Magnification, use of fine interfascicular grafts for repair, and development of intraoperative electrophysiologic measurements of function have had a substantial impact on this field in the last 10 to 20 years. Basic surgical principles established during and since World War II remain the foundation for surgical repair of peripheral nerve injury but have been complemented nicely by these more recent advances. Selection of patients for surgery, as well as the timing of such, has been reviewed with emphasis on the differences between suspected transections and lesions in continuity, as well as comments on serious peripheral entrapments and tumors affecting nerve. The importance of not only preoperative electromyographic studies but also the intraoperative use of stimulation and stimulation and recording of nerve action potentials (NAPs) for lesions in continuity has been stressed. Operative techniques such as neurolysis, NAP recordings, suture, split repair, and interfascicular graft repair have been reviewed and some commentary on results provided. There has been a gradual evolution of centers in this country and abroad for care of the more serious surgical nerve problems. It is anticipated that in the future, such centers will be able to provide improved data concerning results with civilian nerve injuries. Key words: peripheral nerve repair suture grafts NAPs EMG Neurolysis MUSCLE & NERVE 13S43-852 1990
SURGICAL REPAIR OF PERIPHERAL NERVE INJURY DAVID G. KLINE, MD, FACS
During the last quarter of a century, three significant changes in the surgical management of nerve injuries have occurred. Magnification by either an operating microscope or use of eye loupes has become commonplace as has the use of the bipolar coagulator, some microinstruments, and relatively fine suture mate ria^.^"'*^^ Of even greater importance has been the re-emergence of graft repair as a practical method to regain innervation.23’24*28 This technique permits optimal resection of the injury site(s) back to healthy tissue and bridges neural gaps with minimal tension at the repair sites. During this same period of time, electrophysiologic methods for intraoperative assessment of the most frequent and serious injury
From the Department of Neurosurgery, Louisiana State University Medical Center, New Orleans, Louisiana. Presented at AAEE (now AAEM) Annual Meeting, International Symposium on Peripheral Nerve Regeneration, Washington, DC, September 17, 1989. Address reprint requests to Dr. Kline at the Department of Neurosurgery, Louisiana State University Medical Center, 1542 Tulane Avenue, New Orleans, LA 701 12-2822. Accepted for publication February 16, 1990 CCC 0148-639W90/090843-010 504.00 0 1990 John Wiley & Sons, Inc
Surgical Repair of Nerve Injury
to nerve, one leaving it in continuity, have evolved. 12~15,33These techniques permit a physiologic decision for or against resection so that a lesion “looking good” but having poor regeneration is not left unresected or conversely, one “appearing bad” but having good regeneration is not falsely resected. These advances have most likely led to improved functional results at least in some surgeon’s hands. More certain is that these operative changes have increased the incidence of surgical explorations and attempted repairs of lesions previously felt not to be repairable particularly those of the brachial plexus and the proximal arm and leg. Other shifts in management, although less dramatic, have also occurred. Where direct neural repair has been impossible or has failed, substitutive procedures have been further refined, in some situations newly devised, and in most centers increasingly used to restore function. Older recognized issues such as the importance of timing for surgery, the role of preoperative electromyographic evaluation, and the relatively aggressive use of physical and occupational therapy have received renewed e r n p h a ~ i s . ’ ~ ~ ~Related “ ~ ” ’ ~to~ ~ ~ ~ ~ ~ all of the above as well as to trends in medicine as a whole has been an increase in the “center-ori-
MUSCLE & NERVE
September 1990
843
ented” management of at least the more complex and difficult nerve lesions. SELECTION AND TIMING FOR SURGERY
The two major reasons for operating on a nerve injury are to improve motor and sensory function and to ameliorate neural pain. There are a few settings where partial loss following an injury can progress and, of course, surgery and usually urgent surgery is necessary for these. Progressive lesions are associated with vascular injury and are due to pseudoaneurysm, arteriovenous malformation, or clot in a potentially tight space.26 Mechanism is usually a penetrating injury due to gunshot wound or knife but can on occasion be due to contusive forces. Acute decompression with removal of the vascular lesion and neurolysis of involved nerves is necessary. Causalgia, which is usually associated with penetrating injury to proximal nerve(s) may require acute sympathectomy. Less frequently, a missile fragment or foreign body may lodge directly in nerve and may require acute removal to control pain. Severe soft tissue contusion and associated vascular compromise can result in a Volkmann’s in the forearm or an anterior compartment syndrome in the leg. Acute fasciotomy as well as exposure of the nerves is necessary. ACUTE AND SUBACUTE REPAIR OF LACERATIONS
The most frequent reason for relatively acute operation is to repair nerve injury associated with extremity laceration. Such injuries are due to glass, knives, or sometimes metal edges, and transect nerve in a relatively sharp fashion. In this setting, there may be an advantage to acute suture since tissue to be trimmed from each stump is minimal, associated anatomy is not greatly distorted, and nerve can be repaired at the same time as other soft tissue injuries. Such primary or acute repair is especially important for lacerating injuries to plexus or proximal sciatic nerve where the scar and stump retraction associated with delay makes secondary repair more difficult. Results in most transection cases well selected for acute repair are at least as good as in those re aired later and in some cases are even better. 18,30!32 By comparison, blunt transections are best repaired after a delay of several weeks. Mechanisms include laceration due to propellar blades, power saws, auto metal, and some complex fractures. Damage to each stump associated with blunt mechanisms is difficult to delineate acutely. End-
844
Surgical Repair of Nerve Injury
to-end repair of potentially damaged stumps leads inevitably to a poor result. With delay, the amount of neuroma and glioma to be trimmed is quite evident. Injuries lacerating soft tissue and associated with distal neural deficit do not always transect nerve completely and in some cases only contuse nerve. If the nerve is partially divided in a sharp fashion, then acute repair of the divided portion is in order. On the other hand, if it is known that the partial division is blunt or suspected that contusion rather than neural laceration is responsible for deficit, delay in repair is in order. In the interim, there may occur significant clinical or electromyographic (EMG) reversal of loss. If not, then the lesion can be evaluated by intraoperative stimulation and recording of nerve action potentials (NAPS). These electrophysiologic tests along with magnification are used to determine whether a split or total repair or more simply, only a neurolysis is necessary. LESIONS I N CONTINUITY
T h e majority (60 to 70%) of serious injuries to nerve are due to blunt mechanisms such as stretch and contusion and leave the nerve in continuity. Since nerve is not divided, there is a variable degree of intraneural injury. Thus, there may be a mixture of axonotmesis and neurotmesis and even an element of neurapraxia. Other lesions will be totally neurotmetic despite maintenance of gross continuity. Swelling or neuroma formation at the injury site is variable and does not necessarily correlate with the severity of intraneural axonal injury, let alone to the degree of distortion and disruption of the connective tissue framework of the As a result, some of these lesions will improve spontaneously with time and without operation but unfortunately many will not. The most important consideration is whether or not deficit distal to such a lesion is complete or incomplete by clinical as well as electromyographic criteria. With incomplete lesions with useful sparing of distal function, further function will usually result in time. Thus, with nonoperative management including good physical therapy, function at least as good and usually better than that achieved by surgery occurs. When milder degrees of injury occur to the entire cross-section of the nerve, findings include diffuse muscular weakness rather than paralysis, and partial rather than complete sensory loss even in autonomous Function almost always improves with time unless injury site
MUSCLE & NERVE
September 1990
is near that for potential neural entrapment such as the olecranon notch, the carpal tunnel, o r the head of fibula. Of course, associated and especially expanding, lesions such as aneurysm, fistula, or clot can prevent recovery unless they are corrected. A complete lesion in continuity which shows no significant clinical or electrical recovery for 2 to 4 months postinjury is usually a candidate for surgical exploration and intraoperative electrophysiologic assessment. (Fig. 1). How long to delay surgical exploration remains somewhat controversial even on a case-by-case basis. If, however, the clinician waits the many months it may take to be certain that there is no clinical nor EMG evidence of functional reinnervation coming about, then it may be too late for such greatly delayed repair to be effective. This is because innervational input sites, particularly for muscle, will shrink and atrophy. Once such involution has occurred, bringing even thousands of axons back to a potential innervational site is like “sowing good seed on fallow ground.” ELECTRICAL STUDIES FOR LESIONS IN CONTINUITY
For effective electrodiagnostic work-up, the normal innervational pattern of the involved nerve has to be known as well as the time it takes for regeneration to site(s)just distal to that of the injury. Most incontinuity lesions have an area of retrograde axonal loss. This distance as well as the length of the injury site has to be traversed before regenerating axons reach distal stump, and then eventually reach even more distal reinnervational points that can be tested by noninvasive electrical
techniques. Even for an axonotmetic lesion, this regrowth process takes months. For example, with a complete midhumeral level injury to radial nerve, it would be 3 to 4 months before one could expect any significant reversal of deinnervational change in brachioradialis, let alone voluntary contraction of that muscle. With a buttock or thigh level sciatic nerve injury, regenerating axons may not reach even proximal calf or leg muscles for a year to a year and a half. Reinnervational delays after many injuries to brachial plexus elements, especially those at a spinal root or trunk level, can be very great because distances to be made up are great. Axons must not only reach muscle but make some reconstruction of axon to motor end plate to muscle fiber junctions to reverse electromyographic changes such as fibrillations and to show nascent potential^.^ A significant reduction in number of fibrillations and deinnervational potentials occur several weeks before nascent potentials appear. There is further delay before muscle action potentials can be evoked. Thus, even after axons reach muscle there is a further delay (weeks to a month or more) before signs of electrical reinnervation occur. Nonetheless, some reversal of deinnervational change does occur weeks before there is objective contraction of muscle. It has to be remembered that the EMG is a physiologic test. It tests the electrical concomitants of deinnervation andlor reinnervation. Thus, reversal of deinnervational and occurrence of reinnervational changes suggests that, with time, useful function will occur but does not guarantee such. In addition, the EMG is a sampling technique and what is occurring in one area of a muscle may not be happening elsewhere. Thus, a
FIGURE 1. Stimulation and recording from a lesion in continuity involving the proximal median nerve and explored 5 months after the injury. NAPS were recorded from this injury due to contusion and stretch so only a neurolysis was done.
Surgical Repair of Nerve Injury
MUSCLE & NERVE
September 1990
845
thorough sample as well as a discriminating appraisal of electrical activity recorded is a must. Despite these points, EMG in well-trained hands remains a very valuable method for evaluating lesions in continuity even in the early months postinjury. Contraction of muscle distal to the injury in response to neural stimulation is also a valuable harbinger of clinical return. Such a response in an adequately regenerating nerve may antedate voluntary muscle contraction by weeks to several months. 18.2635 Again, for stimulation even directly to the nerve to demonstrate early recovery, axons must have not only reached muscle but there must be some degree of motor end-plate reconstruction. When axons are only in distal stump and have not as yet reached muscle, neither EMG nor stimulation will show evidence of recovery and yet such early regeneration may have great potential for functional reinnervation. Thus, in order to assess relatively early regeneration in a lesion in continuity, it may be necessary to surgically expose the nerve and to do direct, intraoperative NAP as well as stimulation studies. TIMING FOR SURGERY ON LESIONS IN CONTINUITY
The more focal lesions in continuity are associated with fracture, contusion to soft tissue and some gunshot These can be reliably evaluated by intraoperative electrical studies at 2 to 3 months postinjury. Many injuries due to stretch or severe contusion produce lengthier lesions and are associated with vehicular and skiing accidents, falls, crush injuries, and shotgun injuries. These stretches require months of regeneration before a large number of axons can traverse the injury segment and reach distal stump. As a result, these cases are followed for 3 to 5 months.” If deficit remains complete in the distribution of one or more nerves or plexus elements, an exploration is done. Definite signs of spontaneous reversal of loss take longer to appear in this group of patients than in any other. In addition, when early clinical or EMG recovery does not occur, one wants to be in a good position at exploration to record from the stretch segment and distal nerve and to be certain whether or not adequate regeneration is occurring. This is necessary so that the correct decision can be made for or against resection and repair. In the usual stretch injuries to plexus selected for surgery, the majority of involved roots or elements are left in continuity. Some have the possibility for spontaneous and significant re-
846
Surgical Repair of Nerve Injury
growth while many do not.8217By 3 to 5 months, the relative regenerative capacity of the various elements can be accurately evaluated by recording across the stretched segments. Many but not all civilian gunshot wounds involving plexus require operation. Since most of these gunshot wounds are more focal than the majority of stretch injuries, intraoperative assessment is accurate at 2 to 3 months postinjury. l 1 Delay in referral, healing of associated injuries, and management of infection may of course alter this recommended injury to operative interval. Aside from these factors, when one is in doubt about the direction of recovery it is better to look and assess the nerve directly, for undue delay in suture repair leads to poor results.”’* (Table I). TIMING FOR ENTRAPMENTS AND TUMORS
These spontaneous conditions are usually not associated with injury but their course can certainly
Table 1. Management of the neuroma in continuity. Incomplete loss with significant distal sparing 1. Most cases will improve with conservative treatment. They are followed by serial clinical and EMG examinations. Physical therapy is important. 2. Operation may still be required: Partial lesions associated with expanding masses due to hematoma, aneurysm, or A-V fistula usually require urgent operation Partial lesions close to or in areas of potential entrapment may require relatively early operation Lesions where distal loss, although partial, is significant may require later operation Neural pain not amendable to medications and physical therapy may require later operation Complete or near complete loss with little or no distal sparing 1. Relatively focal lesions in continuity due to fracture or gunshot wound. a. Follow by clinical and EMG examinations for 2-3 months. b. If no significant clinical or electrical improvement, explore. c. lntraoperative stimulation and NAP studies used to decide for or against resection. 2. Relatively lengthy lesions in continuity due to contusionlstretch or shotgun. a. Follow by clinical and EMG examinations for 4-5 months. b. If no significant clinical or electrical improvement, explore. c. If no response to stimulation and no NAP across lesion, resection and repair by suture or graft is necessary. d. lntraoperative evoked cortical or somatosensory studies may be necessary to evaluate repairability of proximal spinal nerves. (See Table 3).
MUSCLE & NERVE
September 1990
be complicated by such. Uncomplicated entrapment and tumor associated neural loss tends to be mild. Many or almost all axons may have peripheral connections. EMG signs of deinnervation may thus be mild or absent altogether. On the other hand relative conductive loss or slowing is a cardinal feature especially with entrapments. This, of course, make the risk of further loss of function with operation greater than with a completely injured nerve where loss is already maximal or near maximal. If the entrapped nerve is sufficiently symptomatic in terms of pain and/or loss of function to be bothersome to the patient, then operation without undue delay is certainly indicated. It is seldom that the indication is emergent or even urgent.' Exceptions are occasionally provided by acute carpal tunnel syndrome, and less frequently acute compression of ulnar at elbow level associated with pregnancy or the immediate postpartum period. Almost always a clinically significant median or ulnar entrapment will show electrical abnormalities both by noninvasive conductive studies and, if one goes to the trouble to do so, by direct recording from nerve intraoperatively. The latter is interesting to do during such operations but not as essential as it is in more serious in-continuity lesions where the question of resection is present. If an entrapped nerve with significant clinical and electrical findings is to be followed along without operation, then this should be done very closely and very carefully. Since the operations involved, especially for carpal tunnel, are of relatively low risk, are relatively straightforward and also have a good success rate, the better part of valor is usually to decompress the nerve and where indicated to transpose it. A number of posterior interosseous nerve (PIN) palsies will reverse by 2 to 3 months after onset, and thus, a conservative period of management is certainly in order for these. By sharp contrast, loss associated with peroneal entrapment at the head of the fibula, even when unassociated with contusion, does not usually reverse. If peroneal decompression is delayed too long, reversal even of partial loss may not occur. Results are better with median than ulnar, and better with ulnar than peroneal entrapments, while those with PIN entrapment are excellent.""' Tumors arising in nerves are usually but not always benign. Thus, there is a tendency to follow them if they are not too symptomatic.' At stake with an operation is a variable degree of loss associated with tumor removal. Neural loss is a necessity if the lesion is a malignant intrinsic tumor,
Surgical Repair of Nerve Injury
since wide resection and/or amputation of the limb well proximal to the lesion is usually recommended. Loss may also occur with removal of benign large tumors especially those in the brachial or pelvic plexus, particularly if they are neurofibromas. Progressive loss and pain remain indications for operation but in addition, if a lesion is greater than a few centimeters in diameter, even in a less symptomatic patient, removal is probably indicated. Surgery can be more readily done without loss of function in the smaller tumors. Certainly tumors arising from brachial and pelvic plexus and more proximal major nerves such as sciatic, median, ulnar, and radial should be removed when they are identified, since surgery at a later date when they are larger is more difficult and carries a higher risk of serious loss. Nonetheless, there are examples in our series of 300 nerve tumors suspected to be either neurofibromas or schwannomas which have been followed for years by repetitive clinical and computerized axial tomography (CAT) or magnetic resonance imaging (MRI) examinations and which have not changed appreciably in size. Currently, 68% of neurofibromas operated on have been removed with little significant additional loss of function. If neurofibromas associated with von Recklinghausen's disease (VRD) are analyzed, that figure decreases to 52%. Key to dissection of these difficult lesions is an interfascicular approach including those fascicles entering arid exiting the tumor and those surrounding it. Sequential NAP recording often shows fascicles entering and leaving the tumor which are electrically silent and which can be sacrificed permitting removal of the rest of the tumor with little jeopardy of distal f ~ n c t i o n . By ' ~ comparison, schwannomas can almost always be completely removed with less dissection and little risk of functional loss. COMMON SURGICAL TECHNIQUES IN CURRENT USE External Neurolysis. There are two types of neurolysis, external and internal. The former consists of dissection of nerve free from surrounding tissues including scar and includes exposure of the entire circumference of the nerve. This is the initial neural step in any peripheral nerve operation whether nerve is transected or in continuity. It is usually begun by working from normal to abnormal nerve beginning well distal as well as proximal to the lesion site. Thickened or scarred portion of the epineurium is resected. If carefully done, long lengths of nerves can be mobilized without serious
MUSCLE & NERVE
September 1990
847
interference with blood supply. Although manipulation of the external environment of nerve may sometimes help neuritic pain, it is doubtful that it improves function at least in a direct fashion. lntraoperativeRecordings (Nerve Action Potentials).
External neurolysis is usually a necessary step before intraoperative stimulation and recording studies are done. This permits accurate placement of electrodes both at sites proximal and distal to the lesion. One then moves the recording electrodes onto the injury site to see if a response is recordable distal to the lesion and then if it extends to the distal stump. Stainless steel wire or annealed platinum 18-gauge wire is bent like a shepherd’s crook to accommodate a portion of the circumference of the nerve. Bipolar electrodes, one set for stimulation and one for recording, are soldered to shielded wire and the connections are sealed with surgical cement in the center of drilled out Delrin rods. These materials can be either gas or heat sterilized and can be used many times. A simpler and equally effective technique is to use sterilized EEG needle electrodes, which can be directly inserted into the nerve at both stimulating and recording sites. Stimulus can be provided by a stimulator with a stimulus isolation unit and recordings by any oscilloscope with differential amplification and high- and low-frequency filters. The stimulator must be interfaced to the oscilloscope so that each stimulus triggers a sweep on the scope. Care must be taken not to filter out the evoked response. Most EMG machines can be used in an intraoperative setting to provide acceptable stimulation and recording parameters. Currently used most often is a TECA T D 20. Since distances between stimulating and recording electrodes are relatively short, and shocks are delivered directly to the nerve, very brief stimuli are applied 0.02 to 0.1 msec in duration in order to reduce as much as possible the slope of the shock artifact. As a result, voltage used is higher than that to which the electrodiagnostician may be accustomed. Time base is set at 0.5 to 1.0 msec per division to separate as much as possible any evoked response from the shock artifact. Truly conducted responses are usually recorded at amplification settings of 50 pV to 5 mV per division. Filter settings are adjusted to provide an optimal trace. Other details for this type of approach to intraoperative evaluation are published elsewhere. 12,14,33 It is important to realize that with the serious injury in continuity what counts is whether o r not
848
Surgical Repair of Nerve Injury
there is a recordable response at all and not its form or even latency. Sixty-cycle noise, a problem in some operative settings, can be reduced by taking the bovie ground plate to the ground of the recording setup. Muscle action potential (MAP) artifacts have slower conduction than NAPS, have rounded peaks, and are usually more complex so can almost always be readily differentiated from an NAP. If no NAP is recorded then the lesion is resected beginning where the NAP was last recorded and extended beyond the lesion until good cross-sectional architecture is seen. Repair is then done by end-to-end suture if neurolysis, and thus mobilization of the stumps, provides enough length for repair with minimal tension. If this cannot be done then the gap must be closed by grafts. If an NAP can be recorded through the lesion and just beyond, it is not resected and only an external neurolysis is done. Occasionally a lesion conducting an NAP will look and feel worse over a portion of its circumference than in other areas. Proximal and distal nerve is then split into their fascicular structure and traced as best as possible through and beyond the injury site. Rerecording at a fascicular level identifies those fascicles more severely involved.34 Fascicles conducting are spared resection whereas those that do not are resected and repaired end to end or more commonly by interposition of sural or antebrachial cutaneous grafts. This is called a “split” repair. Magnification either by eye loupe or the microscope is of great help for this type of procedure (Fig. 2).
Split Repair.
internal Neurolysis. This procedure is done when there is incomplete loss distal to the lesion, but the patient has pain of a neuritic nature which does not respond to conservative management. Some surgeons assess lesions in continuity by doing an internal neurolysis and, whether or not one can dissect, each fascicle through injury site to make the decision for or against resection and repair. This clinic prefers use of intraoperative electrical studies to make such decisions. Certainly, fascicular continuity does not guarantee that intrafascicular structure will be sufficient to support signifiThis is especially so cant axonal with not only injection injuries but most contusive and stretch injuries whether due to closed or penetrating (such as gunshot wounds) mechanisms. In these frequent cases, some gross fascicular structure can be maintained and yet damage within fas-
MUSCLE & NERVE
September 1990
T? CI
T.G.
S1-R1 Sl-R2 S1-43 S2+ S3-R2+
C.D. C5,6,7 Loss(C) 6 Months
R3
/A 11 8
\
Nourolyala
I
/
c5
upraacapular
1.
radlal n
mxlllarv n
FIGURE 2. Stretch injury to C5,C6,and C7 spinal roots and upper and middle trunks explored 4 months after injury. By clinical and EMG examination, loss in this distribution remained complete for 4 months. Myelogram revealed a small rneningocele at C7 but also at C8. lntraoperative NAP recording showed a small regenerative potential (NAP) recorded from C5 to the suprascapular nerve. No NAPs could be obtained with stimulation of C5, C6,or C7 and recordingfrom the anterior and posterior divisions of the upper and middle trunks. The suprascapular nerve was split back into its outflow from the C5 and spared. There was good fascicular structure to lead sural grafts from C5 to the lateral and posterior cords but poor fascicular structure at the C6 and C7 proximal root levels so that only one fascicle at each of these levels appeared acceptable to lead a graft from. Evoked cortical activity with stimulation at the C6 and C7 proximal root levels was poor while that obtained from the C5 root was excellent.
cicles will not be sufficient to support significant regeneration (Fig. 3). Internal neurolysis, particularly through an injured segment, is not done without the potential of some loss of regenerated axons so it should be reserved for special circumstances. Suture is still the procedure of choice for transected nerve or focal lesions in continuity where opposition of stumps can be gained without excessive tension. Achievement can be helped by mobilization of the stumps. In addition, some nerves such as the ulnar close to the elbow and the radial in the upper
End-to-End Suture or Epineurial Repair.
Surgical Repair of Nerve Injury
FIGURE 3. Stretch injury was explored 6 months after injury. Loss was, again, complete in the C5,C6,and C7 distributionsof the plexus. A regenerative potential conducting 30 m/sec could be recorded by stimulating C7 and recording from the distal upper trunk so this element had only a neurolysis and was spared resection. Stimulationof C5 and C6 and recording from the distal middle trunk and anterior and posterior divisions gave no NAPs. The damaged segment was resected. Grafts could be led out from C5 while the C6 spinal root was found damaged deep into its foramen to a dural level.
arm can be transposed to help make up length. Less length can be gained for the peroneal nerve by leveling off the head of the fibula so distal nerve is not so angulated as it traverses this bone and divides into superficial and deep branches. Mild flexion of the extremity also helps gain length for median, radial, tibial, and peroneal repairs. l8 Finely spaced interrupted sutures are used to approximate the stumps, first laterally and then along volar and dorsal epineurial surfaces. Fascicular geography changes after 1 to 2 cm of neural trimming, but groups of fascicles can be opposed as closely as possible even though the repair is done at an epineurial level.*l Results depend, of course, on nerve involved as well as level of injury and somewhat less on the relative age of the patient and severity of associated injuries. Many lesions, especially those due to blunt mechanisms, produce lengthy lesions which, if resected, lead to gaps not readily reGraft Repair.
MUSCLE & NERVE
September 1990
849
paired end-to-end. The two leading causes of failure of suture repair are inadequate resection of proximal and distal injuries back to healthy tissue and distraction at the repair site. Graft repair has the potential to correct both of these problems. Small caliber grafts seem to fare better than longer whole nerve grafts, and this is an important ob~ervation.~~.’~ Most popular as a graft technique is an interfascicular grouped fascicular approach.5322Groups of fascicles are isolated both proximally and distally and a number of small caliber nerves are used to span the gap. The ends of each graft are “fish-mouthed” or spread and held by fine suture to a group of fascicles both proximally and distally. As much of the fascicular structure of each stump is covered in this fashion as possible. Common donor nerves include the sural and antebrachial cutaneous nerves and for PIN repair, the superficial sensory branch (SSR) can be harvested and used (Fig. 4). Controversy still exists as to the length of gap A.J. 5 Months
which dictates need for Most surgeons using the technique feel that gaps of approximately an inch or so can be made up by careful mobilization of the stumps. If transposition is possible, then another inch or two of length can be gained. Flexion of the extremity also gains length but may require immobilization for 3 weeks and runs the risk of joint stiffness. At some sites, direct neural repair without use of grafts is difficult. Examples are provided by supraclavicular plexus lesions and most pelvic plexus injuries. Also controversial is the relation between length of grafts and results. Some surgeons feel nerves or elements repaired properly with grafts, even very long ones, fare well. Of course, this depends greatly on the nerve and level of the injury as well as other factors. For example, it is difficult to gain useful foot dorsiflexion even with end-toend suture repair of the peroneal nerve. In my experience it is difficult to gain good results once grafts longer than 3 inches have to be used. Exceptions are provided by outflows going to muscles such as the biceps, brachialis, finger and wrist flexors, and gastrocnemius-soleusmuscles. COMMENTARY ON RESULTS
S1-R1 S2-Rl
S2-R2
Some summaries of results of nerve repair are useful and provide some rough guidelines.5720327 Only careful study of a detailed analysis such as was provided by the Veterans Administration monograph concerning World War I1 American injuries gives a broad enough perspective to select patients for surgery in an informed fashion.35 A similar careful and detailed analysis for civilian injuries remains to be accomplished. To do this well requires a large and diverse number of patients but as the “center concept” for management of such lesions has increased, appropriate numbers are being collected here and abroad. Despite the above, a few generalizations are indicated:
L
L _L,
1. Results with neurolysis are usually good espe“Axlllary
n.
FIGURE 4. The antebrachial cutaneous nerve was used to repair a lengthy stretch injury involving not only the axillary nerve but its posterior cord origin. Posterior cord to radial function was good and an NAP could be evoked through this pathway but not along the axillary nerve. The thorocodorsal branch, although injured when stimulated, gave latissimus dorsi contraction so it was split back through its posterior cord origin while the posterior cord to distal axillary segment was resected and replaced by two 3-inch grafts.
850
Surgical Repair of Nerve Injury
cially if nerves or elements are well selected for this procedure. Inherent in this is the fact that lesser injuries or incomplete injuries to begin with are usually selected for n e u r o l y ~ i s . ~If~ ~ ~ ’ intraoperative electrophysiologic studies show definite evidence of regeneration, then over 90% of neurolysed nerves have good recovery. Under those circumstances, this favorable response applies even to nerves such as the ulnar, medial cord, or even lower spinal roots, for if
MUSCLE & NERVE
September 1990
NAPS are recorded through them and thus resection avoided, recovery will be much more substantial than that provided by the very best of suture or graft repairs (Table 2). 2. Results with split or partial repair are also very good especially if the sparing of some fascicular structure is based on a recordable NAP distal to the lesion. Of interest, even when graft rather than suture repair is necessary for a portion of the lesion, the results are also very good. 3. End-to-end suture tends to do better than graft repair but this relates to the fact that lesser injuries are usually repaired by this technique. Nonetheless, success with grafts in the last decade has been most encouraging and, for many peripheral nerve surgeons, is the principle technique of repair used. Despite this advance, overall results with all types of suture repair still relate strongly to nerve as well as level involved. Thus, radial repairs fare better than median and median better than ulnar. In the lower extremity, posterior tibia1 and femoral repairs fare much better than peroneal no matter which technique of repair is used. In a similar vein, upper and middle trunk or C5, C6, and some C7 spinal root repairs do much better than lower trunk or C8, T1 root repairs. Lateral cord repair results are far superior to those of the posterior cord while once a medial cord lesion has to be resected the results, especially for medial cord to ulnar nerve outflow, is dismal. "," 4. Young children seem to do better with suture repair than adults. Unclear is whether this is due to relatively short lesion to potential innervational site distances or to the willingness of
Table 2. Operative results-upper extremity nerves.
Transections Primary suture Secondary suture Secondary graft In continuity (+) NAP = Neurolysis (-) NAP = Suture (-) NAP = Graft Split repair
Radial
Median
Ulnar
11110 1219 15110
18115 513 1319
1318 1319 512
48/45 14111 22116 212
66/61 20117 20117 313
38/35 1519 1215 111
Operations donelnumber reaching grade 3 or better results. Grade 3 indicates return of at least moderate strength in proximal muscles of element injured and contraction against gravity of distal muscles. For this table results are totals for all levels of injury to each nerve as well as all injury mechanisms.
Surgical Repair of Nerve Injury
most youngsters to use, and to find substitutive mechanisms even for severely disabled extremities. 5. No surgical procedure is without possible complications. Fortunately, the risk of death and serious complications is low but they can and do occur." Additional neural loss, onset of pain or aggravation of preexisting pain, infection and poor wound healing can complicate recovery. 6 . Finally, the need for rehabilitation and good physical and occupational therapy not only postoperatively but preoperatively cannot be stressed enough.36 Nerves do not heal by being put to rest, and immobilization of the limb, unless needed for finite periods for fractures and vascular and tendon repairs, is to be avoided. 7. Indications and timing for operative intervention for lacerations, gunshot wounds, stretch,
Table 3. Selection of plexus stretch injuries for operation. ~~
~
Clinical features 1. Is clinical loss, especially motor loss, complete in the distribution of one or more plexus elements? 2. Does loss by clinical and EMG criteria persist for 4 or 5 months postinjury? 3. If injury is at spinal nerve level, how proximal is it? Features that may make successful operation less likely 1. Winging of scapula-long thoracic nerve. 2. Rhomboid paralysis-dorsal scapular nerve. 3. Diaphragm paralysis-phrenic nerve. 4. Extensive paraspinal deinnervation by EMG- posterior branch of anterior root. (More distal and thus reparable injury to some roots may still be present.) 5. Positive sensory potentials can suggest preganglionic injury at C8, T1, and sometimes C7. Higher levels may still be operable. Somatosensory studies may be used in the same way. (Some investigators feel they can distinguish mixtures of pre- and postganglionic lesions by such studies.) 6. Myelopathy and/or fracture of cervical spine Less certain contraindications to surgery 1. Some recovery of sensation in shoulder and upper arm. 2. Total flail arm (Some can be partially repaired using usually proximal C5 and sometimes C6 outflow(s).) 3. Meningoceles at some (usually lower) levels but not at all levels. a. Meningocele(s) strongly suggests but does not prove proximal root damage. b. Meningoceles at one or more levels suggest but does not prove proximal damage at other root levels without meningoceles. c. Absence of meningoceles does not prove lateral damage nor does presence of a meningocele always mean proximal damage or on occasion even damage at all.
MUSCLE & NERVE
September 1990
851
and tumors involving brachial plexus are complex and vary according to mechanism and level of injury. Careful delineation requires a lengthier treatise. Several examples of stretch
injury involving the brachial plexus are provided by Figures 2, 3, and 4. Some of the criteria for selection of stretch injuries for operation are provided in Table 3.
-
REFERENCES 1. Brown P: Factors influencing the success of surgical repair of peripheral nerves. Surg Clin North Ant 1972;52(5):11371155. 2. DasGupta T: Tumors of the peripheral nerves. Clin Neurosurg 1978;25:574-590. 3. Donaghy R: History of microneurosurgery, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGraw-Hill, 1985, pp 20-26. 4. Gilliatt RW: Physical injury to peripheral nerves: physiologic and electrodiagnostic aspects. Mayo Clin Proc 1981;56:361-376. 5. Haase J, Bjerve P, Semesen K: Median and ulnar nerve transections treated with microsurgical interfascicular cable grafting with autologous sural nerve. J Neurosurg 1980;53:73-84. 6. Hudson A, Hunter D: Timing of peripheral nerve repairs: important local neuropathological factors. Clin Neurosurg 1977;24:340-391. 7. Hudson A, Mayfield F: Chronic injuries of nerves by entrapment, in Youmans J (Ed): Neurological Surgery. Philadelphia, WB Saunders, 1983, pp 2430-2474. 8. Hudson A, Trammer B: B&hial plexus injuries, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGraw-Hill, 1985, pp 1817- 1832. 9. Kempe L: Operative Neurosurgery. New York, Springer-Verlag, 1970, vol. 2, pp 158- 170. 10. Khodadad G: Microsurgical techniques in repair of peripheral nerves. Surg Clin North Am 1972;52(5):11571166. 11. Kline D: Civilian gunshot wounds to the brachial plexus. J Neurosurg 1989;70:166- 174. 12. Kline D, Dejonge B: Evoked potentials to evaluate peripheral nerve injuries. Surg Gynecol Obstet 1968;127:12391248. 13. Kline D, Hackett E: Reappraisal of timing for exploration of civilian peripheral nerve injuries. Surgery 1975;78:5465. 14. Kline D, Hackett E: Management of the neuroma in continuity, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGraw-Hill, 1983, pp 1864- 1871. 15. Kline D, Hackett E, May P: Evaluation of nerve injuries by evoked potentials and electromyography. J Neurosurg 1969;31:128- 136. 16. Kline D, Hudson A: Complications of nerve injury and repair, in Greenfield L (Ed): Complications in Surgery and Trauma. Philadelphia, JB Lippincott, 1984, pp 695-708. 17. Kline D, Judice D: Operative management of selected brachial plexus lesions.J Neurosurg 1983;58:631-649. 18. Kline D, Nulsen F: Acute injuries of peripheral nerve, in J Youmans (Ed): Neurological Surgery. Philadelphia, WB Saunders, 1982, pp 2363-2429.
852
Surgical Repair of Nerve Injury
19. Lusk M, Garcia C, Kline D: Tumors of brachial plexus. Neumsurgely 1987;21:439-453. 20. MacKinnon S, Dellon A: Surgery of the Peripheral Nerve. New York, Thierne Medical Publishers, 1988. 21. McCillicuddy J: Techniques of nerve repair, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGrawHill Co, 1985, pp 1871-1881. 22. McQuarrie I, Idzikowski C: Injuries to peripheral nerves, in Miller T , Rowlands B (Eds): Physiologic Bmis of Modern Surgical Care. St. Louis, CV Mosby, 1988, pp 802-815. 23. Millesi H, Meissl G, Berger A: Further experience with interfascicular grafting of median, ulnar, and radial nerves.J Bone Joint Surg 1976;52B:209-218. 24. Millesi H: Reappraisal of Nerve Repair. Surg Clin North A m 1981;61:32 1-340. 25. Narakus A: The surgical treatment of traumatic brachial plexus lesions. Int Surg 1980;65:521-527. 26. Nulsen F, Slade H: Recovery following injury to brachial plexus, in Woodhall B, Beebe G (Eds): Peripheral Nerve Regeneration, VA monograph. Washington, DC, Government Printing Office, 1957, pp 389-408. 27. Omer G: The evaluation of clinical results following peripheral nerve suture, in Omer G, Spinner M (Eds): Management of Peri$heral Nerve Problems. Philadelphia, WB Saunders, 1980, pp 431-438. 28. Rodkey WG, Cabaud H, McCarroll H: Neuropathy after loss of a nerve segment: comparison of epineural suture under tension vs multiple nerve grafts. J Hand Surg 1980;5:366-371. 29. Samii M: Use of microtechniques in peripheral nerve surgery-experience with over 300 cases, in Handa H (ed): Microneurosurgery. Tokyo, Igaku Shoin Ltd., 1975, pp 8593. 30. Seddon H: Surgical Disorders of Penpheral Nerves. Baltimore, Williams & Wilkins, 1972. 3 1. Spinner M: Injuries of the Major Branches of Peripheral Nerves of the Forearm. Philadelphia, WB Saunders, 1978. 32. Sunderland S: Nerves and Nerve Injuries, ed 2, Edinburgh, Churchill Livingstone Ltd., 1978. 33. Terzis J, Dykes R: Electrophysiological recordings in peripheral nerve surgery. A review.] Hand Surg 1976;1:5266. 34. Williams HG, Terzis J: Single fascicular recording: an intraoperative diagnostic tool for the management of peripheral nerve lesions. P l a t Reconstr Surg 1976;57:562. 35. Woodhall B, Nulsen F, White J, Davis L: Neurosurgical Implications. Peripheral Nerve Regeneration. VA monograph. Washington, DC, U.S. Government Printing Office, 1957, pp 569-638. 36. Wynn Parry CB: Rehabilitation of the Hand, ed 3, London, Butterworths, 1978.
MUSCLE & NERVE
September 1990
SURGICAL REPAIR OF PERIPHERAL NERVE INJURY DAVID G. KLINE, MD, FACS
During the last quarter of a century, three significant changes in the surgical management of nerve injuries have occurred. Magnification by either an operating microscope or use of eye loupes has become commonplace as has the use of the bipolar coagulator, some microinstruments, and relatively fine suture mate ria^.^"'*^^ Of even greater importance has been the re-emergence of graft repair as a practical method to regain innervation.23’24*28 This technique permits optimal resection of the injury site(s) back to healthy tissue and bridges neural gaps with minimal tension at the repair sites. During this same period of time, electrophysiologic methods for intraoperative assessment of the most frequent and serious injury
From the Department of Neurosurgery, Louisiana State University Medical Center, New Orleans, Louisiana. Presented at AAEE (now AAEM) Annual Meeting, International Symposium on Peripheral Nerve Regeneration, Washington, DC, September 17, 1989. Address reprint requests to Dr. Kline at the Department of Neurosurgery, Louisiana State University Medical Center, 1542 Tulane Avenue, New Orleans, LA 701 12-2822. Accepted for publication February 16, 1990 CCC 0148-639W90/090843-010 504.00 0 1990 John Wiley & Sons, Inc
Surgical Repair of Nerve Injury
to nerve, one leaving it in continuity, have evolved. 12~15,33These techniques permit a physiologic decision for or against resection so that a lesion “looking good” but having poor regeneration is not left unresected or conversely, one “appearing bad” but having good regeneration is not falsely resected. These advances have most likely led to improved functional results at least in some surgeon’s hands. More certain is that these operative changes have increased the incidence of surgical explorations and attempted repairs of lesions previously felt not to be repairable particularly those of the brachial plexus and the proximal arm and leg. Other shifts in management, although less dramatic, have also occurred. Where direct neural repair has been impossible or has failed, substitutive procedures have been further refined, in some situations newly devised, and in most centers increasingly used to restore function. Older recognized issues such as the importance of timing for surgery, the role of preoperative electromyographic evaluation, and the relatively aggressive use of physical and occupational therapy have received renewed e r n p h a ~ i s . ’ ~ ~ ~Related “ ~ ” ’ ~to~ ~ ~ ~ ~ ~ all of the above as well as to trends in medicine as a whole has been an increase in the “center-ori-
MUSCLE & NERVE
September 1990
843
ented” management of at least the more complex and difficult nerve lesions. SELECTION AND TIMING FOR SURGERY
The two major reasons for operating on a nerve injury are to improve motor and sensory function and to ameliorate neural pain. There are a few settings where partial loss following an injury can progress and, of course, surgery and usually urgent surgery is necessary for these. Progressive lesions are associated with vascular injury and are due to pseudoaneurysm, arteriovenous malformation, or clot in a potentially tight space.26 Mechanism is usually a penetrating injury due to gunshot wound or knife but can on occasion be due to contusive forces. Acute decompression with removal of the vascular lesion and neurolysis of involved nerves is necessary. Causalgia, which is usually associated with penetrating injury to proximal nerve(s) may require acute sympathectomy. Less frequently, a missile fragment or foreign body may lodge directly in nerve and may require acute removal to control pain. Severe soft tissue contusion and associated vascular compromise can result in a Volkmann’s in the forearm or an anterior compartment syndrome in the leg. Acute fasciotomy as well as exposure of the nerves is necessary. ACUTE AND SUBACUTE REPAIR OF LACERATIONS
The most frequent reason for relatively acute operation is to repair nerve injury associated with extremity laceration. Such injuries are due to glass, knives, or sometimes metal edges, and transect nerve in a relatively sharp fashion. In this setting, there may be an advantage to acute suture since tissue to be trimmed from each stump is minimal, associated anatomy is not greatly distorted, and nerve can be repaired at the same time as other soft tissue injuries. Such primary or acute repair is especially important for lacerating injuries to plexus or proximal sciatic nerve where the scar and stump retraction associated with delay makes secondary repair more difficult. Results in most transection cases well selected for acute repair are at least as good as in those re aired later and in some cases are even better. 18,30!32 By comparison, blunt transections are best repaired after a delay of several weeks. Mechanisms include laceration due to propellar blades, power saws, auto metal, and some complex fractures. Damage to each stump associated with blunt mechanisms is difficult to delineate acutely. End-
844
Surgical Repair of Nerve Injury
to-end repair of potentially damaged stumps leads inevitably to a poor result. With delay, the amount of neuroma and glioma to be trimmed is quite evident. Injuries lacerating soft tissue and associated with distal neural deficit do not always transect nerve completely and in some cases only contuse nerve. If the nerve is partially divided in a sharp fashion, then acute repair of the divided portion is in order. On the other hand, if it is known that the partial division is blunt or suspected that contusion rather than neural laceration is responsible for deficit, delay in repair is in order. In the interim, there may occur significant clinical or electromyographic (EMG) reversal of loss. If not, then the lesion can be evaluated by intraoperative stimulation and recording of nerve action potentials (NAPS). These electrophysiologic tests along with magnification are used to determine whether a split or total repair or more simply, only a neurolysis is necessary. LESIONS I N CONTINUITY
T h e majority (60 to 70%) of serious injuries to nerve are due to blunt mechanisms such as stretch and contusion and leave the nerve in continuity. Since nerve is not divided, there is a variable degree of intraneural injury. Thus, there may be a mixture of axonotmesis and neurotmesis and even an element of neurapraxia. Other lesions will be totally neurotmetic despite maintenance of gross continuity. Swelling or neuroma formation at the injury site is variable and does not necessarily correlate with the severity of intraneural axonal injury, let alone to the degree of distortion and disruption of the connective tissue framework of the As a result, some of these lesions will improve spontaneously with time and without operation but unfortunately many will not. The most important consideration is whether or not deficit distal to such a lesion is complete or incomplete by clinical as well as electromyographic criteria. With incomplete lesions with useful sparing of distal function, further function will usually result in time. Thus, with nonoperative management including good physical therapy, function at least as good and usually better than that achieved by surgery occurs. When milder degrees of injury occur to the entire cross-section of the nerve, findings include diffuse muscular weakness rather than paralysis, and partial rather than complete sensory loss even in autonomous Function almost always improves with time unless injury site
MUSCLE & NERVE
September 1990
is near that for potential neural entrapment such as the olecranon notch, the carpal tunnel, o r the head of fibula. Of course, associated and especially expanding, lesions such as aneurysm, fistula, or clot can prevent recovery unless they are corrected. A complete lesion in continuity which shows no significant clinical or electrical recovery for 2 to 4 months postinjury is usually a candidate for surgical exploration and intraoperative electrophysiologic assessment. (Fig. 1). How long to delay surgical exploration remains somewhat controversial even on a case-by-case basis. If, however, the clinician waits the many months it may take to be certain that there is no clinical nor EMG evidence of functional reinnervation coming about, then it may be too late for such greatly delayed repair to be effective. This is because innervational input sites, particularly for muscle, will shrink and atrophy. Once such involution has occurred, bringing even thousands of axons back to a potential innervational site is like “sowing good seed on fallow ground.” ELECTRICAL STUDIES FOR LESIONS IN CONTINUITY
For effective electrodiagnostic work-up, the normal innervational pattern of the involved nerve has to be known as well as the time it takes for regeneration to site(s)just distal to that of the injury. Most incontinuity lesions have an area of retrograde axonal loss. This distance as well as the length of the injury site has to be traversed before regenerating axons reach distal stump, and then eventually reach even more distal reinnervational points that can be tested by noninvasive electrical
techniques. Even for an axonotmetic lesion, this regrowth process takes months. For example, with a complete midhumeral level injury to radial nerve, it would be 3 to 4 months before one could expect any significant reversal of deinnervational change in brachioradialis, let alone voluntary contraction of that muscle. With a buttock or thigh level sciatic nerve injury, regenerating axons may not reach even proximal calf or leg muscles for a year to a year and a half. Reinnervational delays after many injuries to brachial plexus elements, especially those at a spinal root or trunk level, can be very great because distances to be made up are great. Axons must not only reach muscle but make some reconstruction of axon to motor end plate to muscle fiber junctions to reverse electromyographic changes such as fibrillations and to show nascent potential^.^ A significant reduction in number of fibrillations and deinnervational potentials occur several weeks before nascent potentials appear. There is further delay before muscle action potentials can be evoked. Thus, even after axons reach muscle there is a further delay (weeks to a month or more) before signs of electrical reinnervation occur. Nonetheless, some reversal of deinnervational change does occur weeks before there is objective contraction of muscle. It has to be remembered that the EMG is a physiologic test. It tests the electrical concomitants of deinnervation andlor reinnervation. Thus, reversal of deinnervational and occurrence of reinnervational changes suggests that, with time, useful function will occur but does not guarantee such. In addition, the EMG is a sampling technique and what is occurring in one area of a muscle may not be happening elsewhere. Thus, a
FIGURE 1. Stimulation and recording from a lesion in continuity involving the proximal median nerve and explored 5 months after the injury. NAPS were recorded from this injury due to contusion and stretch so only a neurolysis was done.
Surgical Repair of Nerve Injury
MUSCLE & NERVE
September 1990
845
thorough sample as well as a discriminating appraisal of electrical activity recorded is a must. Despite these points, EMG in well-trained hands remains a very valuable method for evaluating lesions in continuity even in the early months postinjury. Contraction of muscle distal to the injury in response to neural stimulation is also a valuable harbinger of clinical return. Such a response in an adequately regenerating nerve may antedate voluntary muscle contraction by weeks to several months. 18.2635 Again, for stimulation even directly to the nerve to demonstrate early recovery, axons must have not only reached muscle but there must be some degree of motor end-plate reconstruction. When axons are only in distal stump and have not as yet reached muscle, neither EMG nor stimulation will show evidence of recovery and yet such early regeneration may have great potential for functional reinnervation. Thus, in order to assess relatively early regeneration in a lesion in continuity, it may be necessary to surgically expose the nerve and to do direct, intraoperative NAP as well as stimulation studies. TIMING FOR SURGERY ON LESIONS IN CONTINUITY
The more focal lesions in continuity are associated with fracture, contusion to soft tissue and some gunshot These can be reliably evaluated by intraoperative electrical studies at 2 to 3 months postinjury. Many injuries due to stretch or severe contusion produce lengthier lesions and are associated with vehicular and skiing accidents, falls, crush injuries, and shotgun injuries. These stretches require months of regeneration before a large number of axons can traverse the injury segment and reach distal stump. As a result, these cases are followed for 3 to 5 months.” If deficit remains complete in the distribution of one or more nerves or plexus elements, an exploration is done. Definite signs of spontaneous reversal of loss take longer to appear in this group of patients than in any other. In addition, when early clinical or EMG recovery does not occur, one wants to be in a good position at exploration to record from the stretch segment and distal nerve and to be certain whether or not adequate regeneration is occurring. This is necessary so that the correct decision can be made for or against resection and repair. In the usual stretch injuries to plexus selected for surgery, the majority of involved roots or elements are left in continuity. Some have the possibility for spontaneous and significant re-
846
Surgical Repair of Nerve Injury
growth while many do not.8217By 3 to 5 months, the relative regenerative capacity of the various elements can be accurately evaluated by recording across the stretched segments. Many but not all civilian gunshot wounds involving plexus require operation. Since most of these gunshot wounds are more focal than the majority of stretch injuries, intraoperative assessment is accurate at 2 to 3 months postinjury. l 1 Delay in referral, healing of associated injuries, and management of infection may of course alter this recommended injury to operative interval. Aside from these factors, when one is in doubt about the direction of recovery it is better to look and assess the nerve directly, for undue delay in suture repair leads to poor results.”’* (Table I). TIMING FOR ENTRAPMENTS AND TUMORS
These spontaneous conditions are usually not associated with injury but their course can certainly
Table 1. Management of the neuroma in continuity. Incomplete loss with significant distal sparing 1. Most cases will improve with conservative treatment. They are followed by serial clinical and EMG examinations. Physical therapy is important. 2. Operation may still be required: Partial lesions associated with expanding masses due to hematoma, aneurysm, or A-V fistula usually require urgent operation Partial lesions close to or in areas of potential entrapment may require relatively early operation Lesions where distal loss, although partial, is significant may require later operation Neural pain not amendable to medications and physical therapy may require later operation Complete or near complete loss with little or no distal sparing 1. Relatively focal lesions in continuity due to fracture or gunshot wound. a. Follow by clinical and EMG examinations for 2-3 months. b. If no significant clinical or electrical improvement, explore. c. lntraoperative stimulation and NAP studies used to decide for or against resection. 2. Relatively lengthy lesions in continuity due to contusionlstretch or shotgun. a. Follow by clinical and EMG examinations for 4-5 months. b. If no significant clinical or electrical improvement, explore. c. If no response to stimulation and no NAP across lesion, resection and repair by suture or graft is necessary. d. lntraoperative evoked cortical or somatosensory studies may be necessary to evaluate repairability of proximal spinal nerves. (See Table 3).
MUSCLE & NERVE
September 1990
be complicated by such. Uncomplicated entrapment and tumor associated neural loss tends to be mild. Many or almost all axons may have peripheral connections. EMG signs of deinnervation may thus be mild or absent altogether. On the other hand relative conductive loss or slowing is a cardinal feature especially with entrapments. This, of course, make the risk of further loss of function with operation greater than with a completely injured nerve where loss is already maximal or near maximal. If the entrapped nerve is sufficiently symptomatic in terms of pain and/or loss of function to be bothersome to the patient, then operation without undue delay is certainly indicated. It is seldom that the indication is emergent or even urgent.' Exceptions are occasionally provided by acute carpal tunnel syndrome, and less frequently acute compression of ulnar at elbow level associated with pregnancy or the immediate postpartum period. Almost always a clinically significant median or ulnar entrapment will show electrical abnormalities both by noninvasive conductive studies and, if one goes to the trouble to do so, by direct recording from nerve intraoperatively. The latter is interesting to do during such operations but not as essential as it is in more serious in-continuity lesions where the question of resection is present. If an entrapped nerve with significant clinical and electrical findings is to be followed along without operation, then this should be done very closely and very carefully. Since the operations involved, especially for carpal tunnel, are of relatively low risk, are relatively straightforward and also have a good success rate, the better part of valor is usually to decompress the nerve and where indicated to transpose it. A number of posterior interosseous nerve (PIN) palsies will reverse by 2 to 3 months after onset, and thus, a conservative period of management is certainly in order for these. By sharp contrast, loss associated with peroneal entrapment at the head of the fibula, even when unassociated with contusion, does not usually reverse. If peroneal decompression is delayed too long, reversal even of partial loss may not occur. Results are better with median than ulnar, and better with ulnar than peroneal entrapments, while those with PIN entrapment are excellent.""' Tumors arising in nerves are usually but not always benign. Thus, there is a tendency to follow them if they are not too symptomatic.' At stake with an operation is a variable degree of loss associated with tumor removal. Neural loss is a necessity if the lesion is a malignant intrinsic tumor,
Surgical Repair of Nerve Injury
since wide resection and/or amputation of the limb well proximal to the lesion is usually recommended. Loss may also occur with removal of benign large tumors especially those in the brachial or pelvic plexus, particularly if they are neurofibromas. Progressive loss and pain remain indications for operation but in addition, if a lesion is greater than a few centimeters in diameter, even in a less symptomatic patient, removal is probably indicated. Surgery can be more readily done without loss of function in the smaller tumors. Certainly tumors arising from brachial and pelvic plexus and more proximal major nerves such as sciatic, median, ulnar, and radial should be removed when they are identified, since surgery at a later date when they are larger is more difficult and carries a higher risk of serious loss. Nonetheless, there are examples in our series of 300 nerve tumors suspected to be either neurofibromas or schwannomas which have been followed for years by repetitive clinical and computerized axial tomography (CAT) or magnetic resonance imaging (MRI) examinations and which have not changed appreciably in size. Currently, 68% of neurofibromas operated on have been removed with little significant additional loss of function. If neurofibromas associated with von Recklinghausen's disease (VRD) are analyzed, that figure decreases to 52%. Key to dissection of these difficult lesions is an interfascicular approach including those fascicles entering arid exiting the tumor and those surrounding it. Sequential NAP recording often shows fascicles entering and leaving the tumor which are electrically silent and which can be sacrificed permitting removal of the rest of the tumor with little jeopardy of distal f ~ n c t i o n . By ' ~ comparison, schwannomas can almost always be completely removed with less dissection and little risk of functional loss. COMMON SURGICAL TECHNIQUES IN CURRENT USE External Neurolysis. There are two types of neurolysis, external and internal. The former consists of dissection of nerve free from surrounding tissues including scar and includes exposure of the entire circumference of the nerve. This is the initial neural step in any peripheral nerve operation whether nerve is transected or in continuity. It is usually begun by working from normal to abnormal nerve beginning well distal as well as proximal to the lesion site. Thickened or scarred portion of the epineurium is resected. If carefully done, long lengths of nerves can be mobilized without serious
MUSCLE & NERVE
September 1990
847
interference with blood supply. Although manipulation of the external environment of nerve may sometimes help neuritic pain, it is doubtful that it improves function at least in a direct fashion. lntraoperativeRecordings (Nerve Action Potentials).
External neurolysis is usually a necessary step before intraoperative stimulation and recording studies are done. This permits accurate placement of electrodes both at sites proximal and distal to the lesion. One then moves the recording electrodes onto the injury site to see if a response is recordable distal to the lesion and then if it extends to the distal stump. Stainless steel wire or annealed platinum 18-gauge wire is bent like a shepherd’s crook to accommodate a portion of the circumference of the nerve. Bipolar electrodes, one set for stimulation and one for recording, are soldered to shielded wire and the connections are sealed with surgical cement in the center of drilled out Delrin rods. These materials can be either gas or heat sterilized and can be used many times. A simpler and equally effective technique is to use sterilized EEG needle electrodes, which can be directly inserted into the nerve at both stimulating and recording sites. Stimulus can be provided by a stimulator with a stimulus isolation unit and recordings by any oscilloscope with differential amplification and high- and low-frequency filters. The stimulator must be interfaced to the oscilloscope so that each stimulus triggers a sweep on the scope. Care must be taken not to filter out the evoked response. Most EMG machines can be used in an intraoperative setting to provide acceptable stimulation and recording parameters. Currently used most often is a TECA T D 20. Since distances between stimulating and recording electrodes are relatively short, and shocks are delivered directly to the nerve, very brief stimuli are applied 0.02 to 0.1 msec in duration in order to reduce as much as possible the slope of the shock artifact. As a result, voltage used is higher than that to which the electrodiagnostician may be accustomed. Time base is set at 0.5 to 1.0 msec per division to separate as much as possible any evoked response from the shock artifact. Truly conducted responses are usually recorded at amplification settings of 50 pV to 5 mV per division. Filter settings are adjusted to provide an optimal trace. Other details for this type of approach to intraoperative evaluation are published elsewhere. 12,14,33 It is important to realize that with the serious injury in continuity what counts is whether o r not
848
Surgical Repair of Nerve Injury
there is a recordable response at all and not its form or even latency. Sixty-cycle noise, a problem in some operative settings, can be reduced by taking the bovie ground plate to the ground of the recording setup. Muscle action potential (MAP) artifacts have slower conduction than NAPS, have rounded peaks, and are usually more complex so can almost always be readily differentiated from an NAP. If no NAP is recorded then the lesion is resected beginning where the NAP was last recorded and extended beyond the lesion until good cross-sectional architecture is seen. Repair is then done by end-to-end suture if neurolysis, and thus mobilization of the stumps, provides enough length for repair with minimal tension. If this cannot be done then the gap must be closed by grafts. If an NAP can be recorded through the lesion and just beyond, it is not resected and only an external neurolysis is done. Occasionally a lesion conducting an NAP will look and feel worse over a portion of its circumference than in other areas. Proximal and distal nerve is then split into their fascicular structure and traced as best as possible through and beyond the injury site. Rerecording at a fascicular level identifies those fascicles more severely involved.34 Fascicles conducting are spared resection whereas those that do not are resected and repaired end to end or more commonly by interposition of sural or antebrachial cutaneous grafts. This is called a “split” repair. Magnification either by eye loupe or the microscope is of great help for this type of procedure (Fig. 2).
Split Repair.
internal Neurolysis. This procedure is done when there is incomplete loss distal to the lesion, but the patient has pain of a neuritic nature which does not respond to conservative management. Some surgeons assess lesions in continuity by doing an internal neurolysis and, whether or not one can dissect, each fascicle through injury site to make the decision for or against resection and repair. This clinic prefers use of intraoperative electrical studies to make such decisions. Certainly, fascicular continuity does not guarantee that intrafascicular structure will be sufficient to support signifiThis is especially so cant axonal with not only injection injuries but most contusive and stretch injuries whether due to closed or penetrating (such as gunshot wounds) mechanisms. In these frequent cases, some gross fascicular structure can be maintained and yet damage within fas-
MUSCLE & NERVE
September 1990
T? CI
T.G.
S1-R1 Sl-R2 S1-43 S2+ S3-R2+
C.D. C5,6,7 Loss(C) 6 Months
R3
/A 11 8
\
Nourolyala
I
/
c5
upraacapular
1.
radlal n
mxlllarv n
FIGURE 2. Stretch injury to C5,C6,and C7 spinal roots and upper and middle trunks explored 4 months after injury. By clinical and EMG examination, loss in this distribution remained complete for 4 months. Myelogram revealed a small rneningocele at C7 but also at C8. lntraoperative NAP recording showed a small regenerative potential (NAP) recorded from C5 to the suprascapular nerve. No NAPs could be obtained with stimulation of C5, C6,or C7 and recordingfrom the anterior and posterior divisions of the upper and middle trunks. The suprascapular nerve was split back into its outflow from the C5 and spared. There was good fascicular structure to lead sural grafts from C5 to the lateral and posterior cords but poor fascicular structure at the C6 and C7 proximal root levels so that only one fascicle at each of these levels appeared acceptable to lead a graft from. Evoked cortical activity with stimulation at the C6 and C7 proximal root levels was poor while that obtained from the C5 root was excellent.
cicles will not be sufficient to support significant regeneration (Fig. 3). Internal neurolysis, particularly through an injured segment, is not done without the potential of some loss of regenerated axons so it should be reserved for special circumstances. Suture is still the procedure of choice for transected nerve or focal lesions in continuity where opposition of stumps can be gained without excessive tension. Achievement can be helped by mobilization of the stumps. In addition, some nerves such as the ulnar close to the elbow and the radial in the upper
End-to-End Suture or Epineurial Repair.
Surgical Repair of Nerve Injury
FIGURE 3. Stretch injury was explored 6 months after injury. Loss was, again, complete in the C5,C6,and C7 distributionsof the plexus. A regenerative potential conducting 30 m/sec could be recorded by stimulating C7 and recording from the distal upper trunk so this element had only a neurolysis and was spared resection. Stimulationof C5 and C6 and recording from the distal middle trunk and anterior and posterior divisions gave no NAPs. The damaged segment was resected. Grafts could be led out from C5 while the C6 spinal root was found damaged deep into its foramen to a dural level.
arm can be transposed to help make up length. Less length can be gained for the peroneal nerve by leveling off the head of the fibula so distal nerve is not so angulated as it traverses this bone and divides into superficial and deep branches. Mild flexion of the extremity also helps gain length for median, radial, tibial, and peroneal repairs. l8 Finely spaced interrupted sutures are used to approximate the stumps, first laterally and then along volar and dorsal epineurial surfaces. Fascicular geography changes after 1 to 2 cm of neural trimming, but groups of fascicles can be opposed as closely as possible even though the repair is done at an epineurial level.*l Results depend, of course, on nerve involved as well as level of injury and somewhat less on the relative age of the patient and severity of associated injuries. Many lesions, especially those due to blunt mechanisms, produce lengthy lesions which, if resected, lead to gaps not readily reGraft Repair.
MUSCLE & NERVE
September 1990
849
paired end-to-end. The two leading causes of failure of suture repair are inadequate resection of proximal and distal injuries back to healthy tissue and distraction at the repair site. Graft repair has the potential to correct both of these problems. Small caliber grafts seem to fare better than longer whole nerve grafts, and this is an important ob~ervation.~~.’~ Most popular as a graft technique is an interfascicular grouped fascicular approach.5322Groups of fascicles are isolated both proximally and distally and a number of small caliber nerves are used to span the gap. The ends of each graft are “fish-mouthed” or spread and held by fine suture to a group of fascicles both proximally and distally. As much of the fascicular structure of each stump is covered in this fashion as possible. Common donor nerves include the sural and antebrachial cutaneous nerves and for PIN repair, the superficial sensory branch (SSR) can be harvested and used (Fig. 4). Controversy still exists as to the length of gap A.J. 5 Months
which dictates need for Most surgeons using the technique feel that gaps of approximately an inch or so can be made up by careful mobilization of the stumps. If transposition is possible, then another inch or two of length can be gained. Flexion of the extremity also gains length but may require immobilization for 3 weeks and runs the risk of joint stiffness. At some sites, direct neural repair without use of grafts is difficult. Examples are provided by supraclavicular plexus lesions and most pelvic plexus injuries. Also controversial is the relation between length of grafts and results. Some surgeons feel nerves or elements repaired properly with grafts, even very long ones, fare well. Of course, this depends greatly on the nerve and level of the injury as well as other factors. For example, it is difficult to gain useful foot dorsiflexion even with end-toend suture repair of the peroneal nerve. In my experience it is difficult to gain good results once grafts longer than 3 inches have to be used. Exceptions are provided by outflows going to muscles such as the biceps, brachialis, finger and wrist flexors, and gastrocnemius-soleusmuscles. COMMENTARY ON RESULTS
S1-R1 S2-Rl
S2-R2
Some summaries of results of nerve repair are useful and provide some rough guidelines.5720327 Only careful study of a detailed analysis such as was provided by the Veterans Administration monograph concerning World War I1 American injuries gives a broad enough perspective to select patients for surgery in an informed fashion.35 A similar careful and detailed analysis for civilian injuries remains to be accomplished. To do this well requires a large and diverse number of patients but as the “center concept” for management of such lesions has increased, appropriate numbers are being collected here and abroad. Despite the above, a few generalizations are indicated:
L
L _L,
1. Results with neurolysis are usually good espe“Axlllary
n.
FIGURE 4. The antebrachial cutaneous nerve was used to repair a lengthy stretch injury involving not only the axillary nerve but its posterior cord origin. Posterior cord to radial function was good and an NAP could be evoked through this pathway but not along the axillary nerve. The thorocodorsal branch, although injured when stimulated, gave latissimus dorsi contraction so it was split back through its posterior cord origin while the posterior cord to distal axillary segment was resected and replaced by two 3-inch grafts.
850
Surgical Repair of Nerve Injury
cially if nerves or elements are well selected for this procedure. Inherent in this is the fact that lesser injuries or incomplete injuries to begin with are usually selected for n e u r o l y ~ i s . ~If~ ~ ~ ’ intraoperative electrophysiologic studies show definite evidence of regeneration, then over 90% of neurolysed nerves have good recovery. Under those circumstances, this favorable response applies even to nerves such as the ulnar, medial cord, or even lower spinal roots, for if
MUSCLE & NERVE
September 1990
NAPS are recorded through them and thus resection avoided, recovery will be much more substantial than that provided by the very best of suture or graft repairs (Table 2). 2. Results with split or partial repair are also very good especially if the sparing of some fascicular structure is based on a recordable NAP distal to the lesion. Of interest, even when graft rather than suture repair is necessary for a portion of the lesion, the results are also very good. 3. End-to-end suture tends to do better than graft repair but this relates to the fact that lesser injuries are usually repaired by this technique. Nonetheless, success with grafts in the last decade has been most encouraging and, for many peripheral nerve surgeons, is the principle technique of repair used. Despite this advance, overall results with all types of suture repair still relate strongly to nerve as well as level involved. Thus, radial repairs fare better than median and median better than ulnar. In the lower extremity, posterior tibia1 and femoral repairs fare much better than peroneal no matter which technique of repair is used. In a similar vein, upper and middle trunk or C5, C6, and some C7 spinal root repairs do much better than lower trunk or C8, T1 root repairs. Lateral cord repair results are far superior to those of the posterior cord while once a medial cord lesion has to be resected the results, especially for medial cord to ulnar nerve outflow, is dismal. "," 4. Young children seem to do better with suture repair than adults. Unclear is whether this is due to relatively short lesion to potential innervational site distances or to the willingness of
Table 2. Operative results-upper extremity nerves.
Transections Primary suture Secondary suture Secondary graft In continuity (+) NAP = Neurolysis (-) NAP = Suture (-) NAP = Graft Split repair
Radial
Median
Ulnar
11110 1219 15110
18115 513 1319
1318 1319 512
48/45 14111 22116 212
66/61 20117 20117 313
38/35 1519 1215 111
Operations donelnumber reaching grade 3 or better results. Grade 3 indicates return of at least moderate strength in proximal muscles of element injured and contraction against gravity of distal muscles. For this table results are totals for all levels of injury to each nerve as well as all injury mechanisms.
Surgical Repair of Nerve Injury
most youngsters to use, and to find substitutive mechanisms even for severely disabled extremities. 5. No surgical procedure is without possible complications. Fortunately, the risk of death and serious complications is low but they can and do occur." Additional neural loss, onset of pain or aggravation of preexisting pain, infection and poor wound healing can complicate recovery. 6 . Finally, the need for rehabilitation and good physical and occupational therapy not only postoperatively but preoperatively cannot be stressed enough.36 Nerves do not heal by being put to rest, and immobilization of the limb, unless needed for finite periods for fractures and vascular and tendon repairs, is to be avoided. 7. Indications and timing for operative intervention for lacerations, gunshot wounds, stretch,
Table 3. Selection of plexus stretch injuries for operation. ~~
~
Clinical features 1. Is clinical loss, especially motor loss, complete in the distribution of one or more plexus elements? 2. Does loss by clinical and EMG criteria persist for 4 or 5 months postinjury? 3. If injury is at spinal nerve level, how proximal is it? Features that may make successful operation less likely 1. Winging of scapula-long thoracic nerve. 2. Rhomboid paralysis-dorsal scapular nerve. 3. Diaphragm paralysis-phrenic nerve. 4. Extensive paraspinal deinnervation by EMG- posterior branch of anterior root. (More distal and thus reparable injury to some roots may still be present.) 5. Positive sensory potentials can suggest preganglionic injury at C8, T1, and sometimes C7. Higher levels may still be operable. Somatosensory studies may be used in the same way. (Some investigators feel they can distinguish mixtures of pre- and postganglionic lesions by such studies.) 6. Myelopathy and/or fracture of cervical spine Less certain contraindications to surgery 1. Some recovery of sensation in shoulder and upper arm. 2. Total flail arm (Some can be partially repaired using usually proximal C5 and sometimes C6 outflow(s).) 3. Meningoceles at some (usually lower) levels but not at all levels. a. Meningocele(s) strongly suggests but does not prove proximal root damage. b. Meningoceles at one or more levels suggest but does not prove proximal damage at other root levels without meningoceles. c. Absence of meningoceles does not prove lateral damage nor does presence of a meningocele always mean proximal damage or on occasion even damage at all.
MUSCLE & NERVE
September 1990
851
and tumors involving brachial plexus are complex and vary according to mechanism and level of injury. Careful delineation requires a lengthier treatise. Several examples of stretch
injury involving the brachial plexus are provided by Figures 2, 3, and 4. Some of the criteria for selection of stretch injuries for operation are provided in Table 3.
-
REFERENCES 1. Brown P: Factors influencing the success of surgical repair of peripheral nerves. Surg Clin North Ant 1972;52(5):11371155. 2. DasGupta T: Tumors of the peripheral nerves. Clin Neurosurg 1978;25:574-590. 3. Donaghy R: History of microneurosurgery, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGraw-Hill, 1985, pp 20-26. 4. Gilliatt RW: Physical injury to peripheral nerves: physiologic and electrodiagnostic aspects. Mayo Clin Proc 1981;56:361-376. 5. Haase J, Bjerve P, Semesen K: Median and ulnar nerve transections treated with microsurgical interfascicular cable grafting with autologous sural nerve. J Neurosurg 1980;53:73-84. 6. Hudson A, Hunter D: Timing of peripheral nerve repairs: important local neuropathological factors. Clin Neurosurg 1977;24:340-391. 7. Hudson A, Mayfield F: Chronic injuries of nerves by entrapment, in Youmans J (Ed): Neurological Surgery. Philadelphia, WB Saunders, 1983, pp 2430-2474. 8. Hudson A, Trammer B: B&hial plexus injuries, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGraw-Hill, 1985, pp 1817- 1832. 9. Kempe L: Operative Neurosurgery. New York, Springer-Verlag, 1970, vol. 2, pp 158- 170. 10. Khodadad G: Microsurgical techniques in repair of peripheral nerves. Surg Clin North Am 1972;52(5):11571166. 11. Kline D: Civilian gunshot wounds to the brachial plexus. J Neurosurg 1989;70:166- 174. 12. Kline D, Dejonge B: Evoked potentials to evaluate peripheral nerve injuries. Surg Gynecol Obstet 1968;127:12391248. 13. Kline D, Hackett E: Reappraisal of timing for exploration of civilian peripheral nerve injuries. Surgery 1975;78:5465. 14. Kline D, Hackett E: Management of the neuroma in continuity, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGraw-Hill, 1983, pp 1864- 1871. 15. Kline D, Hackett E, May P: Evaluation of nerve injuries by evoked potentials and electromyography. J Neurosurg 1969;31:128- 136. 16. Kline D, Hudson A: Complications of nerve injury and repair, in Greenfield L (Ed): Complications in Surgery and Trauma. Philadelphia, JB Lippincott, 1984, pp 695-708. 17. Kline D, Judice D: Operative management of selected brachial plexus lesions.J Neurosurg 1983;58:631-649. 18. Kline D, Nulsen F: Acute injuries of peripheral nerve, in J Youmans (Ed): Neurological Surgery. Philadelphia, WB Saunders, 1982, pp 2363-2429.
852
Surgical Repair of Nerve Injury
19. Lusk M, Garcia C, Kline D: Tumors of brachial plexus. Neumsurgely 1987;21:439-453. 20. MacKinnon S, Dellon A: Surgery of the Peripheral Nerve. New York, Thierne Medical Publishers, 1988. 21. McCillicuddy J: Techniques of nerve repair, in Wilkins and Rengachary (Eds): Neurosurgery. New York, McGrawHill Co, 1985, pp 1871-1881. 22. McQuarrie I, Idzikowski C: Injuries to peripheral nerves, in Miller T , Rowlands B (Eds): Physiologic Bmis of Modern Surgical Care. St. Louis, CV Mosby, 1988, pp 802-815. 23. Millesi H, Meissl G, Berger A: Further experience with interfascicular grafting of median, ulnar, and radial nerves.J Bone Joint Surg 1976;52B:209-218. 24. Millesi H: Reappraisal of Nerve Repair. Surg Clin North A m 1981;61:32 1-340. 25. Narakus A: The surgical treatment of traumatic brachial plexus lesions. Int Surg 1980;65:521-527. 26. Nulsen F, Slade H: Recovery following injury to brachial plexus, in Woodhall B, Beebe G (Eds): Peripheral Nerve Regeneration, VA monograph. Washington, DC, Government Printing Office, 1957, pp 389-408. 27. Omer G: The evaluation of clinical results following peripheral nerve suture, in Omer G, Spinner M (Eds): Management of Peri$heral Nerve Problems. Philadelphia, WB Saunders, 1980, pp 431-438. 28. Rodkey WG, Cabaud H, McCarroll H: Neuropathy after loss of a nerve segment: comparison of epineural suture under tension vs multiple nerve grafts. J Hand Surg 1980;5:366-371. 29. Samii M: Use of microtechniques in peripheral nerve surgery-experience with over 300 cases, in Handa H (ed): Microneurosurgery. Tokyo, Igaku Shoin Ltd., 1975, pp 8593. 30. Seddon H: Surgical Disorders of Penpheral Nerves. Baltimore, Williams & Wilkins, 1972. 3 1. Spinner M: Injuries of the Major Branches of Peripheral Nerves of the Forearm. Philadelphia, WB Saunders, 1978. 32. Sunderland S: Nerves and Nerve Injuries, ed 2, Edinburgh, Churchill Livingstone Ltd., 1978. 33. Terzis J, Dykes R: Electrophysiological recordings in peripheral nerve surgery. A review.] Hand Surg 1976;1:5266. 34. Williams HG, Terzis J: Single fascicular recording: an intraoperative diagnostic tool for the management of peripheral nerve lesions. P l a t Reconstr Surg 1976;57:562. 35. Woodhall B, Nulsen F, White J, Davis L: Neurosurgical Implications. Peripheral Nerve Regeneration. VA monograph. Washington, DC, U.S. Government Printing Office, 1957, pp 569-638. 36. Wynn Parry CB: Rehabilitation of the Hand, ed 3, London, Butterworths, 1978.
MUSCLE & NERVE
September 1990
