Surgical management and outcome in patients with radial nerve lesions

Transcription

1 J Neurosurg 95: , 2001 Surgical management and outcome in patients with radial nerve lesions DANIEL H. KIM, M.D., ANDREW C. KAM, M.B.B.S., F.R.A.C.S., PADMAVATHI CHANDIKA, M.D., ROBERT L. TIEL, M.D., AND DAVID G. KLINE, M.D. Department of Neurosurgery, Stanford University Medical Center, Stanford, California; and Department of Neurosurgery, Louisiana State University Health Sciences, New Orleans, Louisiana Object. The goal of this paper was to review surgical management and outcomes in patients treated for radial nerve (RN) lesions at Louisiana State University Health Sciences over a period of 30 years. Methods. Two hundred sixty patients with RN injuries were evaluated. The most common mechanisms of injuries involving the RN included fracture of the humerus, laceration, blunt contusions, and gunshot wounds. One hundred and eighty patients (69%) underwent surgery. Lesions not in continuity required primary or secondary end-to-end suture repairs or graft repairs. With the use of direct intraoperative nerve action potential recording, RN injuries in which the lesion was in continuity required external or internal neurolysis or resection of the lesion followed by end-to-end suture or graft repair. A minimum of 1.5 years follow-up review was available in 90% of the patients who underwent surgery. Motor function recovery to Grade 3 or better was observed in 10 (91%) of 11 patients who underwent primary suture repair, 25 (83%) of 30 who underwent secondary suture repair, 43 (80%) of 54 who received graft repair, and 63 (98%) of 64 in whom neurolysis was performed. Sixteen (71%) of 21 patients with superficial sensory RN injury achieved satisfactory pain relief after complete resection of a neuroma or neurolysis. Conclusions. This study clearly demonstrates that excellent functional recovery can be achieved with proper surgical management of RN injuries. KEY WORDS radial nerve injury neurolysis suture repair graft repair outcome I Abbreviations used in this paper: BM = brachialis muscle; BRM = brachioradialis muscle; ECRBM = extensor carpi radialis brevis muscle; ECRLM = ECR longus muscle; EMG = electromyographic; GSW = gunshot wound; LSUHS = Louisiana State University Health Sciences; NAP = nerve action potential; PIN = posterior interosseous nerve; RN = radial nerve; SM = supinator muscle; SRN = superficial RN; SSRN = superficial sensory RN. NJURY to the RN can result in significant functional motor impairment of the arm and wrist. Although intrinsic muscles of the hand are not compromised by RN injury, the ability to extend the forearm, wrist, and fingers may be lost. With loss of wrist extension, the ability to grasp is significantly reduced due to a reduction in the mechanical advantage of flexion at the distal interphalangeal joint by the flexor digitorum profundus muscle. Sensory deficits as a result of RN injury are functionally less significant because they involve the dorsum of the hand and the anatomical snuffbox. The purpose of this study was to evaluate retrospectively 260 patients for injuries involving the RN and its branches. Results in 180 patients who had undergone surgical repair for severe injury were analyzed with respect to the mechanism of injury, the type of lesion (in continuity or not in continuity), and the type of repair that was performed. Pain amelioration following surgical resection of SSRN neuroma was also reviewed. Excellent spontaneous and postsurgical functional recoveries of the RN were observed, although as expected, recovery of proximal muscles supplied by the RN was more encouraging than recovery of the distal muscles. Radial nerve recovery can be attributed to several factors: 1) the predominance of motor fibers that reduce the possibility of cross motor sensory reinnervation; 2) muscles innervated by the RN receive their input relatively proximally within the limb and are not involved in delicate movements that require complex coordinated muscle contraction; and 3) even though the initial loss of the ECRLM results in significant disability, the muscle recovers well. 25 Despite these considerations, surgical repair of the injured RN may not always be sufficient to achieve restoration of finger and thumb extension, and tendon transfer surgery may be necessary. Clinical Material and Methods Patient Population Between 1967 and 1997, a total of 260 patients with lesions of the RN and its branches were evaluated retrospectively at LSUHS. Based on preoperative clinical and electrophysiological evaluations, 180 of the 260 patients with severely damaged or poorly recovering lesions underwent surgical exploration and repair. The minimum follow-up period was 12 months, and the mean follow-up period was 18 months. The age of patients comprising the surgical group ranged from 8 to 64 years of age (mean 42 years). 573

2 D. H. Kim, et al. TABLE 1 The LSUHS grading system for motor and sensory function* at its origin. It descends to be posterior to the pectoralis major muscle and successively in front of the subscapularis, teres major, and latissimus dorsi muscles. Close to its origin, the RN produces a variable number of branches that run obliquely to supply one or more of the three heads of the triceps muscle. Accompanied by the profunda brachii artery, the RN passes laterally and posteriorly, running deep with respect to the long head of the triceps to lie on the posterior aspect of the humerus (Fig. 1). Between the lateral and medial heads of the triceps muscle, the RN runs along the spiral groove under the cover of the lateral head of the triceps, issuing further muscular branches to the medial and lateral heads of the triceps and anconeus muscles. As the RN reaches the lateral border of the medial head of the triceps, it pierces the lateral intermuscular septum and continues its spiral course forward in close association with the lateral side of the BM and BRM. Radial Nerve at the Elbow. The RN is located in the groove between the BM and the BRM; it descends between the BM and the ECRLM and then passes in front of the lateral epicondyle into the forearm. The motor branch to the BRM originates approximately 2 or 3 cm proximal to the elbow and 7 or 8 cm distal to the humeral groove. From its medial aspect, the RN also produces motor branches to the BM, which also receives a supply from the musculocutaneous nerve. Additionally, several branches of the RN in the distal arm proximal to the elbow contribute a motor supply to the BRM. The ECRLM can receive input from the RN before the latter bifurcates into the PIN and SSRN, or sometimes the ECRL bundle will travel in the SSRN for a short distance and then leave it to innervate the ECRLM. In the antecubital fossa, under the cover of the BRM and ECRLM, the RN divides into the PIN and the SRN (Fig. 2). The RN has two cutaneous branches near the lateral intermuscular septum, the inferior lateral brachial cutaneous, and the posterior antebrachial cutaneous. The inferior lateral brachial cutaneous, also known as the upper branch of the posterior antebrachial cutaneous, pierces the deep fascia close to the lower brachial portion of the cephalic vein, supplying the skin that lies over approximately the lower half of the lateral and anterior aspects of the arm. The posterior antebrachial cutaneous, the lower and larger branch, pierces the fascia of the arm a little below the level at which the lateral brachial cutaneous branch emerges and descends to the back of the forearm to supply skin as distal as the wrist or the dorsum of the hand. Posterior Interosseous Nerve. Arising at variable levels below the elbow, the PIN is the deeper terminal branch of the RN in the antecubital fossa. The PIN innervates the extensor muscles of the forearm, except for the ECRLM, and contains afferent fibers from the wrist joint. Just before passing through the SM, the PIN branches off to the ECRBM and SM. The PIN passes between the superficial and deep laminae of the SM. The most proximal part of the superficial head of the SM may be tendinous and forms a fibrous arch known as the arcade of Frohse, which is a possible site of RN entrapment. Also, the ECRBM can sometimes have a fibrous edge that may compress the PIN as it enters beneath the arcade of Frohse. After running obliquely through the SM, the PIN cross- Evalu- Grade ation Description individual muscle grades proximal RN lesion 0 absent absent radial motor function 1 poor trace of contraction in triceps muscle, absence or trace of contraction in BRM against gravity only 2 fair triceps & BRM contract against force, but absence or trace of supination & no wrist extension 3 moderate triceps & BRM contract against force; supination & wrist extension against gravity; absence or trace of finger or thumb extension 4 good triceps & BRM contract against force; supination & wrist extension against force; trace or better finger & thumb extension 5 excellent good triceps & BRM contraction, supination & wrist extension; finger & thumb extension at least against gravity & some resistance individual muscle grades distal RN lesion involving PIN 0 absent absent extensor carpi ulnaris, extensor communis, & extensor pollicis longus muscle function 1 poor trace of contraction in extensor carpi ulnaris muscle against gravity only; absent extensor communis & extensor pollicis longus muscle function 2 fair recovery of extensor carpi ulnaris muscle, absence or trace of extensor communis or extensor pollicis longus muscle, or both 3 moderate recovery of extensor carpi ulnaris muscle, some extensor communis muscle, absence or trace of extensor pollicis longus muscles 4 good full strength of extensor carpi ulnaris muscle; recovery of moderate strength of extensor communis & extensor pollicis longus muscles 5 excellent full recovery of extensor carpi ulnaris, extensor communis, & extensor pollicis longus muscles * The LSUHS muscle grading system was used to evaluate preoperative loss and postoperative recovery. This system is based on earlier British and American systems, but includes some important changes. For a mid-arm RN lesion (such as that caused by fracture of the humerus), the brachioradialis, supinator, and ECR are the proximal muscles, and the extensor carpi ulnaris, extensor communis, and extensor pollicis longus are the distal muscles. For a posterior cord lesion, the proximal muscles would be the latissimus dorsi, deltoid, and triceps, and the more distal muscles would be the brachioradialis, supinator, and ECR. Sensory loss is less significant and more variable with an RN lesion than with a median or ulnar nerve lesion. Therefore, sensory grading is not included when function for the whole RN is graded. A breakdown of patients who underwent surgery according to the level of injury to the RN and its branches follows: 83 patients with lesions at the arm level; 30 patients with injuries at the elbow/forearm level; 37 patients with PIN involvement; nine patients with lesions at the dorsal forearm level; and 21 patients with injuries at the wrist or distal forearm level that involved the SSRN. The mechanisms of injury included: fracture-related contusion; lacerations; contusion caused by blunt trauma; nerve entrapment; GSWs; and iatrogenic injury. Preoperative and postoperative evaluations were performed using the LSUHS grading system, whereby Grade 3 or better recovery is considered a favorable functional outcome (Table 1). Surgical Anatomy Radial Nerve in the Arm. The larger of the two terminal branches of the posterior cord of the brachial plexus, the RN lies posterior to the third portion of the axillary artery 574

3 Radial nerve lesions FIG. 1. Intraoperative photograph showing the RN at the arm level. The right RN is viewed from a posterior aspect, between the bellies of the long and lateral heads of the triceps brachii muscle. Note its close relationship to the accompanying profunda brachii vessels. Unless otherwise indicated, labels apply to muscles. A = artery; N = nerve. es the bare area of the proximal radius and enters the extensor compartment of the forearm. At the distal border of the SM, it fans out into multiple branches in two major components. One component innervates the more superficial layer of muscles (extensor digitorum communis, extensor digiti quinti, and extensor carpi ulnaris) and the second innervates the deeper muscles (abductor pollicis longus, extensor pollicis longus and brevis, and extensor indicis). After leaving the SM, the nerve lies under the extensor carpi ulnaris, extensor digitorum communis, and extensor digiti minimi muscles, with the abductor pollicis longus muscle located deep with respect to the nerve. Joined by the posterior interosseous artery, the PIN runs distally in the forearm, passes superficially with respect to the extensor pollicis brevis, which is located deep in relation to the extensor pollicis longus muscle, and then rests on the interosseous membrane by penetrating the extensor pollicis brevis muscle. Continuing distally along the interosseous membrane, it divides into terminal branches that provide sensory innervation to the wrist (Fig. 3). Superficial Sensory Radial Nerve. The SSRN is a direct continuation of the RN that is easily identified in the forearm as it courses under the edge of the BRM, lateral to the radial artery. At the junction of the middle and distal thirds of the forearm, approximately at the point at which the abductor pollicis longus and ECRL tendons meet, the nerve arrives at the extensor aspect of the limb. The nerve extends below the fascia to the subcutaneous plane, winds around the radius moving deep with respect to the tendon of the BRM, pierces the deep fascia, and splits into four or five branches. The distal branches of the SRN course across the anatomical snuffbox, superficial to the thumb extensor tendons, and are thus posterior to the wrist joint and the scaphoid bone. These branches supply a variable area of skin over the dorsal surface of the radial portions of three and one-half digits. Most commonly, it innervates the dorsoradial aspect of the thumb to the interphalangeal joint level, the web space between the thumb and index finger, the proximal interphalangeal level of the dorsal aspect of the index and middle fingers, and the radial portions of the ring finger (Fig. 4). Surgical Technique External neurolysis is the cornerstone for almost all peripheral nerve surgery. With the aid of magnification, neurolysis is performed in a circumferential fashion, both proximal to and distal from the site of injury, to expose the extent of the injured segment before intraoperative NAPs are recorded. Before suture or graft repair can be performed, any epineural scarring must be resected using fine dissecting scissors, microscissors, or a scalpel, and regions of bleeding at the epineural or subepineural level must be coagulated using an irrigating bipolar forceps. When performing intraoperative recordings of the NAP on a lesion in continuity, stimulating and recording electrodes are placed on the nerve proximal to the lesion to assess the NAP. The recording electrodes are then moved into the region of injury distal from the lesion, and changes in the evoked NAP are observed. If the NAP is present, external neurolysis, with or without internal neurolysis, is sufficient to achieve a favorable outcome. When no NAP is recorded across a lesion in continuity, resection and repair of the injured RN is necessary. 11,14 Internal neurolysis is indicated in cases in which an injury is more severe at one portion of the nerve than an- 575

4 D. H. Kim, et al. FIG. 2. Photograph depicting the RN as it enters the antecubital region. The right RN is exposed in the antecubital fossa with the BRM reflected laterally. Note that the PIN can be traced distally to pass between the two bellies of the SM. other, resulting in pharmacologically resistant severe neuralgia in the presence of an NAP transmitted across the lesion. Split or partial graft repairs may be necessary when individual fascicles or bundles of fascicles do not transmit NAPs. In cases of transected nerves or those in which a nerve segment unable to transmit an NAP requires resection, end-to-end epineural repair can occasionally be achieved with the aid of magnification. Sharp dissection is necessary before suture or graft repair to mobilize the proximal and distal stumps, with adequate resection of the stump to healthy epineurium and fascicular structure. After careful hemostasis and avoidance of excessive tension at the suture site, the repair is performed using an appropriate caliber (No. 6-0 or 8-0) monofilament interrupted nylon or Prolene suture. Meticulous attention must be paid to achieving anatomical alignment of the fascicles during apposition of the two stumps. If the distance between the stumps does not allow for direct end-to-end suture repair without excessive tension despite mobilization, a graft repair is necessary. 14 Results Arm Level Of the 120 patients evaluated for RN injuries at the arm level, 37 displayed spontaneous recovery, thus avoiding the need for surgical repair. The remaining 83 patients required surgical repair for persistent, severe RN deficits. Mechanisms of injury included 36 fracture-related contusions (43%); 13 contusions not associated with fractures (16%); 12 lacerations (14%); 10 GSWs (12%); nine iatrogenic injection injuries (11%); and three compression injuries (4%) (Table 2). In cases in which there was a fracture-related contusion, 36 (63.2%) of 57 patients experienced little or no recovery, necessitating surgical exploration (Fig. 5). In 31 of those patients the lesions were in continuity and in the remaining five there was distraction of the RN as a result of the fracture, which required resection of both stumps to healthy segments and then graft repair. Of the 31 patients in whom the lesions were in continuity, 28 patients (90.3%) achieved a Grade 3 or better functional recovery. Overall, 32 (88.9%) of the 36 achieved a Grade 3 or better functional recovery following surgery. Thirteen other patients suffered contusions related to blunt trauma and underwent surgical repair. All were found to have lesions in continuity, 11 (84.6%) of whom achieved Grade 3 or better functional recovery following external neurolysis. The two patients who did not achieve satisfactory recovery required lengthy graft repair ( 7 cm long) for segments with absent NAP recordings. For those cases in which the RN injury was caused by laceration, all 12 patients who underwent surgery had suffered a lesion not in continuity and required suture or graft repair. A Grade 3 or better functional recovery was observed in 11 (91.7%) of those patients. The sole patient who did not achieve satisfactory recovery underwent delayed secondary graft repair 9 months postinjury and had exhibited severe atrophy of the innervated radial muscles preoperatively. In the majority of patients with RN injuries following a GSW, the lesion was in continuity with varying degrees of intraneural derangement. Of the 15 cases evaluated, five displayed spontaneous clinical and electrophysiolog- 576

5 Radial nerve lesions FIG. 3. Photograph of the PIN as it is observed distal to the inferior margin of the SM. After passing between the two bellies of the SM, the right PIN emerges at the inferior edge of the SM, issuing branches to the extensor digitorum and extensor pollicis longus muscles. The posterior interosseous artery is seen to accompany the PIN. ical improvement at follow-up review, thus avoiding the need for surgical exploration. Ten patients (67%) who displayed little or no improvement during clinical and electrophysiological evaluation underwent surgical management. Ninety percent (nine of 10 patients) achieved Grade 3 or better functional recovery. Of the nine patients who underwent surgery for iatrogenic injection injuries to the RN, in four positive NAP recordings were obtained and the patients only required external neurolysis. In the other five the NAP was absent and the patients underwent resection and suture repair. All nine patients achieved a Grade 3 or better functional recovery, although extension of thumb and fingers remained somewhat weak. When lesions in continuity were caused by trauma, encouraging results were seen in those patients who required only external neurolysis because NAP recordings had been positive. Grade 3 or better functional recovery was observed in 21 (95%) of 22 cases. When no NAP was recorded, resection of the neuroma and suture or graft repair resulted in a favorable outcome in 39 (91%) of 43 cases (Fig. 6). In patients in whom the lesion was not in continuity, whenever primary or secondary end-to-end suture repairs were possible, a Grade 3 or better functional recovery was observed in 80% and 75% of those patients, respectively. By comparison, three (60%) of five of the patients requiring graft repair achieved favorable outcomes. Elbow and Forearm Level Forty-five patients were evaluated for RN lesions located at the elbow and forearm levels, excluding the PIN. Thirty patients (67%) displayed minimal or no clinical and electrophysiological recovery, which necessitated surgical exploration of the RN lesions. Common mechanisms of injury included nine lacerations (30%); nine fracturerelated contusions (30%); six injuries related to GSW (20%); four contusions secondary to blunt trauma (13%); and an additional two were caused by prior suture repairs for lacerations (Table 3). Of the nine patients with lacerations who were surgically treated, three underwent primary end-to-end suture repair, two underwent secondary suture repair, and two received secondary graft repairs. Two of the patients with lacerations at the elbow had partial injuries to the RN, enabling a partial split-graft repair of the nerve. All three patients with sharp lacerations who underwent primary end-to-end suture repair achieved an excellent recovery; one patient obtained a Grade 4 result and the other two achieved a Grade 5 recovery. Overall, eight of the nine patients achieved Grade 3 or better functional recovery (89%). In cases of RN injuries related to distal humeral fractures, eight of nine patients who underwent surgical exploration were observed to have a lesion in continuity for which the NAP was present. All of these patients achieved a Grade 3 or better recovery following external neurolysis. No NAP was present in the only patient who did not achieve a satisfactory result and that patient required resection and a lengthy graft repair of the transected nerve. In five of six patients with GSWs the lesion was in continuity and a positive NAP recording was obtained across the lesion resulting in neurolysis. One patient had suffered transection of the RN, requiring a secondary graft repair. 577

6 D. H. Kim, et al. FIG. 4. Photograph demonstrating the SSRN at the forearm level. The left SSRN is seen passing under the BRM, lateral to the radial artery. br. = branch; digi. = digitorum; Flex. = flexor; V = vein. TABLE 2 Outcomes in 83 patients who underwent surgery for RN lesions at the arm level* No. of No. No. W/ Cases Surgically Grade 3 Factor Evaluated Treated Outcome mechanism of injury laceration GSW fracture contusion compression injection total lesion category & type of surgery not in continuity 1st suture repair 5 4 2nd suture repair 8 6 2nd graft repair 5 3 in continuity NAP present neurolysis NAP absent suture repair graft repair * 1st = primary; 2nd = secondary; = not applicable. Despite that event, all six patients achieved a Grade 3 or better functional recovery. Of the eight patients who suffered an RN injury as a result of blunt trauma and contusion to the nerve at the elbow, half achieved a satisfactory spontaneous recovery. Of the remaining four patients who required surgical exploration, 75% achieved a Grade 3 recovery following external neurolysis, and one patient achieved a Grade 2 recovery following graft repair for an absent intraoperative NAP recording (Fig. 7). Three patients who had undergone previous primary suture repair at another institution were referred to LSUHS for evaluation because of their poor postoperative recovery. One patient showed promising signs of recovery and did not require surgical reexploration. The other two showed little improvement and underwent reoperation. Only one of these two patients regained a Grade 3 functional recovery after external neurolysis of excessive scarring. The patient with the poor result required resection of the neuroma at the previous repair site and a graft repair. Posterior Interosseous Nerve Of the 49 patients who presented with PIN syndrome, 37 (76%) underwent surgery after little or no improvement was observed following conservative nonsurgical management (Table 4). Almost half of these were patients suffering from PIN entrapment at either the arcade of Frohse or the distal margin of the SM (Fig. 8). Extremely encouraging results were obtained following complete division of the volar head of the SM, with 18 patients (94.7%) achieving an overall Grade 3 or better functional recovery. Results were excellent following repair of lacerations and fracture-related contusions to the PIN. Of the seven laceration-related injuries, two patients had PINs that appeared grossly to be in continuity, but displayed no NAPs. Repair was necessary and good results were achieved. The remaining five patients had RN transection as a result of a sharp laceration and were eligible for primary end-to-end suture repair. All seven patients achieved a Grade 3 or better functional recovery following surgical repair. Of the seven fracture-related RN injuries treated by surgery, one was found to have a complete transection with a long segment of neuroma, possibly related to three previous attempts to stabilize radius and ulna fractures. A 6.5-cm graft repair was necessary after the stumps were resected back to healthy fascicles. The remaining six patients had lesions in continuity. In three patients, positive NAP recordings were obtained and they underwent external neurolysis alone; in the other three NAPs were absent, requiring graft repair. Overall, postoperative outcomes were once again encouraging, with all seven patients achieving a Grade 3 or better functional recovery. 578

7 Radial nerve lesions FIG. 6. Photograph demonstrating the RN at the arm level. The RN is identified as it courses between the lateral (1) and medial (2) heads of the triceps muscle and pierces the lateral intermuscular septum to continue along its spiral course in close association with the lateral sides of the BM and BRM. Stimulating (triple electrode) and recording (double electrode) electrodes are placed across the lesion (arrowheads) to record NAPs intraoperatively. suture repairs. Nineteen patients underwent resection of a segment of the SSRN and significant pain relief was seen in 16 (84.2%) of them. The two patients who underwent external neurolysis following contusion to the distal forearm experienced poor amelioration of pain postoperatively (Table 6). FIG. 5. Plain x-ray film demonstrating a fracture of the humerus. Arrowhead points to the spiral fracture through the junction between the middle and distal third of the humerus, a common site for RN injury. Dorsal Forearm Level (Branches of the PIN) Eleven patients were evaluated who had sustained injuries to PIN branches occurring at the dorsal forearm level, distal to the SM (Table 5). Of those 11 patients, three had suffered PIN branch injuries caused by contusion, but two experienced spontaneous recovery and did not need surgical repair. Five of the nine surgical cases had lacerating injuries to PIN branches, four of which were sharp lacerations. Two of these patients underwent primary suture repair and achieved a Grade 3 or better functional recovery. A delay in referral and surgical exploration made it technically difficult to identify the distal stump in another case. It was necessary to identify the PIN more proximally in the forearm and trace it distally to these branches. This patient required a long graft repair with resultant Grade 2 functional recovery. Tendon transfer surgery was later offered to improve finger extensor function. Superficial Sensory Radial Nerve Injuries to the SSRN are known to result in painful hypesthesia over the dorsum of the hand, as well as the forearm level. Twenty-one of the 35 patients we saw underwent surgical management. Mechanisms of injury included 11 lacerations, eight contusions, and two previous Discussion Most of the serious RN injuries resulted from fractures, lacerations, contusions, or GSWs. A number of patients TABLE 3 Outcomes in 30 patients who underwent surgery for RN lesions at the elbow and forearm levels No. of No. No. W/ Cases Surgically Grade 3 Factor Evaluated Treated Outcome mechanism of injury laceration GSW fracture contusion injection Volkmann ischemic contracture prior suture total lesion category & type of surgery not in continuity 1st suture repair 3 3 2nd suture repair 2 2 2nd graft repair 6 4 in continuity NAP present neurolysis NAP absent suture repair 0 0 graft repair

8 D. H. Kim, et al. FIG. 7. Anterolateral view of RN, which is exposed between the BM and BRM. A vascular loop is placed around the radial artery. A lesion in continuity (arrowhead) is identified following blunt trauma. TABLE 4 Outcomes in 37 patients who underwent surgery for PIN lesions No. of No. No. W/ Cases Surgically Grade 3 Factor Evaluated Treated Outcome mechanism of injury laceration fracture contusion compression prior suture sleep compression total lesion category & type of surgery not in continuity 1st suture repair 1 1 2nd suture repair 3 1 2nd graft repair 3 3 in continuity NAP present neurolysis NAP absent suture repair 2 2 graft repair 4 4 FIG. 8. Photograph demonstrating signs of PIN syndrome. Weakness of the extensor digitorum communis (arrowhead), extensor digiti quinti, and extensor carpi ulnaris muscles is observed. This is caused by entrapment of the PIN at the distal edge of the SM. with iatrogenic injection injuries and entrapment syndromes experienced poor spontaneous recovery and thus required surgical intervention. For patients with severe injuries or poorly recovering injuries to the RN or its branches in whom little spontaneous improvement occurs, surgical timing is of vital importance. Injuries caused by sharp transections are best repaired within a 72-hour period. In contrast, transections from blunt injuries are best repaired after a delay of several weeks so that the margin of injury is better delineated, allowing for more accurate resection to normal healthy fascicles. Lesions in continuity cannot be accurately assessed during the acute period when seen during tendon repairs or open reduction of a fracture. The extent of nerve injury can be assessed by recording the NAP 2 to 3 hours following injury and, later sometimes, by stimulation of the nerve alone. The decision whether to resect and repair during the acute stage or perform a delayed neurolysis is worth the delay in time. Excessive delays in surgery, however, increase the risk of permanent loss of important motor function and the onset of muscular atrophy and contractures. Surgical techniques involving tendon transfers may then be necessary to achieve some functional use of the upper extremity. In this relatively large study group of 260 patients with lesions of the RN and its branches, as many as 80 patients (31%) exhibited varying degrees of spontaneous recovery, thus obviating the need for surgical exploration. Of the remaining 180 surgical cases that were analyzed, a high proportion of patients experienced satisfactory recovery following surgical repair. Almost half of the injuries were related to fractures and lacerations involving the upper extremity (Tables 7 and 8). In fracture-related injuries, 70% of RN injuries occur at the level of the arm. In particular, fractures involving the junction of the middle and distal third of the humerus and the midpoint of the humerus at the level of the spiral groove were associated with RN palsy. In their series of patients Bostman, et al., 4 observed that it was more common for the distal third of the humerus to sustain a longitudinal spiral fracture, which was commonly associated with a lacerated or interposed RN. Midshaft humeral fractures were more commonly transverse fractures associated with a neuropraxic injury. 5 Occasionally, the nerve was observed to be trapped in the fracture line after attempted closed reduction of the humeral fracture. 22,23,26 A period of observation is recommended for fracture-related RN injuries because spontaneous recovery rates as high as 76% have been reported in the literature ,23,26 If the fracture is to be reduced openly and internally instrumented or if vascular repair is necessary, the injured RN should be explored and studied by NAP monitoring if enough time has elapsed for a possible suture or graft repair. 2,4,19,20 Other less common sites of fracture-related injury to the RN and its branches include injuries to the proximal radi- 580

9 Radial nerve lesions TABLE 5 Outcomes in nine patients who underwent surgery for RN lesions at the dorsal forearm level No. of No. No. W/ Cases Surgically Grade 3 Factor Evaluated Treated Outcome mechanism of injury laceration fracture contusion prior suture total lesion category & type of surgery not in continuity 1st suture repair 2 2 2nd suture repair 0 0 2nd graft repair 4 2 in continuity NAP present neurolysis 1 1 NAP absent suture repair 1 0 graft repair 1 1 TABLE 6 Outcomes in 21 patients who underwent surgery for SSRN lesions Mechanism of No. of No. No. W/ Sig- Injury & Cases Surgically nificant Re- Type of Surgery Evaluated Treated lief of Pain laceration excision 11 8 external neurolysis 0 0 fracture contusion excision 6 6 external neurolysis 2 0 compression prior suture excision 2 2 external neurolysis 0 0 total excision external neurolysis 2 0 us and ulna as well as to the distal radius at the wrist level. 1,19,22,23 At the fracture site, lesions in continuity were commonly found intraoperatively and were believed to be caused by stretching and contusion. Of the 53 patients who underwent surgical repair for fracture-related injuries to the RN and its branches, 47 (88.7%) achieved a Grade 3 or better recovery. Up to 22% of injuries to the RN and its branches are results of laceration. Functional recovery following repair of these injuries is encouraging. 11,12 In our studies, favorable Grade 3 or better functional recoveries were observed in more than 90% of those patients who had undergone primary surgical repair. In cases in which delays in repair occurred, mobilization of both nerve stumps and secondary end-to-end suture or graft repairs resulted in a less than favorable prognosis of 64.3% for Grade 3 or better functional recovery. Any surgical repair of lacerations to the SSRN is fraught with a high incidence of painful neuroma formation. 7 10,15,17 The established treatment of choice for SSRN laceration injuries has been resection of the nerve, achieving amelioration of pain in approximately 70% of patients. Success with this approach is not universal, and pain syndromes associated with SSRN injury can be difficult to treat. Contusive injuries to the RN and its branches can result in lesions in continuity that may have intraneural derangement over a significant length. This mechanism of injury was seen most often (27.6%) at the wrist or distal forearm levels involving the SSRN. Approximately 40% of the patients in our series who had suffered contusion injuries to the RN and its branches spontaneously improved and did not require surgical management. In the surgical series, 24 (82.8%) of the 29 patients achieved a Grade 3 or better improvement in clinical and electrophysiological studies. Gunshot wounds resulting in injuries to the RN have a higher tendency to occur at the arm and elbow level. Despite an initial severe loss of function, six of the 22 patients we evaluated for GSW-related RN injuries spontaneously recovered to an overall Grade 3 or better function. Injury to the RN is commonly a result of a focal contusion caused by the transfer of kinetic energy from the projectile to the nerve, producing a lesion in continuity with neuroma formation. 25 In only one of the 16 surgical cases did the patient suffer a direct transection of the RN by the projectile. If there was little improvement in the electrophysiological studies and clinical state after a period of 3 to 4 months, exploration and external neurolysis was recommended for patients who experienced a persistent severe deficit. When a neuroma was present, internal neurolysis and intraoperative NAP monitoring was performed on the fascicles to identify and preserve functioning nerve fascicles. The neuroma and nonfunctioning nerve fascicles were resected, and a partial graft repair was performed, achieving a Grade 3 or better recovery of function in 15 patients (93.8%). The majority of iatrogenic injection injuries to the RN occurred at the level of the arm. Injury to the RN may be a result of direct needle-stick trauma, chemical irritation from the injected solution, or neuritis due to progressive inflammatory and fibrotic changes. 9 Unlike patients in whom the injury was sustained by the median nerve, in whom more than 50% recovered spontaneously, nine (75%) of the 12 patients evaluated in this series required TABLE 7 Mechanisms of injury to the RN at various levels in 180 patients No. of Cases Mechanism Upper Elbow/ Dorsal of Injury Arm Forearm PIN Forearm SSRN Total laceration GSW fracture contusion compression injection prior suture total

10 D. H. Kim, et al. TABLE 8 Lesion categories and types of surgery performed in 180 patients with RN lesions Lesion Category & No. of Cases No. W/ Grade 3 Type of Surgery Surgically Treated Outcome (%) not in continuity 1st suture repair (91) 2nd suture repair (77) 2nd graft repair (67) excision (84) in continuity NAP present neurolysis (98) NAP absent suture repair (88) graft repair (86) external neurolysis 2 0 (0) surgical exploration. 3,24 Although neuroma resection and graft repair was necessary in a number of patients, a Grade 3 or better functional recovery was achieved in all nine patients. In cases of SSRN lesions, the subcutaneous location of the nerve in the dorsoradial aspect of the wrist and hand subjects the SSRN to injury by relatively minor contusions and lacerations. The resultant neuroma is commonly painful and recalcitrant to nonsurgical treatment and pharmacological agents, and it can also be difficult to achieve pain amelioration by using surgical intervention. Previous neurolysis and simple excision have been accompanied by poor results in achieving pain amelioration. 9 Mackinnon and Dellon 8,18 reported significant improvement in surgical results after they developed their technique of resecting the neuroma and implanting the proximal end of the nerve away from the denervated skin; here it was free from tension and implanted into the vascularized environment of a muscle. Failure to identify the cutaneous nerves that contribute to the neuroma has been the cause of poor surgical results. Contributions to the painful neuroma from the lateral cutaneous nerve of the forearm and the posterior cutaneous nerve can be identified by selective diagnostic nerve blocks and can later be transected during neuroma resection. 18 Clinical assessment of functional recovery beyond the site of injury is often difficult and may involve a significant delay before any hand muscle recovery can be detected. We used EMG studies during the preoperative and postoperative period to assess the electrophysiological integrity of the nerve. The EMG studies showed fibrillation potentials indicative of denervation within 4 to 8 weeks after injury. During the recovery period following nerve injury, EMG studies often show reinnervation activity in the intrinsic muscles before any clinical evidence of functional recovery. 16 Following nerve repair, evidence of electrophysiological recovery appeared after 4 to 6 months, with clinical activity returning 3 to 5 weeks later. 6 After adequate external neurolysis, the use of intraoperative NAP monitoring is an integral part of surgical decision making. Tiel 27 and Kline 13 and their colleagues showed that intraoperative NAP recording is a reliable indicator of useful electrophysiological recovery in the assessment of nerve injury. It can assist with more precise localization of pathological findings and aids the decision concerning whether to resect the nerve lesion. Conclusions Functional outcomes following surgical repair of injuries to the RN and its branches with complete or severe deficits are encouraging. The type of surgical repair (external/internal neurolysis, primary/secondary suture repair or graft repair, or resection and implantation) is dependent on the type of lesion (in continuity or not in continuity) found and the presence of NAPs. Mechanisms of injury that result in lesions in continuity in which NAPs are present have the most favorable outcomes with Grade 3 or better recoveries. When a lesion not in continuity is caused by a sharp laceration, primary end-to-end surgical nerve repair achieves better favorable functional outcomes than delayed repair in which either suture or graft may be used. With blunt laceration or in cases of a lesion in continuity in which no NAP is present, resection of the neuromatous segment, until normal fascicular structures are identified, and subsequent graft repair also has a relatively favorable functional outcome. Painful neuromas of the SSRN are best managed by identifying the contributing cutaneous nerves other than the SSRN and transecting them, resecting the SSRN neuroma, and placing the proximal nerve stump on the undersurface of the BRM. References 1. Alnot J, Osman N, Masmejean E, et al: [Lesions of the radial nerve in fractures of the humeral diaphysis. Apropos of 62 cases.] Rev Chir Orthop Reparatrice Appar Mot 86: , 2000 (Fr) 2. Amillo S, Barrios RH, Martinez Peric R, et al: Surgical treatment of the radial nerve lesions associated with fractures of the humerus. J Orthop Trauma 7: , Bhatia M, Jindal AK: Injection induced nerve injury: an iatrogenic tragedy. J Assoc Physicians India 44: , Bostman O, Bakalim G, Vainionpaa S, et al: Immediate radial nerve palsy complicating fracture of the shaft of the humerus: when is early exploration justified? Injury 16: , Bostman O, Bakalim G, Vainionpaa S, et al: Radial palsy in shaft fracture of the humerus. Acta Orthop Scand 57: , Boswick JA Jr: Isolated high median nerve injuries. Bull Hosp Jt Dis Orthop Inst 44: , Dellon AL, Mackinnon SE: Pain after radial sensory nerve grafting. J Hand Surg (Br) 11: , Dellon AL, Mackinnon SE: Treatment of the painful neuroma by neuroma resection and muscle implantation. Plast Reconstr Surg 77: , Gaur SC, Swarup A: Radial nerve palsy caused by injections. J Hand Surg (Br) 21: , Jones NF, Shaw WW, Katz RG, et al: Circumferential wrapping of a flap around a scarred peripheral nerve for salvage of endstage traction neuritis. J Hand Surg (Am) 22: , Kline DG: Macroscopic and microscopic concomitants of nerve repair. Clin Neurosurg 26: , Kline DG: Physiological and clinical factors contributing to the timing of nerve repair. Clin Neurosurg 24: , Kline DG, Happel LT: Penfield Lecture. A quarter century s experience with intraoperative nerve action potential recording. Can J Neurol Sci 20:3 10, Kline DG, Nulsen FE: The neuroma in continuity. Its preoper- 582

11 Radial nerve lesions ative and operative management. Surg Clin North Am 52: , Lanzetta M, Foucher G: Entrapment of the superficial branch of the radial nerve (Wartenberg s syndrome). A report of 52 cases. Intl Orthop 17: , Lester RL, Smith PJ, Mott G, et al: Intrinsic reinnervation myth or reality? J Hand Surg (Br) 18: , Lluch AL, Beasley RW: Treatment of dysesthesia of the sensory branch of the radial nerve by distal posterior interosseous neurectomy. J Hand Surg (Am) 14: , Mackinnon SE, Dellon AL: Results of treatment of recurrent dorsoradial wrist neuromas. Ann Plastic Surg 19:54 61, Mast JW, Spiegel PG, Harvey JP Jr, et al: Fractures of the humeral shaft: a retrospective study of 240 adult fractures. Clin Orthop 112: , Packer JW, Foster RR, Garcia A, et al: The humeral fracture with radial nerve palsy: is exploration warranted? Clin Orthop 88:34 38, Pollock FH, Drake D, Bovill EG, et al: Treatment of radial neuropathy associated with fractures of the humerus. J Bone Joint Surg (Am) 63: , Shah JJ, Bhatti NA: Radial nerve paralysis associated with fractures of the humerus. A review of 62 cases. Clin Orthop 172: , Shaw JL, Sakellarides H: Radial-nerve paralysis associated with fractures of the humerus. A review of forty-five cases. J Bone Joint Surg (Am) 49: , Sheu JJ, Yuan RY: Superficial radial neuropathy caused by intravenous injection. Acta Neurol Belg 99: , Taha A, Taha J: Results of suture of the radial, median, and ulnar nerves after missile injury below the axilla. J Trauma 45: , Takami H, Takahashi S, Ando M: Severance of the radial nerve complicating transverse fracture of the mid-shaft of the humerus. Arch Orthop Trauma Surg 119: , Tiel RL, Happel LT Jr, Kline DG: Nerve action potential recording method and equipment. Neurosurgery 39: , 1996 Manuscript received November 3, Accepted in final form May 17, Address reprint requests to: Daniel H. Kim, M.D., Department of Neurosurgery, Stanford University Medical Center, Room R201, Edwards Building, 300 Pasteur Drive, Stanford, California neurokim@stanford.edu. 583