Recurring Brachial Plexus Neuropathy in a Diabetic Patient After Shoulder Surgery and Continuous Interscalene Block

Interscalene brachial plexus block provides effective anesthesia and analgesia for shoulder surgery. Continuous interscalene techniques allow aggressive physical therapy to improve joint range of motion (^1–4). However, the performance of regional blockade on a patient with a preexisting neurologic condition or a history of neurologic complications after regional anesthesia is controversial. We present a patient with a known peripheral neuropathy who developed a mild right brachial plexus neuropathy after shoulder surgery with continuous interscalene analgesia. Four months later, the patient underwent similar surgical and anesthetic procedures on the contralateral side and developed a severe left brachial plexus neuropathy.

Case Report

A 36-yr-old, 84-kg woman presented with severe atraumatic bilateral adhesive capsulitis of her shoulders. She had known insulin-dependent diabetes mellitus complicated by peripheral neuropathy, autonomic neuropathy, and nephropathy. The patient had previously undergone bilateral carpal tunnel release under local anesthesia. On examination, the patient had bilateral numbness in the C8 dermatome. No weakness of the upper extremities was noted. Electromyography (EMG) revealed changes suggestive of an early peripheral neuropathy, without evidence of a plexopathy or radiculopathy.

A right shoulder arthroscopic capsular release and debridement were performed under general anesthesia. Surgical duration was 100 min. After emergence, and confirmation of neurologic integrity, an interscalene catheter (Braun Contiplex^®, B. Braun Medical, Bethlehem, PA) was placed by using a nerve stimulator technique, and an infusion of 0.25% bupivacaine at 12 mL/h was initiated. The catheter remained in place for approximately 72 h, during which the patient underwent intensive physical therapy including continuous passive motion of the shoulder joint. The patient was discharged before complete resolution of the block to allow a more comfortable transport home.

Three weeks later, the patient returned for routine postoperative evaluation and reported persistent numbness of the medial aspect of her right arm and forearm. She also stated that her entire arm had been moderately weak, but that full strength had returned after approximately 2 wk. The patient had resolution of her symptoms by the time of her next visit 4 mo later. At that time, her only deficit on examination was mild numbness in the C8 dermatome bilaterally (which was present before her initial procedure). Because she was satisfied with the surgical outcome after her right shoulder procedure, she underwent a left shoulder arthroscopic capsular release. Surgery was performed under general anesthesia. The operating time was 50 min. After emergence, a neurologic examination confirmed preoperative neurologic function. An interscalene catheter was placed, and an infusion of 0.25% bupivacaine at 12 mL/h was initiated. By the following morning, the catheter had been dislodged, and there was no evidence of sensory or motor block. A second interscalene catheter was placed which provided excellent analgesia. Approximately 96 h postoperatively, the interscalene catheter was discontinued, and complete return of sensory function and partial return of motor function was subsequently noted. The patient underwent physical therapy in the afternoon.

That night, the patient awoke with marked weakness and pain of the entire left upper extremity. Neurologic consultation demonstrated profound deficits in the posterior and lateral cord distributions (absent function in biceps, triceps, wrist, and hand extensors). However, the medial cord was also affected (1 of 4 function in thenar, interossei, and hypothenar). Large fibers (motor, vibration, and proprioception) were affected more than small fibers (pain, temperature). No sensory or motor deficits were noted in the right upper extremity. Deep tendon reflexes were absent in both upper extremities. Magnetic resonance imaging results were negative for a compressive etiology. The diagnosis of idiopathic brachial plexus neuropathy was made, although local anesthetic toxicity and the patient’s underlying peripheral neuropathy were considered possible contributing factors.

At dismissal 48 h later, there was mild improvement in median and ulnar nerve function, but no radial function was noted. Decreased sensation in the axillary nerve distribution was also recorded. The patient continued to have excruciating upper extremity pain and reported only slight improvement in the sensory and motor deficits. Repeat EMG with nerve conduction studies performed 6 wk postoperatively was again consistent with a severe left brachial plexopathy with active denervation and loss of motor units in muscles innervated by both the upper and lower plexus (worse in upper plexus components). The contralateral extremity was also affected, although only mildly. A left upper extremity neuropathic pain condition was diagnosed, and transcutaneous electrical nerve stimulation and gabapentin therapy were initiated.

Four months postoperatively, motor function had recovered significantly, and only mild subjective sensory changes were present. The patient was instructed to gradually taper her dose of gabapentin over several months as tolerated. However, she subsequently developed neuropathic symptoms in the lower extremities, and the medication was continued. One year later, the patient reported additional recovery of sensory and motor function, although mild deficits were still present. Electrophysiologic studies confirmed a severe length-dependent sensorimotor peripheral neuropathy affecting both upper and lower extremities, consistent with progression of her underlying diabetic neuropathy.

Discussion

Perioperative neurologic complications may be related to the surgical procedure or anesthetic technique (^5,6). In addition, patient factors such as sex, body habitus, and preexisting neurologic disorders also contribute to postoperative neural dysfunction (^7–9). The etiology may also be multifactorial. In this patient, direct surgical or needle/catheter-related trauma, vascular injury (expanding hematoma), local anesthetic toxicity, exacerbation or progression of an underlying neurologic condition, and inflammatory neuritis were possible etiologies. However, the delayed presentation and diffuse deficits involving the entire brachial plexus allows several potential mechanisms to be excluded.

Our patient had a grossly normal neurologic examination in the postanesthesia recovery room. Thus, surgical trauma or neuropraxia can be ruled out. Likewise, needle- and catheter-induced injury or local anesthetic toxicity may be excluded as primary etiologies because the severe neurologic deficits occurred acutely 8–12 hours after discontinuation of the interscalene infusion and documentation of partial block resolution. Magnetic resonance imaging eliminated a compressive process, such as hematoma.

Our patient’s presentation is most consistent with brachial plexus neuropathy, also called “idiopathic brachial neuritis,” or Parsonage-Turner syndrome. Brachial plexus neuropathy has an abrupt onset in a previously healthy individual. Although half of the cases appear unrelated to any precipitating event, in others the neuropathy follows an upper respiratory infection or an immunization. Strenuous exercise and surgery may also be inciting factors (¹⁰). Nerve biopsy of affected individuals suggests that these brachial neuropathies have an inflammatory-immune pathogenesis (¹¹). As noted in our patient, pain may awaken the patient at night and is often severe. In as many as one third of cases, the pain may be bilateral, with both sides becoming involved simultaneously or sequentially. The pain precedes severe muscle weakness with subsequent muscle wasting, usually involving the shoulder girdle. Rarely, the wasting may be more distal. Sensory changes are less severe. Occasionally, lower cranial nerves and distal peripheral nerve trunks may also be affected. Most patients improve within several months, and 80%–90% recover within two-three years (¹⁰).

Brachial plexus neuropathy may occur in the postoperative period, regardless of surgical site (nonupper extremity) or anesthetic technique (^10,12). There is a single report of brachial plexus neuropathy after a regional anesthetic; a single-dose interscalene block performed for total shoulder replacement. The patient developed severe pain and motor deficits 18 hours after block resolution. An EMG performed four weeks postoperatively revealed widespread involvement of the brachial plexus bilaterally, with greater severity noted on the operative side. At 12 months postoperatively, neurologic recovery was virtually complete (¹³).

Although brachial plexus neuropathy is the most likely primary etiology, additional mechanisms may have contributed to the severity and duration of neurologic deficits. The “double-crush” hypothesis theorizes that nerves compromised in one region become susceptible to injury at another site (¹⁴). In our patient, the presence of a proximal inflammatory neuritis may have predisposed her nerves to the neurotoxic effects of local anesthetic as well as amplified the preexisting damage associated with her peripheral sensorimotor neuropathy. Although confirmatory human data are lacking, prolonged exposure to local anesthetics may lead to local anesthetic toxicity. Kyttä et al. (¹⁵) reported disruption and vacuolization of myelin sheaths in rat sciatic nerves repeatedly injected with 1 mL of 0.5% bupivacaine. These results suggest a potential risk of nerve injury associated with repeated injection of 0.5% bupivacaine. However, the dose was 7 to 8 times larger than the equivalent human dose, and the concentration larger than that used in continuous infusions. Laboratory data suggest that in diabetes, local anesthetic requirement is reduced, and the risk of local anesthetic-induced nerve injury is increased (⁹). The patient had an underlying peripheral neuropathy that progressed from mild to severe in 18 months. Although she had no clinical or EMG evidence of a preexisting plexopathy or radiculopathy, the already compromised peripheral nerves would be more vulnerable to injury. Theoretically, the combined impact of the inflammatory neuritis, local anesthetic neural toxicity, and diabetic neuropathy could have acted in combination to worsen the neurologic outcome in our patient.

It is also interesting that after a second similar shoulder procedure, our patient developed a similar neurologic disorder. Although recurrence of brachial plexus neuropathy is unusual, it has been reported (¹⁰). However, a subsequent postoperative episode occurring in the previously unaffected extremity has not been noted. Furthermore, the risk of additional neurologic compromise in a patient with a history of postoperative neurologic dysfunction undergoing a regional anesthetic technique is unknown, but the possibility should be discussed. In this case, the need for intensive physical therapy necessitated the placement of an indwelling interscalene catheter and was requested by both the surgeon and the patient.

Our case supports the possibility that brachial plexus neuropathy may occur spontaneously in the postoperative period and confound the diagnosis of surgical neuropraxia or regional anesthesia-related nerve injury (¹⁶). Meticulous documentation of the patient’s preoperative neurologic examination will allow early diagnosis of new deficits. Prompt evaluation of postoperative neurologic dysfunction is critical to rule out reversible etiologies.