Anatomy

The nervous system is a mechanism that alerts the organism to changes within its internal structures or its external environment and enables reaction to these changes. The peripheral nervous system connects the central nervous system (CNS) to the periphery. It includes the cranial nerves, the spinal nerves with their roots and rami, the peripheral nerves, and the peripheral components of the autonomic nervous system. Peripheral nerves contain motor fibers to the endplates of skeletal muscles; sensory fibers from organs with endings in skin, muscle, tendon, and joint; and autonomic fibers to blood vessels, sweat glands, and hair follicle musculature. The concentration of functional capacity is unmatched by any other system. Severance of a 7-mm-diameter median nerve in an adult effectively devastates the function of the hand and forearm.

The essential component of the system is the cell body with its dendrites and elongated axons. The axon is a column of neuronal cytoplasm ( axoplasm ) enclosed by a cell membrane ( axolemma ). Thomas and associates described the axoplasm as a “fluid cytosol in which are suspended formed elements.” The cytoskeleton contains microtubules, neurofilaments, and a matrix; it provides the apparatus for axoplasmic transport.

The nerve fiber is defined as the axon and its enveloping Schwann cell sheath, which are contained within a basal lamina or basement membrane ( Figure 30.2 ). Nerve fibers are either myelinated or nonmyelinated. Larger axons are wrapped along their length by a continuous series of Schwann cells and a myelin sheath, which is laid down in spiral layers by the surface of the cell. Smaller, nonmyelinated fibers are contained in bundles by similar columns of Schwann cells. The nodes of Ranvier represent the points of contiguity of adjacent Schwann cells.

Proximal stump of the fifth cervical nerve 24 h after traction rupture. Nonmyelinated fibers are enveloped in clusters by Schwann cell cytoplasm; large myelinated fibers are present (original magnification ×19,500).

The caliber of nonmyelinated axons varies from 0.4 to 1.25 µm. The smallest and slowest conducting nonmyelinated nerve fibers are approximately 1 µm in diameter and subserve autonomic activity, thermal sense, and delayed pain sensibility. Myelinated nerve fibers range from 2 to 22 µm in diameter and are classified according to their caliber, speed of conduction, and function. The largest and fastest conducting elements are somatic afferent and efferent myelinated fibers that are approximately 20 µm in diameter. An outstanding review of this field is provided by Lawson.

The specialized property of nerve fibers is their ability to propagate action potentials. In nonmyelinated nerve fibers, a wave of depolarization spreads continuously along the axon and is attenuated by the large capacitance of the axolemma such that the velocity of conduction is restricted to about 1 m/s. In myelinated nerve fibers, myelin restricts electrical activity to the nodes of Ranvier so that the impulse has to travel in leaps from one node to the next by saltatory conduction. The speed of conduction is actually much greater. Demyelination leads to a decrease in conduction velocity and, if severe, to a complete conduction block.

Voltage-Gated Ion Channels

Exchange of ions across the axon’s membrane occurs through voltage-gated ion channels. In nonmyelinated nerve fibers these channels are distributed along the length of the axon membranes. In myelinated nerve fibers they are concentrated at the node of Ranvier. Voltage-gated potassium channels are densely concentrated at the paranode, under the cover of myelin lamellae. These channels act to stabilize membrane potential and inhibit conduction. Voltage-gated sodium channels are exposed at the node and permit the influx of sodium ions, which lowers the potential difference across the membrane and thus facilitates conduction. A number of ion channels have been described, and some sodium channels are specific to nociceptor fibers. Changes in the density and expression of such ion channels may underlie some of the phenomena of pain occurring after injury to nerves. This is a field of fundamental importance, and significant new work is appearing all the time. The matter has been extensively reviewed by Chiu.

Axonal Transport

The axon functions to transport materials to and from the cell body. Two forms of transport are recognized: fast and slow. Slow transport is centrifugal and is associated with the transport of cytoskeletal elements, including membrane proteins, secretory proteins, and peptides. Rates of transport are 1 to 4 mm/day, about the same as the rate of peripheral regeneration after axonotomy. The centripetal or retrograde fast component acquires elements from the terminal by endocytosis, such as nerve growth factor (NGF) and other neurotrophins, and conveys them to the cell body within multifascicular bodies. Rates of fast transport are 200 to 400 mm/day. Axonal transport is oxygen dependent, and it is sensitive to temperature. Interference with the process leads to marked slowing or cessation of conduction and, when prolonged, to degeneration of nerve cells.

The Schwann cells, skin, and other target organs are rich sources of NGF and other neurotrophins that are essential for development, maturation, and maintenance of cell bodies. Deprivation of this supply by transection of a peripheral nerve can lead to central cell death. This is most apparent when the injury is violent and close to the neural axis and even more so in the immature CNS. This matter has been reviewed in detail by Windebank and McDonald.

Connective Tissue Elements

Nerve fibers are embedded in the endoneurium, which contains longitudinally orientated cells, abundant collagen fibrils, and blood vessels. Nerve fibers are aggregated into fascicles by the perineurium , a connective tissue layer composed of flattened cell processes alternating with layers of collagen; it is a diffusion barrier. The perineurium is a strong membrane with high resistance to compression from without, distention from within, and longitudinal traction.

The epineurium is the most abundant of all the connective tissue elements of a nerve trunk and occupies between 60 and 85% of its cross-sectional area. It is more loosely applied than the perineurium and contains numerous longitudinally orientated blood vessels. The epineurium is condensed at its surface as a glistening, translucent membrane, and in an uninjured nerve, individual fascicles can be seen within the perineurium surrounded by an opaque pearly white membrane. Jabaley describes the epineurium as internal and external. Epineurium ensheathes the entire nerve and lies between the fascicles, where it blends with the perineurium and aggregates a number of these fascicles into discrete bundles. Jabaley cautions: “In both dissection and suturing the principal reason for avoiding perineurial injury is to avoid injury to the internal milieu, the conducting portion of the nerve.”

There is one further connective tissue structure that is of great clinical significance, the adventitial “paraneurium” or “mesoneurium.” This loose areolar tissue permits the nerve to glide; at intervals, vascular pedicles pass through this layer to the nerve trunk. Lundborg described a rich blood supply, the intrinsic epineurial, perineurial, and endoneurial plexus and the extrinsic regional vessels that course through the mesoneurium ( Figure 30.3 ). These two systems form “separate but extensively interconnected microvascular systems.” Some of these extrinsic pedicles are so well formed (e.g., those arising from the superior ulnar collateral vessels) that they provide the basis for free vascularized nerve grafts.

Blood supply of the ulnar nerve seen at surgery. The nerve lies in a foamy, glistening adventitia containing longitudinal vessels running in the external epineurium. Deeper vessels run in the epineurium between individual fascicles (original magnification ×40).

A number of nerves are more richly supplied than others. At the knee, the tibial nerve has a far richer blood supply than the common peroneal nerve. This fact may explain the notoriously poor outlook for the peroneal nerve when it is ruptured at the knee. The radial nerve in the axilla and upper part of the arm is supplied by branches from the adjacent profunda arteries. After high transection of the nerve and the adjacent artery, the distal segment of it becomes relatively ischemic; this may explain the rather poor results seen after repair of such high lesions. It must be noted that the blood supply to the roots of the spinal nerves within the spinal canal is far less robust.

It is the mesoneurium that permits normal nerve gliding. Wilgis and Murphy found that the brachial plexus has an excursion of at least 15 mm in relation to the position of the arm and that the median and ulnar nerves at the elbow move through an average distance of 7.3 and 9.8 mm with full motion. The greatest excursion of peripheral nerves occurs at the wrist, proximal to the carpal tunnel; here, the median and ulnar nerves have 15.5 and 14.8 mm of longitudinal gliding, respectively. Injury or surgical intervention that compromises gliding of the nerve impairs function. Tethering of a nerve also may cause severe neuropathic pain. Wilgis commented on the effects of injury on gliding: “In other words, the length of the nerve substance has been shortened because of the injury. Because of this overall shortening, excursion cannot take place.”

The importance of retaining or restoring gliding of peripheral nerves and preventing their tethering or entrapment cannot be overemphasized. The connective tissue of a nerve trunk consists of a collection of sliding interphase zones that permit considerable excursion of the epineurium within the adventitial mesoneurium and of the individual bundles within the deep layer of the epineurium. Among the most important of Millesi and associates’ many contributions is their recognition of the importance of this function. They argue for preservation of the “gliding apparatus” in a paper that should be widely read.

Conduction Block (Also Called Neurapraxia)

There are a number of causes of CB. The distinction between them is somewhat artificial as ischemic anoxia is a common factor to all. Compression of a nerve sufficient to cause mechanical deformation of the myelin sheath also must interfere with its circulation. There are some well-defined patterns as described in the following subsections.

Wallerian Degeneration

Wallerian degeneration affects the nerve distal and proximal to the point of axonal interruption, and it extends to affect the cell body. It can be seen as a process by which the environment of a normal nerve, so inimical to regeneration, is transformed into one that is actively receptive to regenerating axons, at least for a limited period ( Figure 30.5 ). This comforting thought must be set against the reality of clinical practice: Wallerian degeneration cannot occur in an ischemic limb, for it is energy-dependent. The nerve simply dies.

Wallerian degeneration: The appearance of the distal stump of the ulnar nerve 3 weeks after transection shows axonal disintegration (original magnification ×3400).

The harmful consequences of Wallerian degeneration are certainly much less severe in cases (i.e., axonotmesis) where the basal lamina remains intact. This is especially so on the cell body of the neuron because the regenerating axon is able to reestablish contact with the distal Schwann cells rapidly, restoring the flow of neurotrophins from these and from the target tissues. The extent of central cell death is much less.

The changes are briefly summarized in the following subsections.

The Distal Segment

This list describes the outcomes of the distal segment of the nerve:

• •
  

  Conduction is lost. Landau found that the interval between injury and the last observation of the neuromuscular response ranged from 66 to 121 h. Our own observations of motor conductivity after preganglionic injury to the brachial plexus suggests that the motor response ceases about 3 days after injury. In one case stimulation of the avulsed ventral roots of the spinal nerves evoked a motor response 132 h after injury. Such persisting conduction in the distal segment of a nerve during the first 2 days excludes a second, more distal lesion, or ischemia of the distal segment.

  
  
• •
  

  There is an enormous increase in the number of Schwann cells, both fibroblasts and macrophages. The macrophages are responsible for clearing the debris of axoplasm and myelin. The Schwann cells change their phenotype and downregulate expression of genes important for the maintenance of myelin, nodes, and paranodes. The fibroblasts deposit endoneurial collagen—a process that is apparent in clinical specimens by 3 weeks.

  
  
• •
  

  There are substantial changes in the expressions of neurotrophins, such as the brain derived neurotrophic factor (BDNF) and neurotrophins 3 and 4, in muscle and in the distal nerve. These reach a peak at 7 days after injury.

As time passes, more and more collagen is deposited and the number of Schwann cells decreases, and these change their phenotype to one less receptive to regenerating axons. These changes occur at varying intervals according to the nature of the injury. Ischemia hinders, indeed it may prevent, the completion of Wallerian degeneration and concomitant regeneration ( Figure 30.6 ). The evidence from the cellular and molecular responses suggests that the peak of “receptivity” to regenerating axons in the distal segment lies between 1 and 3 weeks after injury.

Example of a biopsy taken 8-mm proximal to the stump of a ruptured median nerve 3 months following injury. The associated rupture of the axillary artery was not repaired. Widespread fibrosis, disintegration of axons, persistent myelin fragments that were not cleared by macrophages, and very scanty evidence of cellular and regenerative activity (original electron microscopy ×4880).

Stretch Injury

Peripheral nerves outside the spinal canal have considerable tensile strength, but function is damaged by elongation of 12% or greater. Lundborg and Rydevik showed that venous flow is blocked when a nerve is stretched by 8% of its resting length and that stretching by 16% produces ischemia. In the early stages of stretching of a nerve, elongation is enabled by stretching of the epineurium and straightening of the irregular course of the fibers within the fascicles.

Haftek observed: “Before rupture of the epineurium the damage to the nerve fibers is either neurapraxia or axonotmesis, because the endoneurial sheath and Schwann fibers remain intact.” With continued traction, the caliber of the fibers is diminished, the endoneurial space narrows, and myelin is disrupted. Finally, rupture occurs when all elements, including the epineurium, are torn. Clinicians engaged in the correction of long-standing deformity of a limb or seeking to lengthen such limbs must bear the preceding facts in mind. The authors have seen some very serious injuries to nerves and to arteries that complicate such operations.

Physical Examination of Nerve Injury

Electrodiagnosis

According to Smith, “[n]erve conduction studies and electromyography should be considered as an extension of the clinical examination of nerve, muscle, and the neuromuscular junction.” 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, whether there is continuity between the exposed nerve and the spinal cord. Conduction across the nerve lesion indicates that at least some of the axons are intact.

After transection of a nerve, axons become unexcitable and neuromuscular transmission fails. Direct stimulation of the nerve distal to the level of the lesion elicits no response. However, some conduction is maintained for several days after nerve transection before Wallerian degeneration is complete.

Fibrillation potentials are one of the earliest electromyographic signs of muscle denervation, but their onset depends on the distance between the site of the nerve lesion and the muscle. There may be an interval of 10 to 14 days before fibrillations are seen. Reappearance of voluntary motor unit potential activity indicates that reinnervation is taking place, and electromyographic evidence of it usually precedes clinical evidence of recovery. The finding of a few motor units showing reinnervation, even at an early stage after injury, does not, however, imply that full recovery of a nerve will take place.

Electrodiagnostic evaluation needs to be done and interpreted by an expert and further considered by an informed surgeon. The findings of these investigations must be interpreted with great care during the first 10 to 14 days after transection of a nerve. In an analysis of incomplete lesions of large nerve trunks (e.g., those seen in the sciatic nerve after hip arthroplasty), the clinician may be lulled into a false sense of security by electrodiagnostic evidence of an incomplete lesion. Such evidence should not be taken to imply that full recovery can be anticipated. Unless the nerve has been wholly transected, it is likely that there will be mixed elements of neurotmesis, axonotmesis, and prolonged CB.

The subject is comprehensively reviewed by Smith, from whose work Tables 30.3 and 30.4 are taken. This study correlates electrodiagnostic findings with Seddon’s classification of nerve injury. As Smith says, however:

The distinction between neurapraxia and axonal degeneration of partial or complete degree is difficult in the acute stages of nerve injury, prior to evidence of denervation in the form of fibrillation potentials on electromyography. Therefore, electrodiagnostic tests cannot reliably differentiate a neurapraxic lesion from one with Wallerian degeneration in the first few days after nerve injury.

Dellon emphasized that such studies are not a substitute for obtaining a careful history and physical examination, that normal values do not necessarily indicate absence of neurologic abnormality, and that abnormal electrodiagnostic findings do not mean that the patient requires operative treatment!

It is within the operating theater that electrodiagnostic effort is of particular value. Bonney and Gilliatt demonstrated persisting conduction in sensory fibers when the DRG has become separated from the dorsal horn of the spinal cord with traction injuries of the brachial plexus. This principle was extended to intraoperative investigation. Landi and associates recorded cortical evoked potentials from nerve stumps stimulated at surgery with scalp electrodes.

Operative neurophysiologic studies have been performed in more than 5000 operations on adults and children since 1977. Application of the technique to diagnosis of birth lesions of the brachial plexus has been described. For simple stimulation and observation of motor responses, only the simplest unipolar or bipolar stimulator is necessary. For stimulation and recording from muscle and nerve and for recording conduction across a lesion, a more elaborate apparatus is required. We formerly used Medelec MS91 (Viasys Health Care, Madison, WI), which features a bipolar stimulator with platinum electrodes, but now uses the Medelec/TEAC Synergy (Peachtree City, GA) monitoring system; its electrodes are available from Ambu (Ballerup, Denmark).

In recording of somatosensory-evoked potentials (SSEPs), the reference electrode is placed on the forehead at the hairline, the ground electrode at the temple, and the recording electrode on the skin surface overlying the second or third intervertebral space. A sterile handheld bipolar stimulator is used to stimulate the nerve directly. Before preparing the limb, the stimulator is used to record signals from the median and ulnar nerves. The uninjured side acts as a control. The median and ulnar nerves are stimulated at the wrist on the injured side and the signals recorded. SSEPs are recorded at a stimulus rate of 3 to 5 pulses/s, a duration of 2.0 ms, and an intensity of 150 to 300 V, with signals averaged from between 50 and 200 sweeps. The stimulus rate for the handheld bipolar stimulator is 3 to 5 pulses/s at an intensity of 3 to 15 V.

For measuring conduction across a lesion, sterile handheld bipolar stimulators are placed on either side of it, and the ground electrode is placed in a convenient adjacent area. The skin is first prepared with abrasive paste and alcohol wipes to lower resistance, and the aim for reference and recording electrode impedance is balanced and should be below 2.0 kΩ. Once the surface electrodes are positioned, they should be secured with tape.

The quality of the traces may be adversely affected by a number of factors: ambient noise interference from electrical equipment in the theater, mobile phones, and deep fibrosis of the nerve; an operating site that is either too wet or too dry also can cause interference with intraoperative recordings. SSEPs are relatively unaffected by anesthetics, but prolonged use of muscle relaxants must be avoided because they prevent observation of muscle activity.

Kline and Hudson have related their extensive experience in the use of electrodiagnostic techniques for the analysis of injured peripheral nerves. Compound nerve action potentials are used to measure regeneration into the distal nerve by stimulating and recording across the site of a lesion. The outcome of the investigation, as it relates to clinical outcome, was described for nearly 1000 nerves with serious lesions in continuity in both the upper and lower limbs. Four hundred thirty-eight nerves showed a recordable nerve action potential traversing the lesion. Neurolysis was performed, and of these nerves, 404 (92%) improved to at least a useful functional result. In another 428 nerves no nerve action potential could be recorded and repair was performed. Of these, 240 (56%) eventually regained useful function. Other important data were collected on nerves for which partial repair was done, with preservation of individual bundles that had recordable nerve action potentials.

Imaging

High-resolution ultrasonography (i.e., 17 MHz) has great potential in the early detection of ruptures or other serious injuries to nerves ( Figure 30.9, A, B ). It has been shown to be reliable in identifying the nerve, confirming the level of injury, and demonstrating continuity or interruption of fascicles. The authors have found that interpretation of ultrasound findings is much more difficult in cases evaluated late, where there is abundant fibrosis around the nerves.

High-resolution ultrasound and clinical correlation of infraclavicular closed brachial plexus traction injury of axillary and musculocutaneous nerves. A, Ultrasound axial image at 17 MHz demonstrates severe thickening and loss of fascicular architecture in posterior and lateral cords, with preservation of normal appearance in the medial cord. B, Operative exposure (proximal right) demonstrating retraction of scar encasement (Kocher clamp) around radial (RN) and axillary (Ax) nerve divisions of posterior cord and scarred but conducting lateral cord after neurolysis. Ax nerve recovered following nerve transfer from pectoralis major branches, and musculocutaneous nerve recovered spontaneously. LC , Lateral cord; PM , pectoralis minor, retracted; SA, subclavian artery.

( A, © Ogonna Nwawka, MD, Hospital for Special Surgery, New York; B, © Scott W. Wolfe, MD.)

Lesions in Continuity

Whether to leave a lesion in continuity alone or whether to resect and bridge the gap is a most difficult decision, made more so when there is clinical evidence of some recovery. The decision is easier when a number of intact fascicles can be seen traversing the lesion. The consistency and the diameter of the neuroma are helpful. The firmer and the larger, indeed the more florid the neuroma, the less likely recovery will be good.

The most informative method is to stimulate above the lesion and to record from the nerve or from individual fascicles below it; a response with good amplitude may well indicate a good prognosis. A response in individual fascicles may allow separation of an intact part of a nerve from the damaged portion.

It is important to remember the potential limitations of intraoperative studies of nerve conduction. Detection of conduction across a lesion does not guarantee good recovery of function. We were always able to record conduction across the neuroma of the ruptured upper trunk in birth lesions of the brachial plexus. It is usual to find that there is some conduction across lesions of the sciatic nerve inflicted during arthroplasty of the hip; however, it has been found that this does not guarantee adequate recovery or reduction of pain. On several occasions, a lesion of the spinal accessory nerve has been left alone because conduction through it was demonstrable. This was a mistake and it proved necessary to return to the area, resect the lesion, and graft it.

The integrity of the perineurium is a valuable guide to the surgeon. The presence of intact bundles (fascicles) crossing the lesion is, for us at any rate, a strong deterrent to resectioning. No one should forget the lesson of case 3 in the series of Birch and St. Clair Strange. There, exploration plus decompression of the sciatic nerve in the notch 3 years after the injury was rapidly followed by relief of pain and recovery from a paralysis affecting the common peroneal component. The lesion was, of course, a CB prolonged by external causes.

Montgomery and coauthors reported a similarly gratifying outcome after neurolysis of the sciatic nerve in a patient who had endured severe pain for a number of years. The lesion was predominantly a conduction block, prolonged and maintained by fibrosis tethering of the nerve after total hip arthroplasty.

Even more remarkable is the case described by Camp and associates. The patient had been suffering intractable and increasing pain for 18 years. Pain was eliminated by neurolysis of the ulnar nerve, which had become adherent to the pulsatile vein graft used to repair the brachial artery. That lesion was a CB prolonged by external causes; the pain was neurostenalgia.

Considerations Before Surgical Intervention

Taking all of the preceding in account, the following should be considered before performing surgery:

• •
  

  Lifesaving or limb-saving measures come first. The surgeon has a duty to assess a patient’s ability to undergo a prolonged intervention.

  
  
• •
  

  For cases reviewed late, indolent wounds and infections must be cleared. The texture of the skin may require massage and oiling.

  
  
• •
  

  Nonunion of a long bone can be dealt with at the same time as a nerve repair. A torn rotator cuff is repaired either at the same time as the circumflex nerve or after it is repaired.

  
  
• •
  

  Deep scars from a penetrating projectile injury, or from burns, present a most hostile bed for nerve grafts. These areas will need replacement with healthy full-thickness skin flaps, pedicled or free, before nerve repair.

  
  
• •
  

  The timing for treatment of a severe fixed deformity from uncorrected paralysis or ischemic fibrosis should be adapted to the individual patient’s needs. Serial plaster of Paris splinting is particularly useful in overcoming fixed flexion deformity of the wrist, the proximal interphalangeal (PIP) joints, and the elbow. Immobilization of the part is necessary after elongation of tendons, or after muscle slide or capsulotomy, and in such cases the authors prefer to repair the nerve at the same time.

  
  
• •
  

  Is it worthwhile? Are other, simpler measures available? Birch et al note that:

  
  
  
  

  By the time the changes of degeneration are present the patient is a better candidate for the examination halls than for restorative treatment. The object of the clinician must be to make the diagnosis before the signs of peripheral degeneration have appeared: that is, before the best time for intervention has passed.

  
  
  
  
• •
  

  Paralysis caused by neglected high radial, high ulnar, or common peroneal nerve lesions may be treated by the appropriate musculotendinous transfer.

  
  
• •
  

  Static or dynamic splinting is helpful to patients by diminishing their disability; by giving an indication of what is expected to be achieved; and, of course, by ensuring that these modalities are ready for a course of postoperative treatment.

  
  
• •
  

  Patients appreciate a comprehensible statement of what has happened, what can be done, and when it can be completed. It is helpful for them to know for how long they must plan to be away from work, doing curtailed daily activities, and driving.

  
  
• •
  

  A quarter of all of the authors’ patients have suffered iatrogenic nerve injuries. In our opinion, it is the clinician responsible for treating an iatrogenic situation who should take charge, give clear advice, set out a clear-cut plan of action, and avoid a partisan approach. Advise patients that appropriate records of operative findings will be released promptly to their legal advisors and that clinicians do not prepare medicolegal reports.

Physical Examination of Nerve Injury

Electrodiagnosis

According to Smith, “[n]erve conduction studies and electromyography should be considered as an extension of the clinical examination of nerve, muscle, and the neuromuscular junction.” 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, whether there is continuity between the exposed nerve and the spinal cord. Conduction across the nerve lesion indicates that at least some of the axons are intact.

After transection of a nerve, axons become unexcitable and neuromuscular transmission fails. Direct stimulation of the nerve distal to the level of the lesion elicits no response. However, some conduction is maintained for several days after nerve transection before Wallerian degeneration is complete.

Fibrillation potentials are one of the earliest electromyographic signs of muscle denervation, but their onset depends on the distance between the site of the nerve lesion and the muscle. There may be an interval of 10 to 14 days before fibrillations are seen. Reappearance of voluntary motor unit potential activity indicates that reinnervation is taking place, and electromyographic evidence of it usually precedes clinical evidence of recovery. The finding of a few motor units showing reinnervation, even at an early stage after injury, does not, however, imply that full recovery of a nerve will take place.

Electrodiagnostic evaluation needs to be done and interpreted by an expert and further considered by an informed surgeon. The findings of these investigations must be interpreted with great care during the first 10 to 14 days after transection of a nerve. In an analysis of incomplete lesions of large nerve trunks (e.g., those seen in the sciatic nerve after hip arthroplasty), the clinician may be lulled into a false sense of security by electrodiagnostic evidence of an incomplete lesion. Such evidence should not be taken to imply that full recovery can be anticipated. Unless the nerve has been wholly transected, it is likely that there will be mixed elements of neurotmesis, axonotmesis, and prolonged CB.

The subject is comprehensively reviewed by Smith, from whose work Tables 30.3 and 30.4 are taken. This study correlates electrodiagnostic findings with Seddon’s classification of nerve injury. As Smith says, however:

The distinction between neurapraxia and axonal degeneration of partial or complete degree is difficult in the acute stages of nerve injury, prior to evidence of denervation in the form of fibrillation potentials on electromyography. Therefore, electrodiagnostic tests cannot reliably differentiate a neurapraxic lesion from one with Wallerian degeneration in the first few days after nerve injury.

Dellon emphasized that such studies are not a substitute for obtaining a careful history and physical examination, that normal values do not necessarily indicate absence of neurologic abnormality, and that abnormal electrodiagnostic findings do not mean that the patient requires operative treatment!

It is within the operating theater that electrodiagnostic effort is of particular value. Bonney and Gilliatt demonstrated persisting conduction in sensory fibers when the DRG has become separated from the dorsal horn of the spinal cord with traction injuries of the brachial plexus. This principle was extended to intraoperative investigation. Landi and associates recorded cortical evoked potentials from nerve stumps stimulated at surgery with scalp electrodes.

Operative neurophysiologic studies have been performed in more than 5000 operations on adults and children since 1977. Application of the technique to diagnosis of birth lesions of the brachial plexus has been described. For simple stimulation and observation of motor responses, only the simplest unipolar or bipolar stimulator is necessary. For stimulation and recording from muscle and nerve and for recording conduction across a lesion, a more elaborate apparatus is required. We formerly used Medelec MS91 (Viasys Health Care, Madison, WI), which features a bipolar stimulator with platinum electrodes, but now uses the Medelec/TEAC Synergy (Peachtree City, GA) monitoring system; its electrodes are available from Ambu (Ballerup, Denmark).

In recording of somatosensory-evoked potentials (SSEPs), the reference electrode is placed on the forehead at the hairline, the ground electrode at the temple, and the recording electrode on the skin surface overlying the second or third intervertebral space. A sterile handheld bipolar stimulator is used to stimulate the nerve directly. Before preparing the limb, the stimulator is used to record signals from the median and ulnar nerves. The uninjured side acts as a control. The median and ulnar nerves are stimulated at the wrist on the injured side and the signals recorded. SSEPs are recorded at a stimulus rate of 3 to 5 pulses/s, a duration of 2.0 ms, and an intensity of 150 to 300 V, with signals averaged from between 50 and 200 sweeps. The stimulus rate for the handheld bipolar stimulator is 3 to 5 pulses/s at an intensity of 3 to 15 V.

For measuring conduction across a lesion, sterile handheld bipolar stimulators are placed on either side of it, and the ground electrode is placed in a convenient adjacent area. The skin is first prepared with abrasive paste and alcohol wipes to lower resistance, and the aim for reference and recording electrode impedance is balanced and should be below 2.0 kΩ. Once the surface electrodes are positioned, they should be secured with tape.

The quality of the traces may be adversely affected by a number of factors: ambient noise interference from electrical equipment in the theater, mobile phones, and deep fibrosis of the nerve; an operating site that is either too wet or too dry also can cause interference with intraoperative recordings. SSEPs are relatively unaffected by anesthetics, but prolonged use of muscle relaxants must be avoided because they prevent observation of muscle activity.

Kline and Hudson have related their extensive experience in the use of electrodiagnostic techniques for the analysis of injured peripheral nerves. Compound nerve action potentials are used to measure regeneration into the distal nerve by stimulating and recording across the site of a lesion. The outcome of the investigation, as it relates to clinical outcome, was described for nearly 1000 nerves with serious lesions in continuity in both the upper and lower limbs. Four hundred thirty-eight nerves showed a recordable nerve action potential traversing the lesion. Neurolysis was performed, and of these nerves, 404 (92%) improved to at least a useful functional result. In another 428 nerves no nerve action potential could be recorded and repair was performed. Of these, 240 (56%) eventually regained useful function. Other important data were collected on nerves for which partial repair was done, with preservation of individual bundles that had recordable nerve action potentials.

Imaging

High-resolution ultrasonography (i.e., 17 MHz) has great potential in the early detection of ruptures or other serious injuries to nerves ( Figure 30.9, A, B ). It has been shown to be reliable in identifying the nerve, confirming the level of injury, and demonstrating continuity or interruption of fascicles. The authors have found that interpretation of ultrasound findings is much more difficult in cases evaluated late, where there is abundant fibrosis around the nerves.

High-resolution ultrasound and clinical correlation of infraclavicular closed brachial plexus traction injury of axillary and musculocutaneous nerves. A, Ultrasound axial image at 17 MHz demonstrates severe thickening and loss of fascicular architecture in posterior and lateral cords, with preservation of normal appearance in the medial cord. B, Operative exposure (proximal right) demonstrating retraction of scar encasement (Kocher clamp) around radial (RN) and axillary (Ax) nerve divisions of posterior cord and scarred but conducting lateral cord after neurolysis. Ax nerve recovered following nerve transfer from pectoralis major branches, and musculocutaneous nerve recovered spontaneously. LC , Lateral cord; PM , pectoralis minor, retracted; SA, subclavian artery.

( A, © Ogonna Nwawka, MD, Hospital for Special Surgery, New York; B, © Scott W. Wolfe, MD.)

Lesions in Continuity

Whether to leave a lesion in continuity alone or whether to resect and bridge the gap is a most difficult decision, made more so when there is clinical evidence of some recovery. The decision is easier when a number of intact fascicles can be seen traversing the lesion. The consistency and the diameter of the neuroma are helpful. The firmer and the larger, indeed the more florid the neuroma, the less likely recovery will be good.

The most informative method is to stimulate above the lesion and to record from the nerve or from individual fascicles below it; a response with good amplitude may well indicate a good prognosis. A response in individual fascicles may allow separation of an intact part of a nerve from the damaged portion.

It is important to remember the potential limitations of intraoperative studies of nerve conduction. Detection of conduction across a lesion does not guarantee good recovery of function. We were always able to record conduction across the neuroma of the ruptured upper trunk in birth lesions of the brachial plexus. It is usual to find that there is some conduction across lesions of the sciatic nerve inflicted during arthroplasty of the hip; however, it has been found that this does not guarantee adequate recovery or reduction of pain. On several occasions, a lesion of the spinal accessory nerve has been left alone because conduction through it was demonstrable. This was a mistake and it proved necessary to return to the area, resect the lesion, and graft it.

The integrity of the perineurium is a valuable guide to the surgeon. The presence of intact bundles (fascicles) crossing the lesion is, for us at any rate, a strong deterrent to resectioning. No one should forget the lesson of case 3 in the series of Birch and St. Clair Strange. There, exploration plus decompression of the sciatic nerve in the notch 3 years after the injury was rapidly followed by relief of pain and recovery from a paralysis affecting the common peroneal component. The lesion was, of course, a CB prolonged by external causes.

Montgomery and coauthors reported a similarly gratifying outcome after neurolysis of the sciatic nerve in a patient who had endured severe pain for a number of years. The lesion was predominantly a conduction block, prolonged and maintained by fibrosis tethering of the nerve after total hip arthroplasty.

Even more remarkable is the case described by Camp and associates. The patient had been suffering intractable and increasing pain for 18 years. Pain was eliminated by neurolysis of the ulnar nerve, which had become adherent to the pulsatile vein graft used to repair the brachial artery. That lesion was a CB prolonged by external causes; the pain was neurostenalgia.

Considerations Before Surgical Intervention

Taking all of the preceding in account, the following should be considered before performing surgery:

• •
  

  Lifesaving or limb-saving measures come first. The surgeon has a duty to assess a patient’s ability to undergo a prolonged intervention.

  
  
• •
  

  For cases reviewed late, indolent wounds and infections must be cleared. The texture of the skin may require massage and oiling.

  
  
• •
  

  Nonunion of a long bone can be dealt with at the same time as a nerve repair. A torn rotator cuff is repaired either at the same time as the circumflex nerve or after it is repaired.

  
  
• •
  

  Deep scars from a penetrating projectile injury, or from burns, present a most hostile bed for nerve grafts. These areas will need replacement with healthy full-thickness skin flaps, pedicled or free, before nerve repair.

  
  
• •
  

  The timing for treatment of a severe fixed deformity from uncorrected paralysis or ischemic fibrosis should be adapted to the individual patient’s needs. Serial plaster of Paris splinting is particularly useful in overcoming fixed flexion deformity of the wrist, the proximal interphalangeal (PIP) joints, and the elbow. Immobilization of the part is necessary after elongation of tendons, or after muscle slide or capsulotomy, and in such cases the authors prefer to repair the nerve at the same time.

  
  
• •
  

  Is it worthwhile? Are other, simpler measures available? Birch et al note that:

  
  
  
  

  By the time the changes of degeneration are present the patient is a better candidate for the examination halls than for restorative treatment. The object of the clinician must be to make the diagnosis before the signs of peripheral degeneration have appeared: that is, before the best time for intervention has passed.

  
  
  
  
• •
  

  Paralysis caused by neglected high radial, high ulnar, or common peroneal nerve lesions may be treated by the appropriate musculotendinous transfer.

  
  
• •
  

  Static or dynamic splinting is helpful to patients by diminishing their disability; by giving an indication of what is expected to be achieved; and, of course, by ensuring that these modalities are ready for a course of postoperative treatment.

  
  
• •
  

  Patients appreciate a comprehensible statement of what has happened, what can be done, and when it can be completed. It is helpful for them to know for how long they must plan to be away from work, doing curtailed daily activities, and driving.

  
  
• •
  

  A quarter of all of the authors’ patients have suffered iatrogenic nerve injuries. In our opinion, it is the clinician responsible for treating an iatrogenic situation who should take charge, give clear advice, set out a clear-cut plan of action, and avoid a partisan approach. Advise patients that appropriate records of operative findings will be released promptly to their legal advisors and that clinicians do not prepare medicolegal reports.

Nerve Resection and Neurolysis