The management of a patient suffering from an injured peripheral nerve requires an understanding of the mechanics of injury, the pathological response, and subsequent regenerative capacity. Decisions concerning whether to operate, when to operate, and what to do once the lesion is exposed must be based upon not only a firm understanding of the pathology of the repair but also some acceptance of the limitations for neural regeneration in terms of practical functional recovery. Clinical examination, electrodiagnostic studies, and radiologic studies are helpful in making such decisions. Patient selection for operation as well as timing, type(s) of operation, and the value of operation persist as controversial issues.
Guidelines for Injury Evaluation
The major determinants for deciding whether or not to operate on injured nerves are (1) the mechanism of injury, (2) the severity of the neurological loss, and (3) the presence of severe pain. Sharp or blunt lacerations involving soft tissues and nerve(s) with severe distal loss will require operation. Blunt injuries associated with stretch, fracture, contusion, compression, and even gunshot wound (GSW) are more likely to preserve some physical continuity of the involved nerve and mayor may not improve without operation. If loss is complete distal to the injury, complete improvement with time is less likely but can, on occasion, still occur. When loss is incomplete and continuity of the nerve is likely because of the mechanism of injury, function will usually improve with time. There are, of course, exceptions: (1) when the partially injured nerve, although in continuity, is compressed by a pseudoaneurysm or expanding clot, and (2) when the site of nerve injury is close to an area of potential entrapment -e.g. ulnar nerve at the olecranon notch, median nerve at the wrist, posterior interosseous nerve in the region of the supinator, or peroneal nerve at the head of the fibula. Although regeneration following proximal nerve lesions is faster than that which follows distal injury, axons must traverse great distances to reach distal target sites. Thus, in most cases, gaining good results is more difficult following proximal lesions than distal ones and delays in their repair should be avoided.
Axonal Regeneration Considerations
The injured peripheral nerve has characteristic neuronal and axonal responses. The severity of injury will partly determine the degree of axonal regeneration. Although the rate of axonal growth and maturation of motor function is slow, the rate of regeneration is predictable. Regeneration proceeds at the rate of 1 mm per day or 1 inch per month. This helps the physician establish approximate deadlines in relationship to time of injury or previous repair in expecting clinical signs of reinnervation. If the first target muscle begins to show function at the expected time and power improves over the next 1-2 months, the decision against surgery is clear. If the expected time schedule is not met, or the subsequent early quantitative extent of motor activity in the first target muscle does not match the expectancy after repair, operative intervention is indicated. Unfortunately, too much time is required for many nerve lesions to reach even early regenerative milestones. Under these circumstances, if repair is delayed until after these deadlines are met, results are not as good as with earlier repair.
The time required for regeneration involves the following considerations:
1. There is a delay before regenerating axons reach the nerve distal to either injury or suture repair. The segment of retrograde degeneration proximal to the injury must first be overcome, and then there is usually a delay of 1-2 weeks before axons penetrate the injury or repair site and reach the distal stump. This period of delay may be 2-4 weeks.
2. Once the fibers have reached the distal stump, the rate of axonal growth decreases as the distance of the injury from the neuron increases.
3. A terminal delay of weeks to several months takes place between the time when axons reach their distal targets and when sufficient maturation of the axons and their receptors occurs to allow maximal function. Thus, it is not enough for axons to reach their distal targets; they must do so in sufficient number and with enough caliber and myelination to produce acceptable function.
Evidence of regeneration, as gauged by return of nerve function, can help guide the initial management of such lesions. Positive evidence for some significant nerve function either initially or within 6 weeks post-injury, implies a favourable result. When a significant proportion of axons have escaped initial dysfunction or have suffered only a minor degree of nerve fiber injury, regeneration occurs, exceeding the best that nerve repair could yield.
The more frequent clinical situation is tb3t of total nerve dysfunction in which the lesion has not been operatively inspected, or in which exposure at surgery has revealed a neuroma in continuity. If a nerve repair has been performed elsewhere under uncertain circumstances, a similar management dilemma arises. In these cases, delayed surgical exploration with intraoperative nerve action potential recordings is invaluable in making the final decision regarding resection and repair of the damaged nerve.
Clinical Evaluation Motor Examination
A point to stress regarding clinical motor examination for specific nerve injuries is that the single most important step in management of any nerve injury is a detailed examination of the limb, with careful grading of all motor and sensory function. The examination must then determine whether loss is complete or incomplete distal to the injury site. Only in this fashion can one can tell on subsequent examination whether or not function has changed. Motor examination is sufficient by itself as proof of regeneration when recovery is obvious. Clinically observed voluntary motor function can also be confirmed by motor response to nerve stimulation. Nerve stimulation is especially helpful in early recognition of adequate peroneal recovery and avoidance of a needless operation. Patients with injury to the peroneal nerve are unable to initiate voluntary action in the peroneus and anterior tibial muscles (eversion and dorsiflexion of the foot). This may continue for several weeks after electrophysiologic recovery has been demonstrated by strong muscle contraction on peroneal nerve stimulation: (1) just behind the head of the fibula or, (2) just inside the lateral hamstring, where the nerve trunk is readily palpated. Importantly, one must be certain that the muscle observed to contract is in the distribution of the nerve presumed to be stimulated.
If paresthesias are obtained by percussion of nerve distal to the injury, there is a suggestion that some sensory axons are continuous from the point percussed through the lesion to the central nervous system. If the response moves further distally with time, and especially if this is associated with diminished paresthesias in response to tapping over the injury site, evidence of continued sensory fiber regeneration down the distal stump is present (positive Tinel's sign). A positive Tinel's sign, however, implies only fine fiber regeneration and tells the examiner nothing about the quantity and eventual quality of the new fibers. On the other hand, total neural interruption is strongly suggested by an absence of distal sensory response (negative Tinel's sign) after adequate time has elapsed for fine fiber regeneration to occur (4-6 weeks). A negative Tinel's sign is more valuable in clinical evaluation than a positive Tinel's sign.
Return of sweating in an autonomous zone signifies sympathetic nerve fiber regeneration. This return may antedate sensory or motor return by weeks or months, since autonomic fibers regenerate rapidly. Return of sweating does not necessarily mean that sensory or motor function will follow.
True sensory recovery is a useful sign, especially when it occurs in autonomous zones where overlap from adjacent nerves is minimal. Autonomous zones for the median nerve include the volar and dorsal surfaces of the forefinger and volar surface of the thumb. The radial nerve does not have a reliable autonomous zone. If there is any sensory loss in its distribution, it will usually involve the region of the anatomic snuff box. The autonomous zone for the ulnar nerve includes the palmar surface of the distal 11/2 phalanges of the little finger. Autonomous zones for the tibial nerve include the heel and a portion of the sole of the foot, while for the peroneal nerve it includes mid-dorsum of the foot. Unfortunately, sensory recovery, even in an autonomous zone, does not ensure subsequent motor recovery.
A thorough baseline electromyographic (EMG) study 2-3 weeks following the injury will document the extent of denervation and will confirm the pattern or distribution of the injury. EMG studies should be done serially to search for signs of reinnervation or persistence of denervation. With regeneration, insertional activity will begin to return and the fibrillation and denervation potentials will decrease in number and sometimes be replaced with occasional nascent motor action potentials. Such changes indicate that some regenerating fibers have reached muscle and that some axon-to-motor end plate connections have been reconstructed. These signs tell nothing, however, of the eventual extent or quality of regeneration. Nonetheless, when decreased numbers of fibrillations as well as nascent potentials are found in muscles in the distribution of an injured nerve, a short interval of further conservative management is suggested. The EMG is important because it can give evidence of regeneration weeks or months before voluntary motor function is detectable. It can also detect retained motor units to indicate a partial lesion early after injury. The EMG is particularly helpful in defining the level of injury in a brachial plexus lesion and thus in selecting patients for operation as well as the type of operation to be used. Paraspinal muscle denervation suggests a proximal lesion(s) to one or more roots and thus is a negative finding. Proximal damage to the lower three roots can result in extensive paraspinal denervation while the C5 and even the C6 roots may be more laterally injured and are, thus, repairable. The electromyographer has difficulty sampling distinct spinal levels within the paraspinal muscle because there is so much overlap.
An operation is usually indicated in brachial plexus lesions if complete loss in the distribution of one or more upper roots (C5,C6,C7) and their distal outflows does not begin to reverse clinically or electrically in the early months postinjury. The presence of EMG changes suggesting reinnervation does not guarantee recovery of function, and the test must be weighed in conjunction with clinical findings and other electrical data. Because the EMG can continue to show quite severe denervational changes even though the muscle contracts voluntarily, the EMG should never be substituted for a careful clinical examination. Rather, it should supplement the clinical examination. EMG is especially valuable in identifying anomalous innervation, such as occurs frequently in the forearm and hand.
Sensory Nerve Action Potential (SNAP)
SNAP studies can be helpful in evaluating the level of brachial plexus stretch injuries. Lesions at a root level that are restricted to the preganglionic region and do not extend into the postganglionic region produce complete distal sensory loss and preservation of distal sensory conduction. The latter is preserved because sensory fibers damaged distal to the dorsal root ganglion do not degenerate. This retention of sensory conduction from an anesthetic area can be tested by stimulating fingers in the C6 (thumb and index finger), C67-8 (long finger), and C8- T1 (little and ring fingers) distributions and recording from the median, radial, and ulnar nerves proximally. The presence of a compound sensory nerve action potential substantiates a preganglionic injury in the distribution of one or more roots. Since even distal sensory distributions of roots overlap with one or more other roots, it is difficult to be certain by these studies that one root, C6 for example, has a preganglionic injury.
Stimulation of an anesthetic forefinger (or even thumb) can produce a SNAP in the median nerve distribution if either C6 or C7, or c6 and C7 roots, are damaged at a preganglionic level. This makes it difficult to determine by SNAP studies whether or not the C6 root has incurred a preganglionic injury. The situation is even less favourable for the CS root since there are no specific noninvasive stimulation or recording sites for this outflow: Detailed evaluation of upper roots by SNAP recordings is not possible at this level.
Somatosensory-Evoked Potential (SSEP)
SSEP study has been used in evaluating the level of injury: i.e. pregangliooic versus postganglionic, in brachial plexus lesions, These studies have limited value in the early months following injury. Somatosensory studies can, however. be used at the time of surgery for stretch/ contusion brachial plexus injuries. If the injury is postganglionic, stimulation of the root proximal to the level of the injury should evoke a somatosensory potential over the cervical spine (SSP) and an evoked cortical response over the contralateral cranium (ECR). If the injury is preganglionic or pre- and postganglionic, stimulation of the root, even within or close to the intervertebral foramen, will evoke no such responses. Repair of at least that element is unlikely to be successful. Unfortunately, production of an SSP or ECR probably requires only a few hundred or so intact fibers between site stimulated and site recorded, so a positive response only ensures minimal continuity of spinal nerve or root. A negative ECR is of more importance than a positive ECR.
Intraoperative Nerve Action Potential (NAP)
NAP study involves operative exposure of the nerve trunk on either side of the lesion. Since one ideally seeks to decide whether to repair a nerve by 8 weeks after injury, NAP becomes an important definitive test when gross appearance of a neuroma in continuity is equivocal and the first target muscle is more than 3 inches downstream. The important considerations with NAP recordings are:
1. The gross appearance of a neuroma in continuity does not necessarily correlate with the internal architecture.
2. If axons have been given an opportunity to traverse the lesion, their presence may be recorded by the NAP long before those axons have had an opportunity to reach their end target.
3. This technique is particularly useful in lower-extremity nerve lesions in which the first target muscle may lie 6-8 inches below the lesion. Thus, neither nerve stimulation nor EMG can settle the issue for 6-8 months or more, but it is important that decisions regarding resection be made before that time.
4. NAP recordings are also very helpful in defining the extent of brachial plexus lesions and provide a useful index of how much of the proximal stump of the lesion in continuity to resect. Most brachial plexus injuries selected for operation will have one or more elements in continuity, but with a variable amount of intraneural damage. Intraoperative NAP recording helps sort out the need for resection. At surgery, the critical observation is whether or not there is a recordable response, and not its form or even its velocity. Regenerative NAP responses are small and usually slow, while those due to partial sparing may be small but are usually faster or have conduction in a normal range. Where there has been preganglionic without postganglionic injury, more distal recording will show a rapid conducting, large NAP, which is just as diagnostic as absence of an SSP or ECR when the root is stimulated at that level.
Cervical Spine and Other X-rays
Cervical spine fractures are frequently associated with severe proximal, irreparable stretch injuries, at least at the root levels associated with those vertebrae. Fractures of other bones such as the humerus, clavicle, scapula, and/or ribs, when observed, give rough estimates of the forces brought to bear on the shoulder, arm, and neck, but do not necessarily help localize the level or document the extent of the injury. Damage to the plexus is usually more proximal than the fracture site would indicate, frequently at the root level. Midhumeral fractures are especially associated with radial nerve injuries. Comminuted fractures of the radius and ulna at midforearm level can also be associated with combined median and ulnar nerve injuries, and on occasion with posterior interosseous nerve palsy. The peroneal component of the sciatic nerve is often, but not always, selectively involved in hip dislocation or fracture. Lower femur fractures as well as tibial and fibular fractures may involve the peroneal and/or tibial nerves. Once again, the nerve injury may be more proximal than the fracture(s) site(s) may suggest. A midshaft femur fracture may be associated with a more proximal sciatic stretch injury at the buttock level. Chest radiographs may reveal elevation of a nonfunctioning diaphragm, which denotes phrenic nerve paralysis. This is a relatively poor prognostic sign for repairability of the C5 nerve root following closed injuries, because it usually implies proximal damage at that level of the neck.
Myelography may be an important part of the work-up in a patient with severe brachial plexus stretch injury. It is usually not indicated for infraclavicular or axillary level plexus lesions (most gunshot wounds to the plexus), unless there is radiologic evidence of damage to the cervical spine or a medial supraclavicular trajectory. A meningocele at a given level indicates that enough force was applied at a proximal root level to tear the arachnoid and produce a leakage of contrast agent. It does not necessarily mean that the root is avulsed out of the spinal cord. More commonly, the presence of a meningocele implies that, although the root may still be in gross continuity, it has significant internal damage at a very proximal level. A number of patients have had successful repair of roots at levels where meningoceles were absent (usually upper root levels), despite meningoceles on other roots (usually lower levels). Nonetheless, if a meningocele is present, it is most likely that the root has proximal and thus irreparable damage. This finding also makes it more likely that damage at other levels without meningoceles is very proximal. Modern myelography with water-soluble contrast agents may delineate the rootlets in the subarachnoid space, and comparison of the affected and unaffected sides may delineate sites of root disruption. Myelography is still a useful adjunct in the decision-making process concerning plexus injuries.
Magnetic Resonance Imaging (MRI & MRV) and Computed Tomography (CT)
Computed tomography (CT) scanning with intrathecal contrast is of interest in stretch injuries, although an abnormality may still be missed because slices are usually not thin enough to cover all of the root regions at each level. As a result, myelography still remains the preferred radiologic study. Magnetic resonance imaging (MRI) help visualize the nerve root. Cerebrospinal fluid (CSF) within meningoceles can be seen on MRI and MRV, which non-invasively showing the pseudomeningocele very clear and it is considered the perfect choice to demonstrate the severity of the lesion.
Guidelines for Timing of Repair
In deciding when to repair, the surgeon must define: (1) when the time for useful recovery by spontaneous regeneration has passed, and (2) the elapsed time when a nerve repair has little to offer. When the duration of total muscle denervation exceeds 24 months ("24-month rule"), most muscles are subject to relatively severe time limitations for the return of useful function. This is less likely to be so for large bulky muscles, such as biceps and gastrocnemius-soleus, than for smaller muscles, such as those of the forearm and hand. An exception to this guideline are the facial muscles which, although relatively small, may benefit from late reinnervation by facial nerve repair or neurotization procedures. Other exceptions to the "24-month rule" may occur in a few lesions that have maintained some nerve fiber continuity. If some fibers traverse the lesion, even though their number is insufficient to produce useful function distally, they may promote distal stump architecture preservation. Very late repair after resection of the lesion in continuity can occasionally produce function. Distance from the site of nerve injury to the desired muscle influences the timing of surgery. When the site of injury is a long distance from an important muscle, it is essential to perform the repair within a few months postinjury. This is especially so with sciatic nerve and brachial plexus injuries.
Relatively early repair of other nerve injuries also is beneficial. For example, when the radial nerve is injured in association with a closed midhumeral fracture, the probability for good spontaneous recovery is high. Exploration should be undertaken if there is no recovery by 4 months. By this time, a midhumeral axonotemetic injury of the radial nerve should have regenerated to the next muscle downstream, the brachioradialis. If the radial nerve is seriously damaged between the fracture fragments, repairing it much later would begin to yield less satisfactory results for return of motor function.
In contrast, there exist cases when the distance between the nerve injury and the muscle to be reinnervated is such that repair, early or late, will not accomplish a useful degree of motor recovery. For example, repair of the ulnar nerve near the axilla or peroneal nerve above the midthigh may accomplish little in the way of function to important distal muscles. Other high repairs may be indicated, however, either because there are useful proximal muscles (such as the triceps and proximal forearm extensors in case of radial nerve) or because sensory recovery is valuable (such as that in the distribution of the median nerve).
Time limitation is less severe in sensory recovery than in motor recovery. This is an important consideration in favour of median nerve repair as high as the axillary level, even though it may only contribute minimal motor function at the hand level. A high median nerve repair is especially important if a mechanically useful hand can be provided by substituting or transferring some of the ulnar and radial motor function that does exist. Similarly, repair of the tibial component of the sciatic nerve at a level as high as the sciatic notch may be indicated. Protection to weightbearing plantar areas can be given by even a low-grade recovery of sensory function. Thus, restoration of protective sensation to the sole of the foot is important enough to warrant a proximal repair. Some degree of useful plantar flexion will usually come about as well.
Finally, motor recovery from spontaneous regeneration (without nerve repair) also has limitations in time. High nerve lesions will usually result in lack of useful distal motor function if the muscle is over 24 inches distal to the lesion and is totally denervated by the injury. Relatively early evidence for some recovery of motor function, even if only detectable by EMG, will greatly improve the prognosis. These findings must be evident by 2-3 months postinjury. Therefore, in some high lesions, measures to compensate for lost motor function can be taken quite early. Tendon transfers can be performed without awaiting the effect of possible late regeneration from a high or proximal ulnar nerve lesion.
Early (Primary) versus Delayed (Secondary) Repair
Early operative intervention is infrequently needed in most peripheral nerve stretch injuries. There are notable exceptions. An enlarging hematoma or aneurysmal sac will convert a partial incontinuity nerve injury into a complete and, with time, irreversible lesion unless the mass is removed as early as possible. A severely contused forearm or a distal humeral fracture associated with brachial artery injury predisposing to Volkmann's ischemic contracture are other exceptions regarding delay of an operation. In this instance, early fasciotomy, treatment of the vascular injury, and in many cases neurolysis of one or more nerves are necessary.
A similar syndrome can involve the anterior compartment of the lower leg, requiring urgent intervention if irreversible neural as well as muscular changes are to be avoided. A severe noncausalgic pain syndrome secondary to a missile embedded in nerve also can benefit from an early operation and removal of the missile fragment. Injury to nerve in areas of potential entrapment may also require early release of the nerve and section of the connective tissue structures likely to cause entrapment. This is done to avoid potentially irreversible changes in the nerve. Early (primary) repair is a valid option for the repair of simple, clean lacerating injuries such as those caused by glass and knives. In civilian injuries, primary repair is best for sharply transected supraclavicular and axillary level brachial plexus and sciatic nerve injuries; immediate exploration provides the best opportunity for both accurate identification and end-to-end repair without need for grafts. This is especially so with sharp plexus injuries in which there is vascular damage that must be repaired at once. If such a wound site is explored some weeks later, one is usually confronted by heavy scar that makes dissection and identification of involved neural elements difficult. At the time of exploration, one must make sure that the transection is sharp and clean before a primary repair is done. When confronted with a transected nerve, the following factors favour primary repair:
1. Nerve stumps are easy to locate and their relationships to other injured structures usually are preserved.
2. Nerve stumps are minimally retracted.
3. A single operative procedure is definitive and may be the only operation necessary to repair soft tissue as well as nerve injury.
Primary repair should only be undertaken by surgeons who have total mastery of the anat0my of the injured region and who have been trained in macro- and microperipheral nerve surgical techniques. Not all transecting injuries lend themselves to primary repair. If the ends are ragged or contused, a delayed repair is preferable. In this case, the surgeon cannot know how much of either stump to resect in order to get back to healthy neural tissue. Even with injuries caused by sharp objects, a contused rather than transected nerve can result, and delayed repair then becomes mandatory. If clinical or substantial EMG recovery does not occur in the first 2-3 months, reoperation to evaluate the lesion in continuity and to make a decision for or against resection and repair is indicated.
In summary, the arguments favouring delayed or secondary repair include:
1. Damage to the proximal and distal stumps has had time to be defined by visible intraneural scarring on cross-section. The surgeon can then be certain that resection back to normal neural tissue is accomplished.
2. Associated injuries have had a chance to heal, infection has been minimized, and the patient has learned to use the extremity before being subjected to operation and sometimes to immobilization and its attendant discomfort.
3. The epineurium has thickened so as to allow easier placement of epineural sutures.
4. Operation is elective and can be performed accurately.
5. The distal stump is cleared of degenerating axoplasm and myelin.
6. Approximately 15 % to 20% of lacerations have not transected the nerve or nerves in a limb with associated complete loss of function in the distribution of one or more nerves. It is impossible to decide immediately postinjury whether or not to resect such a lesion in continuity.
When the nerve is not known to be transected (closed injury) but is without function, especially following high-velocity missile wounding, delayed (secondary) repair is indicated. The majority of closed injuries to the nerve are due to stretch/ contusion. The nerve is not divided and there is a variable degree of intraneural damage. This may be a mixture of axonotmesis, neurotemesis and neuropraxia, or may be due to complete neurotemesis. Thus, a delay of several months is necessary, since this will permit (1) any element of neuropraxia to resolve, (2) associated injuries to heal, and (3) most importantly, physiologic evaluation of the lesion at the operating table. If adequate regeneration is occurring, spontaneous activity can be detected by means of intraoperative NAP recording techniques by 8-10 weeks postinjury.
If a NAP is present, the nerve will fare well with simple neurolysis. Most often, external neurolysis is performed. This consists of freeing the nerve from surrounding tissue, including scar, and exposing the entire circumference of the nerve. Internal neurolysis, involving resection of scar tissue away from the nerve fascicles, is usually reserved for certain partial nerve lesions requiring split repair and management of refractory neuritic pain. Neurolysis may or may not assist in continued regeneration and hasten recovery. Some authors believe that adhesions and scar tissue can obstruct or delay the growth of regenerating axons and even block conduction in nerve fibers. Other investigators say that recovery under these circumstances would have occurred even if neurolysis had not been done. Neurolysis may also relieve or ameliorate noncausalgic neuralgic pain by removing adhesions or constricting scar that fix and at times deform the nerve. Neurolysis or nerve repair is less likely to ameliorate pain in nonfocal injuries, such as stretch-contusion, particularly of the plexus. A daily regimen of carbamazepine and amitriptyline may help relieve pain. Vigorous physical therapy is essential in pain management. Early mobilization of the involved limb should be stressed to the patient and family. Reassurance of the temporary nature of the pain, at least in patients with acute nerve injury, is also helpful. Occasionally, patients may benefit from transcutaneous peripheral nerve stimulation devices.
If a NAP is not present and 8 weeks have elapsed, recovery will not occur unless resection back to healthy neural tissue and repair are performed. An end-to-end repair is preferred. Use of autologous grafts, usually using the sural nerve, is the method of choice for bridging a gap that cannot be closed without tension by an end-to-end union. The success of nerve grafting declines as the length of the graft increases, usually because the lengthier injuries requiring longer grafts are more severe.
Grafting very long defects under unfavourable conditions is not worthwhile in some nerves, because the chances of obtaining any useful function are remote. In such cases, alternative methods such as neurotization procedures should be considered. These include use of the cervical plexus, accessory nerve, or intercostal nerves as proximal outflow to attach to sural grafts. Such procedures have sometimes provided either shoulder abduction or biceps/brachialis contraction. Neurotization has difficulty substituting for loss of more than one function, although a recent report in a small number of patients suggests that this may still be a possibility.
After it is clear that recovery is unlikely, an injured nerve should be repaired with the least possible delay. This minimizes distal nerve trunk and fascicular atrophy which will lead to poor results. According to Sunderland, such atrophy is evident by the end of the first month postinjury, reaches a peak somewhere between the third and fourth months, and then. levels out. As a general guideline, focal lesions in continuity (those associated with fracture, soft tissue contusion, and some gunshot wounds) can be accurately evaluated intraoperatively at 2-3 months postinjury.
Stretch or severe contusion injuries (those associated with vehicular and skiing accidents, falls, crush injuries, and shotgun pellets) produce lengthier lesions and need to be followed longer to assess full regenerative capacity. These lesions can usually be accurately evaluated intraoperatively by electrical recordings at 3-5 months postoperatively. Delay in referral, healing of associated injuries, and management of infection may alter the timing of operation. Despite these guidelines for lesions in continuity, when one is in doubt about the direction of recovery, it is better to assess the nerve directly. Excessive delay in nerve repair leads to poor results.
Selection of Patients for Surgery
Despite the guidelines outlined above, controversy concerning the value and timing of surgery as well as patient selection continues to exist. This is especially evident in the management of brachial plexus stretch/contusion injuries. Some authors consider few or none of these injuries suitable for surgery, while others believe that all stretch injuries should be explored. This brings one to the question: How does one select patients with brachial plexus stretch lesions for surgery? Table 2 summarizes some of the criteria currently used for selection.
Management of the Neuroma in Continuity
|Incomplete loss with significant distal sparing|
|Most cases will improve with conservative treatment. They are followed by serial clinical and EMG examinations. Physical therapy is important.|
|Operation may still be required:|
|Partial lesions associated with expanding masses due to hematoma, aneurysm, or A-V fistula usually require urgent operation.|
|Partial lesions close to or in areas of potential entrapment may require relatively early operation.|
|Lesions where distal loss, although partial, is significant may require later operation.|
|Neural pain not amenable to medications and physical therapy may require later operation.|
|Complete or near complete loss with little or no distal sparing|
|Relatively focal lesions in continuity due to fracture or gunshot wound.|
|Follow by clinical and EMG examinations for 2-3 months.|
|If no significant clinical or electrical improvement, explore.|
|Intraoperative stimulation and NAP studies used to decide for or against resection.|
|Relatively lengthy lesions in continuity due to stretch/contusion or shotgun|
|Follow by clinical and EMG examinations for 4-5 months.|
|If no significant clinical or electrical improvement, explore.|
|If no response to stimulation and no NAP across lesion, resection and repair by suture or graft are necessary|
|Intraoperative evoked cortical or somatosensory studies may be necessary to evaluate repairability of proximal spinal nerves.|
From Kline D.G. et al. 1969.
Very proximal injury involving the roots close to the spinal cord and/or the cord itself is the most frequent reason for not attempting direct repair on these injuries. However, most patients requiring surgical treatment have (1) proximal lesions close to and, in some cases, involving the spinal cord, and (2) lengthy lesions requiring long grafts for repair. If the sensory root has evidence of a very proximal (preganglionic) injury, successful repair of the motor root at such a proximal level is technically difficult, although not impossible. Direct repair is impossible if the roots are avulsed from the spinal cord, or if secondary cord damage makes regeneration through the grafts unlikely. Some of these patients may be candidates for neurotization procedures.
Also not candidates for operation are stretch lesions confined to the lower plexus elements, such as C8 and Tl nerve roots, or the lower trunk to the medial cord. Results with repair of these lesions are poor, except in children. Adults seen 1 year or later after injury do not usually benefit from direct neural repair, and thus are not good operative candidates for such an approach.
Other relative contraindications ("stops") to surgery include flail arm, Horner's syndrome, and meningoceles and other myelographic abnormalities. Patients with total arm paralysis due to severe brachial plexus stretch injury are very difficult to salvage by direct repair. It is especially difficult to recover distal forearm and hand function in patients with flail arms. In these patients, every attempt is made to regain significant shoulder abduction and flexion of the forearm. Presence of Horner's syndrome, although an indication of proximal Tl and/or C8 root injury, does not necessarily mean that roots at higher levels are damaged at such a proximal level.
Once the data derived from the clinical examination and the electrophysiologic and radiologic studies are assembled, the managing physician should sit down with the patient and the family to explain the sites and nature of the nerve injuries. Most patients will be familiar with electrical cables, and this may serve as a good analogy, as long as they understand that restoration of continuity does not, alone, produce function. It is useful to explain to the patient that the delay following injury is according to plan, and has allowed distinction of those elements which show evidence of recovery from those nerve elements which do not.
The patient and the family should also understand that the patient will be under the care of the managing physician for 2-6 years, during which time spontaneous recovery and recovery following surgery is observed and further management decisions are made. By 2-3 years after repair, sufficient return will have occurred so as to allow the experienced evaluator to give a prognosis regarding the ultimate function. This sets the stage for appropriate reconstructive operations, such as muscle transfers and joint fusions, which will further facilitate functional recovery of the limb.
Selection of Plexus Stretch Injuries for Operation
|Is lesion complete or incomplete in distribution of one or more elements?|
|Does significant motor (not sensory) improvement occur in first 4 months?|
|If injury is at a nerve root level, how proximal?|
|Relative "stops" (contraindications to direct neural repair):|
|Winging of the scapula-long thoracic nerve|
|Rhomboid paralysis-dorsal scapular nerve|
|Diaphragm paralysis-phrenic nerve|
Extensive paraspinal denervation by EMG-posterior branch of anterior root (more distal and, thus, repairable injury to some roots may still be present)
|Positive sensory potentials can suggest preganglionic injury at C8, T1, and sometimes C7; higher roots may still be operable|
|Myelopathy and/or fracture/dislocation of spine|
|Less certain "stops":|
|Total flail arm|
|Sensory improvement without motor improvement|
|Meningoceles at some (usually lower) but not all levels|
|False positive and negative rates are significant.|
|Meningoceles strongly suggest but do not prove proximal root damage.|
|Meningoceles at one or more levels suggest but do not prove proximal damage at other root levels without meningoceles.|
|Absence of meningoceles does not prove lateral damage nor does presence of a meningocele always mean proximal damage or any damage at all.|
From Kline D.G. et al. 1983.
Patients lacking in motivation following head injury will usually not have the drive to successfully complete the entire course of rehabilitation during the period of nerve regeneration.
Peripheral nerve operations should only be performed in operating rooms with appropriate facilities that include excellent illumination and the provision of magnification during surgery. A range of instruments is required from conventional forceps and scissors to delicate microinstruments. An appropriate range of sutures from 10-0 through 1-0 is required. Cases vary in duration, and the duration of the operation is often unpredictable as the final decisions for and against grafting are only made during the operation itself. Because of the uncertainty of the nature of the surgery, these operations should be undertaken by surgeons who have the ability to perform external neurolysis, nerve suture, nerve grafting, and neurotization procedures. Facility with vein and arterial repair is also essential. Operating rooms must have appropriate electrophysiologic equipment and personnel so that intraoperative NAP and SEP studies can be repeatedly conducted during the course of an operation.
Failure to achieve useful regeneration to the extent that is expected after nerve repair may be due in part to excessive conservatism on the part of the surgeon, who may be reluctant to resect scarred nerve ends back to normal nerve segments when this creates an extensive gap to be bridged (Figure 8).19,25 Incisions for exposure of the lesion should be large enough to fully mobilize the injured neural element(s). Mobilization of the nerve is necessary to overcome the gap that results from retraction of the stumps and resection of the damaged nerve ends.19,35 Maximal nerve Iengd1 is obtained by dissecting to and somewb2t beyond the distal and the proximal joints.
Neurophysiologic studies have shown that extensive mobilization does not affect the subsequent function of a normal nerve or the regenerative process in one that is injured. Mobilization allows the surgeon to resect back to healthy-appearing neural tissue, to shorten the gap between resected stumps, and to join separated nerve ends. Nonetheless, grafts are a useful alternative and most large gaps can be closed by a combination of nerve mobilization and graft repair. Repair of a nerve will yield the best results when there is no cross-sectional area of scar to block a maximal downgrowth of axons from above or to prevent the maximum availability of receiving tubules distally. The appearance of healthy nerve ends usually coincides with brisk bleeding. Healthy nerve ends facilitate maximum fascicular apposition when the nerve ends are brought together end-to-end or are joined by multiple interfascicular grafts. This helps to promote the entry of regenerating axons into fasciculi of the distal stump and not into interfascicular epineural tissue.
Distraction in nerve repair should be avoided. This makes mobilization, transposition, and accurate apposition of the nerve stumps all the more important. If the nerve is under too much tension, distraction is likely to occur, particularly if flexion of the extremity is needed to gain an end-to-end apposition. In this case, the surgeon should resort to relatively short interfascicular nerve grafts, using an autologous nerve such as the sural nerve. The importance of tension as a contributing factor for failure of nerve repair has been demonstrated experimentally in the primate model. Other factors leading to failure of nerve repair include tissue manipulation that is not gentle, sacrifice of longitudinal vessels deep to the epineural level, and intraoperative or postoperative stretch of the nerve.
Prior to discharge from hospital, the patient should have a very clear understanding of what was performed at operation, A simple sketch may assist in explaining. Sutures and staples are conventionally removed between the eighth and tenth postoperative day. Those removing the sutures should understand the required duration of immobilization. Following simple neurolysis, the managing physician may want the patient to move into a vigorous passive and active physical therapy program without delay. If a direct nerve suture has been performed, immobilization of the joints may be required for 3 or more weeks, with progressive extension of the joints for up to 6 weeks. Premature mobilization may result in suture-line disruption. Those with graft repairs are permitted more motion in the early weeks but should avoid hyperextension or hyperabduction of the joints.
If possible, it is appropriate to reassess the patient at approximately 6 weeks following nerve repair. This allows the physician to be certain that the patient understands what is required and that the patient is not adopting an overly passive stance. Third-party insurance agencies should understand the significance of the injury and the likely duration of disability. If the patient can return to the labor force, even in a reduced capacity, this should be encouraged immediately. The next follow-up visit should coincide with the anticipated time of initial motor recovery of the first downstream muscle. At this stage, the patient should be encouraged to exercise those muscles which have received axonal regrowth. This is also a good time to observe whether or not there is evidence of malingering. medical-legal neurosis, or prolongation of disability by compensation payments. In these cases, evidence of sympathetic dystrophy may become more and more obvious as prolonged overprotection and immobilization of limbs leads to secondary, sympathetically maintained pain syndromes and joint contracture. Decision-making with regard to patients suffering from peripheral nerve injury is straightforward if there is either clear-cut evidence of progressive improvement or obvious evidence of no improvement. The management of patients demonstrating partial improvement requires considerable experience. On occasion, re-exploration may be judged too risky, as the patient has too much to lose and not enough to gain. On Other occasions, reoperation may be indicated.
Treatment strategies for the injured peripheral nerve are guided by an accurate evaluation of the initial injury. A reasonable approach to these injuries is presented as a simplified flow chart. The level particularly the proximal extent, of the injury should be determined very carefully using clinical, electrodiagnostic, and radiologic criteria. This assessment is especially critical in selecting patients for direct brachial plexus repair, which provides the best opportunity for some functional recovery. Signs of regeneration should be sought by using clinical examination and electrophysiologic studies. Initial neural recovery is best demonstrated by return of motor activity. Nerve stimulation may be required to establish with certainity that motor function has returned. Definite sensory recovery in autonomous zones can be important early evidence of regeneration in distaI lesions of major sensory nerves. Sensory recovery usually occurs relatively late, however, and this may be misleading, especially in the case of radial and peroneal nerves.
The ability to obtain NAP recordings across the nerve lesion provides additional early evidence for adequate regeneration. This usually requires an operative exposure of the involved nerve. Other qualitative evidence for neural regeneration is provided by Tinel's sign and EMG. If each of these sources fails to give even slight evidence favouring regeneration, and the deadline for significant reinnervation is many months away, surgical exploration will be necessary. Should a neuroma in continuity be found at exploration, the surgeon is best guided with regard to resection and nerve repair by NAP recordings. If nerve resection and repair is needed, adequate mobilization of the nerve and apposition of healthy nerve ends either end-to-end or with autologous nerve grafts is indicated.
Results with all types of nerve repair are related to the specific nerve involved as well as level involved. Radial repairs fare better than median, and median better than ulnar. In the lower extremity, tibial and femoral nerves fare better than peroneal nerve. The need for rehabilitation and good physical and occupational therapy pre- and postoperativdy is very important for useful recovery. Even this will fail to produce function in a stiff and often painful hand or foot.