (1) * There were 14 proximal and 86 distal recommended nerve transfers. common presentation of Erb palsy (C5 7 nerve root injury) in patients who experience some early recovery in elbow flexion only to reach a plateau months later, when the level of recovery is less than acceptable. Based on the survey results, it appears that in addition to the onset of recovery, the subsequent temporal pattern and degree of recovery were equally critical in the decision-making process concerning treatment. In the absence of considerable elbow flexion and wrist extension by 4 to 6 months of age, many surgeons will undertake exploration of the brachial plexus. 12,16,18,56 Among the surgeons who decided to operate, there was disagreement concerning management of a neuroma-incontinuity. Opinion was nearly equally divided between those favoring resection of the neuroma (17 respondents) and those who did not (14 respondents). This division reflects results published in the available literature on the management of a neuroma-in-continuity, for which some have recommended resection and nerve grafting 6 and eschew internal neurolysis, whereas others maintained that neurolysis is beneficial, especially if intraoperative electrophysiological testing reveals significant conduction across the neuroma. 50 Surgeons electing not to resect the neuroma may also have been swayed by the amount of TABLE 9 Summary of the most commonly recommended donor nerves (and their targets) for Case 4* No. of Donor Nerve Surgeons (%) Targets (no.) intercostal 13 (54) median (6), ulnar (4), radial (2), medial cord (2), lower trunk (2) spinal accessory 13 (54) C-7/middle trunk (9), median (3), C-8 (2), axillary (1), radial (1) contralat C-7 4 (17) median (3), C-7/middle trunk (1), lower trunk (1) cervical plexus 3 (13) C-8 (3), C-7/middle trunk (2), T-1 (2) phrenic 2 (8) median (1), radial (1) other 4 (17) C-7/middle trunk (4), lower trunk (1), median (1), musculocutaneous (1) * Of 30 surgeons indicating that they would undertake surgery in this case, the 24 respondents represented in this table were those indicating that nerve repair would be the primary procedure. The intercostal nerves can be coapted to more than one target in a single reconstruction. Other nereve transfers included: dorsal scapular to C-7/middle trunk (two cases); C-7 to C-7/middle trunk (one case); C-6 to C-7/middle trunk (one case); C-5 to lower trunk (one case); C-7 to musculocutaneous (one case); C-7 to median (one case). partial C-5 and C-6 recovery and were unwilling to sacrifice the function already present. 4 Intraoperative electrical stimulation of the upper trunk proximal to the neuroma would certainly result in some elbow flexion, making it particularly uncomfortable for the surgeon now faced with dividing a functioning nerve. Furthermore, the child would possess less function postoperatively and would require several months to experience baseline or better function, a difficult situation for parents. An alternative management scheme, chosen by four of the respondents (see other in Fig. 3), requires internal neurolysis of the neuroma and identification (through stimulation) and preservation of functioning motor fascicles that traverse the neuroma, while dividing and reconstructing electrically silent and nonfunctioning sensory fascicles. 28 Although this is the most eloquent of the repairs in that functional fascicles are spared and damaged ones are grafted, it has not gained much popularity. It is technically demanding, there is potential for excessive scarring that limits regeneration, and iatrogenic damage to functioning fascicles is possible. In a case such as this, in which a patient exhibits partial recovery of biceps and deltoid function but minimal recovery of supra- and infraspinatus muscle function, consideration of the suprascapular nerve is required. If the origin of the nerve is distal to the neuroma, repair of the latter will provide innervation to the suprascapular nerve. If, however, the origin is part of the neuroma, specific repair of the suprascapular nerve is required. A graft can be placed from the C-5 or C-6 nerve root to the suprascapular nerve. Alternatively, the spinal accessory nerve can be transferred to the suprascapular nerve without disturbing the neuroma-in-continuity. Four respondents chose the latter repair. Case 2 Case 2 depicted the most devastating type of plexus 8 Neurosurg. Focus / Volume 16 / May, 2004

Surgical repair of brachial plexus injury TABLE 10 Summary of the most commonly recommended nerve transfers for Case 4* Donor Nerve Grouped by Target (no.) Neurosurg. Focus / Volume 16 / May, 2004 Transfer Target proximal spinal accessory (9) cervical plexus (2) dorsal scapular (2) C-7/middle trunk C-7 (1) contralat C-7 (1) C-6 (1) cervical plexus (3) C-8 spinal accessory (2) C-5 (1) lower trunk contralat C-7 (1) cervical plexus (2) T-1 medial cord distal intercostal (6) spinal accessory (3) contralat C-7 (3) median C-7 (1) spinal accessory (1) radial intercostal (4) ulnar spinal accessory (1) axillary C-7 (1) musculocutaneous * There were 29 proximal and 24 distal nerve transfers recommended. injury, one of complete avulsion with panplexus loss of function. Because of the poor prognosis for any spontaneous recovery, nearly all (96%) of the respondents elected to operate. Most respondents chose to undertake relatively early surgical treatment with a mean observation period of 2.4 months. Of the four cases, the observation period in this one was the shortest consistent with a variety of recommendations in the literature, which propose an observation period ranging from 2 weeks 4,34 to less than 5 or 6 months. 37,42,49 Clearly, the timing of surgery relates to the confidence of the surgeon in determining the presence of complete avulsion. Clear evidence of complete brachial plexus avulsion indicates that there will be no spontaneous recovery and that surgical repair is indicated as soon as possible. Although the authors of older reports considered amputation an option in such a case, 1,29,39,41 all of the respondents proposed procedures were consistent with the contemporary approach to peripheral nervous system repair. Nerve transfer based reinnervation was the method of choice. Consistent with the priorities for motor recovery, the two most commonly recommended nerve transfer procedures were designed to restore elbow flexion (intercostal nerve-to-musculocutaneous nerve transfer) and shoulder abduction (spinal accessory nerve-to-suprascapular nerve transfer). The spinal accessory nerve suprascapular nerve procedure has been widely reported; 1,33,36 however, some authors have maintained that better abduction is achieved when performing axillary nerve transfer. 15,44 An advantage of the suprascapular nerve transfer is that some shoulder external rotation may be regained with reinnervation of the infraspinatus muscle. Successful reinnervation of the musculocutaneous nerve with intercostal nerves has been reported, 34,44 and some authors have proposed that it is the best of all extraplexal transfers. 49 Some surgeons, however, have maintained that spinal accessory nerve transfer is a better source for reinnervation of the musculocutaneous nerve in terms of motor recovery, whereas the intercostal transfer achieves better pain relief. 54 Use of the intercostal nerves also provides protective sensory reinnervation of the extremity, an advantage over use of the spinal accessory nerve. For this to occur, both the sensory and motor portions of the intercostal nerve have to be transferred. The sensory nerves are directed at the lateral antebrachial cutaneous nerve, the distal termination of the musculocutaneous nerve. An alternative approach is to direct the intercostal sensory nerves to the lateral cord component of the median nerve (sensory C-5 and C-6). Reinnervation of the median nerve was also considered an option. In addition to the intercostal nerves, other reported nerve transfers to the median nerve include the cervical plexus 43 and the contralateral C-7 nerve root. 21,47,53 The phrenic nerve has also been used as a donor, most frequently transferred to the musculocutaneous nerve. This is consistent with evidence that this transfer procedure successfully produces elbow flexion without causing respiratory compromise. 22 This nerve transfer has been shown to stabilize the shoulder and prevent subluxation, but it has been unimpressive in restoring shoulder abduction. 11 Case 3 Case 3 depicts a patient with traumatic loss of upper brachial plexus function in which there was no evidence of recovery at 4 months. Nearly all (90%) of the surgeons elected to operate. The observation period, however, was longer than in previous cases, averaging approximately 5 months. This likely reflects the presence of a partial injury and the hope for further spontaneous recovery. As with the previous cases, regaining elbow flexion and shoulder abduction remained the chief priorities. These were largely achieved by performing intercostal nerve musculocutaneous nerve and spinal accessory nerve suprascapular nerve transfers, respectively. Unlike the patient in Case 2, some intraplexal nerves were available as donor nerves in this patient, and it has been suggested that intraplexal donor nerves are superior sources for motor neurotization. 49 The respondents would have used both the medial pectoral and ipsilateral C-7 nerves in their transfer procedures. Nevertheless, even though these and other intraplexal donor nerves were available, the most common donors were still the spinal accessory and intercostal nerves (both extraplexal). This may reflect the belief that, for these particular extraplexal donor nerves, the results are at least as good as those obtained with intraplexal donors. 34,44,49 Impressive results have been reported when performing the medial pectoral nerve transfer to the musculocutaneous nerve. 43 Recently, several authors have conducted the Oberlin procedure, 27,51 which involves a fascicle of the 9

A. J. Belzberg, et al. ulnar nerve to the musculocutaneous nerve. Because of the short distances, reinnervation of the muscle occurs quickly, thus limiting the degree of denervation muscular atrophy. Case 4 The patient in Case 4 suffered a lower plexus injury with avulsion of the C-7, C-8, and T-1 nerve roots, as well as injury of the lower trunk and fibrosis. Because elbow flexion and shoulder abduction were intact, but motor function distal to the elbow was absent, the respondents sought mainly to restore finger flexion and wrist extension. This is consistent with the ranking of motor priorities listed in Table 4. Compared with the previous cases, considerably fewer surgeons would undertake surgery in this case because of the poor prognosis associated with surgical intervention. Overall, the nerve transfers used in Case 4 differ in important ways from those conducted in Case 3. First, in the former there was a greater reliance on proximal transfers (those terminating within the brachial plexus). The reason for this is unclear and although advocated by some physicians, 45 others have argued that the site of distal coaptation should be closer to the target muscle. 2,49 Second, in Case 4 more respondents used the contralateral C-7 nerve root as a source for neurotization. Consistent with reports advocating its transfer directly into the median nerve, 2,49 three of the five proponents of the contralateral C-7 root transfer chose a distal transfer to the median nerve over a proximal intraplexal site of coaptation (transfer to C-7/middle trunk or lower trunk). Respondents made less use of intraplexal donor sites despite reports that intraplexal donor nerves often yield superior results, 49 probably because they considered elbow flexion and shoulder abduction the highest priorities and were unwilling to risk upper plexus function in the hope of restoring more distal function. Third, we did not allow the respondents to consider use of tendon or muscle transfers, a technique that would have great potential in this particular case. CONCLUSIONS In this study a multinational and -disciplinary group of experienced peripheral nerve surgeons were asked how they would manage four patients who had suffered BPIs. The physicians decisions underscore the many areas of disagreement regarding the treatment of brachial plexus injury including the following; the diagnostic approach to defining the injury, timing of and indications for surgical intervention in birth-related palsy, management of neuroma-in-continuity, the choice of nerve transfers to achieve elbow flexion and shoulder abduction, the use of intra- or extraplexal donors for neurotization, and use of distal or proximal coaptation during the nerve transfer surgery. These areas of disagreement are likely to persist until prospective randomized studies are conducted to determine the relative merit of each competing form of treatment. Acknowledgment We thank Dr. James Campbell for his input on study design. References 1. Allieu Y, Cenac P: Neurotization via the spinal accessory nerve in complete paralysis due to multiple avulsion injuries of the brachial plexus. Clin Orthop 237:67 74, 1988 2. Alnot JY: Traumatic brachial plexus lesions in the adult: indications and results. Microsurgery 16:22 29, 1995 3. Bentolila V, Nizard R, Bizot P, et al: Complete traumatic brachial plexus palsy. Treatment and outcome after repair. J Bone Joint Surg 81:20 28, 1999 4. Boome RS: General discussions on the brachial plexus, in Boome RS (ed): The Brachial Plexus. Hand and Upper Extremity, Vol 14. London: Churchill Livingstone, 1997, pp 1 7 5. Bufalini C, Pescatori G: Posterior cervical electromyography in the diagnosis and prognosis of brachial plexus injuries. J Bone Joint Surg Br 51:627 631, 1969 6. Capek L, Clarke HM, Curtis CG: Neuroma-in-continuity resection: early outcome in obstetrical brachial plexus palsy. Plast Reconstr Surg 102:1555 1564, 1998 7. Carlstedt T, Anand P, Hallin R, et al: Spinal nerve root repair and reimplantation of avulsed ventral roots into the spinal cord after brachial plexus injury. J Neurosurg (Spine 2) 93: 237 247, 2000 8. Carlstedt T, Grane P, Hallin RG, et al: Return of function after spinal cord implantation of avulsed spinal nerve roots. Lancet 346:1323 1325, 1995 9. Carvalho GA, Nikkhah G, Matthies C, et al: Diagnosis of root avulsions in traumatic brachial plexus injuries: value of computerized tomography myelography and magnetic resonance imaging. J Neurosurg 86:69 76, 1997 10. Chow BC, Blaser S, Clarke HM: Predictive value of computed tomographic myelography in obstetrical brachial plexus palsy. Plast Reconstr Surg 106:971 979, 2000 11. Chuang DC, Lee GW, Hashem F, et al: Restoration of shoulder abduction by nerve transfer in avulsed brachial plexus injury: evaluation of 99 patients with various nerve transfers. Plast Reconstr Surg 96:122 128, 1995 12. Clarke HM, Al-Qattan MM, Curtis CG, et al: Obstetrical brachial plexus palsy: results following neurolysis of conducting neuromas-in-continuity. Plast Reconstr Surg 97:974 984, 1996 13. Doi K, Kuwata N, Muramatsu K, et al: Double muscle transfer for upper extremity reconstruction following complete avulsion of the brachial plexus. Hand Clin 15:757 767, 1999 14. Doi K, Sakai K, Kuwata N, et al: Double free-muscle transfer to restore prehension following complete brachial plexus avulsion. J Hand Surg Am 20:408 414, 1995 15. Dolenc VV: Contemporary treatment of peripheral nerve and brachial plexus lesions. Neurosurg Rev 9:149 156, 1986 16. Duclos L, Gilbert A: Obstetrical palsy: early treatment and secondary procedures. Ann Acad Med Singapore 24:841 845, 1995 17. Friedman AH: Neurotization of elements of the brachial plexus. Neurosurg Clin N Am 2:165 174, 1991 18. Gilbert A, Brockman R, Carlioz H: Surgical treatment of brachial plexus birth palsy. Clin Orthop 264:39 47, 1991 19. Gilbert A, Whitaker I: Obstetrical brachial plexus lesions. J Hand Surg Br 16:489 491, 1991 20. Goddard NJ, Fixsen JA: Rotation osteotomy of the humerus for birth injuries of the brachial plexus. J Bone Joint Surg Br 66: 257 259, 1984 21. Gu YD, Chen DS, Zhang GM, et al: Long-term functional results of contralateral C7 transfer. J Reconstr Microsurg 14: 57 59, 1998 22. Gu YD, Ma MK: Use of the phrenic nerve for brachial plexus reconstruction. Clin Orthop 323:119 121, 1996 23. Hems TE, Birch R, Carlstedt T: The role of magnetic resonance imaging in the management of traction injuries to the adult brachial plexus. J Hand Surg Br 24:550 555, 1999 10 Neurosurg. Focus / Volume 16 / May, 2004

Surgical repair of brachial plexus injury 24. Jones S: Diagnostic value of peripheral and spinal somatosensory evoked potentials in tractions injuries of the brachial plexus, in Terzis JK (ed): Microreconstruction of Nerve Injuries. Philadelphia: WB Saunders, 1987, pp 463 471 25. Kawabata H, Masada K, Tsuyuguchi Y, et al: Early microsurgical reconstruction in birth palsy. Clin Orthop 215:233 242, 1987 26. Laurent JP, Lee R, Shenaq S, et al: Neurosurgical correction of upper brachial plexus birth injuries. J Neurosurg 79:197 203, 1993 27. Loy S, Bhatia A, Asfazadourian H, et al: [Ulnar nerve fascicle transfer onto to the biceps muscle nerve in C5-C6 or C5-C6-C7 avulsions of the brachial plexus. Eighteen cases.] Ann Chir Main Memb Super 16:275 284, 1997 (Fre) 28. Mackinnon SE, Glickman LT, Dagum A: A technique for the treatment of neuroma in-continuity. J Reconstr Microsurg 8: 379 383, 1992 29. Malone JM, Leal JM, Underwood J, et al: Brachial plexus injury management through upper extremity amputation with immediate postoperative prostheses. Arch Phys Med Rehabil 63: 89 91, 1982 30. Marshall RW, De Silva RD: Computerised axial tomography in traction injuries of the brachial plexus. J Bone Joint Surg Br 68:734 738, 1986 31. Meyer RD: Treatment of adult and obstetrical brachial plexus injuries. Orthopedics 9:899 903, 1986 32. Michelow BJ, Clarke HM, Curtis CG, et al: The natural history of obstetrical brachial plexus palsy. Plast Reconstr Surg 93: 675 681, 1994 33. Millesi H: Brachial Plexus Injuries: Management and Results, in Terzis JK (ed): Microreconstruction of Nerve Injuries. Philadelphia: WB Saunders, 1987, pp 347 360 34. Nagano A: Treatment of brachial plexus injury. J Orthop Sci 3:71 80, 1998 35. Narakas AO: Obstetrical brachial plexus injuries, in Lamb DW (ed): The Paralyzed Hand. Edinburgh: Churchill Livingstone, 1987, pp 116 135 36. Narakas AO, Hentz VR: Neurotization in brachial plexus injuries. Indication and results. Clin Orthop 237:43 56, 1988 37. Penkert G, Carvalho GA, Nikkhah G, et al: Diagnosis and surgery of brachial plexus injuries. J Reconstr Microsurg 15:3 8, 1999 38. Petras AF, Sobel DF, Mani JR, et al: CT myelography in cervical nerve root avulsion. J Comput Assist Tomogr 9:275 279, 1985 39. Ransford AO, Hughes SP: Complete brachial plexus lesions: a ten-year follow-up of twenty cases. J Bone Joint Surg Br 59- B:417 420, 1977 40. Rapoport S, Blair DN, McCarthy SM, et al: Brachial plexus: correlation of MR imaging with CT and pathologic findings. Radiology 167:161 165, 1988 41. Rorabeck CH: The management of the flail upper extremity in brachial plexus injuries. J Trauma 20:491 493, 1980 42. Ruch DS, Friedman A, Nunley JA: The restoration of elbow flexion with intercostal nerve transfers. Clin Orthop 314: 95 103, 1995 43. Rutowski R: Neurotizations by means of the cervical plexus in over 100 patients with from one to five root avulsions of the brachial plexus. Microsurgery 14:285 288, 1993 44. Samardzic M, Rasulic L, Grujicic D, et al: Results of nerve transfers to the musculocutaneous and axillary nerves. Neurosurgery 46:93 103, 2000 45. Sedel L: The results of surgical repair of brachial plexus injuries. J Bone Joint Surg Br 64:54 66, 1982 46. Sherburn EW, Kaplan SS, Kaufman BA, et al: Outcome of surgically treated birth-related brachial plexus injuries in twenty cases. Pediatr Neurosurg 27:19 27, 1997 47. Songcharoen P, Wongtrakul S, Mahaisavariya B, et al: Hemicontralateral C7 transfer to median nerve in the treatment of root avulsion brachial plexus injury. J Hand Surg Am 26: 1058 1064, 2001 48. Tada K, Tsuyuguchi Y, Kawai H: Birth palsy: natural recovery course and combined root avulsion. J Pediatr Orthop 4: 279 284, 1984 49. Terzis JK, Vekris MD, Soucacos PN: Outcomes of brachial plexus reconstruction in 204 patients with devastating paralysis. Plast Reconstr Surg 104:1221 1240, 1999 50. Tiel RL, Happel LT Jr, Kline DG: Nerve action potential recording method and equipment. Neurosurgery 39:103 109, 1996 51. Tung TH, Novak CB, Mackinnon SE: Nerve transfers to the biceps and brachialis branches to improve elbow flexion strength after brachial plexus injuries. J Neurosurg 98: 313 318, 2003 52. Volle E, Assheuer J, Hedde JP, et al: Radicular avulsion resulting from spinal injury: assessment of diagnostic modalities. Neuroradiology 34:235 240, 1992 53. Waikakul S, Orapin S, Vanadurongwan V: Clinical results of contralateral C7 root neurotization to the median nerve in brachial plexus injuries with total root avulsions. J Hand Surg Br 24:556 560, 1999 54. Waikakul S, Wongtragul S, Vanadurongwan V: Restoration of elbow flexion in brachial plexus avulsion injury: comparing spinal accessory nerve transfer with intercostal nerve transfer. J Hand Surg Am 24:571 577, 1999 55. Walker AT, Chaloupka JC, de Lotbiniere AC, et al: Detection of nerve rootlet avulsion on CT myelography in patients with birth palsy and brachial plexus injury after trauma. AJR 167: 1283 1287, 1996 56. Waters PM: Comparison of the natural history, the outcome of microsurgical repair, and the outcome of operative reconstruction in brachial plexus birth palsy. J Bone Joint Surg Am 81: 649 659, 1999 Manuscript received March 25, 2004. Accepted in final form April 2, 2004. Address reprint requests to: Allan Belzberg M.D., Department of Neurological Surgery, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, Maryland 21287-7509. email: belzberg@jhu.edu. Neurosurg. Focus / Volume 16 / May, 2004 11