Introduction

Lesions of the brachial plexus may be evaluated and localized not only by careful neurological examination and imaging studies (CT, MRI) but also by direct inspection of the lesion and by neurophysiological studies. Direct inspection alone in the operating room has been shown repeatedly to be unreliable in assessing the severity of peripheral nerve lesions, especially brachial plexus injuries.^1–3 Neurophysiologic evaluation, both pre-operatively and intra-operatively, is a very valuable adjunct in the diagnosis of, and decision making in, peripheral nerve lesions-in-continuity. Most plexus injuries are stretch injuries leaving the extraspinal plexus unruptured and grossly in continuity but internally disrupted with variable damage to axons and their connective tissue sheaths – the neuroma-in-continuity.

When considering neurophysiological studies of the peripheral nervous system, particularly the brachial plexus, one must keep in mind the entire anatomy of the system as related to the brachial plexus. This complex system extends from the peripheral motor and sensory end organs to the dorsal and ventral rootlets of each spinal nerve as well as the spinal cord and the motor and sensory cortex of the brain. For example, nonoperative, surface stimulation of peripheral nerves or their sensory terminals may be used to assess brachial plexus function. This can be done specifically to assess preganglionic nerve root injury or lesser plexus injuries, or entrapment such as thoracic outlet syndrome. In these instances, recordings – somatosensory evoked potentials (SEPs) – are made from tissues overlying the spinal cord or scalp superficial to parietal cortex.⁴ These recordings, however, test the entire length of the system from sensory terminals to brain and are not specifically localizing. A simpler method to determine if the injury is preganglionic is to stimulate the anesthetic fingers and attempt recordings from peripheral nerves leading from them.^5,6 Presence of a sensory nerve action potential (sNAP) there, despite distal anesthesia in the sensory dermatome, indicates a preganglionic injury at its corresponding proximal spinal nerve; for example at C8, if an anesthetic fifth finger is stimulated and a SNAP is recordabe over the ulnar nerve then a preganglionic injury is presumed. Such studies are less diagnostic for upper elements such as C6 because of potential dermatomal overlap with C7 which may also have a preganglionic injury. Distal peripheral stimulation studies to evaluate C5 alone may also be very difficult to achieve. Some of these problems are helped by direct operative stimulation of spinal nerves with recordings from at skin level over the spinal cord and/or scalp but there can be drawbacks to those studies as well.

For serious plexus injuries the gold standard for preoperative electrophysiological study remains the EMG.⁷ Such studies are essential for the accurate work-up of plexus injuries but have pitfalls. The operative equivalent to EMG includes direct stimulation of proximal or more peripheral neural structures and recording from the muscles they supply, to not only test continuity of nerves but also to provide evidence for distal regeneration after injury or repair.⁸ However, even these relatively straightforward approaches can have serious limitations.

Nerve action potentials (NAP)

The most accurate operative neurophysiological assessment of brachial plexus injury is obtained after a through exposure of the plexus elements and, where possible, the injury sites. This includes a neurolysis – a 360° exposure of these elements. For the frequent extensive stretch injuries of the plexus, this may mean an infraclavicular as well as a supraclavicular exposure. Such exposure can be more limited for more local lesions such as some gunshot wounds, lacerating injuries, iatrogenic injuries, tumors and even thoracic outlet syndrome cases. Those lesions are often, but not always, lateral to the spinal foramina and more readily lend themselves to less complex simpler stimulation and recording studies across their lesions-in-continuity. In some cases, the plexus element can be stimulated proximal to the lesion (neuroma) and the recording electrode is placed distal to the lesion. This may not be possible for some proximal plexus stretch lesions in which case the stimulating electrode is placed as close as possible to the foramen of the injured element and the recording electrode placed more distally on the plexus. To obtain a reliable recording of a nerve action potential (NAP) in an adult, the distance between the electrodes should be 4 cm or greater.

For the frequent stretch injuries in the plexus, leading to a neuroma-in-continuity, a variety of operative electrophysiologic tests are available to guide clinical decision making.^5,9,10 First, expose the spinal nerves and their trunks up to and somewhat into their foramina taking care to preserve the phrenic and long thoracic nerves. Use of a simple disposable operative stimulator suffices to test and preserve their integrity during the operation. For C5, C6, and C5, C6 and C7 lesions, it suffices to expose only those spinal nerves and their more distal outflows, and it is not usually indicated to dissect out C8, T1 and the lower trunk. Even for adult flail arms with loss of C8 and T1 function, only these same upper elements are usually dissected. Often preoperative CT myelography or MRI studies have indicated irreversible avulsion injuries at those lower levels. Resection and direct repair of those lower elements, at least in the adult, is pointless.

Once the relevant parts of the brachial plexus are exposed, NAP studies are done. Tripolar electrodes shaped like “shepherds crooks” are placed beneath the more proximal portion of the exposed spinal nerve which is gently elevated away from surrounding tissue. For evaluation of C5, a bipolar recording electrode is then placed 3 to 4 cm distally beneath distal upper trunk or, in sequence, on the suprascapular nerve and on both the anterior and posterior divisions of the upper trunk. This electrode is also gently elevated away from the surrounding soft tissues to reduce aberrant conduction. In plexus stretch injuries, it is not unusual to be unable to place the stimulator proximal to an injured segment, in which case stimulation and recording must be done distal to the lesion and results interpreted accordingly.

An EMG machine is used to both stimulate and record. Stimulus intensity is gradually increased with the stimulus duration set at 0.05 msec to 0.1 msec at a frequency of 2 to 3 Hz until a response is either visualized or not.¹¹ Stimulus intensity is 3 to 15 V for healthy nerves although significantly higher levels (usually less than 100 V) are needed for damaged nerves. Amplification settings are incrementally increased when needed, from 50 µv/division to 5 mv/division with the time base usually set at 0.5 msec per division. A variety of high and low frequency filters are tried until it is apparent that there is either a response or not. A positive response, a visible regenerative NAP, is evidence of the presence of at least 4000 partially myelinated axons of greater than 5µ diameter between the stimulating and recording electrodes. Small responses can be further amplified by repetitive summation. This scenario is then repeated at C6 with recordings on distal upper trunk and its divisions and then at C7 to distal middle trunk and its divisions. More detailed descriptions of the equipment, technical requirements and procedures involved in NAP testing as well as other intraoperative electrodiagnostic tests are available.^{5,6,10,12,13,14,15}

Providing the studies are done after several months but before a year or so post injury, a positive trace indicates one of several possibilities in the stretch injured patients¹¹ (Figure 16.1):