Between September 1994 and January 1996, we managed seven patients who had an injury of the spinal accessory nerve that had occurred during an operative procedure on the neck. The average age of the two men and five women was twenty-nine years (range, twenty-four to forty-four years) (Table I). The injury had occurred during removal of a bullet in one patient (Case 6) and during biopsy of a cervical lymph node in the remaining six patients. The injury was on the left side in two patients (Cases 1 and 6) and on the right side in five. The average interval between the time of the injury and the repair was eight months (range, three to fourteen months). The average duration of follow-up was thirty-two months (range, twenty-four to thirty-nine months).
The biopsy of the lymph node was performed, with the use of local anesthesia, by an orthopaedic surgeon in one patient (Case 1), by general surgeons in four patients (Cases 2, 4, 5, and 7), and by an otolaryngologist in one patient (Case 3). At the time of the biopsy, the patients reported acute pain that felt like an electrical current, and four patients (Cases 1, 3, 4, and 5) also noted involuntary movement of the shoulder. The patients had stiffness, pain, weakness, and drooping of the shoulder. The patient who had a bullet embedded in the neck (Case 6) noted weak flexion of the elbow and numbness over the left shoulder. The bullet had gone through the chin and had migrated into the left side of the neck. Initially, the physical examination and the electrodiagnostic studies suggested that she had an incomplete injury of the brachial plexus. Seven days later, the bullet was removed by a general surgeon with the patient under general anesthesia. After removal of the bullet, severe pain and stiffness developed in the shoulder girdle.
All of the patients had reported the symptoms to the operating surgeons. However, the injury of the spinal accessory nerve was recognized in only two patients (Cases 2 and 7). Four patients (Cases 2, 5, 6, and 7) were referred to an orthopaedic surgeon. One patient (Case 1) was seen by the orthopaedic surgeon who had performed the biopsy. The diagnoses that were considered were a herniated nucleus pulposus (Case 1), recurrent lymphadenitis (Case 3), psychological reaction (Case 4), and cervical spondylosis (Case 5). In one patient (Case 6), the symptoms had been attributed to injury of the brachial plexus and the cervical muscles by the bullet as it traversed the neck. The patients were managed with cervical traction, use of a cervical orthosis, physical therapy, or medications that included steroids and anti-inflammatory or antidepressant drugs. Two patients (Cases 2 and 7), who were thought to have paralysis of the trapezius, were managed with physical therapy because the electrodiagnostic studies had showed continuity of the nerve. However, these two patients had no signs of clinical improvement. The seven patients came to us for a second opinion at an average of six months (range, one to twelve months) after the injury.
When they were first seen by us, the patients reported pain and stiffness (rather than weakness) in the shoulder girdle that became intolerable with prolonged use of the upper extremity. The ability to lift heavy objects or to perform overhead activities was limited. Physical examination revealed marked atrophy of the trapezius, asymmetry of the neck line, drooping of the shoulder, and lack of visible contraction of the trapezius. They had weakness on abduction or forward elevation of the shoulder. Active abduction was 90 degrees or less in six patients (Cases 1 through 6) and 120 degrees in one patient (Case 7). The patients had winging of the scapula during these active movements. None of the patients reported any paresthesia over the neck or shoulder girdle suggestive of an injury of the great auricular nerve or the cervical nerves. Because of an associated injury of the brachial plexus, one patient (Case 6) had weakness on external rotation of the shoulder (grade 3+ on manual muscle-testing), flexion of the elbow (grade 4), and abduction of the shoulder (grade 1+ [due to weakness of the deltoid and complete paralysis of the trapezius]). This patient also had mild numbness and paresthesia over the shoulder and arm.
Electromyography revealed denervation potentials (fibrillations and positive sharp waves) in the trapezius without voluntary potentials in six patients (Cases 1 through 6) and denervation potentials and a few voluntary potentials in one (Case 7). In one patient (Case 6), a reduced number of voluntary potentials and a few denervation potentials were seen in the deltoid, infraspinatus, biceps, and brachioradialis, indicating a lesion of the brachial plexus but continuity of the nerves. Before one of the patients (Case 2) was seen by us, electromyography was thought to have shown continuity of the spinal accessory nerve, but we believe that the levator scapulae muscle had been erroneously tested in that patient.
We recommended operative management to all of the patients as there was no evidence of clinical recovery. The patients were informed that the nerve may be unidentifiable intraoperatively and therefore irreparable. They were also told that, when more than three months have elapsed since the injury, the prognosis for recovery may be poor even if the nerve can be repaired6,10. We recommended simultaneous repair of the nerve and muscle transfers to the four women who were seen more than six months after the injury (Cases 1, 3, 4, and 5)13. However, they were concerned about the additional scar resulting from the muscle transfers and preferred to have the repair of the nerve first and the transfers later, if necessary. To restore function, we suggested the Dewar-Harris procedure3 for the men and we recommended the Eden-Lange procedure2 for the women because of its cosmetic advantage (no scar over the lateral aspect of the thigh). In one patient (Case 6), the lesion of the brachial plexus was not explored because the nerve was in continuity.
Despite our recommendation that the nerve be repaired as soon as possible, the patients waited one to four months after the consultation. The patients discussed the issue of misdiagnosis and intraoperative injury of the nerve with the initial surgeons, and three patients (Cases 2, 3, and 4) filed a malpractice lawsuit.
Operative Technique
After the induction of general anesthesia, the patient was placed in a lateral position and the involved upper extremity was draped in a fashion that permitted free movement of the extremity. The ipsilateral lower extremity also was prepared in case a free sural-nerve graft was needed. The previous incision in the posterior cervical triangle, which was transverse in all of the patients, was extended one to two centimeters cephalad from the anterior end and caudad from the posterior end to create a z-shaped incision. Soft-tissue dissection proceeded from the two ends of the z-shaped incision and extended toward the site of the previous operation. The first structures to be exposed were the posterior margin of the sternocleidomastoid and the anterior margin of the trapezius, which represent the anterior and posterior sides of the posterior cervical triangle, respectively. The great auricular nerve was seen wrapped around the margin of the sternocleidomastoid.
The cephalad part of the spinal accessory nerve was identified 0.5 to 1.5 centimeters cephalad to the point of exit of the great auricular nerve at the posterior margin of the sternocleidomastoid. The caudad part of the nerve usually was visible through the thin superficial fascia between the trapezius and the sternocleidomastoid after dissection of the scar-free area anterior to the trapezius. During dissection, branches of the nerve were identified and, if possible, repaired. The cut ends of the nerve were trimmed until normal fasciculi were exposed. This procedure was performed even if a more cephalad anastomosis was necessary or if there was a longer defect. Although poor mobility of the nerve has been reported12, we were able to perform an end-to-end repair in our patients. None of our patients needed nerve-grafting.
The nerve was repaired with use of an operating microscope set at a low magnification (six times). The nerve was found to be completely severed in five patients (Cases 1 through 5). We were able to perform an end-to-end repair with 9-0 nylon sutures without placing undue tension on the nerve and with the cranium held in a neutral position6. The nerve appeared to be in continuity in two patients. One of them (Case 7) needed only a neurolysis because intraoperative stimulation of the nerve revealed electroconductivity. The other patient (Case 6) had an end-to-end repair of the nerve as there was no evidence of electrical activity.
Intraoperatively, three patients (Cases 1, 2, and 5) were found to have a branch arising from the caudad part of the spinal accessory nerve. In another patient (Case 7), in whom the nerve was in continuity, there was a branch to the superior portion of the trapezius arising cephalad to a neuroma at the site of the injury. Three patients (Cases 3, 4, and 6) had no branches from the spinal accessory nerve. In three patients (Cases 4, 5, and 6), the external jugular vein was found to be thrombosed. The patient who had been injured by a bullet (Case 6) had scarring of the cervical, trapezius, and sternocleidomastoid muscles.
Postoperative Management and Assessment
Postoperatively, a Philadelphia collar was worn to restrict the motion of the neck. A sling was used to keep the forearm flexed at the elbow in order to prevent the shoulder from being pulled downward. After removal of the sling and the collar at three and six weeks, respectively, the patient was allowed full activity. Electromyography was performed every month until there was evidence of reinnervation of the trapezius muscle and regeneration of the nerve.
The patients were usually examined every month for one year and then every three months. At each examination, they were questioned about pain and stiffness in the shoulder girdle. Functional assessment included measurement of active abduction of the shoulder and manual muscle-testing of the trapezius. The contour, function, and strength of the trapezius were assessed with the patient sitting. The examiner held the arm of the patient adjacent to the patient's body in external rotation and with the elbow extended. The patient attempted to abduct the shoulder against resistance provided by the examiner. (Because of weakness of the deltoid in one patient [Case 6], the trapezius was tested in that patient by applying resistance to the tip of the shoulder during active abduction.) The contraction of the trapezius and its ability to stabilize the scapula during a complete arc of abduction were assessed and compared with those of the normal extremity. A grade of 0 was given if muscle contraction was clinically absent; 1, if muscle contraction was clinically evident but abduction of the shoulder was not greater than the preoperative level; 2, if strength was between grades 1 and 3; 3, if a full arc of abduction against gravity was possible; 4, if strength was between grades 3 and 5; and 5, if strength was equal to that on the contralateral, normal side. The ability to shrug the shoulder was not used as an indicator of function of the trapezius, as shrugging can be performed by the levator scapulae and rhomboid muscles6.
At the latest follow-up evaluation, the patients were asked whether they were very satisfied, satisfied, unsatisfied, or very unsatisfied with the clinical outcome after the operation. The results of the operation, including the residual symptoms, the strength of the trapezius on manual muscle-testing, and the level of the patient's satisfaction, were evaluated. A score of 3 points was given when the patient was very satisfied, there were no residual symptoms even during strenuous activities, or the trapezius was grade 5 on manual muscle-testing. A score of 2 points was given when the patient was satisfied, there were no symptoms except during strenuous activities, or the trapezius was grade 3 or 4 on manual muscle-testing. A score of 1 point was given when the patient was unsatisfied, the symptoms were decreased but still intermittent, or the trapezius was grade 1 or 2 on manual muscle-testing. A score of 0 points was given when the patient was very unsatisfied, there was no decrease in the symptoms, or the trapezius was grade 0. A total score of 9 points indicated an excellent result; 6, 7, or 8 points, a good result; 3, 4, or 5 points, a fair result; and 0, 1, or 2 points, a poor result.
Nerve regeneration was confirmed by electromyography. One patient (Case 7) had an increase in voluntary potentials one month after the repair of the nerve. Polyphasic regenerative potentials were first noted at two months in three patients (Cases 2, 3, and 6) and at three months in the remaining three (Cases 1, 4, and 5).
On physical examination, contraction of the trapezius could be visualized at two months in one patient (Case 7), at four months in two (Cases 3 and 6), and at five months in four (Cases 1, 2, 4, and 5). At these time-periods, the patients reported considerable relief of pain and stiffness in the shoulder girdle. Full active abduction against gravity was regained at three months by one patient (Case 7), at six months by two (Cases 3 and 6), at seven months by three (Cases 2, 4, and 5), and at eight months by one (Case 1). At the same time-periods, the patients were able to tolerate any residual pain and stiffness, except when lifting heavy objects or performing overhead activities. Manual muscle-testing revealed the strength of the trapezius to be equal to that on the normal side at six months in one patient (Case 7), at ten months in one (Case 6), at eleven months in two (Cases 3 and 5), at twelve months in two (Cases 2 and 4), and at thirteen months in one (Case 1). The patients did not have pain or stiffness in the shoulder girdle after they had regained normal strength of the trapezius.
At the latest follow-up evaluation, four patients (Cases 3, 4, 5, and 7) had regained normal function of the shoulder and had no symptoms even when lifting heavy objects or performing overhead activities. The other three patients (Cases 1, 2, and 6), who had almost no difficulties with activities of daily living, reported fatigue when lifting heavy objects or performing overhead activities. In addition, one (Case 1) had a stretching sensation in the neck, and another (Case 2) had periscapular discomfort during such activities. The third patient (Case 6) had grade-4+ strength for flexion of the elbow and grade-4 strength for abduction of the shoulder (to 90 degrees) and external rotation on manual muscle-testing. The numbness and paresthesia over the shoulder and arm as well as the pain and stiffness that had developed in the shoulder girdle after removal of the bullet had resolved. This patient reported periscapular discomfort, occasional discomfort in the neck, and fatigue and weakness of the upper extremity when lifting heavy objects or performing overhead activities. All three patients (Cases 1, 2, and 6) reported that these residual symptoms were mild and tolerable and thus different in nature from the preoperative pain and stiffness. Two of the patients (Cases 1 and 2) had mild atrophy of the trapezius.
Four patients (Cases 3, 4, 5, and 7) were very satisfied with the outcome, and the results were rated as excellent (total score, 9 points); the three remaining patients (Cases 1, 2, and 6) were satisfied with the outcome, and the results were rated as good (total score, 7 points). The result of the repair of the spinal accessory nerve was difficult to rate in one patient (Case 6) because of the associated injury of the brachial plexus and the damage of the cervical muscles. The residual symptoms, especially the discomfort in the neck and the generalized weakness of the upper extremity, could be attributed to the coexisting conditions. Still, the result in that patient was rated as good because the periscapular discomfort and fatigue with strenuous activities were similar to those of another patient (Case 2) who had a good result.
The spinal accessory nerve is susceptible to iatrogenic injury8,11,12,14,15,17,19,21,24. Recognition of this injury is important not only for medical reasons but also because of the legal implications. The diagnosis is often delayed, possibly because of a lack of familiarity with the anatomy of the posterior cervical triangle. Orthopaedic surgeons usually assume that patients who have pain and stiffness in the shoulder girdle have problems related to the shoulder and cervical spine. However, injury of the spinal accessory nerve should be considered in the differential diagnosis when there is a history of an operation in the posterior triangle of the neck.
The spinal accessory nerve follows a superficial course, and anatomical landmarks are helpful in identifying it. The cephalad part of the nerve is visualized within two centimeters cephalad to the point of exit of the great auricular nerve at the posterior margin of the sternocleidomastoid1,18, or cephalad to the level where the external jugular vein crosses the posterior margin of the sternocleidomastoid24. The spinal accessory nerve is accompanied by a branch of the transverse cervical artery14,23. Another landmark, which can be used when the great auricular nerve is injured, is a line from the second cervical transverse process, palpated beneath the sternocleidomastoid, to the tip of the shoulder4. The great auricular nerve was intact in all of our patients, and we believe that it is the most reliable landmark for identification of the cephalad part of the spinal accessory nerve.
Occasionally, it is difficult to identify the caudad part of the nerve4,24, which usually enters the anterior margin of the trapezius five centimeters medial to the clavicular insertion1,6,18. The spinal accessory nerve runs in the same direction as, and superficial to, the cervical nerves, which are separated only by a fascial layer1. It may be difficult to identify the spinal accessory nerve when the branches of the cervical nerves are also injured. Ando1 noted that the cervical nerves ran almost in the same direction as, but entered the anterior margin of the trapezius cephalad to, the spinal accessory nerve in seven of fifty-six cadavera. We found that the caudad part of the nerve was most easily visualized in the scar-free posteroinferior area of the posterior cervical triangle, anterior to the trapezius, through the superficial thin fascia between the trapezius and the sternocleidomastoid.
Electromyography of the trapezius is helpful in assessing the function of the nerve; however, it should be performed with caution. The trapezius is a thin muscle, and denervation results in atrophy, which makes insertion of a needle electrode into the muscle more difficult. After penetration of the skin, it is important to advance the electrode slowly. The examiner should feel the electrode come into contact with the fascia over the trapezius and perform the study when the tip of the electrode just penetrates the fascia. We found that advancement of the electrode for more than a few millimeters after penetrating the fascia resulted in assessment of the underlying muscles.
When the diagnosis has been made, nonoperative treatment is indicated if there is evidence of recovery within three months after the onset of symptoms6,10,12,15. The recovery should be clinically evident because voluntary potentials on electromyography alone do not necessarily predict a good prognosis15, as was evident in one patient (Case 7). Patients who have no signs of a clinical recovery should be managed operatively12. The weaknesses of our study are that it consisted of relatively young patients and that the longest interval between the injury and repair was fourteen months. We do not know whether the same results can be obtained in elderly patients or after a longer delay before the repair. On the basis of the excellent result in the oldest patient in our series, who was forty-four years of age, we suggest that an attempt should be made to repair the nerve unless the trapezius is fibrotic.
Earlier reports have indicated that repair of the spinal accessory nerve usually results in the relief of symptoms and a return of function, although the methods for evaluating the results have varied widely4,5,7,9,11-16,20-22,24. Also, the results have not always been satisfactory, and this has been assumed to be related to the timing and the quality of the repair as well as to the magnitude of the injury of the nerve and other structures. There is little information on the effects of the degree and the magnitude of the injury of the nerve and other structures, although there have been reports on patients in whom the caudad nerve stump was not found4,24. Generally, patients who had repair of the nerve within three months after the injury had an excellent result. The results, although beneficial in some patients, were poorer when the repair was delayed for more than three months. Nordén13 suggested that the prognosis is poor if the repair is done more than six months after the injury. Gabel and Nunley6 as well as Leffert10 recommended repair of the nerve within three months after the injury for the best results. However, patients are often seen much later. In our series, only one patient (Case 6) had the operation within three months after the injury, but the results in our patients were satisfactory. Several investigators have made recommendations regarding the duration of observation after the injury and before intervention. The suggested durations were three or four months16, six months13, and twelve months5,11,12,14,24. Muscle transfers have been recommended for patients who are seen more than twelve months after the injury12,14,24. In our series, however, repair of the nerve at fourteen months relieved the pain, stiffness, and weakness. We do not believe that a duration of twelve months is the borderline between nerve repair and muscle transfers; we think that the spinal accessory nerve should be repaired as long as denervation potentials are detectable on electromyography and the muscle fibers of the trapezius are not fibrotic secondary to denervation.
In the postoperative period, our patients had considerable relief of pain and stiffness before full recovery of the trapezius. The patients reported considerable relief of symptoms when contraction of the trapezius became evident on physical examination. In addition, after they had regained full active abduction against gravity, they were able to tolerate the symptoms except when lifting heavy objects and with overhead activities. Our findings suggest that mild recovery of the muscle is sufficient to alleviate symptoms and that pain and stiffness disappear when the trapezius has normal strength.
Injury of the spinal accessory nerve should be suspected if a patient has pain or stiffness in the shoulder girdle and a history of an operation in the region of the neck. Repair of the nerve is a simple procedure involving only a small extension of the previous operative scar and superficial dissection associated with little operative risk. The dissection should be atraumatic and proceed from the scar-free areas to the site of the previous operation. We recommend exploration and repair of the nerve if clinical recovery does not occur.