This paper is an introduction to the principles of nerve repair as it relates to the trigeminal nerve. The anatomy, the mechanisms of damage to its branches, the mechanisms of healing and the problems which face the surgeon in attempting to repair the branches of this nerve by either primary, delayed primary repair or by nerve grafting will be discussed. Further papers will discuss the repair of the inferior dental and lingual nerves.
Nerve surgery in the maxillofacial region is largely confined to the trigeminal and facial nerves and their branches. The trigeminal nerve and branches can be damaged as a result of fractures to the facial skeleton, during tumour resection, or implant placement but most often as a result of the removal of teeth, particularly the mandibular third molar.
The trigeminal nerve arises with a larger sensory root and a smaller motor root from the ventral surface of the pons. The motor root is inferior to the sensory root and passes laterally over the deep surface of the trigeminal ganglion to enter the third mandibular division and thence to the muscles of mastication while the sensory fibres constitute the trigeminal ganglion and distribute to their various locations from this ganglion.
The first division, the ophthalmic division, passes anteriorly within the lateral wall of the cavernous sinus to the medial part of the superior orbital fissure and into the orbit almost immediately splitting into the nasocilliary, frontal supraorbital and supratrochlear branches and the lacrimal branch.
The second division, the maxillary division, courses downward and forward and enters the upper part of the pterygopalatine fossa through the foramen rotundem and then into the orbit through the inferior orbital fissure and out via the infraorbital foramen to supply the anterior part of the face and nose. The posterior, middle and anterior superior alveolar branches of the infraorbital nerve enter the maxillary alveolus from above to supply the maxillary teeth.
The third division, the mandibular division, travels through the foramen ovale and into the infratemporal fossa, the lingual nerve separating away from the mandibular division close to the base of the skull, just after leaving the foramen ovale. It lies anterior and medial to the inferior alveolar nerve and descends between the lateral and medial pterygoid muscles. At the lower end of the lateral pterygoid muscle, it receives fibres from the chorda tympani and thus the fibres of taste. The nerve then follows the lateral surface of the medial pterygoid muscle. At the upper end of the mylohyoid muscle, it curves sharply medially over this muscle and into the mouth. At this point it releases fibres to the submandibular ganglion, and then passes into the tongue. This is where it is vulnerable to the removal of the third molar teeth and it is in this area that repair is required if needed.
The inferior alveolar nerve winds around the lower border of the lateral pterygoid muscle and then turns sharply lateral to reach the inner aspect of the mandible and into the body of the mandible via the mandibular foramen, passing laterally within the mandibular canal and exits via the mental foramen. However, before leaving the mandible, it loops anteriorly and then superior and posteriorly in the premolar area. The two areas of vulnerability of this nerve are therefore at the third molar site as it courses laterally in the mandible and in the premolar area, where it loops around before leaving the mandible by the mental foramen. Each division supplies part of the face, and in combination, the three branches supply sensation to most of the face with the exception of the soft tissues around the angle of the mandible, which is supplied by branches of the upper cervical nerves.11
Histologically, the cell body of a spinal nerve is located in the spinal cord or the posterior root ganglion, with the axons extending into the sensory receptors. The cell body of the trigeminal nerve is found within the trigeminal ganglion, with the axons extending out into the sensory receptors which include the pulps of the teeth, the periosteum of the jaws, oral mucosa, temporomandibular joint (TMJ) and the facial skin. Stimulation from these sensory organs synapse in the trigeminal ganglion and pass on to the nucleus caudalis in the medulla, pons and upper cervical segments of the spinal cord, and then on to the cortex after crossing over in the central trigeminal tract.
These axons and their Schwann cells are covered by a thin endoneurium, with groups of them joining together to form the fascicles. The fascicles in turn are supported by the perineurium, which separates them into functional units. The epineurium is the connective tissue sheath surrounding the perineurium and forms the nerve trunk. The mesoneurium surrounds the entire structure and contains the blood vessels which provide the nutrition for the framework.12
Damage to the trigeminal nerve can be of various magnitudes and may be a simple contusion or bruising to complete transection. Commonly, branches of the trigeminal nerve are damaged as a result of trauma and in particular during the removal of teeth, most often the mandibular third molar.
According to Tay and Zuniga,13 other common causes include local anaesthetic injections, orthognathic surgery and the placement of dental implants. These authors went on to say that 60% of injuries to the inferior dental nerve were due to dental extractions and that the inferior dental nerve was the most common branch damaged, followed by the lingual nerve. Pogrel,14 in a cadaveric study, concluded that other causes may include scalpel blade injury, transection, crush and stretch injuries.
Various classifications for nerve injuries have been proposed but the most commonly used has been proposed by Seddon.15 He classified nerve damage into three main categories: (1) neuropraxia, which is due to a local conduction block with a decrease in axonal conduction, or a localized injury as a result of indirect pressure, i.e., retraction during surgery or oedema after surgery. This injury resolves rapidly once the pressure has been released or the swelling has subsided; (2) axonotmesis, which occurs as a result of complete destruction of the axonal conduction and degeneration of the distal segments, without disruption of the supporting structures. Vigorous retraction or compression will result in this type of injury. Recovery depends upon the distance between the site of reinnervation and the damage to the nerve; (3) neurotmesis occurs as a result of total disruption of axonal conduction, as well as the supporting structures of the nerve. This type of damage can be subclassified into (a) epineural disruption and (b) perineural disruption.
Rood16 further refined this classification into:
|1.||Neuropraxia||Grade 1||Functional conduction block|
|Ischaemia or oedema|
|Grade 2||Demyelinization no axon loss|
|2.||Axonotmesis||Grade 1||Axon degeneration wallerian|
|Grade 2||Axon degeneration endoneural|
|3.||Neurotmesis||Grade 1||Perineural disruption|
|Grade 2||Perineural and epineural|
Sunderland17 considered nerve damage and subdivided them into five separate headings depending upon severity and these may be considered under the following headings: (1) conduction block; (2) transection of the axon with intact endoneurium; (3) transection of the nerve fibre axons and sheath inside an intact perineurium; (4) transection of the fascicles, nerve trunk continuity being maintained by the epineural tissues; (5) transection of the entire nerve trunk.
The gate theory of pain can answer some of the problems relating to nerve injury and repair. Input from the sensory organs of the face is subject to various conscious and subconscious stimuli which can modify this sensory input. The nucleus caudalis acts as a “gate” for the sensory input from the trigeminal nerve, and if the input is pain it can then be modified within the nucleus caudalis by either inhibitory or stimulatory influences and therefore modify the sensations reaching the cortex. This gate theory of pain described by Melzack and Wall18 in part explains the clinical observations of pain in most injuries and diseases.
Pain is transmitted by the unmyelinated fibres and touch. Temperature and proprioception are transmitted by the myelinated fibres. The myelinated fibres act as a counterbalance to pain and are inhibitory to the “gate” while the unmyelinated fibres are not. The unmyelinated fibres are more primitive, slower in activity, are more resistant to injury and are more easily regenerated. The faster myelinated fibres have evolved later, and because of their complexity are less regenerative and die as a result of injury. These two influences usually remain in balance but this can change either due to trauma, disease, or other influences. In the case of pain associated with herpes zoster, the virus affects the myelinated fibres which alter the balance at the “gate” in the pons and as such the unmyelinated fibre activity is greater and the pain is also greater.
Therefore, if the trigeminal nerve is damaged following third molar removal and there has been a complete severance of the nerve, i.e., neurotmesis (according to Seddon) or a level-five injury (according to Sunderland), all conduction ceases and there is no sensory input to the brain, and the area will feel anaesthetic. Regeneration of the proximal end then takes place, which is at maximum at about three months19 but the distal end is subject to wallerian degeneration and dies away. The proliferating axons attempt to bridge the gap and if they find a perineural tube to grow down, gradually function will occur.
However, this is not the usual situation and many factors can influence the problem as mentioned by Wolford and Stevao.10 Often a neuroma will form in the gap between the two ends of the nerve, which is a frustrated attempt at repair; the regenerating axons from the proximal end cannot find a perineural tube to grow down, and as such an enlargement grows at the proximal end. The neuroma consists of dense fibrous tissue intermingled with nerve fascicles, which are often painful due to the proliferation of the more primitive unmyelinated pain fibres. The myelinated fibres are more complex, are slower to recover and require more nutrients and oxygen to synthesize myelin, which can then lead to an imbalance of the fibres ascending to the gate.
If, however, the injury is not a complete transection of the nerve, the more primitive, and therefore hardier unmyelinated fibres recover in greater number while the myelinated fibres die off and a neuroma may develop. More often though, as the area heals, avascular scar tissue is formed which further favours the loss of the myelinated fibres, for the same reasons as mentioned above. Therefore, this imbalance leads to an increase in pain from the partially resected nerve. Because of the partial injury, the cell bodies in the trigeminal ganglion do not die but remain as injured cells and produce pain mediators which can increase pain levels. This deafferentation neuropathy is not able to be treated by peripheral surgery, such as resection of the neuroma and repair by delayed primary anastomosis or nerve grafting but instead is more amenable to treatment by central acting medications such as Tegretol. The unmyelinated, or C-fibre neuroma and deafferentation neuropathies are the two most common responses of a nerve to injury and can cause long-term pain for the patient.20
This is not always the case and some nerve injuries recover spontaneously neuropraxia and others require repair or graft axonotmesis and neurotmesis in order to regain function. Therefore, the prime reasons for surgical intervention are: (1) an injury which produces a defect or damage which does not recover spontaneously; and (2) a loss of function resulting in anaesthesia, paraesthesia or dysaesthesia which cannot be corrected by other means.
Documentation of sensory nerve injury is important in determining the nature of the problem and the type of injury. The first step is to obtain the patient’s main complaint, whether it is a loss of sensation, pain or some other abnormal sensation or functional impairment.21 In taking the history, the important things to determine are the nature of the injury, the date of the incident and the progress of the symptoms. Return of sensation within the first four weeks indicates a neuropraxia and implies an excellent prognosis, whereas a late onset in the return of function indicates a more severe injury, such as an axonotmesis, and if there is no return of sensation by three months, neurotmesis should be considered.
The purpose of neurosensory testing is to determine the outline of the sensory deficit, quantify the magnitude and character of the deficit and record it for comparison in an objective manner over time. The clinician should look for objective signs of damage, such as oedema, erythema, ulceration or signs of hyperactivity of the sympathetic nervous system, such as blanching, flushing and changes in temperature or sweating over the anatomical distribution of the nerve. Palpation directly over the accessible portion of the injured nerve, such as the lingual aspect of the mandible, or in the buccal vestibule in the mental nerve area, or over the skin of the face supplied by the infraorbital nerve may elicit a tinel sign or tingling over the distribution of the nerve. It may also elicit a painful sensation indicating possible neuroma formation, particularly on the lingual aspect of the mandible with the lingual nerve. If the pain follows an anatomical pattern, an in-continuity neuroma may be suspected, and if there is pain without radiation, neurotmesis and neuroma formation may be suspected. Radiographs and particularly CT scans may indicate foreign bodies, such as screws, implants or other alloplastic materials which may be causing the problem. In addition, fine-slice CT or MRI scans may define the problem further.
Neurosensory testing is designed to evaluate as objectively as possible, the nature of the problem and to standardize the tests in order to follow the patient over time and to determine improvement or change. Different tests will differentiate which of the fibres are recovering. As mentioned above, there are two basic fibres we are dealing with and these are the myelinated and non-myelinated fibres. The more primitive or non-myelinated, slow-acting fibres are the pain fibres and recover more quickly than the myelinated ones. Common tests include light touch, brush directional touch, two-point discrimination, temperature change and pin-prick. The first sensation to return is pain and this can be determined by pin-prick; the others are slower and determine the finer points of the recovery as the myelinated fibres recover. Photographs or diagrams are helpful in documenting the injury and its recovery.22 More definitive and sophisticated tests (somatosensory evoked potentials) can be used in the research laboratory but simple testing will suffice to determine the nature of the problem, to record the return of function and monitor the return after surgery (Fig 1).
Timing of repair
Early exploration and repair of motor nerves, such as the facial nerve, offers the best result. Motor end-plate function is easily lost and is critical for recovery. However, sensory function and the sensory receptors are not so critical. It has been traditional in the past to wait and see what happens following damage to the divisions of the trigeminal nerve. Alling23 cast some doubt about the concept of early repair of the lingual and inferior alveolar nerves when he reported on a retrospective study carried out on 103 members of the American Association of Oral and Maxillofacial Surgeons (AAOMS). He found that damage to the lingual nerve occurred in 0.06% of cases and to the inferior alveolar nerve in 0.43% of cases. He also reported that few of these presented a permanent problem. However, Pogrel,24 in a chapter in Complications in Oral and Maxillofacial Surgery, indicated that most authors recognize that there is an approximately 0.6% to 5% incidence of sensory damage to the trigeminal nerve following third molar surgery but most recover according to Alling.23 Hillerup and Stolze25 found that most lingual nerve injuries were closed injuries, were not known to the operator and as such the nature of the injury was unknown. They also recommended a period of observation and record of three months before intervention. As most of the injuries to the branches of the trigeminal nerve are closed and the operator is unsure of the nature of the injury, it is sensible to have a period of observation and record before surgical intervention. Therefore, it would seem logical to repair the nerve as early as possible once it was clear that it was not going to recover. It is generally felt that 3–6 months is the optimal time to wait between injury and attempted repair. This concept has also been advocated by many other authors.6,9,10,19–25
Theoretically, the best time to repair a nerve is at a time when the regeneration is most active. Holmes and Young27 report that the proliferative power of the Schwann cells is at a peak 2–3 weeks after injury and is said to decline after about three months. Coupled with this, it is felt that shrinkage of the distal perineural tubes, which increases with time after injury, has a great effect on the advancing axons ability to grow down the distal tubules. Therefore, it is critical to consider surgery at this time, i.e., around three months after injury if one is to capture this activity.10
According to Donoff,28 if no spontaneous recovery has occurred by six months it is unlikely to occur and therefore surgery should be carried out at this time or earlier which will allow for maximum recovery power of the Schwann cells and also allow the operator time to assess the injury and allow for any spontaneous recovery.
Wolford and Stevao10 have indicated that early repair also circumvents many of the problems relating to wallerian degeneration, which includes atrophy and fibrosis of the distal portion of the nerve. This atrophy and shrinkage creates many of the problems of mismatch which become apparent when attempting to repair the nerve, particularly if a graft is required and the timing is late. They also state that if the injury is primarily a traumatic neuroma, timing of the repair is not as important as it will have little bearing on the final outcome.
According to Sunderland,29 complete recovery after nerve repair will not occur and this will be less if a graft is required. It is, therefore, apparent that the best time to repair the nerve is at the time of injury but this is often not practical because of the access necessary to repair the nerve and the need for informed consent. Also, because many of these lesions are closed, the operator is unaware of the problem until after the surgery has been completed.
Robinson et al.9 have offered algorithms for the management of inferior alveolar and lingual nerve injuries which are useful. These algorithms offer a flow chart for the management of these nerve injuries.
Principles of nerve repair
It is accepted30 that perineural repair will offer the best chances of success. However, it is not always possible to perform a perineural repair and so an epineural repair will have to suffice.4 Although perineural repair will offer the best anatomic restoration of the nerve, it should be looked upon with some reality.
The inferior alveolar nerve is a neurovascular bundle which lies within the mandible and has numerous fascicles which vary in size and number. According to Svane,31 there may be as many as 18 or more fascicles at the third molar site. Therefore, it is not practical to suture each perineural tube together as each suture will evoke an inflammatory response and produce a fibrous tissue reaction which reduces the chance of recovery by inhibiting the growth of axons. It is therefore prudent to carry out an epineural repair. According to Wolford and Stevao,10 perineural repair also produces more trauma by dissecting each individual fascicle and suturing them together. They state that the trigeminal nerve branches are polyfascicular, i.e., multiple fascicules of different size and non-grouped, and therefore the epineural suture technique is the most appropriate method of repair in this group.
According to O’Brien and Morrison,32 there is no convincing evidence that a perineural repair is significantly better than an epineural repair if magnification is used as magnification will allow more accurate alignment of the fascicles with the repair.
Hausaman30 recommended that a tension-free anastamosis is also important for functional return. Miyamoto33 found that tension across the repair of greater than 23 gms will inhibit axonal growth as the gap produced within the nerve structure will be too great for the axons to grow down the distal tubes. Hausaman30,34 also stated that if the nerves cannot be placed together in a passive, tension-free way, then a nerve graft should be considered.
There are many factors which affect the results of nerve repair. These can be considered under a number of headings. According to Wolford and Stevao,10 the factors affecting the success of the procedure are multiple and include: (1) the time between injury and repair; (2) the type and extent of injury; (3) the vascularity of the site; (4) the orientation of the nerve and graft; (5) the length of graft required; (6) the quality of the repair; (7) the tension of the repair; (8) the preparation of the graft; and (9) the age and general health of the patient. These factors can be split into those external to the nerve and those local to the nerve.
The factors external to the nerve include the type and extent of injury, the amount of contusion, bruising and vascularity of the surrounding tissues, the age of the patient and the presence or absence of infection. Of these factors, age and general health are the most important, and in all reported cases the best results occur in children.35
The factors local to the nerve include those proximal to the injury, at the injury, those distal to the injury, the timing of repair, the technique and quality of the repair and tension at the site of the anastamosis.
Factors proximal to the injury relate to wallerian degeneration, which usually relates to the distal segment but degeneration will also occur along the proximal end, usually to the first node of Ranvier. In severe injury this damage may be seen as far proximally as the ganglion.36 Thus, in severe cases there may be damage along the proximal stump, which is beyond control of the surgeon even if a meticulous repair is performed.
The level of injury will directly influence the result of regeneration. In general, the more proximal the injury, the more rapid will be the initial response but the final outcome may be poorer due to the longer distance the axons have to travel to their destination and the higher incidence of fascicular cross-over. In addition, there is a greater chance of the distal fascicles collapsing before the axons reach their final destination. This is probably not as important with the trigeminal nerve as the distances involved are relatively short. There is also the possibility of end-organ collapse, e.g., muscle wasting, although this is less likely with peripheral sensory nerves than motor nerves. Therefore, damage to the mental nerve will have a greater chance of recovery and over a shorter time than damage at the third molar site. Also, the more proximal the injury, the more difficult it becomes to repair because of the problems of access. This is particularly so with the inferior alveolar nerve, which has to be removed from its position within the body of the mandible. The lingual nerve is also difficult to approach because of its posterior position within the mouth and the need for retraction of the tongue during repair, which is often difficult.
The type of nerve will also make a difference because a pure motor or pure sensory nerve will recover more quickly than a mixed nerve. The trigeminal nerve is mixed as it supplies sensation to the face and also motor function to the muscles of mastication, and in the case of the lingual nerve, taste and stimulus to the submandibular salivary gland.
The nature and extent of injury will also affect recovery, and the more extensive the injury, the less chances of recovery. Associated tissue damage will also play a role because of disruption to the blood supply, necrosis of tissue, infection and possible scar tissue formation.
Factors distal to the injury are also important and once again relate to wallerian degeneration. Generally, the distal tubules will lose 70% of their original diameter over a period of weeks after injury.31 They will, however, retain their capacity to enlarge when axons enter the tubule. The relative sizes of the proximal and distal tubules will influence the repair and will also influence the orientation of the fascicles.10 The use of magnification will, therefore, help with this alignment. This shrinkage factor adds further evidence for early repair when the proximal and distal tubules are more correctly aligned.
Considerations for nerve grafting
Nerve grafts are required when there is a break in the continuity of a nerve and at the time of repair. After preparation of the proximal and distal stump, the two ends cannot be brought together without tension, in which case a graft will be required to bridge the gap. The two most common nerves used for grafting the trigeminal nerve are the sural nerve in the leg and the greater auricular nerve in the neck. Other nerves can be used for grafting but these are the most convenient. Each nerve has its own problem from the point of view of morbidity from the surgery as the patient will be left with a sensory loss over the lateral aspect of the foot if the sural nerve is used, and to the ear and lateral face if the greater auricular nerve is used. The patient requires an informed discussion about the problems involved so they can make a decision as to which graft donor site they wish to choose. Miloro and Stoner36 subjectively assessed outcomes following sural nerve harvest and found that most patients tolerate sural nerve harvest without significant donor site morbidity.
The donor nerve and the damaged nerve need to approximate one another in diameter, fascicular size and number if the graft is to be successful. The average diameter of the inferior alveolar nerve is 2.4 mm30,38 and the lingual nerve is 3.2 mm.39 The greater auricular nerve is 1.5 mm in diameter and the sural nerve is 2.0 mm in diameter which is less than the diameter of the inferior alveolar nerve. Svane30 also indicated the fascicular size and numbers varied in the inferior alveolar nerve and in some instances there were up to 18 fascicles at the third molar site, reducing down to 12 at the mental foramen. The fascicular size and number in the sural and greater auricular nerves are much less with 8–9 fascicles40 in the greater auricular nerve and 1239 in the sural nerve. The lingual nerve has approximately 15–18 fascicles41 at the third molar site, reducing down to 9 as it enters the tongue. Therefore, there is a considerable mismatch between the diameter of the nerves to be repaired and the proposed grafts, including the number of fascicles. However, this can be overcome with multiple cable grafts in parallel and in this way the fascicular patterns can be more fully utilized.10
The length of the graft can occasionally be an issue as the maximum length one can obtain from the greater auricular nerve is approximately 2–4 cm, while that of the sural nerve is 20–30 cm. In addition, the inferior alveolar and lingual nerves are polyfascicular in pattern, with the fascicles varying in size and shape.30,41 The fascicular pattern is more uniform in the sural nerve but less in number.30 The greater auricular nerve is polyfascicular in pattern and therefore more accurately approximates the inferior alveolar and lingual nerves, particularly if parallel cable grafts are used for repair.10
The cross-sectional shape of the sural nerve is again a little different from that of the inferior alveolar and lingual nerves. These nerves are round and the sural nerve is more flattened. The greater auricular nerve is round and therefore more compatible with the inferior alveolar and lingual nerves. The cross-sectional area of the inferior alveolar nerve30 and the lingual nerve41 are much greater than the sural nerve.
All of these parameters indicate a mismatch between the various nerve grafts which are often used for grafting purposes and will possibly influence the outcome of the repair.
Nerve grafting with other materials
Alternative conduits have been used for nerve grafting. These include autologous vein grafts42 and the use of various allografts such as Gore-tex (WL Gore & Associates Inc, Flagstaff, AZ, USA)43 and polyglycolic acid (PGA), such as Neurotube (Neurogen LLC, Bel Air, MD, USA).10 The use of the posterior facial vein or the long saphenous vein in the leg have also been used for grafting purposes.42 Although the long saphenous vein has valves within it, which delays the ingress of axons, this graft can be turned inside-out when used, to overcome these valves within the graft but also to utilize the effects of the nerve growth factors which are thought to be in the adventitia.43 Pogrel41 has reported good success with the use of vein grafts, particularly for the inferior alveolar nerve where the nerve is protected by the mandible.
The inherent strength of the vein is of importance, and the possible collapse of the conduit, and therefore kinking, may lead to an inhibition of the advancing axons. The success of vein grafts for the lingual nerve is not as good as it is for the inferior alveolar nerve as the lingual nerve is subject to movement and the grafts, of any length, are thought to kink and therefore be non-viable. The allograft conduits appear to overcome this problem as their inherent stiffness is better but reports indicate that Gore-tex grafts were not particularly successful.44 However, according to Wolford and Stevao,10 the resorbable conduits have been found to be effective in preliminary studies. In addition, these conduits could be filled with nerve growth factors, similarly the resorbable conduits. These conduits could also be used to protect a nerve graft, as could a vein, and this would overcome possible collapse.