Nerve Architecture

Most peripheral nerves are mixed and contain motor, sensory, nociceptive, and autonomic fibers.¹ The autonomic fibers are responsible for sudomotor control, vasoreactivity, and piloerection. The main constituent of a peripheral nerve is connective tissue, which can vary from 85% in the sciatic nerve at hip level to about 25% for the peroneal nerve at fibular level.² The external sheath, the epineurium, is constituted from connective tissue with an amount of collagen and elastic fiber; it contains abundant blood vessels and protects the nerve against compression. The endoneurium provides connective tissue support for the nerve fibers. A certain degree of fiber undulation enables a physiologic amount of nerve stretch.^3,4 Nerve elongation below 12% of its resting length can be compensated for without functional damage, but venous flow is already blocked when a nerve is stretched by 8%, and a stretch of 16% produces ischemia.⁵ Attached to the epineurium is another capillarized outer gliding layer, by some referred to as the mesoneurium or paraneurium, that is traversed by vascular channels supplying the nerve. This mesoneurium also allows the nerve to glide between tissue planes during physiologic movement.^6–8 It is important to restore and retain gliding of injured nerve in order to prevent its tethering and entrapment.^4,9 The epineurium includes the internal epineurium or interfascicular epineurium, which ensheathes the fascicles. This is less compact than the external epineurium. Directly around each fascicle is the perineurium, which is constituted by perineurial cells in between circular, oblique, and longitudinal strands of collagen fibers. The perineurium is responsible for maintaining the physiologic milieu of the conducting elements, and it acts as a diffusion or blood-nerve barrier. Breaching the perineurium interferes with conduction and provokes demyelination of the contained nerve fibers.¹⁰ The role of perineurial fibroblasts in trauma appears to be underestimated because they initiate and maintain the connective tissue reaction and as such contribute to neuroma development.^11–13 Neuroma formation should be regarded as nondirected and a futile attempt at regeneration, where axons do not manage to reach the appropriate distal basal lamina and cannot reconnect to their natural target organ.

A further connective tissue layer is the endoneurium, which surrounds each myelinated fiber and groups of unmyelinated fibers. The endoneurial layer’s role in injury and repair is less well understood.

The vascular supply of peripheral nerves is abundant. At regular intervals, collateral vessels enter through the mesoneurium to connect to the nerve’s central artery.¹⁴ If a nerve is circumferentially mobilized, several of these “segmental” arteries are disconnected, and the central vessel maintains the blood supply.¹⁰ It is therefore possible to circumferentially mobilize a nerve over a relatively long stretch without putting the nerve at ischemic risk.¹⁰ Some of these extrinsic pedicles sustain vascularized nerve grafts, as in the case of the ulnar nerve.^15,16 Nerves, however, differ extensively in their vascular supply; the tibial, for example, is more richly supplied than the common peroneal nerve at knee level.¹⁷

Timing of Operative Therapy

The Schwann cells, the skin, and other target organs are rich sources of neurotrophins that are essential for the development, maturation, and maintenance of the cell bodies. Transection of a peripheral nerve leads to central cell death by deprivation of neurotrophins (i.e., nerve growth factor). Progressive neuronal death, ischemia, and fibrous proliferation are important limiting factors for useful recovery.³⁵ There is extensive evidence for progressive neuronal death to occur after a critical time window has passed. Sensory and motor neurons appear to differ in the onset of such cell loss. There is also evidence that early nerve repairs can stop this process of neuron loss.³⁶ Early repair improves the reconstructed nerve’s regenerative capacity. The efficacy of axonal regeneration is significantly affected by the amount of cell loss already present at the time of repair.^37–40 Central changes after peripheral nerve injury are more extreme in more proximal, more extensive, and more violent injuries and thus are an important factor in prognosis after repair.^22,41

When a nerve lesion is deep enough to interrupt the axons and to separate them from the cell nucleus, wallerian degeneration¹⁹ (anterograde degeneration) is the consequence. This is not a consequence exclusively of complete transection but also of profound crush or severe traction, as happens in violent blunt injuries. When there is no regeneration of axons into the distal stump, changes will occur in the target organ that over time become irreversible:^42,43 Motor end plates disappear, and the denervated muscle becomes fibrosed. Studies of central and peripheral conduction are of inestimable value in the analysis of injuries to the brachial plexus when these can be operated within 60 hours of the injury. This implies that injured nerves should be explored and repaired as soon as possible, unless there are realistic chances that the lesion bears enough potential for spontaneous regeneration to a functional level. In those cases, observation is justified. However, a favorable lesion might also progress to a less favorable one if the original cause is not addressed. Recovery is likely for the nerve accidently encircled by a suture or crushed under a plate if the cause is urgently removed. If that cause remains for days or weeks, a much more unfavorable lesion develops. One also cannot overemphasize the significance of the worsening pain and the deepening of nerve lesion caused by expanding hematoma or ischemia. The vulnerability of the lumbosacral plexus to expanding hematoma was discussed by Donaghy.⁴⁴ Nerves crushed in a swollen ischemic limb or in a tense compartment progress from conduction block to much deeper and much less favorable degenerative lesions.

Some crucial clinical insights regarding timing have been known for a long time. In 1908, Sherren examined 50 cases of acute suture performed at the London Hospital.^44a Pain sensibility recovered before touch sensibility; tactile localization continued to improve for more than 2 years. Recovery of power was rather slower. There was no recovery in only one patient, whose wound had become infected. Sherren recommended primary suture “because the prognosis after secondary suture is more unfavourable.”^44b In 1924, Platt and Bristow reviewed the late results of nerve injuries treated in the First World War and noted the “extreme perfection attained after so-called primary suture,” acknowledging that the more severe gun shot wounds of nerves were treated by secondary suture.^44c In 1937, Platt and Bristow Platt wrote, “in primary sutures performed under ideal conditions, complete recovery of motor power and recovery of protopathic sensibility at least, is to be expected … however, in more extensive wounds with widespread bruising and multiple tendon injuries, and in wounds in which infection had already secured a hold, partial or complete failure after primary suture is almost inevitable.”^44d Nine years later, Zachary and Holmes compared 55 cases of primary sutures of nerves referred to the Peripheral Nerve Injury Centre at the Wingfield-Morris Hospital, Oxford, England, during the years 1940 to 1944, with 36 “early” secondary sutures. The results of secondary suture were distinctly better. The primary repair was resected and the nerve resutured in 16 patients. Histologic examination of the resected material revealed a number of causes for the failure of the first operation: poor matching of the proximal and distal stumps; coarse suture material lodged between the stumps; separation of the stumps; and dense scar between the stumps or within the distal stump. Zachary and Holmes concluded, “Formal nerve suture should be undertaken at the earliest moment when it is possible to recognise the extent of damage to the nerve, excise the injured segment, and bring together the mobilised nerve ends without the prospect of undue post operative tension.” These early contributions came at a time when antibiotics were unavailable and when grafting of nerves was rarely performed in Great Britain.

Glasby used the freeze-thawed muscle graft (FTMG) in a series of experiments that provided evidence about the factors governing the quality of regeneration after nerve repair.⁴⁵ The findings were decisive. Regeneration was worsened by delay and also by association with arterial injury, fractures, or cavitation and hematoma.^46,47

Improved results after urgent repair have been reported for the median and ulnar nerves,^48–51 the radial nerve,⁵² the musculocutaneous nerve,⁵³ the sciatic nerve,⁵⁴ the common peroneal nerve,^55,56 and the closed traction lesion of the brachial plexus.^51,57 Kato and coworkers⁵¹ showed that recovery of function and relief of pain were decisively better in urgent repairs of the closed traction lesion of the brachial plexus. The spinal accessory nerve appears to be the one exception to the rule because it exhibits good recovery of function even after substantially delayed repairs. Uncomplicated open wounds and nerves can be left for 24 hours until an experienced surgeon can address them. When, however, a nerve injury is associated with damage to a major blood vessel, with an impending threat of peripheral ischemia, or with increasing pressure within a fascial compartment, delay is not permissible. The case is an emergency, as it is with cases of open fracture or fracture-dislocation.

Thus, the clinician must always bear in mind that the sooner the distal segment is connected to the cell body and proximal segment the better the result will be.

The primary limitation of very early repair is that there is a significant risk that even a very experienced surgeon will be unable to determine how much of the proximal and distal stump must be resected. Leaving in place any length of nerve that is irretrievably damaged is often the cause of failed repairs—such as those observed by Zachary and Holmes that had dense scar between the stumps. Because of this, when there is any component of blunt injury involved, a delay of a few weeks may be necessary.¹⁰ On the other hand, urgent operation enables detection of rupture of the perineurium, and studies of proximal and distal conduction ease the difficulties relating to the appropriate level of section.

The World War II experience with the urgent suture of nerves was dismal.^58,59 The results of delayed repair, in special units, were surprisingly good, and they fell a little short of modern series.^59A Urgent repair of nerves injured in current conflicts show promise.^66A

Classifications and Grading

Seddon distinguished among three types of nerve injury in his 1943 classification²⁰:

The implications of this classification with regard to chances for spontaneous recovery are seemingly straightforward, and the classification is easy to apply and remember. That is why it still is the most popular classification system used. In neurapraxia and axonotmesis, the prognosis for recovery is favorable, given that the cause is removed. With neurotmesis, however, there is no recovery without nerve repair. An important factor that frequently is forgotten by clinicians, who have discovered electrophysiologic signs of regeneration in an injured nerve, is that unless the cause of the lesion is removed, an initially favorable prognosis can change to a less favorable one (e.g., if an injured nerve is still compressed). And for this and various other reasons, “electrophysiologic signs of life” should never preclude surgical exploration, when it is necessary.

Sunderland’s classification system, which was introduced in 1951, is more elaborate and recognizes five degrees of severity.⁶⁰ A neurapraxic lesion is the equivalent to a Sunderland type I lesion. In principle, lesions with axonolysis (as equivalent to Seddon’s axonotmesis) are further differentiated between those that have an intact endoneurium (Sunderland type II) and those that have not (Sunderland type III). A type IV lesion denotes complete disruption of the contents of the epineurium, both connective tissue elements and axons, but with the nerve still in continuity, and a type V lesion is a severed or pulled-apart nerve that is not in continuity.

The question remains of what way classifications can actually really depict the clinical picture and as such have relevance for clinical decision making. In theory, falling in the continuum between need to graft a lesion and leaving the lesion alone after decompression and external neurolysis is a Sunderland type III lesion, in which there is considerable internal fibrosis that effectively precludes that enough sprouting axons overcome this soft tissue barrier to sprout distally in an orderly fashion. Some patients with Sunderland type III lesions recover, and some do not. This is why nerve action potential (NAP) recording is valuable for type II, III, and IV lesions. In this Sunderland classification threshold situation, there is axonolysis, the endoneurium is destroyed and fibrotic,⁶¹ and the perineurium theoretically could still be intact—and yet such a lesion has a poor chance for recovery without grafts.

In actuality, the different sectors of a lesion of the whole nerve can depict an array of injury grades, unless the lesion is neurapraxic or neurotmetic. Part of the lesion might show a gap (Sunderland type V) abutted by former nerve tissue that transformed into fibrous tissue (Sunderland type IV), and the rest of the nerve’s cross section might have axonolysed fascicles with fibrotic endoneurium (Sunderland type III). A tiny portion of the cross section, however, might conduct a compound nerve action potential because the axonal integrity of a minor portion of the fascicles is still intact. Thus the assignment of the whole nerve grades does not really reflect what should be done surgically in such a case (which would be a split repair after careful internal neurolysis in the case presented).

It helps clinical decision making with regard to surgery to simplify and distinguish a nondegenerative (focal conduction block) from a degenerative nerve^62,63 lesion with axonal division and ongoing wallerian degeneration (axonal discontinuity in all its forms: axonotmesis and neurotmesis, or the Sunderland equivalent spectrum from types II to V).⁶⁴

In the case of a complete nerve palsy in conjunction with vasomotor and sudomotor paralysis (dry, red, and warm skin) and a Tinel sign at the level of the lesion, a mere conduction block (neurapraxia, nondegenerative lesion) is highly unlikely. Neuropathic pain indicates partial destruction of a nerve’s integrity. A very vivid and strong Tinel sign at the level of the lesion points toward axon rupture.

Closed Traction Injury

A combination of the blunt mechanisms of stretch/traction and contusion/compression is a frequent cause of acute nerve injury. Traction can lead to rupture of nerve and artery. Most frequently affected by traction are the following:

If complicated by an arterial lesion, this group of injuries has an unfavorable prognosis.

The most common type of civilian nerve trauma most likely is a stretch-related injury in context with a motor vehicle crash.^66,74,75 Three percent (2.8%) of a Canadian trauma population of 5777 patients treated between 1986 and 1996 sustained associated peripheral nerve injuries (162 of 5777 patients).⁷⁴ Motor vehicle crashes were the predominant cause (46%). Eighty-three percent of the patients were male, and the mean age was 34.6 years. Seventy-five percent affected the upper extremities (121 of 162 injuries). The radial nerve was the most frequently injured nerve, with 36% (58 of 162 injuries), and the peroneal nerve was the most frequently injured lower limb nerve with 24% (39 of 162 injuries). Surgery was required in 54%. Brachial plexus injuries were identified in 1.2% of multitrauma patients presenting to the same major regional trauma facility (54 of 4538 patients seen at Sunnybrook Hospital, Toronto, Canada).⁷⁵

Open Injury with Transection

With open, complete lesions (Sunderland type V, neurotmetic), a primary reconstruction with an epineurial end-to-end suture is the objective, provided the injury is sharp. Most open lesions are caused by a knife, glass, or a surgeon’s scalpel wound.

In contrast, blunt injuries (Web Fig. 240-1), commonly caused by open fractures or penetrating missile injuries, are better approached with a delayed end-to-end suture. There is extensive tissue damage with a high risk for sepsis. Retraction of stumps is usually performed. If possible, it is helpful to gently approximate nerve stumps to each other or to adjacent structures at the first operation. The nerve stumps are readapted after resection of the stump neuromas within a 2- to 3-week interval. This form of intentionally delayed repair due to unfavorable soft tissue condition at the time of injury is termed delayed end-to-end suture.

WEB FIGURE 240-E1 Example of delayed secondary repair of blunt open sciatic nerve injury with soft tissue loss, infection, and compound femur fracture (“untidy wound”) after motorcycle accident. A, The primary wound before débridement and fracture nailing. B, Delayed secondary nerve reconstruction at middle third of the femur. Exposure of bluntly ruptured and retracted nerve, with nerve stumps on yellow rubber dam background material. C, Interpositional graft reconstruction. D, Wound closure with tissue substitute.

(A, Courtesy of the Trauma Unit “Weiden-Germany.”)

Unfortunately, the bulk of lacerating injuries (Web Fig. 240-2) are blunt and also have ragged, contused, and compressed margins, necessitating more nerve resection and graft interposition.

WEB FIGURE 240-E2 Primary repair. A, An 80% section of the femoral nerve after iatrogenic injury. The lesion was repaired 48 hours after the iatrogenic injury. B, A “tidy” laceration at the wrist that was repaired primarily.

(Copyright Rolfe Birch.)

Sharp transections can often be sutured end to end, and their prognosis is excellent.

Penetrating Missile Wounds

Penetrating missile wounds are blunt injuries and mainly result in contusion and stretch. Most are caused indirectly because the projectile trajectories do not strike the nerve in 85% of cases. However, many of the recited data in the literature have been carried over from older reports⁷⁶ and thus describe injuries caused by older and completely different firearms.^10,77–79 More current reports are related to the Serbo-Croatian wars and Middle Eastern combat activity.^66,66a,80 For example, brachial plexus lesions associated with complete loss of function have been demonstrated to benefit from surgery, particularly in those associated with injuries to C5, C6, C7, and lateral and posterior cord.^80,81 High-velocity missiles can have a direct impact but also can cause impact by the produced shock wave and by cavitation. A “near miss” creates a shock wave with instant high-pressure damage (explosion) on the way to the nerve, followed by low-pressure damage (implosion) on the way out.^10,81–84 A serious problem is the open wound. The damage with a near or close miss can range from mere neurapraxia to a complete degenerative lesion. Lesions that show no spontaneous recovery are thus usually explored with delay and then often need graft repair. Of course, pain syndromes and associated other injuries are a frequent concern. The Red Cross wound classification system is helpful for describing this type of injury.^66a,85

Iatrogenic Injury

A whole chapter of this edition is devoted to iatrogenic injury lesion complex (see Chapter 246), and there is a reason: in some centers, these lesions make up 20% to 25% of operated nerve injuries.^{67–70,86,87} Because of its importance for a nerve surgeon’s practice, some facts about this entity deserve reiteration. Despite the fact that cases of inadvertently injured nerves are not declining, and all mechanisms have been described, the appropriate treatment usually still is delayed. Treatment regimens should follow exactly those for traumatic nerve injuries of other origin. One element of delay certainly is that clinicians frequently embark on false hopes whenever a nerve injury has been induced through the hands of a physician. Whenever there is a wound over a nerve that is nonfunctional, chances come close to certainty that this nerve would benefit from exploration because it is probably cut.

History, Symptoms, and Signs

When and how an injury occurs are important aspects that guide our decision making. High-velocity injury; compound fracture and wounding; and accidental, criminal, or surgical history are likely to mean that there has been a serious lesion. The use of a knife, often enough in the hand of a surgeon, is an indication that a nerve is likely to have been partly or completely severed. The subclavian artery is ruptured in 10% of complete lesions of the brachial plexus and in as many as 30% of cases of violent traction injury of the infraclavicular brachial plexus. The incidence of arterial lesion is high after fracture-dislocations of the elbow and higher still after fracture-dislocations of the knee. Knowledge about the mechanism of an injury, the initial degree of dysfunction, and the course or lack of functional regeneration are most important aspects, which in conjunction with a detailed and systematic examination help us to localize and assign level and depth to an injury, when we cannot assess the patient initially. It is of utmost importance to judge the extent of injury, distinguishing between degenerative and nondegenerative injury, and to determine the length of the nerve and severity of the injury. The diagnosis of the extent of injury depends on history and signs on electrical examination and may involve the use of magnetic resonance imaging (MRI).

The inexperienced surgeon usually is enlightened by the opportunity to observe what a thorough and nevertheless quick systematic examination by an expert can yield with regard to exact branch localization, level and extent of injury, and potential for recovery. To detect the level, thorough knowledge of branching pattern and supplied muscles and sensory area is crucial. A helpful guide is a systematic table and scheme that is routinely used to document the findings. It is valuable to develop an individual systematic sequence of muscles to examine for each nerve, which usually follows the innervated areas and thus branches from proximal to distal. The actual steps in testing the individual muscles is something that is best learned from an expert. By far, the best guide to examination is still O’Brian’s Aids to the Examination of the Peripheral Nervous System.⁹¹ There are some “trick movements” that can compensate for loss of nerve and muscle function by activating other muscle groups supplied by different and more functional nerves.¹⁰ Thus, sometimes a complete lesion is deemed partial and followed instead of being repaired. Examples of trick movements include the following:

Muscle power is graded by most examiners according to the British Medical Research Council (BMRC) system, which has 5 grades of strength (0—complete loss of function, to 5—normal muscle power). This is documented individually for each muscle. Evaluation schemes that account for more functional aspects are useful, but they are also more complex. Such schemes need to be applied uniformly if crude interobserver differences are to be avoided. This usually requires documentation forms, which also list descriptions of the different functional grades.

The BMRC was developed by Highet for war wounds in 1943¹⁰⁰ and was later used in poliomyelitis. Thus, in the BMRC system, grade 3 is only contraction against gravity, 4 mild or moderate pressure, and 5 normal. In contrast, the Louisiana State University Hospital (LSUH) system¹⁰ grades are as follows: 1—observable contraction but not enough to overcome gravity, 2—against gravity only, 3—against gravity and mild pressure, 4—against gravity and moderate pressure, and 5—close to normal.

In the acute setting, the radial, median, and ulnar nerves are tested by asking the patient to form an O between the thumb and little finger, to give the thumbs-up sign, and to open and close the fingers like a fan.

Sensory loss is determined by response to light touch and pinprick and by the ability to localize stimuli.

Sensitivity and sympathetic function give precious clues to the completeness or extent of functional loss. Apart from weakness or paralysis of muscles, the early signs of nerve injury are alteration or loss of sensibility, vasomotor and sudomotor paralysis in the distribution of the affected nerve, and an abnormal sensitivity over the nerve at the point of injury. After severe injury of a nerve with a cutaneous sensory component, the skin in the distribution of the affected nerve is warm and dry starting within 48 hours of trauma. If possible, sensation to light touch and pinprick, vibration sense, position sense, and ability to localize stimuli should be tested and the affected area of skin recorded. Anhidrosis can easily be checked with loupes or an ophthalmoscope set on ±20 if in doubt. Warming of the skin, color change, and capillary pulsation in the fingertips indicate vasomotor paralysis. Ischemia affects the large fibers first, and thus discriminative sensibility and vibration sense are lost early.⁹² The palmar and plantar skin is scrutinized for changes in color and in sweating.

The Hoffmann-Tinel sign, as simple as it might be, is an effective means to detect the point of lesion and to monitor, or more likely rule out, any progress of recovery (see earlier).

A painful nerve is injured or compressed, or both. The occurrence of pain after injury often means that the noxious process is continuing (Web Fig. 240-3). Acute neuropathic pain is characterized by loss of sensation; by painful, spontaneous sensory symptoms throughout the nerve’s territory (dysesthesia); and by lancinating or shooting pain irradiating into the distribution of a main nerve. A constant crushing, bursting, or burning pain in the otherwise undamaged hand or foot indicates serious and continuing injury to major trunk nerves. Progression of sensory loss with a deep bursting or crushing pain within the muscles of the limb, often accompanied by allodynia, can indicate impending critical ischemia. A regular feature of injury caused by critical ischemia is neurostenalgia, which indicates continuation of the noxious process and sometimes also deepening of the lesion. Partial nerve injuries are at times excruciatingly painful. Causalgia is uncommon, and it responds well to the correct operation. Deafferentation pain is related to the death of neurons on the dorsal root ganglion (herpes zoster) or to lesions of the dorsal root of the spinal nerve.^90a

WEB FIGURE 240-E3 Bone fragment (small square and circles) that has been dissected out from brachial plexus at the retroclavicular and infraclavicular level months after the accident. The fragment maintained a “neurostenalgic picture.” The patient was complaining of ongoing severe neuropathic pain since the accident with an electrifying element, without typical features of deafferentation and without radiologic suggestion of root avulsion. Despite the clinical picture of a severe extended upper brachial plexus lesion (C5, C6, and C7), the patient was referred with considerable delay.

Joint function is extremely important to assess. The passive range of motion must be assessed and recorded.

Any soft tissue or bony injury, tendon avulsion, muscle tear, or associated vascular injury needs to be assessed and documented or ruled out.

Electrophysiology

The study of nerve conduction advises the clinician about the health of axons, about their myelination, and, when applied to the proximal stump of a nerve, about whether there is continuity between the exposed nerve and the spinal cord. Conduction across a nerve lesion indicates that at least some of the axons are intact.

After transection of a nerve, axons become inexcitable, and neuromuscular transmission fails. Direct stimulation of the nerve distal to the level of lesion elicits no response chronically. However, some conduction is maintained for some days after nerve transection. Fibrillation potentials appear as muscles are denervated, but their onset depends on the distance between the site of nerve lesion and the muscle, so there may be an interval of 2 to 3 weeks before fibrillations are seen. The reappearance of voluntary motor unit potential activity indicates that reinnervation is taking place, and the electromyographic evidence of this usually precedes clinical evidence of recovery. However, it is important to understand that “some recovery” is often not good enough to restore function. The finding of a few motor units showing reinnervation even at an early stage after injury does not imply that full recovery of a nerve will take place. In analysis of incomplete lesions of large nerve trunks, the clinician may be lulled into a sense of false security by electrodiagnostic evidence of an incomplete lesion. Such evidence should not be taken to imply that full recovery could be anticipated. Unless a nerve has been completely transected, it is likely that there will be mixed elements of neurotmesis, axonotmesis, and prolonged conduction block.

After nerve exposure, electrodiagnostic work is of inestimable value to assess whether a lesion in continuity has a chance for spontaneous recovery or will fare better with graft repair.^10,93,94 Somatosensory evoked potentials (SSEPs) can help to check for root avulsion.^95–97 NAPs for significant conduction require about 3000 to 4000 fibers of at least moderate size and some myelination across a lesion in continuity.^10,73

Imaging

The role of imaging in nerve trauma is to rule out root avulsion,^95,98,99 musculotendinous injury or tear (MRI, ultrasound), or bony injury (computed tomography [CT]), or to assess implanted metal (radiograph) in secondary nerve reconstructions. Ultrasound will gain increasing importance as an excellent means to assess whether a nerve has been torn apart or a neuroma-in-continuity has formed. Modern high-frequency ultrasound devices allow one to recognize fascicles within the nerve and, even more so, the lack of fascicles in case of internal scar (neuroma). However, fascicular continuity does not ensure, although it does favor, axonotmesis and future recovery of function.

Ultrasound also is of invaluable help in depicting the course of the nerve in cases in which transection and dislocation are anticipated. We predict that this means of examination will soon be part of a routine surgical nerve practice.

Indications for Operation

The primary reasons to operate on a nerve include confirming or establishing the diagnosis, restoring continuity to a severed or ruptured nerve, repairing a lesion in continuity that has no chance for functional recovery, and relieving a nerve of an agent that is compressing, distorting, or occupying it. The decision for operating is usually straightforward in the acute case of an open wound or when the nerve injury is associated with damage to long bones, joints, and blood vessels. The indications for operating on nerves after an injury include the following:

This list probably covers most contingencies except that of delay between injury and consultation. Evidently, there is a time limit for successful intervention, except perhaps in the case of persistent pain. Also of note are the following indications:

The aim of the operation is to preserve or restore function. When a nerve is severed, repair offers the only opportunity to do this, and the repair may be done by suture or graft depending on the findings. If the proximal stump is hopelessly damaged or if there is no continuity between the proximal stump and the spinal cord, transfer of other nerves to the distal stump may be possible. If the distal stump is hopelessly damaged, direct implantation of nerves into the muscle (muscular neurotization) may be possible. When the neurologic injury is, by whatever means, irreparable, palliation may be achieved by musculotendinous transfer or other reconstruction.

The sooner the distal segment is reconnected to the cell body and to the proximal segment, the better the result will be. For the extreme case of replantation after traumatic amputation, O’Brian⁷² has indicated that primary suture of nerves provides the only chance of recovery.

Reasons not to proceed to repair of a transected nerve include the following:

Concerning nerves injured in the arm or the elbow, Seddon¹⁰⁰ thought that recovery could be awaited if two conditions were met: (1) reasonable apposition of bony fragments, and (2) “complete certainty that there is no threat of ischemia of the forearm muscles.”

A strongly positive Hoffmann-Tinel sign over a lesion soon after injury indicates rupture of the axons. The senior author (RB) regularly finds this on the day of injury in closed traction rupture.

Recognition of Extent of Injury

An argument that has been used far too often when nerve exploration and repair has been delayed in an undue fashion is that of potential for spontaneous recovery.

Common sense, critical evaluation of injury mechanism and involved impact, as well as associated injuries, in conjunction with a thorough examination supported by the more simple electrodiagnostic studies, more often than not reveals that the severity of the lesion precludes useful spontaneous recovery.

Severance of a nerve with a cutaneous sensory component leads to well-defined loss of sensibility and to complete motor, sudomotor, and vasomotor paralysis in the distribution of the nerve. Conduction block is more likely to result in a patchy motor loss. Further, it is more likely to affect large axons more heavily than small ones. As such, vibration sense and sensibility to light touch are likely to be impaired, whereas pain sensibility may be unaffected. In the case of conduction block, axons are intact, and stimulation distal to the lesion will elicit a motor response. If, in contrast, the axons are damaged (degenerative lesion), stimulation distal to the lesion more than 6 days after the injury will not elicit a motor response.¹⁰¹ A diagnosis of neurapraxia cannot be made unless that time has passed after the injury.

Primary Repair: Urgent Surgery

Definition of Terms

The term primary or acute repair is reserved for operations performed within 3 to 5 days of injury, and delayed secondary repair is reserved for those done between 5 days and 3 weeks. Some resection, of 1 mm or so of the nerve stump, is always necessary after the passage of a few days, even in cases of clean section with knife or razor. Conduction in the distal trunk disappears after about 3 days, and things become more difficult with the passing of every day.

The earlier the exploration is done, the easier the diagnosis is. Not only is the field free from scar tissue, but also the axons of the distal stump continue to conduct and the neuromuscular junctions continue to function for a few days. Stimulation of the distal stump produces a motor response, which permits the surgeon to identify bundles with a predominantly motor function. This is nowhere better shown than in the case of the brachial plexus (Fig. 240-3A); it is easy to diagnose rupture up to 3 days after injury but progressively more difficult after that time. With early operation, it is possible to match fascicular arrangements of the stumps; as time goes by, the matching becomes progressively more difficult. This is not all: with delay, there is progressive intraneural collagenization (Fig. 240-3B).

FIGURE 240-3 A, Early plexus exposure (within 72 hours). In early exposures, it is easier to separate ventral and dorsal roots because maintained distal conduction matching of motor and sensory fascicles is possible. The lack of scar facilitates fascicle match. B, Late exposures. The depicted brachial plexus injury is completely scarred more than 2 months after injury, so diagnosis is more difficult. Depictions of a sciatic nerve and an ulnar nerve give further examples of scarring found at late exposure.

(Copyright Rolfe Birch.)

Whenever a nerve in a clean wound is sharply transected, an attempt at primary or acute repair should be made.

From 1956 on, primary or relatively acute suture of severed nerves was practiced at St. Mary’s Hospital whenever it was possible. There, in 1965, Michael Lawrence achieved an outstanding result from primary repair of all flexor tendons, radial and ulnar arteries, and median and ulnar nerves sectioned at the wrist. The policy was extended to revascularization of damaged limbs and to replantation of amputated hands and severe injuries of the brachial plexus.¹⁰⁴

Delayed primary repair after 2 to 3 weeks will still yield a good result (the term early secondary repair is used synonymous by other authors). However, the advantages of dissection in unscarred tissue, incomplete nerve retraction, and preserved distal conduction for matching individual fascicle bundles will be lost by that time (motor versus sensory fascicles).

The main advantage of early brachial plexus exploration and repair within 60 hours remains the prospect of better functional gain with primary repair compared with secondary repair of brachial plexus traction injuries.

One disadvantage of potential urgent nerve repair is that the clinical condition of the patient might preclude nerve repair measures and another is that diagnostic studies to rule out root avulsion might not be available or feasible, in which case one has to rely on SSEP studies and other intraoperative findings.

The major problem is that 40%, and in some series more, of stretch-traction injuries to the C5, C6 portion of the brachial plexus recover spontaneously. This figure reduces to 15% for C5, C6, C7 stretches and to 5% for flail arms. When recovery is evident, it usually takes 3 to 4 months for such to be evident clinically or by electromyography. Infraclavicular stretch-contusion injuries may also recover spontaneously, so exploration, when continuity is suspected, is best done at 3 months.

Having said that, most of the traumatic adult brachial plexus lesions unfortunately come to the attention of the nerve surgeon only with considerable delay, a fact that profoundly influences and worsens functional outcome.¹⁰³

Contraindications to Early Repair

Chief among contradictions to early repair is the general condition of the patient. The treatment of injuries that threaten life or limb must always come first. Soft tissue damage and wound contamination may be so severe that it is wiser to wait until the skeleton and the soft tissues have been stabilized before dealing with the nerve. It is helpful to mark the nerve stumps with sutures.

Of course, there is no need to explore the nerve if the clinician expects spontaneous recovery. Seddon had this to say about nerve lesions complicating closed fractures of the shaft and the lower end of the humerus: “however, on balance, it would seem proper to await spontaneous recovery of the nerve, provided the two conditions are satisfied. The first is reasonable apposition of the bony fragments, and the other complete certainty that there is no threat of ischaemia of the forearm muscles.”¹⁰⁰ If however, the clinician elects to convert a closed fracture to an open one by internal fixation, it is wisest to expose nerves that are not working. The fracture surgeon who does not do this is asking for trouble.

Aspects of Diagnosis

On the first encounter, the history is taken from the patient, with particular attention to mechanism and course of injury. The patient should be asked to express and describe the main problem as a consequence of the injury (e.g., pain, motor deficit, handicap). The injury and deficits should be related to the patient’s occupation, and the patient’s expectations need to be comprehended to put them in perspective. After such a conversation, one should already have a provisional diagnosis.

A thorough history enables the clinician to focus the examination. The affected nerves are evaluated with particular attention to level of injury, severity of the lesion, and associated injuries with a focus on joint mobility and passive range of motion, fractures, vascular status, soft tissue loss and contractures, and previous or ongoing infections. Trick movements need to be ruled out so that functional recovery is not falsely attested when in fact the lesion is complete and needs to be repaired.

During the examination, tendon avulsion and muscle tear must be ruled out; classic examples are upper trunk brachial plexus or circumflex nerve lesions with associated rotator cuff injuries that need to be addressed, along with long biceps tendon tears.

Assessment of associated injury and vascular lesions is of particular importance in brachial plexus lesions because associated injuries are frequently encountered here. The prognosis is worse, with severe associated long bone fractures and vascular lesions especially in brachial plexus injuries because this is an indirect sign of a high impact of the initial injury. Again, ideally the nerve would have been repaired at the same time as the associated injuries. However, frequently we do not have this opportunity and still want to improve the patient’s situation and hence repair the nerve.

The clinician needs to be well prepared and warned if there has been prior vascular injury or reconstruction because chances are high that one will encounter exactly this segment of a reconstructed vessel. It is good practice, therefore, to document whether a reconstructed vessel is actually patent. Vascular reconstructions in the emergency setting are not always successful. This implies that a reconstructed subclavian or axillary artery might in fact not be patent, and the extremity is perfused by collateral supply. It is not a good idea to inadvertently cut through already flimsy collateral supply of an extremity while dissecting the nerve scar. In such cases, we like to have a good angiographic study before we reexplore. If the reconstructed vessel is buried within the nerve scar, one needs to be prepared to potentially also manage vascular injury and necessary reconstruction.

Web Figure 240-3 demonstrates a bone fragment that we dissected out of a patient’s retroclavicular and supraclavicular brachial plexus during secondary reconstruction. The patient had complained of severe, electrifying pain shooting down into the hand at times and also locally, without having had signs of root avulsion (classic sign of impalement and also of the type of pain that occurs when a suture is inadvertently placed in nerve). It is desirable to have detailed reports about the early and preceding procedures and encountered problems. It is also necessary to inquire about other surgery and to coordinate this with the other surgical teams, especially in secondary nerve repair of complex trauma cases.

In making a decision about supplementary examinations, the diagnosis and current status of recovery should be confirmed by NPI. If root avulsion has to be ruled out, postmyelographic thin-slice CT is necessary, unless a devoted radiologist can perform sequential MRI root levels at the spinal root entry and exit zones.⁹⁵ If there is a bony element and fractures have been addressed at prior surgery, current radiographs and CT scans are reviewed to assess whether bone needs to be redressed or whether plates, screws, or wires can be removed during nerve exploration. One certainly does not like the idea of having the first-year trauma resident remove the “minimally invasive” metal after one has done a nerve reconstruction coursing over a plate (e.g., radial nerve at humerus). Sometimes, bony exostoses and overzealous callus formation need to be drilled or rongeured off, and we have encountered median, ulnar, radial, and suprascapular nerve and brachial plexus elements that had to be freed and drilled out of bone.

After conclusion of the preoperative work-up, the patient should be provided with a diagnosis and potential prognosis for an attempt at repair. This should consider the initial impact of the injury, the patients’ age, time delay, level of injury, and associated injuries. It is important to relate this evaluation to the patients’ expectations to avoid falsely raising the patient’s hopes.

The treatment plan includes other necessary surgery, which should be done either at the same session before grafting or at a separate operation before nerve reconstruction. Plan surgery timely, taking into consideration that the overall time delay will be the time between trauma and arrival of the axons at the target organ, after the nerve repair. We organize for follow-up examinations beginning at 6-month intervals if the nerve has been grafted. We follow the patients for at least 3 years but make clear that changes can be observed for up to 6 years or longer.

Aspects of Microsurgical Nerve Repair

The object in nerve repair is to accurately coapt healthy conducting elements without tension. In practical terms, this means accurate coaptation of the bundles.

When a decision to reconstruct has been made, the nerve ends are cut back progressively until healthy pouting bundles show in the cut surfaces. Resection is less in early primary repairs. No more than 1 to 2 mm of nerve is removed in tidy wounds. Finding the right level of section in the closed traction rupture or in the untidy wound with transection is much easier in urgent repairs in which there is still conduction in the distal trunk. In ruptures of the spinal nerves, a stimulator is moved slowly from the rupture towards healthy tissue until an SSEP becomes detectable (proximal stump) or until muscular activity returns (distal stump). This usually coincides with a healthy looking nerve face. In traction ruptures, the amount of tissue resected is usually between 5 and 10 mm. More nerve must be resected in late cases. When the case has been complicated by infection, as much as 4 cm of both stumps are irretrievably fibrosed. Ideally, it is best to match bundle to bundle, sensory fibers to sensory fibers, and motor fibers to motor fibers. It is often easy to match bundle to bundle simply by looking at the nerve end under the microscope or with the loupes, although the changing architecture of the nerve along its length makes this difficult when a long gap has to be bridged.

The technical aspects of nerve repair are detailed in Chapter 244; certain aspects are emphasized here.

General Principles and Use of Landmarks

It is paramount to have a thorough understanding of the normal anatomy, if one approaches distorted and scarred anatomy. As always in peripheral nerve surgery, good use should be made of anatomic landmarks. It is important to approach the lesion from virginal tissue and planes that are not scarred. This is why the lesion is approached from proximally and distally in a way that the nerve’s pristine ends are dissected and followed toward the lesion. This prevents inadvertent laceration of the nerve segments that are buried in scar. When the two “good sides” are dissected out, a much better understanding of the nerve’s current course can be developed. Frequently, the nerve will be completely displaced from its natural location. For lesions in continuity, it is especially important to not cut inadvertently into nerve during scar dissection. Often, nerve needs to be sculpted out of scar, such as by use of a 15-blade knife and scissors. Vascular loops or Penrose drains for larger nerves can help to lift the already developed nerve parts up in order to ease circumferential dissection. This also explains why these approaches usually necessitate larger incisions. Working in planes facilitates overview and prevents digging deep holes that necessitate aggravated retraction. When the nerve ends are completely dissected out, one has to resect back to normal fascicular tissue on both nerve ends by use of a razor blade or a fresh 15-blade knife. The sequentially cut nerve slices will show a gradual change from scar, to scar with some fascicles, and then to completely normal fascicular tissue, with a glossy, moist cross-sectional area and pouting fascicles. If in doubt, a frozen section can help to rule out scar tissue in the section.¹⁰⁵ This is usually not necessary in the periphery but sometimes is more difficult at a brachial plexus spinal nerve level when the section comes close to, or is within, the unroofed foramen.

Decompression

The term decompression refers to the release of a nerve from an externally compressive structure with a macroscopic technique. An example would be to transect a constrictive fibrotendinous band. This should be differentiated from external neurolysis (alternatively termed neuroplasty), which implies a direct manipulation of the nerve itself because the tissue released is directly adherent to the nerve (scar).

Neurolysis

In external neurolysis (or neuroplasty), the nerve is set free from scar, organized hematoma, or bony fragments and callus. Usually, the nerve is released out of such an encasement in a circumferential manner. The epineurium is minimally breached with external neurolysis. This is a partial step to prepare the nerve for recordings and reconstruction if needed. It is done macroscopically and microscopically, depending on the findings. A step further, which at times is necessary, is internal neurolysis (or neuroplasty) (Fig. 240-4), which implies opening of the epineurium, or sometimes resection, in order to lyse certain fascicles or fascicle groups out of a scarred bed within the nerve. The dissection is within the interfascicular epineurium. As such, it is important to avoid damage not only to the fascicle but also to the ensheathing structure of functional fascicles, which is the perineurium. Damage to the perineurium leads to functional loss of the contained fascicle, or fascicular bundle (depending on the nerve, it can be monofascicular, oligofascicular, or polyfascicular). This step needs to be done with proper illumination and magnification. Internal neurolysis is used in the preparation of nerve ends for grafting, in the dissection of a neuroma-in-continuity (in which the neuromatous tissue is dissected away from intact fascicles), and in benign nerve sheath tumors (schwannomas and neurofibromas).

FIGURE 240-4 Internal neurolysis of the right femoral nerve at suprainguinal and infrainguinal levels. An iatrogenic nerve lesion is shown after “uneventful” hip arthrodesis, which resulted in unresolving quadriceps paresis and pain of nerve origin with an element of dysesthesia and algohypesthesia. Intraoperative findings demonstrated a partial nerve lesion with internal scarring and signs of a previous contusional (versus stretch) lesion. The picture sequence demonstrates several microsurgical principles facilitating effective and clean dissection. A, Proper positioning to “open” the inguinal fold, and use of a ring retractor system enabled an approach through a horizontal incision in the bikini line (versus traditional vertical incision crossing the inguinal fold). B, Use of sequential retraction with suprainguinal and infrainguinal exposure without the need to transect the inguinal ligament. C, Key to microsurgical maneuvers of any kind are a comfortable position, perfect image quality (illumination and magnification), and proper hand rest. D, View after lifting the inguinal ligament above the femoral nerve (above the Penfield rubber dams). E, A small area of internal scarring was detected, which was lysed using a diamond knife. F, Femoral nerve after partial internal neurolysis. Thicker fascicles were conducting nerve action potentials, and there was no indication for a split graft repair.

Nerve Suture and Grafting

If tensionless end-to-end suture is possible, the fascicles and the fascicular groups of the two stumps should be matched as precisely as possible. The longitudinal course of arteries in the epineurium greatly helps with correct alignment of the two stumps, if it is a fresh and smooth injury without a contusion element. The coaptation is preceded by preparation of the nerve stumps. Usually, despite sharp cut ends, both of the stumps need to be cut back for 1 mm or less to healthy-appearing fascicular structure, the hallmark of which is a glossy, moist-looking surface with pouting fascicles. The coaptation is started with two opposing lateral sutures, completed by three or four more sutures. If after nerve stump preparation and cutting back, a tensionless repair is not feasible, the defect is bridged with interposed autologous grafts.¹⁰⁶ The grafts are sutured to the epineurium or individual fascicle (groups) of the recipient depending on the nerve caliber, type, and location. The nerve ends are prepared so that the fascicles protrude and are laid in the prepared bed orthodromically as a rule. Fixation is by glue or suture. The sutures unite the fascicles of the stumps with the grafts. Because the individual strands of the grafts are often the same size as the individual bundles, the suture unites epineurium of graft to perineurium of bundle. No discernible difference in outcome could be demonstrated between epineural and fascicular (or perineural) sutures.¹⁰⁷ Orgell¹⁰⁸ described a modified fascicular suture termed group fascicular suture and concluded that epineural suture was the “technique of choice for most acute nerve lacerations.” He pointed out that it was easier and faster and entailed less manipulation of the internal structure of the nerve than did fascicular suture. Kline and Hudson indicated a restriction of fascicular suture to “some oligofascicular nerves.”¹⁰ This is in contrast to Millesi, who prefers interfascicular suture and proposed that the epineurium was the main source of the infiltration of the suture line by fibroblasts.¹⁰⁹

We prefer to combine bundle (fascicular) suture with epineurial suture on an individual basis, guided by the nerve stump structure. Dissection within the nerve is avoided because this surely leads to fibrosis. In the early days after division of a nerve, bundles are mobile within the epineurium, and epineurial suture increases the risk for malalignment.⁵⁰ That is why the needle is passed through the condensed inner epineurium and the perineurium to secure accurate coaptation of larger bundles. The repair is completed by epineurial suture. In delayed repair, fibrosis within the epineurium stabilizes the bundles so that they cannot rotate within the epineurium. In these, epineurial suture may be adequate. The atrophy of the distal stump and the extent of fibrosis in both proximal and distal stumps increases the difficulties in neglected cases. In both primary and secondary suture, the areolar adventitial tissue is pushed back from each stump to expose the true epineurium.

Tensionless suture repair, irrespective of interfascicular or epineural graft coaptation, is the crucial component for successful microsurgical nerve suture. Trumble advised narrowing the gap by careful mobilization and by joint positioning.^101a There are reported series of sutures of digital nerves with some tension.^{102,110–112}

The principle steps of autologous interfascicular nerve grafting are removal of nerve stump neuromas of the recipient (Fig. 240-5) and cutting back to normal fascicular structure on both nerve ends; these are dependent on the findings and individual technical preferences for dissection of fascicle groups that match the individual graft pieces, followed by sequential insertion and coaptation of the grafts.

FIGURE 240-5 Interpositional grafting procedure in peroneal nerve repair. A, Nerve gap after dissection of stumps. B, Cutting sequentially back to fascicular structure in proximal stump, a complete neuromatous slice (white star) is grabbed by forceps. C, Positioning, alignment, and in situ shortening of the first autologous, interpositional graft.

The extent of resection of a damaged stump in closed traction rupture is remarkably small if done primarily and is sufficient once an orderly and recognizable architecture of bundles is displayed. To dissect every fascicle out, as once was proposed, is not advised because this will create unnecessary additional damage to the recipient nerve and potentially stimulate fibrosis. Creating “fascicular fingers” in the form of a grouped fascicular repair sometimes enables a better match and alignment. A freshly cut or “raw” muscle is not a good bed for grafts, but fat flaps and synovium are. The technique to be applied at times even differs between proximal and distal coaptation side and is adjusted to the need for maximal, tensionless alignment. The objective, which should not be forgotten, simply is to have as many guiding rails for sprouting axons at disposal as possible. At the same time, these rails need to be reachable by external vascular blood supply from the wound bed; a perfect match is not feasible, but too generous use of interfascicular dissection risks increased fibrous reaction and scarring.¹¹³ Many sprouting axons will be misdirected and get lost on their way.¹¹⁴ The more rails that are available, the better the chances will be for axons to reach the target organ and thus for good functional recovery.

Interposition of an autologous nerve graft is achieved by one or multiple strands of a cutaneous nerve. The senior author (RB) uses nerves from the damaged upper limb whenever possible. If suture is not substituted by glue, the surgeon should attempt to minimize the number of sutures needed. When the graft is splayed out perfectly and not contorted along its long axis (in other words, well aligned), the adhesive forces cause the donor to stick to the recipient. In such cases, only one or two securing sutures are usually needed on each end.^115,116 When using glue, a clump should be avoided because it can interfere with revascularization. Young and Medawar introduced the concept of fibrin suture in 1940 in rabbits, after a sutureless fibrin glue technique for securing union between divided intracanalicular facial nerve ends had been described as early as 1932 by Ballance and Duel.¹¹⁷ Narakas had popularized and refined its application for nerve repair in the 1980s.¹¹⁵

More frequently, however, the damaged nerve’s cross section is much larger than that of the donor or donor bundle. In these cases, as many grafts as necessary are interposed to cover the recipient’s cross section. Unfortunately, because of the restricted overall graft length, this is not always feasible. At times, fish-mouthing the graft’s tip to cover a larger recipient area is a compromise.

When only a segment of the nerve’s cross section needs to be grafted because of a partial complete nerve lesion, a split repair is done. Because the nerve is still held in continuity, it is easier to fit nerve autografts of exactly matching length; their coaptation is easier because usually the ones on the inner circumference are splinted and held in place by the intact nerve portion, which means not all of them need to be fixed with a suture. Those on the outer circumference can be secured with one stitch from the epineurium or fascicle group on the recipient side to the epineurial graft side. Fibrin glue can be an excellent alternative. The surgeon can dissect “fingers” of fascicular bundles for each graft strand on the recipient side, or just leave the recipient portion undissected after it has been freshened (cut to pouting fascicular structure). In general, it is always a good idea to manipulate intact nerve tissue as little as possible in order not to create additional trauma with potential for fibrotic reactions. Overzealous nerve suture is never a good idea because every suture induces fibrosis. The key is to have the grafts spread out and connected in perfect alignment, in an attempt to prevent contortion. The more challenging part of the split repair is to dissect the neuromatous portion away from functional fascicles and to prepare the nervous graft bed by excising the neuroma; for this, an interfascicular dissection is needed, which has a substantial inherent risk to damage the functional portion of the injured nerve. Interfascicular dissection is achieved with a fresh blade or a diamond knife under microscopic magnification.

The prevailing principle for all graft coaptations should be to place as few sutures as possible, but as many as necessary to ensure a persistently correct orientation. We do not use a suction drain in proximity to a nerve suture or graft because it jeopardizes the grafting site.

A clean wound bed, in which hemostasis has been achieved without destroying vascularity by charring the tissue, provides the best “soil” for graft revascularization. Gliding planes and pedicled fat flaps should be maintained whenever possible to prevent adhesions. In addition, this approach prevents the creation of cavities, which are predestined to fill with seroma and hematoma and also increase the risk for infection.

Harvesting Graft

The sural nerve is used frequently^118,119; other cutaneous nerves that are in the operative field can be harvested provided that the produced donor site deficit is acceptable (e.g., medial cutaneous nerve of the forearm). The lateral cutaneous nerve of the forearm (LCNF) is found just lateral to the biceps tendon, and it can be displayed, in the lower part of the arm, between the biceps and brachialis muscles. About 15 cm of graft is available. The medial cutaneous nerve of the forearm (MCNF) is taken from the arm through a curvilinear incision. About 30 cm of graft is available, and this is the nerve of choice because the donor defect is so trivial. The terminal part of the posterior interosseous nerve has been used.¹²⁰ The superficial radial nerve and lateral cutaneous nerve of the forearm should only be used when the parent nerve is irreparably damaged or when C5 and C6 are ruptured or avulsed. The sural nerve offers up to 45 cm of graft, but it should not be used in the ipsilateral tibial nerve in the leg because this seriously adds to the loss of sensation in the foot and ankle; it is better to use the MCNF. It is important to always avoid the formation of a painful neuroma, and donor nerves should be cut cleanly deep to the deep fascia of the limb.

If the cutaneous nerve caliber matches that of the nerve to be reconstructed (often with monofascicular or oligofascicular nerves), one interposed graft can be bridged with epineural sutures, after the mesoneurium and soft tissue have been cleaned from the recipient nerve ends. Various methods for harvest of the sural nerve have been proposed: single long longitudinal incision, multiple stair-step incisions, and more recently, endoscopic methods.^{118,121–124} The major drawback of minimal access methods is that usable branches may be sectioned and not removed as graft material. With a longitudinal incision, one can harvest under direct vision and in the prone and supine position. Certainly, it is also possible to use several transverse incisions to prevent a large open wound. The latter technique potentially exposes the nerve to the risk for traction injury and of damage to communicating branches from the common peroneal nerve. When maximal lengths need to be harvested (sural nerve), we like to lift the skin at the proximal and distal incisional ends and gently grasp the proximal nerve, cut distal to it (nontraumatized end), and keep the proximal end grasped to be able to coagulate its free fascicles in an attempt to “seal” the end. The proximal end of the cut nerve needs to be buried deep to fascia to prevent neuroma pain. We routinely make use of a longer-lasting local anaesthetic directly at the nerve end and along the skin incision. In using a long incision, it is essential to maintain meticulous hemostasis and to close layer by layer (the proximal part for example will be underneath gastrocnemius fascia). A bandage from foot to below knee will help to prevent subcutaneous hemorrhage. At the end of the operation, the lower leg is slightly elevated to ease venous drainage and prevent swelling. After its excision, a donor site defect in the form of a hypoesthetic area of variable size at the lateral aspect of the lower leg and foot will remain, mainly around the area posterior and anterior to external malleolus and at the proximal lateral aspect of the forefoot. With time, the area will diminish slightly in size owing to collateral branch reinnervation. The patient needs to be aware of the foreseeable donor site deficits.

The grafts, tenderly handed, are laid between saline-soaked swabs and are later denuded of excess mesoneurial tissue. After measuring the gap to be filled, the nerve is sectioned into appropriate lengths with a new blade or with vascular scissors.

Tunneling of Grafts

Nerve grafts can easily be tunneled between two skin incisions if longer stretches need to be covered and it is not desired to enlarge the skin incision (e.g., a formerly open wound that had been superficially compromised before; Web Fig. 240-4). In such a case, the grafts can be bundled, sewn together to a vessel loop, and carefully pulled through the tunnel. For shorter distances, it is sufficient to bluntly create a tunnel and pull a vessel loop with the grafts attached by suture. The sutured graft tips need to be resected once they are in place.

WEB FIGURE 240-E4 Secondary repair in a case of right upper extremity amputation below the shoulder. The accident was due to detachment by a train wheel, following an inadvertent fall during a train show. Successful outpatient initial limb replantation and vascular anastomosis (trauma unit at the Städtische Klinik Karlsruhe in Germany) was complicated and followed by wound infection (“untidy wound”) and severe lymphedema. Exploration and grafting were performed using several incisions and tunneling of grafts in order not to compromise soft tissue and lymph drainage further across the previously compromised areas.

Gaining Nerve Length with Transposition

Sometimes it is helpful to gain nerve length, thus shortening the potential gap that needs to be covered, by transposing the nerve out of its natural course. Examples are anterior transpositions of the radial or ulnar nerves, which can add up to 3 cm.

Nerve Transfers

In principle, an uninjured nerve is used as an axon donor and is therefore transected, usually distally, and transferred to the distal stump of an injured nerve to reinnervate the recipient nerve’s target organ (Fig. 240-6). The first transfers were described decades ago^125–127: Harris and Low suggested the principle in 1903; Tuttle¹²⁸ recommended the use of the accessory nerve and portion of the upper cervical plexus in 1913, Vulpius and Stoffel¹²⁹ the pectoral nerve in 1913, and Foerster¹³⁰ the long thoracic nerve in 1929; Yeoman and Seddon¹³¹ (1961) and Fantis and Slezak¹³² (1967) suggested the intercostal nerves. Yeoman proposed reinnervation of the biceps muscle by intercostal nerves, and Seddon reported these cases in 1963.¹⁰⁰ Nagano and colleagues developed this method for the treatment of avulsion injuries of the brachial plexus.¹³³ The principle has been extended to the reinnervation of free muscles,¹³⁴ and other applications have been described.^135,136 Oberlin and coworkers¹³⁷ pioneered and successfully employed one or two fascicles of the ulnar nerve to reinnervate the motor branch of the biceps in brachial plexus palsies, with outstanding restoration of muscle strength, a concept further applied by Mackinnon and colleagues,¹³⁸ who also reported earlier with Brandt and associates¹³⁹ on use of medial pectoral branch to reinnervate the musculocutaneous nerve. Gu and associates¹⁴⁰ described the use of the contralateral seventh cranial nerve to reinnervate the limb after complete lesion of the brachial plexus. According to Chen and Gu,¹⁴¹ the contralateral seventh nerve is a more effective axon donor than the phrenic nerve, and a vascularized interposed ulnar nerve graft was superior to a nonvascularized graft to the recipient.

FIGURE 240-6 Examples for common nerve transfers. Upper row shows spinal accessory nerve transfer. Nerve is located at posterior triangle of neck, after it has emerged beneath the lateral margin of the sternocleidomastoid muscle and is followed distally beneath the upper ascending margin of the trapezius muscle. Regularly, the nerve gives off another branch to the trapezius of smaller caliber. We intend to spare this branch and disconnect the main trunk as distally as possible. In dependance, where the small branch takes off, one gains an accessory axon donor of variable length. This often necessitates interposition of another autologous graft in order to reach the recipient nerve. The right upper image shows the accessory nerve axon donor on the left (square), coapted to a sural nerve on the right (circle). The lower row shows the basic steps of Oberlin’s transfer. The ulnar nerve at a proximal upper arm level is dissected at the approximate level, where the musculocutaneous nerve branches into biceps and brachial and cutaneous nerves. This is about 12 cm distal to the acromion. The thicker ulnar nerve is encircled with a yellow vessel loop. One motor (stimulation) fascicle is dissected out (held up with jeweler’s forceps) and then coapted to the musculocutaneous branch to the biceps. The biceps branch had been split earlier in a proximal direction and disconnected to gain enough length.

In cases of high common peroneal nerve palsy, the nerve bundle to the lateral head of the gastrocnemius and the soleus has been transferred to the anterior tibial branch of the common peroneal nerve by Gousheh and Babaei.¹⁴²

The senior author (RB) has performed more than 2500 nerve transfers in 1500 patients since 1979, and most were for patients with severe injuries to the brachial plexus. Some of these transfers should be emphasized because they are particularly reliable and important. First comes reinnervation of the paralyzed serratus anterior by transfer of the deep division of the intercostal nerves of T3 and T4 to the nerve to serratus anterior. This operation has been successful in about 90% of cases. Next comes Oberlin’s operation, transfer of a fascicle or fascicular bundle from the ulnar nerve to the musculocutaneous nerve branch to biceps, which was successful in more than 80% of cases. The spinal accessory nerve is a powerful motor donor. Functional elevation and lateral rotation at the shoulder was achieved in more than 70% of 250 cases in which the lesion was confined to the fifth and sixth spinal nerves. In cases of complete preganglionic injury, the spinal accessory nerve is transferred to the ventral root of one of the avulsed spinal nerves. Some useful function followed in more than 50% of these operations. Accessory nerve to biceps branch of musculocutaneous is a reliable operation but requires an intervening graft. The experience at Stanmore with intercostal transfer does not match that of Japanese groups; this experience was shared by Narakas and Allieu.⁵⁷ Functional flexion at the elbow was regained in just fewer than 40% of 300 cases, but some of these patients experienced a welcome improvement in pain. Success has followed in a smaller number of operations where intercostal nerves were transferred to such nerves as the thoracodorsal, medial pectoral, and axillary (circumflex) nerves to the medial head of triceps and to extensor carpi radialis brevis. More recent experience with transfer of intact triceps branches to axillary nerve for deltoid paralysis has also been favorable.

We have little experience with nerve transfers within the distal upper limb. Exceptions are cases of transfer of the dorsal branch of the ulnar cutaneous nerve onto the median nerve for restoration of sensation. Battiston and Lanzetta¹⁴³ reported seven cases of unfavorably high ulnar nerve palsies treated by transfer of the palmar cutaneous branch of the median and of the anterior interosseus nerve before it entered the quadratus into the ulnar nerve at the wrist. Good results were achieved in six cases. Ozkan and coworkers¹⁴⁴ related impressive results from transfer of a digital nerve in irreparable median or ulnar nerve lesions. Eighteen of 20 patients experienced improvement after transferring nerves from inner to outer digits; sensory retraining was used in these patients.

Alternative Methods of Grafting

Conduits

So far, conduits have only been useful for defects up to 3 cm and for cutaneous nerves (Fig. 240-7).^150–155 Research is aimed at tissue-engineered, bioartificial¹⁵⁶ tubes or other autograft surrogates that supply a matrix in concert with neurotrophic factors and cultured cells.^157–159 Schlosshauer and associates gave a detailed review on the current status of synthetic nerve guide implants.¹⁵⁷ Autologous Schwann cells have already been cultured from nerve stump and neuroma areas and implanted into predegenerated sural grafts, tendons, and freeze-thawed muscle grafts.

FIGURE 240-7 Digital nerve repair with conduit. The right-handed patient had a painful neuroma and thus problems holding a pen and writing. A, Proximal nerve and distal nerve stump encircled by vessel loops. B, Nerve gap after neuroma resection. C, Insertion of nerve tube (conduit) with one stay suture at conduit fixed to proximal stump. D, Thumb after wound closure. The procedure was performed under a bloodless field.

Closure and Decision on Splinting

Great care must be taken not to inadvertently disrupt the reconstruction at this stage. Local anesthetic instilled around the proximal stumps in general, and the median and ulnar nerve stumps in particular, can ease the postoperative pain. In long operations, a second dose of antibiotics (after the first one has been given before skin incision) should be administered, and some prefer to extend antibiotic treatment until the second or third postoperative day in such cases. If the tissue layers are readapted meticulously, this will help to prevent hematoma formation. An attempt is made to close the sheaths as physiologically as possible, unless they would impinge on nerve, nerve grafts, or suture points. We do not use suction drains in proximity to grafted nerve. This necessitates meticulous hemostasis by use of bipolar coagulation. The subcutaneous skin layers are sutured with interrupted stitches, and for the skin, there are several options depending on location and surgeon’s preference. We like to support hemostasis with bandages on the respective extremity when possible; however, as a rule, we change these 2 hours after surgery to avoid missing hematomas or progressive swelling.

If grafts course over joints such as the elbow, wrist, and the joints of the hand, movements are restricted so that the flexor tendons and the nerve repairs are protected. Extremes of joint positioning must be avoided. Splints will hold the elbow at 90 degrees of flexion, the wrist at 30 to 40 degrees of flexion, the metacarpophalangeal (MCP) joints at about 70 degrees of flexion, and the proximal interphalangeal (PIP) joint at no more than 30 degrees of flexion. A dorsal finger splint extends to the tips of the fingers, but a palmar splint extends to the PIP joints only. The splints are bandaged such that there is restriction but no rigid immobilization, and gentle, active flexion of the fingers and the thumb must be encouraged from the outset. For nerve reconstructions crossing the volar wrist, dorsal Velcro splints are a practical solution. After brachial plexus reconstruction, the arm is held in a sling. Pendular movements of the shoulder are possible, as are active (as far as possible) lateral rotation and shoulder abduction. Even small flexion-extension motions of the elbow can be usefully done. Such movements must be encouraged from the beginning to counteract the development of a stiff shoulder. Splints are to be removed after 3 weeks. Depending on wound type and location, we remove sutures after 10 to 15 days. If extensive splints were used, range of motion is increased every week until the splint can be discarded after 6 weeks.

Postoperative Considerations

After decompression and neurolysis, functional treatment starting the day after surgery is advised to prevent adhesions and minimize swelling, if the wound allows. Immobilization should be prevented whenever possible, and active use by the patient is encouraged. Intermittent positioning helps to reduce swelling. If swollen, the extremity should be elevated (e.g., over several pillows), especially overnight.

The surgeon checks the postoperative functional status for compromise of other nerves in the vicinity, explains the findings and the procedure, provides the patient with a perspective and follow-up regimen, and most importantly explains the further mobilization regimen to the patient instead of just adding it to the discharge summary.

A nerve suture is considered tear proof after 3 weeks. Most likely this process occurs even more quickly, but this timeframe contains a security interval, during which the patient is cautioned not to overzealously move the affected limb or body part if grafts have been placed passing over a joint region. We rarely see the need to completely immobilize, and therefore we do not routinely use splints after every grafting procedure or nerve suture, and rather make such a decision dependent on location, pathology, and patient (TK). However, the coaptation site must be protected from suture disruption, and the potential risk must be assessed during surgery and balanced against early movement. Certainly, early movement of fingers must be encouraged to prevent a vicious circle of swelling and development of adhesions and contractures. If there is potential for undue tension on the repaired nerve by too vigorous range of motion, we limit the range of motion or immobilize accordingly. The necessity of this depends on the joint and nerve affected. After the security interval of usually 3 weeks (TK), we encourage incremental exercise and range of motion to prevent a fixed deformity, contracture, disuse atrophy, and neuropathic pain to the extent possible. The occasional patient who adheres to relief postures and holding fixed positions should be reminded that the best rehabilitation is return to normal life and activity and that despite ongoing functional deficit, many other physical activities are possible. The patient’s physical therapist needs to be provided with a timeframe defining the periods of immobilization, limited range of motion, and limited weight bearing as applicable, and when the extremity is cleared for full active and passive exercise.

At follow-up, a clear point needs to be determined when work can be taken up again. It does not make sense to have the patient off work for extended periods. No doubt should be left about the physical activities that are deemed necessary for functional recovery, and the patient is reminded that activity is the best physiotherapy. It is essential to address pain of nerve origin. In summary, the patient must be embedded in a postoperative regimen of therapy and needs to be followed at regular intervals and encouraged to take up work again. This is essential for some of the most challenging nerve lesions, which are the ones of the brachial plexus.

No musculotendinous transfer will ever match good nerve regeneration. As such, these operations are adjuncts to improve or replace lost functionality in cases of partial or failed nerve recovery.

Outcome

The most important determinants of outcome are the violence of injury to the nerve and the affected limb, delay in repair, and compromised vascularity.

Other important factors^177–179 are the type and level of nerve affected, associated injuries, extent of lesion, age of patient, graft length, type of necessary repair (neurolysis versus graft), and applied technique.

Overall, proper repair can improve functionality in 50% to 70% of cases.

The functionality and usefulness of recovery depend mainly on the number of axons reaching the correct target organ and on progressive myelination of those axons.¹⁰ Therefore, the number of potential sprouting channels provided by grafts or more direct repair is important, as is the length of the gap after resection and the extent of fibroblastic infiltration of the interposed grafts or sutured nerves. In addition to delay between injury and repair, this regenerative process is influenced by the quality of repair, the degree of damage to the nerve not only during injury but also during repair, and the speed of axonal growth and regeneration.

Every week that passes in delay of repair means progressive atrophy of target tissues as well as cell body degeneration, progressive cell number loss, and central nervous system changes. Omer¹⁸⁰ once confirmed the early findings of Woodhall and Beebe¹⁷⁷: “delay in suture induces a loss of, on average, about 1% of maximal performance for every six days of delay.” An exponential decay of best obtainable regeneration over time most likely reflects this process more precisely. The prognostic effect of level of lesion can be appreciated in those nerves with a long course: the radial, median, and ulnar nerves. A technically sound ulnar repair at wrist level can bring the function of the hand intrinsics back. This is rather exceptional, however, even for primary sutures of the ulnar nerve at the elbow, let alone at the axilla. Repair of the posterior interosseus nerve can restore the function of fingers and thumb; however, this functional gain is rather unusual after repair of the radial nerve at the spiral groove level. In contrast, urgent repairs of C5, C6, and C7, as well as the upper and middle trunk, can achieve results comparable to those seen after repair of combined injuries to more terminal branches of the brachial plexus. Reconstruction of the peroneal nerve will yield worse results than that of the tibial nerve. Ruptures of the axillary nerve are often associated with rotator cuff tears in older patients, which have a bad effect on the recovery of function.

These outcome factors can be distinguished as those that can be influenced by the surgeon and those that cannot. A substantial set of outcome prognosticators, however, cannot be influenced by the surgeon: the type, mechanism, and level of the lesion; nerves affected; associated injuries; and patient age. Other prognosticators are in the hands of the surgeon, and we hope that the this chapter has helped to detail these.