Perioperative Bilateral Median Neuropathy

A 60-yr-old man (86 kg, 180 cm) with a history of ulcerative colitis had an operation for recurrent urethroperineal fistula. His routine preoperative blood tests were normal. Intravenous access was through the right hand and right internal jugular vein. A total of 7 l was given during the procedure. A lumbar epidural was placed followed by a standard general anesthetic with tracheal intubation. The epidural was test dosed with 2% lidocaine with epinephrine followed by 12 ml of the same solution; the patient was positioned supine. His arms were placed on padded arm boards that were abducted to less than 90°. A bubble wrap was positioned around the elbows bilaterally, and the arms were affixed to the arm boards with tape. The upper limbs were checked to ensure that the shoulders were not over-abducted, that the elbows were not over-extended, and that the arms were loose underneath the tape but secure enough to not be able to fall off of the arm boards. The surgery lasted 6 h. During the surgery, the patient was moved several times from supine to extreme Trendelenburg as well as into extreme reverse Trendelenburg. The patient tolerated the procedure without hypotension or other incident and was extubated and taken to the intensive care unit. Later that night, the patient complained of numbness and tingling in bilateral digits 1–3 and the palm. Examination 22 h after surgery showed a reduction of light touch, pin prick, and vibratory sensation in the median nerve distribution and weakness of abductor pollicis brevis (Medical Research Council 4/5) in both hands; Tinel's sign was not present over the median nerve at the wrist. There was swelling of hands and forearms.

A nerve conduction study was performed 26 h after surgery. Antidromic sensory nerve conduction studies were performed recording at the digit 2 and stimulating at the palm, wrist, forearm, and elbow segments. Distal median sensory nerve action potential amplitudes stimulating at the wrist were normal (52 μV on the right and 47 μV on the left; normal > 10 μV). Inching studies with stimulation at the palm and calculating conduction velocity in the palm to wrist segment failed to show focal reduction in conduction velocity in the carpal tunnel segment. Proximal stimulation in the forearm on the right and at the elbow on the left revealed complete conduction block (fig. 1).

Median motor conduction studies were done with recording at the thenar eminence and stimulating at precisely the same sites as the sensory conduction study. Wrist stimulation showed normal distal latencies and compound muscle action potential amplitudes (11.3 mV on the right and 11 mV on the left; normal > 4 mV). Proximal stimulation in the right forearm, 140 mm above the wrist, showed partial motor conduction block (34% decrease in negative peak area) and reduced conduction velocity (40 m/s; normal > 50 m/s) (fig. 2). On the left, proximal stimulation showed a partial motor conduction block (25% decrease in amplitude and negative peak area) and mild reduction in conduction velocity (48 m/s) with stimulation at the elbow. Right sided ulnar sensory and motor conduction studies were normal. Symptoms resolved by 48 h, and a repeat neurologic examination was normal.

Median neuropathy is a distinctly uncommon perioperative neuropathy. When reported, the usual vulnerable site for the median nerve is in the antecubital fossa where the nerve lies adjacent to the medial cubital and basilic veins. At this site, the nerve may be damaged during injection into or cannulation of a vein or artery. 6A preexisting median neuropathy at the wrist, subclinical carpal tunnel entrapment, may become symptomatic. Motor fibers that contribute to the median nerve may also be injured with stretch at the level of the lower trunk or medial cord of the brachial plexus. Perioperative anterior interosseous nerve damage has been recently described. The possible mechanisms of injury for this nerve seem to be needle stick, hematomas, and traction in the antecubital fossa. 7Lastly, although not directly related to surgery, selective damage to the median or anterior interosseous nerve fibers may be a part of idiopathic brachial neuritis. 8 

We describe a man who developed bilateral median neuropathy in the distal forearm following prolonged abdominal surgery. Nerve conduction studies precisely localized the area of abnormality. Given the conduction block and the rapid recovery of symptoms, the injury was neuropraxia rather than axonotmesis. Corresponding to his clinical symptoms, physiologically sensory fibers were more severely affected than the motor fibers.

Given the bilateral symmetrical findings, the documented conduction block, and the rapid recovery, we postulate that the injury to the median nerves was from compression rather than nerve ischemia–infarction or transection. In the forearm, the median nerve descends along the midline in a relatively deep position lying on the anterior face of flexor digitorum profundus, beneath the posterior face of flexor digitorum superficialis. For this reason stretch of the nerve or a direct penetrating trauma to the nerve seems unlikely, and there was no evidence of either. The possible causes of the compression include the tape over the volar forearm used to affix the arms to the padded arm board, an unexpected venous tourniquet effect in the face of third spacing, or generalized soft tissue swelling. A case of bilateral symmetric transient radial nerve injury in an infant after cardiac surgery has been described. In that case, the injury occurred from tight wrapping of hand splints for many days after surgery, which the authors speculate, caused prolonged compression to the posterior interosseus nerve. 9In addition to defining the mechanism of injury, this case raises questions regarding the underlying pathology. Although the presence of “true” conduction block is pathognomic of “demyelination,” given the recovery in less than 24 h, it is unlikely that structural demyelination and remyelination took place in that short time period. Studies on acute compression block in animal models have shown that under certain circumstances, nerve fibers can sustain local compression damage, without Wallerian degeneration and from which recovery is relatively rapid. 10This kind of injury results from mechanical distortion of fibers with the node of Ranvier and the attached myelin sheath being displaced from the compression site toward the uncompressed site, leading to invagination of one internode into the next. In this situation, the degree of conduction block (percent of fibers that are electrically inexcitable proximal to the lesion) is directly proportional to both the amount and duration of the compression. The time of recovery is related to the force and duration of initial compression, to the length of nerve compressed, and to the number of nodes of Ranvier affected. 10The relatively fast recovery in our patient is best explained by a narrow compression force inducing paranodal changes in a limited number of nodes of Ranvier.

The possible role of ischemia in determining the nerve lesion in this case still is questionable. Animal studies have shown mammalian nerve fibers can withstand 4–6 h of ischemia without development of structural damage. Moreover, there is no evidence that ischemia for 4 h exaggerates the effect of mechanical pressure on nerves. 11Ischemia generally produces a lesion of Wallerian-like degeneration. It has also been shown that if pressure is applied at extremely high levels or for a long period of time, the blood vessels may undergo ischemic necrosis prior to similar changes in the nerve. 12 

There is a limited understanding of the relationships between conventional perioperative care and the genesis of peripheral nerve lesions. 13In our experience, most patients with perioperative nerve injuries are referred for electrophysiological evaluation late and when spontaneous recovery is delayed or absent. This is because the majority of these nerve injuries is due to focal demyelination, and as in our case, recover quickly. In selected cases, nerve conduction studies at the onset of symptoms may provide insights into localization, mechanism, and prognosis.