Electrophysiologic Testing for the Diagnosis of Peripheral

� CLINICAL CONCEPTS AND COMMENTARY

Richard B. Weiskopf, M.D., Editor

Anesthesiology 2004; 100:1298–303 © 2004 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc.

Electrophysiologic Testing for the Diagnosis of PeripheralNerve InjuriesMichael J. Aminoff, M.D., D.Sc., F.R.C.P.*

ANESTHESIOLOGISTS have an important role in pre-venting perioperative nerve injury, monitoring nervefunction to minimize damage, and diagnosing peripheralnerve lesions at an early stage to optimize their manage-ment. The purpose of the current article is therefore toclarify the use and limitations of electrophysiologic test-ing in the diagnosis and management of anesthesia-related nerve injuries.

The occurrence of perioperative nerve injuries is welldescribed. In the American Society of AnesthesiologistsClosed Claims Database (a standardized collection of casesummaries from the closed malpractice claims of a numberof insurance companies), 16% of the 4,183 claims havebeen for anesthesia-related nerve injury.1 Regional nerveblock may lead to a focal nerve deficit. During surgeryitself, direct injury or tourniquet compression to insure abloodless field may be responsible. In rare instances, thecompression is seemingly innocuous, as from a blood pres-sure cuff that inflates automatically at periodic intervals.2

Malpositioning of patients during surgery may lead to com-pression or entrapment neuropathies, especially in the up-per limbs and involving particularly the ulnar or radialnerve; less commonly, the median, musculocutaneous, ax-illary, or other nerves are affected. In the legs, peroneal orsciatic neuropathy may lead to foot drop, which may mis-takenly be attributed to a radiculopathy; an obturator orlateral femoral cutaneous neuropathy may also occur,sometimes in relation to a prolonged period in the lithot-omy position.3 In other instances, the mechanism of nerveinjury is not apparent, and symptoms of nerve involvementmay not develop until several days after anesthesia.4,5 Insuch circumstances, the etiology may be multifactorial,relating, for example, to minor degrees of compression inconjunction with a preexisting subclinical lesion,6 meta-bolic derangements, or an increased susceptibility to dam-age,7 or injury may have occurred after the patient has left

the operating room. Regardless of the underlying mecha-nism, anesthesia-related nerve injury most commonly in-volves the ulnar nerve (28% of nerve-injury cases in theClosed Claims Database) or brachial plexus (20%).1 Me-chanical stretch or elongation is probably the most com-mon cause of anesthesia-related brachial plexopathy.

In all these various circumstances, electrophysiologictesting is important in defining the neurogenic basis ofweakness and localizing the site of the lesion. It is also ofhelp in determining the severity of injury and thus inguiding prognostication. Electrodiagnostic testing doesnot, however, indicate the etiology of the neuropathy.For example, it may confirm the presence of an ulnarneuropathy, localize the lesion to the elbow, suggestwhether it is new or of long standing, and indicate itsseverity, but it does not indicate its cause, which must beinferred on clinical grounds. The precise mechanism ofperioperative ulnar neuropathy may not be obvious, butlocation at the elbow provides some support for a com-pressive basis, perhaps related to malpositioning.

With mild injuries, any clinical deficit relates primarilyto a block in the conduction of nerve impulses throughthe affected segment of nerve (neurapraxia), with pre-served conduction in neighboring segments. When theoffending cause has been removed, recovery occurs overa variable time that may be as long as several weeks if theinjury was severe enough to cause structural changes ofthe myelin sheath encasing axons. Complete recovery,however, can generally be anticipated. By contrast, se-vere nerve injuries lead to axonal degeneration, in whichcase recovery does not occur except by axonal regener-ation or sprouting from surviving neighboring axons andis likely to be prolonged and incomplete. The prognosisis influenced particularly by the integrity of the support-ing structures in the nerve. Axonotmesis is the termused to designate such an injury when axons are dis-rupted, but the epineurium (and usually the peri-neurium) remains intact. More severe injury, in whichthe epineurium is disrupted, is designated neurotmesis,and recovery does not occur without surgical repair;even then, it is usually incomplete.8

Electromyography

The electromyographic examination involves record-ing the electrical activity of muscle from a needle elec-

* Professor of Neurology.

Received from the Department of Neurology, School of Medicine, Universityof California, San Francisco, California. Submitted for publication July 31, 2003.Accepted for publication September 9, 2003. Supported was provided solelyfrom institutional and/or departmental sources.

The illustrations for this section are prepared by Dimitri Karetnikov, 7 Ten-nyson Drive, Plainsboro, New Jersey 08536.

Address reprint requests to Dr. Aminoff: Department of Neurology, School ofMedicine, University of California, San Francisco, California 94143-0114. Addresselectronic mail to: [email protected].

Anesthesiology, V 100, No 5, May 2004 1298

trode inserted within it. After amplification and signalprocessing, the activity can be displayed on the screen ofan oscilloscope or a video monitor for visual analysis andfed to a loudspeaker system so it can be monitoredacoustically. The presence and nature of abnormalitiesdepend on the affected component within the motorunit (which consists of the anterior horn cell, its axonand neuromuscular junctions, and the muscles fibersthat it innervates); the distribution of abnormalities indi-cates the likely site of involvement when denervationhas occurred.

Certain findings are suggestive of denervation. Suchfindings include the presence of abnormal spontaneousactivity in the resting muscle (especially fibrillation po-tentials and positive sharp waves, which result frommuscle fiber irritability) and increased insertion activity(i.e., activity induced by insertion or movement of theneedle electrode in the muscle). Insertion activity in-creases within a few days of muscle denervation,whereas abnormal spontaneous activity takes 1–4 weeksto develop, depending on the distance between thenerve lesion and the muscle. The electrical features thatdefine such activity are not discussed here but can befound in standard textbooks.9–13 It should be noted,however, that abnormal spontaneous activity and in-creased insertion activity are not pathognomonic of de-nervation but may also occur in certain disorders ofmuscle or the neuromuscular junction. The electromyo-graphic findings are therefore interpreted in the clinicalcontext in which they are obtained. Furthermore, abnor-

mal spontaneous activity is not always found in dener-vated muscle, and it disappears with reinnervation.Therefore, the electromyographic findings must be re-lated to the temporal profile of nerve injury (table 1).

Abnormalities of motor unit recruitment are also im-portant. Slight activation of a muscle normally causes anumber of motor units to discharge, depending on thedegree of voluntary contraction. With increasing con-traction, more motor units are activated, and they fire ata higher rate. In disorders with neurogenic weakness(regardless of whether this is due to conduction block oraxon loss), a reduced number of motor units are acti-vated for a given degree of voluntary effort, and theirfiring frequency is increased relative to the number ofmotor units activated. In very weak muscles, only a fewmotor units are activated; in paralyzed muscle, no motorunits may fire during attempted contraction.

Therefore, electromyographic findings are helpful inindicating whether weakness has a neurogenic basis andin defining the extent of nerve injury. Depending on thepattern of affected muscles, it is possible to distinguishbetween radiculopathies, plexopathies, and neuropa-thies and also to determine whether a neuropathy in-volves one or several nerves. A specific etiologic diagno-sis cannot be made by the electrophysiologic findings,however.

The configuration of motor unit potentials is helpful indetermining the duration of nerve injury and in indicat-ing whether reinnervation is occurring. In the normallimb muscle, most motor unit potentials are biphasic or

1299ELECTRODIAGNOSIS OF NERVE INJURIES

Anesthesiology, V 100, No 5, May 2004

triphasic, with a duration of 7–15 ms and an amplitudebetween 200 �V and 3 mV. Immediately after acutedenervation, the number of motor units is reduced to alevel that depends on the extent of the lesion (no motorunits are activated if the muscle is denervated com-pletely), but surviving motor unit potentials are un-changed in configuration. If axon loss has occurred (asopposed to conduction block) with moderate nerve in-jury, subsequent reinnervation occurs first by nonin-jured axons sending out collateral sprouts. Thus, over 1to several months after injury, motor units becomelarger. With severe nerve injuries or transection, subse-quent reinnervation requires axonal regeneration; rein-nervated motor units initially appear as small, short-duration, polyphasic motor unit potentials, which thenevolve to longer, larger potentials as more muscle fiberscome to be reinnervated. Long-duration, high-amplitude,polyphasic motor unit potentials are characteristic ofreinnervation after a denervating lesion (table 1).

The electromyographic findings may thus provide aguide to the time of onset of the lesion and to itschronicity, and this may have medicolegal implications.If a patient reports that a wrist drop has developedimmediately after an operative procedure and needleelectromyography performed shortly thereafter revealsabnormal spontaneous activity (fibrillation potentialsand positive sharp waves) in the extensor muscle of thewrist, it is likely that the lesion is at least 1–3 weeks oldand therefore that it preceded the surgery. Similarly, thepresence of long-duration, large-amplitude, polyphasicmotor unit potentials indicates that the denervation oc-curred several months or more before surgery becausesome reinnervation has occurred.

Nerve Conduction Studies

Nerve conduction studies permit assessment of func-tion in motor and sensory nerves. For motor studies, thenerve is stimulated supramaximally at two points (ormore) along its course, and a recording is made of theelectrical response of one of the muscles that it inner-vates. This permits conduction velocity to be deter-mined in the fastest-conducting fibers to that muscle.The size of the muscle response (i.e., the compoundmuscle action potential) provides an estimate of thenumber of motor axons and muscle fibers that are acti-vated by the stimulus. An abnormal reduction in size ofthe response with stimulation of the nerve at one pointalong its course, compared with stimulation at a moredistal site, may be indicative of conduction block,acutely evolving axon loss, or anomalous innervation (inwhich some nerve fibers follow an aberrant course toreach their target).

Sensory conduction studies typically involve stimulat-ing supramaximally the nerve fibers at one point and

recording the nerve action potentials from them at an-other. The latency of the response can be measured and,if desired, converted to a conduction velocity, and thesize of the sensory nerve action potential can also berecorded as a reflection of the number of functioningsensory axons.

Nerve conduction studies are an important means ofevaluating the functional integrity of peripheral nerves.They enable a focal nerve lesion to be localized in pa-tients with a mononeuropathy. Localized peripheralnerve damage leads to evoked motor or sensory re-sponses that are reduced or change abnormally in am-plitude depending on the site of stimulation and record-ing; conduction velocity may also be slowed. Nerveconduction studies combined with needle electromyo-graphy can determine whether a nerve injury is com-plete or incomplete and thus guide prognosis and thelikely course of recovery. With a complete lesion, motorunits cannot be activated volitionally in a distal muscle,and, if axonal loss has occurred, fibrillation potentialsand positive waves are found on needle examinationafter an appropriate interval (that varies with the site ofinjury and recording); electrical stimulation of the nerveabove the lesion does not elicit a response in musclessupplied by branches arising distal to a complete lesion,or it elicits a smaller response with a partial injury.Electrical stimulation below the site of, for example,complete nerve transection continues to elicit a distalresponse until wallerian degeneration of the distal nervestump has occurred (usually in 5–10 days),14 as indicatedin table 1 and figure 1.

In patients presenting with a mononeuropathy, nerveconduction studies may reveal the presence of a subclin-ical polyneuropathy that has made the individual nervesmore susceptible to injury. In patients with multipleaffected nerves, such studies can distinguish between apolyneuropathy (in which there is symmetrical involve-ment of multiple nerves at the same time, usually in alength-dependent manner) and mononeuropathy multi-plex (in which involvement of several nerves occurs,usually noncontiguously and at different times), which isimportant because different causes are likely to be re-sponsible. Finally, nerve conduction studies may suggestwhether the underlying pathologic process is axon lossor demyelination, which has important implications re-garding clinical course and prognosis. Axon loss is char-acterized electromyographically by signs of denervation,and nerve conduction studies reveal small (or absent)compound muscle or sensory nerve action potentials,with little or no change in conduction velocity while thiscan be measured. Demyelination, by contrast, is manifestby markedly slowed nerve conduction velocities. Con-duction block may also occur: Some or all of the axonsin the nerve become unable to transmit impulsesthrough a segment of nerve but can function more dis-tally. Stimulation proximal to the block then leads to a

1300 MICHAEL J. AMINOFF


smaller muscle response (or no response at all) thanwhen the nerve is stimulated distally.

Other Electrophysiologic Techniques

Other techniques for evaluating neuromuscular func-tion have been developed over the years. These includerepetitive nerve stimulation or single-fiber electromyo-graphy to evaluate neuromuscular transmission, quanti-tative electromyographic techniques, late-response stud-ies (F-wave or H-reflex studies) and recording ofsomatosensory evoked potentials to detect proximal pa-thology, and various techniques to evaluate reflex func-tion. These are beyond the scope of the current article,but monitoring of somatosensory evoked potentials issometimes helpful for preventing intraoperative damageto neural structures, especially the spinal cord.15

Clinical Applications

Electromyography and nerve conduction studies pro-vide helpful information for anesthesiologists in severalsettings. They are helpful in determining the basis of anyclinical deficit, in localizing the responsible lesion, andin defining its severity and prognosis. They do not indi-

cate directly the cause of the injury, although the loca-tion and age of the lesion and underlying pathologicprocess (axon loss or demyelinative changes) may helpto distinguish between various possibilities.

As mentioned previously, the mechanism of perioper-ative nerve injury is sometimes obscure. Injury may cer-tainly result from compression of nerves occurring whilethe patient is anesthetized and receiving muscle relax-ants, and proper positioning of patients is thereforeimperative. Ulnar or radial neuropathies in the arm areparticularly common in this context, and the peronealnerve may be compressed against the fibular head. Othernerves are involved less commonly. Individual periph-eral nerves may also be injured by direct injury, as fromintraneural injection of local anesthetics or other sub-stances, or by the placement of a tourniquet to limitblood flow to the limb. In these situations, electrodiag-nostic studies are important in localizing the lesion anddefining the prognosis. Mechanical damage is probablythe major cause of injury in tourniquet paralysis, butischemia may be contributory. In the upper limb, severalnerves are usually affected by tourniquet injuries, withthe radial occasionally affected in isolation; in the legs,the sciatic nerve is affected most often. Electrodiagnosticstudies typically reveal a focal conduction block in af-

Fig. 1. Compound muscle action potential recorded from a muscle with stimulation of its motor nerve immediately after (uppertraces) and 7 days after a focal nerve injury characterized by conduction block (A) and complete axonal loss (as may follow nervetransection) (B).



fected nerves and have sometimes localized the lesion tothe upper or lower edge of the tourniquet.

It may be difficult, particularly for nonneurologists, todistinguish clinically between, for example, a peronealor sciatic neuropathy and lumbar radiculopathy, all ofwhich may lead to foot drop in the perioperative period,or between a radial neuropathy and a cervical radiculop-athy that is causing wrist drop. Clinical definition andlocalization of a peripheral nerve lesion may be espe-cially difficult when selective nerve fascicles are injured,leading to an atypical or incomplete presentation.16 Theelectrodiagnostic findings can help to distinguish be-tween various causes of a particular clinical presenta-tion. In particular, they indicate which muscles havebeen affected, clarify the site of the lesion, and maylocalize any dysfunction with precision to a short seg-ment of peripheral nerve. Weakness in patients with acompressive neuropathy typically relates to demyelinat-ing conduction block, and this can be localized accu-rately by the so-called inching technique,17 in which thesite of stimulation of a peripheral nerve is altered in small(1-inch) steps, while the muscle response is recorded;conduction block at a specific point leads to an abruptreduction in size and increase in latency of the muscleresponse with stimulation at this point, compared tomore distal stimulation (fig. 2). The electrophysiologicfindings are also helpful in determining the underlyingpathologic process and thus the prognosis. In patientswith mild lesions, segmental demyelination is typicallyresponsible, and recovery is then likely to occur quicklyand completely. By contrast, if axonal loss has also oc-curred, evidence of denervation can be found (if theexamination is conducted at a suitable time after onset ofthe lesion, as indicated above), and recovery may bedelayed and incomplete. With mixed lesions, the neura-

praxic component typically recovers quickly, but theaxonal-loss component requires longer for recovery tooccur.

The location of a lesion may be important in determin-ing the likely underlying cause. For example, the devel-opment of acute foot drop may be attributed clinically toa nerve injury as a result of sciatic nerve block, butelectrophysiologic evidence of a focal lesion at the headof the fibula would make this unlikely.

Timing of the ElectrophysiologicExamination

The optimal timing of the electrodiagnostic examina-tion depends on the reason that it is undertaken. In apatient with postoperative reports of weakness or sen-sory changes, electrophysiologic evaluation even in thefirst 2 or 3 days may provide useful information. At thisearly time, the examination can help to determinewhether a nerve lesion is indeed present as evidenced bya reduced recruitment of motor units in involved mus-cles. The presence of at least some motor units undervoluntary control shows that any such lesion is incom-plete; this implies a more favorable prognosis than oth-erwise in patients with an apparently complete lesionclinically. The presence of abnormal spontaneous activ-ity (fibrillation potentials and positive waves) at this timeindicates that a long-standing lesion is present, as does asmall muscle response to distal nerve stimulation (table1). This is of medicolegal importance, suggesting eitherthat perioperative nerve injury is not responsible for thefindings (and perhaps also for the clinical deficit) or thatany perioperative injury was superimposed on a long-standing lesion that may have made the nerve moresusceptible to injury.

Fig. 2. Recordings from the abductor dig-iti minimi muscle to show the likelychanges with an ulnar nerve lesion at theelbow. (A and B) Normal responses tostimulation 2 inches and 1 inch below theelbow, respectively. (C) Small delayed re-sponse expected with stimulation at theelbow. (D and E) Similar small responseswithout additional abnormal delay tostimulation 1 inch and 2 inches proximalto the elbow, respectively.

1302 MICHAEL J. AMINOFF


More information is provided if the examination isrepeated approximately 4 weeks after injury, when ad-equate time has elapsed for the electrophysiologicchanges to have evolved more fully. At this time, moredefinitive information can be obtained about the site,nature, and severity of the lesion, which can guideprognostication.

Serial studies are generally not required becauseprogress can be followed clinically, unless patients havea clinically complete axon-loss lesion that is seeminglynot improving and surgical repair is a consideration. Inthis latter circumstance, serial electrophysiologic studiesevery 3 months may then be worthwhile: Needle elec-tromyography may indicate whether recovery is occur-ring, because voluntary motor unit activity reappearsbefore any clinical signs of recovery.

Intraoperative Electrophysiologic Studies

In recent years, intraoperative recordings from periph-eral nerves by similar techniques to those used in nerveconduction studies have proved useful in the surgicalmanagement of nerve injuries. Recording intraopera-tively has facilitated the identification of individualnerves, the determination of whether they are in conti-nuity, and the localization of damage to a specificsite.18,19 Electrophysiologic identification of nerveswhen their identity is uncertain, e.g., because of scarring,is accomplished by stimulating the tissue under consid-eration and recording the electrical responses of appro-priate muscles. When a nerve has been identified but itscontinuity is uncertain, the failure of stimulation to elicita muscle response may reflect conduction block ornerve transection (or a lack of proximity to the nerve).Mechanical stimulation (e.g., by manipulation or irriga-tion of the nerve) typically causes a brief burst of motorunit potentials, indicating that the nerve is in continuitydistal to the site of stimulation. Loss of such responsesmay indicate that the nerve has been injured, and in thiscircumstance, the response to electrical stimulationshould be assessed. By stimulating or recording at differ-ent sites along the course of a nerve, the site of damagecan be localized precisely. For example, the ulnar nervecan be monitored by stimulating it directly and recordingaction potentials from the nerve itself or from a musclesupplied by the nerve.

Intraoperative monitoring has also helped in the earlyrecognition of nerve damage caused by surgery close tolimb or cranial nerves so that the ongoing surgical pro-cedure can be modified before damage is irreversible.Therefore, it is common to monitor the facial nerveduring surgery in the cerebellopontine angle (e.g., foracoustic neuroma) to prevent injury to the nerve, whichmay be difficult to identify by inspection, especially

when it is caught up in the tumor. Depending on theoperative field, other cranial or spinal nerves may also bemonitored: the cranial nerves to the extraocular musclesduring surgery on the cavernous sinus, the lower cranialnerves during surgery on the skull base, and the spinalnerve roots during spinal surgery. Similarly, electromyo-graphic monitoring may detect early injury to the axillaryand musculocutaneous nerves during shoulder surgeryor to the femoral, obturator, and sciatic nerves duringhip surgery.20 If at-risk nerves are monitored duringsurgery, warning of damage during the operative proce-dure is provided by the development of prolonged neu-rotonic electromyographic discharges or changes in sizeof compound muscle action potentials. Nerve functionhas also been monitored by somatosensory evoked po-tentials, and the incidence of postoperative nerve injurieshas been reduced thereby, but whether the technique hasany advantage over electromyographic monitoring is un-clear. Monitoring by either technique may help to definethe mechanism of intraoperative nerve injury and therebylead to improved surgical technique.21

References

1. Cheney FW, Domino KB, Caplan RA, Posner KL: Nerve injury associatedwith anesthesia: A closed claims analysis. ANESTHESIOLOGY 1999; 90:1062–9

2. Bickler PE, Schapera A, Bainton C: Acute radial injury from use of anautomatic blood pressure monitor. ANESTHESIOLOGY 1990; 73:186–8

3. Warner MA, Warner DO, Harper CM, Schroeder DR, Maxson PM: Lowerextremity neuropathies associated with lithotomy positions. ANESTHESIOLOGY 2000;93:938–42

4. Warner MA: Perioperative neuropathies. Mayo Clin Proc 1998; 73:567–745. Warner MA, Warner DO, Matsumoto JY, Harper CM, Schroeder DR, Maxson

PM: Ulnar neuropathy in surgical patients. ANESTHESIOLOGY 1999; 90:54–96. Alvine FG, Schurrer ME: Postoperative ulnar-nerve palsy: Are there predis-

posing factors? J Bone Joint Surg Am 1987; 69:255–97. Sawyer RJ, Richmond MN, Hickey JD, Jarratt JA: Peripheral nerve injuries

associated with anaesthesia. Anaesthesia 2000; 55:980–918. Robinson LR: Traumatic injury to peripheral nerves. Muscle Nerve 2000;

23:863–739. Aminoff MJ: Electromyography in Clinical Practice, 3rd edition. New York,

Churchill Livingstone, 199810. Neuromuscular Function and Disease: Basic, Clinical, and Electrodiagnos-

tic Aspects, 2 vols. Edited by Brown WF, Bolton CF, Aminoff MJ. Philadelphia,Saunders, 2002

11. Clinical Neurophysiology, 2nd edition. Edited by Daube JR. New York,Oxford University Press, 2002

12. Kimura J: Electrodiagnosis in Diseases of Nerve and Muscle: Principles andPractice, 3rd edition. New York, Oxford University Press, 2001

13. Electrodiagnosis in Clinical Neurology, 4th ed. Edited by Aminoff MJ. NewYork, Churchill Livingstone, 1999

14. Chaudhry V, Cornblath DR: Wallerian degeneration in human nerves:Serial electrophysiological studies. Muscle Nerve 1992; 15:687–93

15. Nuwer MR: Spinal cord monitoring. Muscle Nerve 1999; 22:1620–3016. Stewart JD: Peripheral nerve fascicles: Anatomy and clinical relevance.

Muscle Nerve 2003; 28:525–4117. Campbell WW: The value of inching techniques in the diagnosis of focal

nerve lesions: Inching is a useful technique. Muscle Nerve 1998; 21:1554–618. Brown WF, Veitch J: Intraoperative monitoring of peripheral and cranial

nerves. Muscle Nerve 1994; 17:371–719. Daube JR, Harper CM: monitoring peripheral nerves during surgery, Neu-

romuscular Function and Disease: Basic, Clinical, and Electrodiagnostic Aspects,2 vols. Edited by Brown WF, Bolton CF, Aminoff MJ. Philadelphia, Saunders,2002, pp 1857–65

20. Helfet DL, Anand N, Malkani AL, Heise C, Quinn TJ, Green DS, Burga S:Intraoperative monitoring of motor pathways during operative fixation of acuteacetabular fractures. J Orthop Trauma 1997; 11:2–6

21. Jellish WS, Blakeman B, Warf P, Slogoff S: Somatosensory evoked potentialmonitoring used to compare the effect of three asymmetric sternal retractors onbrachial plexus function. Anesth Analg 1999; 88:292–7



Documents

Electrophysiologic Testing for the Diagnosis of Peripheral