lawrence r.robinson,md jeffrey g.jarvik,md,mph david … · university of washington ... washington...

��American Association of Neuromuscular & Electrodiagnostic Medicine

Lawrence R. Robinson, MD

Jeffrey G. Jarvik, MD, MPH

David G. Kline, MD

ASSESSMENT OF TRAUMATIC

NERVE INJURIES

2005 AANEM COURSE GAANEM 52nd Annual Scientific Meeting

Monterey, California

2005 COURSE GAANEM 52nd Annual Scientific Meeting

Monterey, California

Copyright © September 2005American Association of Neuromuscular & Electrodiagnostic Medicine

421 First Avenue SW, Suite 300 EastRochester, MN 55902

PRINTED BY JOHNSON PRINTING COMPANY, INC.

Lawrence R. Robinson, MDJeffrey G. Jarvik, MD, MPH

David G. Kline, MD

Assessment of Traumatic Nerve Injuries

��


Faculty

ii


Professor

Department of Rehabilitation Medicine

University of Washington

Seattle, Washington

Dr. Robinson attended Baylor College of Medicine and completed his res-idency training in rehabilitation medicine at the Rehabilitation Institute ofChicago. He now serves as professor and chair of the Department ofRehabilitation Medicine at the University of Washington and is theDirector of the Harborview Medical Center Electrodiagnostic Laboratory.He is also currently Vice Dean for Clinical Affairs at the University ofWashington. His current clinical interests include the statistical interpreta-tion of electrophysiologic data, laryngeal electromyography, and the studyof traumatic neuropathies. He recently received the DistinguishedAcademician Award from the Association of Academic Physiatrists and thisyear is receiving the AANEM Distinguished Researcher Award.

Jeffrey G. Jarvik, MD, MPH

Professor

Department of Radiology and Neurology

Adjunct Professor

Department of Health Services

University of Washington

Seattle, Washington

As Director of Neuroradiology at the University of Washington, Dr.Jarvik’s clinical work encompasses the entire range of neuroradiology. Hisclinical and research focus has been on spinal and peripheral nerve imag-ing. His academic focus has been on health services as it relates to diagnos-tic imaging, e.g., how diagnostic imaging influences therapeutic decisionmaking and patient outcomes. Dr. Jarvik is a member of the AmericanSociety of Neuroradiology, the American College of Radiology, and theRadiological Society of North America, among others.

David G. Kline, MD

Boyd Professor and Head

Department of Neurosurgery

Louisiana State University Medical Center

New Orleans, Louisiana

Dr. Kline is currently a Boyd Professor and Head of the Department ofNeurosurgery at Louisiana State University (LSU) Medical Center in NewOrleans. He earned his medical degree from the University ofPennsylvania, then performed his internship at the University of Michigan.He performed residencies at the University of Michigan and Walter ReedGeneral Hospital and Institute of Research. Dr. Kline has served on sever-al editorial boards including Neurosurgery, Microsurgery, and the Journalof Reconstructive Microsurgery, among others. He is also active in manymedical societies. Dr. Kline was recently named a Boyd Professor at LSU,which is only given to faculty members who have attained national orinternational distinction for outstanding teaching, research, or other cre-ative achievement.

Course Chair: Alan R. Berger, MD

The ideas and opinions expressed in this publication are solely those of the specific authors and do not necessarily represent those of the AANEM.

Please be aware that some of the medical devices or pharmaceuticals discussed in this handout may not be cleared by the FDA or cleared by the FDA for the spe-cific use described by the authors and are “off-label” (i.e., a use not described on the product’s label). “Off-label” devices or pharmaceuticals may be used if, in thejudgement of the treating physician, such use is medically indicated to treat a patient’s condition. Information regarding the FDA clearance status of a particulardevice or pharmaceutical may be obtained by reading the product’s package labeling, by contacting a sales representative or legal counsel of the manufacturer of thedevice or pharmaceutical, or by contacting the FDA at 1-800-638-2041.

Faculty had nothing to disclose.

iii


Contents

Faculty ii

Objectives iii

Course Committee iv

Traumatic Nerve Injury to Peripheral Nerves 1Lawrence R. Robinson, MD

Peripheral Nerve Magnetic Resonance Imaging: The Median Nerve in Carpal Tunnel Syndrome 11Jeffrey G. Jarvik, MD, MPH

Surgical Management of Nerve Injuries 19David G. Kline, MD

CME Self-Assessment Test 29

Evaluation 31

Member Benefit Recommendations 33

Future Meeting Recommendations 35

O B J E C T I V E S —This course will provide an overview of the evaluation and management of traumatic nerve injuries. After attendingthis course, the participant will (1) understand the importance and time frame for the electrodiagnosis of nerve trauma, (2) learn the “state-of-the-art” concepts and future potential of neuroimaging of peripheral nerve injuries, and (3) learn the most important principles of theclinical and surgical management of peripheral nerve injuries.

P R E R E Q U I S I T E —This course is designed as an educational opportunity for residents, fellows, and practicing clinical EDX physiciansat an early point in their career, or for more senior EDX practitioners who are seeking a pragmatic review of basic clinical and EDX prin-ciples. It is open only to persons with an MD, DO, DVM, DDS, or foreign equivalent degree.

AC C R E D I TAT I O N S TAT E M E N T —The AANEM is accredited by the Accreditation Council for Continuing Medical Education toprovide continuing medical education (CME) for physicians.

CME C R E D I T —The AANEM designates attendance at this course for a maximum of 3.25 hours in category 1 credit towards theAMA Physician’s Recognition Award. This educational event is approved as an Accredited Group Learning Activity under Section 1 of theFramework of Continuing Professional Development (CPD) options for the Maintenance of Certification Program of the Royal Collegeof Physicians and Surgeons of Canada. Each physician should claim only those hours of credit he/she actually spent in the activity. TheAmerican Medical Association has determined that non-US licensed physicians who participate in this CME activity are eligible for AMAPMR category 1 credit. CME for this course is available 9/05 - 9/08.

iv

Thomas Hyatt Brannagan, III, MDNew York, New York

Timothy J. Doherty, MD, PhD, FRCPCLondon, Ontario, Canada

Kimberly S. Kenton, MDMaywood, Illinois

Dale J. Lange, MDNew York, New York

Subhadra Nori, MDBronx, New York

Jeremy M. Shefner, MD, PhDSyracuse, New York

T. Darrell Thomas, MDKnoxville, Tennessee

Bryan Tsao, MDShaker Heights, Ohio

2004-2005 AANEM PRESIDENT

Gary Goldberg, MDPittsburgh, Pennsylvania

2004-2005 AANEM COURSE COMMITTEE

Kathleen D. Kennelly, MD, PhDJacksonville, Florida

EPIDEMIOLOGY OF PERIPHERAL NERVE TRAUMA

Traumatic injury to peripheral nerves results in considerabledisability everywhere in the world. In peacetime, peripheralnerve injuries commonly result from trauma due to motorvehicle accidents, and less commonly from penetrating trau-ma, falls, and industrial accidents. Out of all patients admit-ted to Level I trauma centers, it is estimated that roughly 2-3% have peripheral nerve injuries.30,36 If plexus and rootinjuries are also included, the incidence is about 5%.30

In the upper limb, the most commonly reported nerveinjured is the radial nerve, followed by the ulnar and mediannerves.30,36 Lower limb peripheral nerve injuries are less com-mon, with the sciatic nerve most frequently injured, followedby the peroneal and rarely tibial or femoral nerves. Fracturesof nearby bones are commonly associated, such as humeralfractures with radial neuropathy.

In wartime, peripheral nerve trauma is much more commonand most of physician’s knowledge about peripheral nerveinjury, repair, and recovery comes from experience derived inWorld Wars I and II, and subsequent wars.20,35,40

Peripheral nerve injuries may be seen as an isolated nervoussystem injury, but may also accompany central nervous sys-tem (CNS) trauma, not only compounding the disability, butmaking recognition of the peripheral nerve lesion problemat-ic. Of patients with peripheral nerve injuries, approximately

60% have a traumatic brain injury.30 Conversely, of thosewith traumatic brain injury admitted to rehabilitation units,10-34% have associated peripheral nerve injuries.7,14,39 It isoften easy to miss peripheral nerve injuries in the setting ofCNS trauma. Since the neurologic history and examinationis limited, early hints to a superimposed peripheral nervelesion might be only flaccidity, areflexia, and reduced move-ment of a limb.

Peripheral nerve injuries are of significant import as theyimpede recovery of function and return to work, and carryrisk of secondary disabilities from falls, fractures, or other sec-ondary injuries. An understanding of the classification,pathophysiology, and electrodiagnosis of these lesions is crit-ical to the appropriate diagnosis, localization, and manage-ment of peripheral nerve trauma.

CLASSIFICATION OF NERVE INJURIES

There are two predominant schemes that have been proposedfor classification of peripheral nerve traumatic injuries; thatof Seddon35 and that of Sunderland40 (Table 1). The formeris more commonly used in the literature. Seddon has used theterms “neurapraxia,” “axonotmesis,” and “neurotmesis” todescribe peripheral nerve injuries.35 Neurapraxia is a compar-atively mild injury with motor and sensory loss but no evi-dence of Wallerian degeneration. The nerve distally conductsnormally. Focal demyelination and/or ischemia are thought

Traumatic Injury to Peripheral Nerves


Professor and ChairDepartment of Rehabilitation Medicine

University of WashingtonSeattle, Washington

to be the etiologies of the conduction block. Recovery mayoccur within hours, days, weeks, or up to a few months.Axonotmesis is commonly seen in crush injuries. The axonsand their myelin sheaths are broken, yet the surroundingstroma (i.e., the endoneurium, perineurium, and epineuri-um) remains partially or fully intact. Wallerian degenerationoccurs, but subsequent axonal regrowth may proceed alongthe intact endoneurial tubes. Recovery ultimately dependsupon the degree of internal disorganization in the nerve aswell as the distance to the end organ.

Sunderland’s classification further divides peripheral nerveinjuries category. Neurotmesis describes a nerve that has beeneither completely severed or is so markedly disorganized byscar tissue that axonal regrowth is impossible. Examples aresharp injury, some traction injuries, or injection of noxiousdrugs. Prognosis for spontaneous recovery is extremely poorwithout surgical intervention.

Sunderland40 uses a more subdivided scheme to describeperipheral nerve injuries, with five groups instead of three. Firstdegree injury represents conduction block with completely

intact stroma and corresponds to Seddon’s classification ofneurapraxia. Prognosis is good. Second degree injury involvestransection of the axon but with intact stroma. Recovery canoccur by axonal regrowth along endoneurial tubes. Thirddegree injury represents transection of the axon andendoneurial tubes, but the surrounding perineurium isintact. Recovery depends upon how well the axons can crossthe site of the lesion and find endoneurial tubes. Fourthdegree injury involves loss of continuity of axons, endoneur-ial tubes, and perineurium. Individual nerve fascicles aretransected and the continuity of the nerve trunk is main-tained only by the surrounding epineurium. Traction injuriescommonly produce these types of lesions. Prognosis is usual-ly poor absent surgical intervention because of the markedinternal disorganization of guiding connective tissue ele-ments and associated scarring. Fifth degree injury describestransection of the entire nerve trunk and is similar toSeddon’s neurotmesis.

Some authors have described another “degree” of injury,known as sixth degree injury.25 This is a mixed lesion withboth axon loss and conduction block each occurring in some

2 Traumatic Injury to Peripheral Nerves AANEM Course

Table 1 Classification Systems for Nerve Injury

Seddon SunderlandClassification Classification Pathology Prognosis

Neurapraxia First degree Myelin injury or ischemia Excellent recovery in weeks to months

Axonotmesis Axons disrupted Good to poor, depending upon integrity ofVariable stromal disruption supporting structures and distance to muscle

Second degree Axons disrupted Good, depending upon distance to muscleEndoneurial tubes intactPerineurium intactEpineurium intact

Third degree Axons disrupted Poor Endoneurial tubes disrupted Axonal misdirectionPerineurium intact Surgery may be requiredEpineurium intact

Fourth degree Axons disrupted Poor Endoneurial tubes disrupted Axonal misdirectionPerineurium disrupted Surgery usually requiredEpineurium intact

Neurotmesis Fifth degree Axon disrupted No spontaneous recoveryEndoneurial tubes disrupted Surgery requiredPerineurium disrupted Prognosis after surgery guardedEpineurium disrupted

Adapted from Dillingham.8

fibers. This type of lesion is probably quite common andrequires skillful electrodiagnostic data collection and analysisto separate from pure axon loss lesions.

EFFECTS OF NEURAPRAXIA ON NERVE AND MUSCLE

As noted above, neurapraxic injuries to peripheral nerves maybe due to ischemia or focal demyelination. When ischemiafor a brief period (i.e., up to 6 hours) is the underlying cause,there are usually no structural changes in the nerve, thoughthere may be edema in other nearby tissues.18

On the other hand, in neurapraxic lesions due to focaldemyelination, there are anatomic changes predominantlyaffecting the myelin sheath, but sparing the axon. Tourniquetparalysis has been used to produce an animal model of a neu-rapraxic lesion, though it is recognized that acute crushinjuries may be different in mechanism than prolongedapplication of a tourniquet.31 In this model, anatomicchanges along the nerve are most marked at the edge of thetourniquet where a significant pressure gradient existsbetween the tourniquet and nontourniquet areas. The pres-sure gradient essentially “squeezes” out the myelin withresulting invagination of one paranodal region into the next.As a result, there is an area of focal demyelination at the edgeof the tourniquet.31 Larger fibers are more affected thansmaller fibers.

In this area of focal demyelination, impulse conduction fromone node of Ranvier to the next is slowed as current leakageoccurs and the time for impulses to reach threshold at succes-sive nodes of Ranvier is prolonged. Slowing of conductionvelocity along this nerve segment ensues. More severedemyelination results in complete conduction block. Thishas been reported to occur when internodal conductiontimes exceed 500-600 ms.33 Since there are very few sodiumchannels in internodal segments of myelinated nerves, con-duction in demyelinated nerves cannot simply proceed slow-ly as it would for normally unmyelinated nerves. Thus, suffi-cient demyelination results in block of conduction ratherthan simply more severe slowing.

There are relatively few changes in muscle as a result of neu-rapraxic lesions. Disuse atrophy can occur when neurapraxiais more than transient. There remains debate as to whethermuscle fibrillates after a purely neurapraxic lesion.

EFFECTS OF AXONOTMESIS ON NERVE AND MUSCLE

Soon after an axonal lesion, the process of Wallerian degen-eration begins to occur in nerve fibers. This process is well

described elsewhere10,27 and will be only briefly reviewed.There are changes in both the axon and the nerve cell body.In the axon, a number of changes occur in the first 2 daysincluding leakage of axoplasm from the severed nerve,swelling of the distal nerve segment, and subsequently disap-pearance of neurofibrils in the distal segment. By day three,there is fragmentation of both axon and myelin with thebeginning of digestion of myelin components. By day eight,the axon has been digested and Schwann cells are attemptingto bridge the gap between the two nerve segments. Nervefibers may also degenerate for a variable distance proximally;depending upon the severity of the lesion, this retrogradedegeneration may extend for several centimeters.

If the lesion is sufficiently proximal, there are also a numberof changes at the nerve cell body level occurring after nervetrauma. Initially, within the first 48 hours, the Nissel bodies(the cell’s rough endoplasmic reticulum) break apart into fineparticles. By 2 to 3 weeks after injury, the cell’s nucleusbecomes displaced eccentrically and the nucleolus is alsoeccentrically placed within the nucleus. These changes mayreverse as recovery occurs.

ELECTRODIAGNOSIS: TIMING OF CHANGES ANDDETERMINING DEGREE OF INJURY

The Compound Motor Action Potential

Neurapraxia

In purely neurapraxic lesions, the compound muscle actionpotential (CMAP) will change immediately after injury,assuming one can stimulate both above and below the site ofthe lesion (Figure 1a-c). When recording from distal musclesand stimulating distal to the site of the lesion, the CMAPshould always be normal since no axonal loss and noWallerian degeneration has occurred. Moving stimulationproximal to the lesion will produce a smaller or absentCMAP, as conduction in some or all fibers is blocked. Itshould be remembered that amplitudes normally fall withincreasing distance between stimulation and recording; hencethere is some debate about how much of a drop in amplitudeis sufficient to demonstrate conduction block. Amplitudedrops exceeding 20% over a 25 cm distance or less are clear-ly abnormal; smaller changes over smaller distances are likelyalso suggestive of an abnormality. In addition to conductionblock, partial lesions also often demonstrate concomitantslowing across the lesion. This slowing may be due to eitherloss of faster conducting fibers or demyelination of survivingfibers. All these changes in the CMAP will generally persistuntil recovery takes place, typically by no more than a fewmonths postinjury. Most importantly, the distal CMAP will

AANEM Course Assessment of Traumatic Nerve Injuries 3

never drop in amplitude in purely neurapraxic injuries, sinceno axon loss or Wallerian degeneration occurs and the distalnerve segment remains normally excitable.

Axonotmesis and Neurotmesis

Electrodiagnostically, complete axonotmesis (equivalent toSunderland grades 2, 3, and 4) and complete neurotmesislook the same, since the difference between these types oflesions is in the integrity of the supporting structures, whichhave no electrophysiologic function. Thus these lesions canbe grouped together as axonotmesis for the purpose of thisdiscussion.

Immediately after axonotmesis and for a “few days” there-after, the CMAP and motor conduction studies look thesame as those seen in a neurapraxic lesion. Nerve segmentsdistal to the lesion remain excitable and demonstrate normalconduction while proximal stimulation results in an absent orsmall response from distal muscles. Early on, this picture

looks the same as conduction block and can be confused withneurapraxia. Hence neurapraxia and axontomesis cannot bedistinguished until sufficient time for Wallerian degenerationin all motor fibers has occurred, typically about 9 dayspostinjury.6

As Wallerian degeneration occurs, the amplitude of theCMAP elicited with distal stimulation will fall. This starts atabout day three and is complete by about day nine.6

Neuromuscular junction transmission fails before nerveexcitability.15,16 Thus in complete axonotmesis at day nine,one has a very different picture than neurapraxia. There areabsent responses both above and below the lesion. Partialaxon loss lesions will produce small amplitude motorresponses, with the amplitude of the CMAP roughly propor-tional to the number of surviving axons. Side-to-side CMAPamplitudes can be compared to estimate the degree of axonloss, though inherent side-to-side variability of up to 30-50%limits the accuracy of the estimate. Using the CMAP ampli-tude to estimate the degree of surviving axons is also most


Figure 1 Representation of changes in the compound muscle action potential (CMAP) after neurapraxia (A), axonotmesis (B),neurotmesis (B), and mixed lesions (C).

A B

C

reliable only early after injury, before axonal sprouting hasoccurred. Use of this technique later after injury will tend tounderestimate the degree of axon loss.

Mixed Lesions

Lesions which have a mixture of axon loss and conductionblock provide a unique challenge. These can usually be sort-ed out by carefully examining amplitudes of the CMAPelicited from stimulation both above and below the lesionand by comparing the amplitude with distal stimulation tothat obtained from the other side. The percentage of axonloss is best estimated by comparing the CMAP amplitudefrom distal stimulation with that obtained contralaterally. Ofthe remaining axons, the percentage with conduction blockare best estimated by comparing amplitudes or areas obtainedwith stimulation distal and proximal to the lesion. Thus if a1 mV response is obtained with proximal stimulation, a 2mV response is obtained distally, and a 10 mV response isobtained with distal stimulation contralaterally, the cliniciancan deduce that probably about 80% of the axons are lost,and of the remaining 20%, half are blocked (neurapraxic) atthe lesion site. As previously mentioned, this analysis is mostuseful only in the acute phase, before reinnervation by axon-al sprouting occurs.

F Waves

F waves may change immediately after the onset of a neu-rapraxic lesion. In a complete block, responses will be absent.However, in partial lesions, changes can be more subtle sinceF waves are dependent upon only 3-5% of the axon popula-tion to elicit a response.12 Thus partial lesions may have nor-mal miminal F-wave latencies, and mean latencies, withreduced or possibly normal penetrance. While F waves areconceptually appealing for detecting proximal lesions (e.g.,brachial plexopathies) it is in few instances that they trulyprovide useful additional or unique information. They aresometimes useful in very early proximal lesions when conven-tional studies are normal since stimulation does not occurproximal to the lesion, but they are not good at distinguish-ing axon loss lesions from conduction block.

Compound or Sensory Nerve Action Potentials

Neurapraxia

The sensory nerve action potential (SNAP) and compoundnerve action potential (CNAP) will show changes similar tothe CMAP after focal nerve injury. In the setting of neu-rapraxia, there is a focal conduction block at the site of thelesion, with preserved distal amplitude. However, the criteriafor establishing conduction block in sensory nerve fibers aresubstantially different than that for the CMAP. When record-

ing nerve action potentials, there is normally a greater dropin amplitude over increasing distance between stimulatingand recording electrodes, due to temporal dispersion andphase cancellation.22 Amplitude drops of 50-70% over a 25cm distance are not unexpected and it is less clear just whatchange in amplitude is abnormal. A large focal change over asmall distance is probably significant. Slowing may alsoaccompany partial conduction blocks, as for the CMAP.Responses elicited with stimulation and recording distal tothe lesion are normal in pure neurapraxic injuries.


Immediately after axonotmesis, the SNAP looks the same asseen in a neurapraxic lesion. Nerve segments distal to thelesion remain excitable and demonstrate normal conductionwhile proximal stimulation results in an absent or smallresponse. Hence neurapraxia and axontomesis can not be dis-tinguished until sufficient time for Wallerian degeneration inall sensory fibers has occurred, typically about 11 days postinjury.6 It takes slightly longer for sensory nerve studies todemonstrate loss of amplitude than for motor studies (i.e., 11days versus 9 days), due to the earlier failure of neuromuscu-lar junction transmission compared to nerve conduction.

Needle Electromyography

Neurapraxia

The needle electromyography (EMG) examination in purelyneurapraxic lesions will show neurogenic changes in recruit-ment with debatable abnormalities in spontaneous activity.As mentioned earlier, there is debate as to whether fibrillationpotentials are recorded after a purely neurapraxic lesion. Onestudy of peripheral nerve lesions in baboons has failed todemonstrate fibrillations in purely neurapraxic lesions.17 Onthe other hand, study of purely neurapraxic lesions in rats5

has suggested fibrillations occur in blocked, but not dener-vated, muscle fibers. There are limited reports of fibrillationsin humans with apparently predominantly neurapraxic nervelesions,43,48 but it is difficult to know whether or not anyaxon loss had occurred in these patients, since nerve conduc-tion studies (NCSs) are not sensitive for detecting minimalaxon loss. Needle EMG is more sensitive for detecting motoraxon loss than NCSs, and hence it is easy to imagine situa-tions in which NCSs are within normal limits, but needleEMG detects minimal or mild axon loss.

Independent of whether or not the needle EMG demon-strates fibrillation potentials in neurapraxia, the most appar-ent change on needle EMG will be changes in recruitment.These occur immediately after injury. In complete lesions(i.e., complete conduction block) there will be no motor unitaction potentials (MUAPs). In incomplete neurapraxic


lesions, there will be reduced numbers of MUAPs firing morerapidly than normal (i.e., reduced or discrete recruitment).Recruitment changes alone are not specific for neurapraxia oraxon loss.

Since no axon loss occurs in neurapraxic injuries, there willbe no axonal sprouting and no changes in MUAP morphol-ogy (e.g., duration, amplitude, or phasicity) anytime afterinjury.


A number of days after an axon loss lesion, needle EMG willdemonstrate fibrillation potentials and positive sharp waves.The time between injury and onset of fibrillation potentialswill be dependent in part upon the length of distal nervestump. When the lesion is distal and the distal stump is short,it takes only 10-14 days for fibrillations to develop. With aproximal lesion and a longer distal stump (e.g., ulnar-inner-vated hand muscles in a brachial plexopathy), 21-30 days arerequired for full development of fibrillation potentials andpositive sharp waves.42 Thus, the electrodiagnostic (EDX)physician needs to be acutely aware of the time since injury,so that severity is not underestimated when a study is per-formed early after injury, and also so that development ofincreased fibrillation potentials over time is not misinterpret-ed as a worsening of the injury.

Fibrillation and positive sharp wave density are usually grad-ed on a 1-4 scale. This is an ordinal scale, meaning that asnumbers increase findings are worse. However, it is not aninterval or ratio scale, i.e., 4+ is not twice as bad as 2+ or 4times as bad as 1+. Moreover, 4+ fibrillation potentials doesnot reflect complete axon loss, and in fact may represent onlya minority of axons lost.3,9 Evaluation of recruitment andparticularly of distally elicited CMAP amplitude are neces-sary before a determination can be made whether or notcomplete axon loss has occurred.

Fibrillation potential size will decrease over time since injury.Kraft24 has demonstrated that fibrillations initially are sever-al hundred microvolts in the first few months after injury.However, when lesions are more than 1 year old, they areunlikely to be over 100 µV in size. Fibrillations will alsodecrease in number as reinnervation occurs, however thisfinding is not usually clinically useful for two reasons. First,since a qualitative or ordinal scale of fibrillation density istypically used and an accurate quantitative measurement offibrillation density is not available, comparison of fibrillationnumbers from one examination to the other is not reliable.9

Second, even in complete lesions, fibrillation density willeventually reduce since the muscle becomes fibrotic and thenumber of viable muscle fibers falls; in this case, reduction in

fibrillation numbers does not predict recovery, but rathermuscle fibrosis.

Fibrillations may also occur after direct muscle injury, as wellas nerve injury. Partanen and Danner32 have demonstratedthat patients after muscle biopsy have persistent fibrillationpotentials starting after 6-7 days and extending for up to 11months. In patients who have undergone multiple trauma,coexisting direct muscle injury is common and can be poten-tially misleading when trying to localize a lesion.

When there are surviving axons after an incomplete axonalinjury, remaining MUAPs are initially normal in morpholo-gy, but demonstrate reduced or discrete recruitment. Axonalsprouting will be manifested by changes in morphology ofexisting motor units. Amplitude will increase, duration willbecome prolonged, and the percentage of polyphasic MUAPswill increase as motor unit territory increases.9,11 This processoccurs soon after injury. Microscopic studies demonstrateoutgrowth of these nerve sprouts starting at 4 days after par-tial denervation.21,27 Electrophysiologic studies utilizing sin-gle-fiber EMG demonstrates increase in fiber density startingat 3 weeks postinjury.26

In complete lesions, the only possible mechanism of recoveryis axonal regrowth. The earliest needle EMG finding in thiscase is the presence of small, polyphasic, often unstableMUAPs previously referred to as “nascent potentials.”

Observation of these potentials is dependent upon establish-ing axon regrowth as well as new neuromuscular junctionsand this observation represents the earliest evidence of rein-nervation, usually preceding the onset of clinically evidentvoluntary movement.9 These potentials represent the earliestdefinitive evidence of axonal reinnervation in completelesions. When performing the examination looking for newMUAPs, the clinician must be sure to accept only “crisp,”nearby MUAPs with a short rise-time, since distant potentialsrecorded from other muscles can be deceptive and could erro-neously suggest intact innervation.

Mixed Lesions

When there is a lesion with both axon loss and conductionblock, needle EMG examination can be misleading if inter-preted in isolation. If, for example, a lesion results in destruc-tion of 50% of the original axons and conduction block ofthe other 50%, then needle EMG will demonstrate abundant(4+) fibrillation potentials and no voluntary MUAPs. TheEDX physician should not then conclude that there is a com-plete axonal lesion, but should instead carefully evaluate themotor NCSs to determine how much of the lesion is neu-rapraxic and how much is axonotmetic. The important point


here is to not take the presence of abundant fibrillations andabsent voluntary MUAPs as evidence of complete denerva-tion.

LOCALIZATION OF TRAUMATIC NERVE INJURIES

The localization of peripheral nerve injuries is sometimesstraightforward but is potentially complicated by a variety ofpossible pifalls. Localization is usually performed by twomethods: (1) detecting focal slowing or conduction block onNCSs, or (2) assessing the pattern of denervation on needleEMG.

Localizing peripheral nerve lesions by NCSs usually requiresthat there be a focal slowing or conduction block as the EDXphysician stimulates above and below the lesion. To see sucha change there must either be focal demyelination orischemia, or the lesion should be so acute that degenerationof the distal stump has not yet occurred. Thus lesions withpartial or complete neurapraxia (due to either demyelinationor ischemia) can be well localized with motor NCSs, as canvery acute axonal injuries.

In pure axonotmetic or neurotmetic lesions, it is more diffi-cult if not impossible to localize the lesion using NCSs. Insuch a case, there will be mild and diffuse slowing in theentire nerve due to loss of the fastest fibers, or there will beno response at all. Conduction across the lesion site will beno slower than across other segments. In addition, providedenough time for Wallerian degeneration has elapsed (i.e., atleast 9 days for motor fibers or 11 days for sensory fibers),there will be no change in amplitude as one traverses the siteof the lesion. Thus, pure axon loss lesions are not well local-ized along a nerve by NCSs.

There are some cases in which indirect inferences can bemade about the location of purely axonal lesions. Forinstance, if the ulnar motor response is very small or absentand the median motor response is normal, this implies anulnar neuropathy rather than a lower brachial plexus lesion.However, in such an instance, the site of pathology along theulnar nerve may not be well defined.

Another indirect inference that can be made based upon sen-sory NCSs is placement of the lesion at a pre- versus postgan-glionic location. Lesions that are proximal to the dorsal rootganglion, i.e. at the preganglionic level (proximal root, cauda

equina, spinal cord) tend to have normal SNAP amplitudes,even in the setting of reduced or absent sensation.1,41 This isa particularly bad prognostic sign when seen in the setting ofpossible root avulsion. On the other hand, lesions occurringdistal to the dorsal root ganglion have small or absent SNAPs(when these are recorded in the appropriate distribution).Thus, SNAPs may be useful to differentiate root versusplexus or other pre- versus postganglionic locations. A limi-tation, particularly in partial lesions, is the wide variability inSNAP amplitudes seen in normal individuals. Mixed pre-and postganglionic lesions are also potentially difficult tointerpret.

The other major EDX method of determining the site ofnerve injury is by needle EMG. Conceptually, if the branch-ing order to various muscles under study is known, the clini-cian can determine that the nerve injury is between thebranches to the most distal normal muscle and the mostproximal abnormal muscle. There are, however, a number ofpotential problems with this approach. First, the branchingand innervation for muscles is not necessarily consistent fromone person to another. Sunderland40 has demonstrated agreat deal of variability in branching order to muscles in thelimbs, variability in the number of branches going to eachmuscle, and variability in which nerve or nerves supply eachmuscle. Thus, the typical branching scheme may not apply tothe patient being studied and consequently the lesion site canbe misconstrued.

Second, the problem of muscle trauma and associated needleEMG findings can be misleading. As mentioned earlier,direct muscle trauma can result in positive sharp waves andfibrillations for months or longer after injury.32 Practicallyspeaking, this can result in erroneously proximal lesion sites,or error in diagnosing more than one lesion. For example, inthe setting of humeral fracture with radial neuropathy, thetriceps not infrequently demonstrates fibrillation potentials,due to direct muscle trauma. However, a clinician could bemisled to localize the lesion to the axilla or higher rather thanspiral groove, if the triceps findings are not recognized tocome from direct muscle rather than nerve injury.

Third, the problem of partial lesions can make for misdiag-nosis to more distal sites. In partial ulnar nerve lesions at theelbow, for example, the forearm ulnar-innervated muscles areoften spared.4 This is thought at least partially to be due tosparing of the fascicles in the nerve that are preparing tobranch to the flexor digitorum profundus and the flexor carpi


ulnaris, i.e., they are in a relatively protected position. Thisfinding could lead a clinician to inadvertently localize thelesion distally to the distal forearm or wrist. Similarly, a lesioninvolving the median nerve in the arm (above the elbow) hasbeen reported to cause findings only in the anteriorinterosseous distribution.46 Intraneural topography needs tobe considered when making a diagnosis based on branch-ing.45

MECHANISMS OF RECOVERY

There are several possible mechanisms of recovery after trau-matic nerve injury; knowledge of these mechanisms, alongwith the type of nerve injury, allows estimation of the prob-able course of recovery.

For motor fibers, resolution of conduction block (in neu-rapraxic lesions), muscle fiber hypertrophy (in partiallesions), distal axonal sprouting of spared axons, and axonalregeneration from the site of injury, may contribute to recov-ery of strength.

Resolution of conduction block, whether based uponischemia or demyelination, is probably the first mechanismto promote recovery of strength after nerve injury.Improvement after a solely ischemic lesion is relatively quick.Demyelinating injuries take longer as remyelination over aninjured segment may take up to several months,13 dependingupon the severity of demyelination and the length of thedemyelinated segment.

In normal adults performing strengthening exercises, thereare generally two mechanisms of increasing force production:initial neural mechanisms followed by later muscle fiberhypertrophy. The initial neural mechanisms are thought toinvolve improved synchronization of motor unit firing,28,29

and they result in increased efficiency (defined as muscleforce per unit of electrical activity) in the absence of musclefiber changes. After several weeks, there is muscle fiber hyper-trophy, which results in further increases in strength. Inpatients with partial nerve lesions, it is unclear how muchneural changes alone (i.e., increased efficiency of firing) cancontribute to increased strength since there is loss of nervefibers. However, it is likely that working the existing musclefibers to fatigue in the setting of partial nerve injuries doesproduce enlargement of muscle fibers and consequentincreases in force production.

Partial axonotmesis of motor nerves also produces distalsprouting of motor fibers from intact axons. It has beenobserved that within 4 days after nerve injury, sprouts arestarting to form from intact axons, typically from distal nodesof Ranvier (nodal sprouts) or from nerve terminals (terminal

sprouts) near denervated muscle fibers.21 Partial recovery intwitch tension has been reported as early as 7-10 days postinjury,2 though electrophysiologic correlates (e.g., polyphasiclong duration motor units) usually take longer. Sometimes,when axonal regeneration occurs, those muscle fibers reinner-vated by distal sprouting become dually innervated, i.e., byboth the sprout and the newly regenerated fiber.19,21 It is notwell understood how multiple synapses are reduced.

Axonal regeneration contributes to recovery in both partialand complete axonotmesis and, with surgical approximation,neurotmesis. In complete axon loss lesions, this is the onlymechanism for muscle recovery. It is noted that in the 24-36hours after injury, the proximal nerve stump has started tosprout regenerating axons and these have started to penetratethe area of injury. The recovery that results from this processdepends upon the degree of injury, presence of scar forma-tion, approximation of the two nerve ends, and age of thepatient.

In relatively more minor axonotmetic lesions, in which theendoneurial tubes are preserved (i.e., Sunderland 2nd degreeinjuries), the axons can traverse the segment of injury in 8-15days and then regenerate along the distal nerve segment at arate of 1-5 mm/day,40 slightly faster for crush injuries thanfor sharp laceration, slightly faster for proximal than distalinjuries, and slightly faster for younger individuals.

In more severe axonotmetic lesions in which there is distor-tion of endoneurial tubes with or without perineurial disrup-tion (Sunderland 3rd and 4th degrees), prognosis for sponta-neous regrowth is worse. Extensive scarring reduces the speedat which regenerating axons can traverse the lesion and moreimportantly reduces the likelihood that they will ever reachtheir end organs. When regrowth occurs, it may also be mis-directed to the wrong end organ. In some of these cases, par-ticularly when a large neuroma is present, surgical interven-tion is required.

In complete neurotmesis (Sunderland 5th degree), axonalregrowth will usually not occur unless the nerve ends arefreed from scar and surgically re-approximated. After surgicalintervention, using either direct approximation or cablegrafting, nerve growth will often occur along the endoneurialtubes of the distal segments. Use of cable grafts (e.g., suralnerve graft) does not provide axons directly since these dieafter harvesting; the graft simply provides a pathway foraxonal regrowth to occur.34,47

In complete lesions, recovery of motor function will alsodepend upon integrity of the muscle when the axon reachesit. Muscles remain viable for reinnervation for 18-24 monthspost injury. However, past this time, due to fibrosis and atro-phy, motor axon regrowth makes little difference since muscle


fibers (the end organ), are no longer viable. For example, incomplete lower trunk brachial plexus lesions, recovery ofhand function is usually not expected no matter how goodthe surgical grafting might be; it simply takes too long foraxons to reach the muscle.

Recovery of sensory function is dependent upon differentmechanisms than motor recovery. There may be redistribu-tion of sensory distribution after an axonal injury, such thatintact fibers provide cutaneous sensation to a larger area thanpreviously.38,44 The mechanisms of axonal regeneration aresimilar to those mentioned above for motor axons. Animportant difference, however, is that one does not have endorgans that may degenerate after 18-24 months as muscledoes; hence sensory recovery may continue for a longer peri-od of time than motor recovery does.

ELECTRODIAGNOSTIC EVALUATION OF PROGNOSIS

Determining the pathophysiology of a peripheral nerve trau-matic injury can help with estimating prognosis. Thoseinjuries that are completely or largely neurapraxic have agood prognosis for recovery within a few months (usually upto 3 months postinjury). Resolution of ischemia and remyeli-nation should be complete by this time.

Mixed injuries typically have two or more phases of recovery(Figure 2). The neurapraxic component resolves quickly asabove and muscle fiber hypertrophy can provide additionalrecovery, but the axonal component is slower, since itdepends upon distal axonal sprouting and on axonal regener-ation from the site of the lesion. Thus patients usually expe-

rience a relatively rapid partial, but incomplete, recovery fol-lowed a slower further recovery. Sensory recovery may pro-ceed for a longer time than motor (Figure 3).

Partial axon loss lesions usually represent axonotmesis,though a partial neurotmesis (e.g., a laceration through partof the nerve) cannot always be excluded in such cases. Inaxonotmesis, recovery will depend upon axonal sproutingand regeneration. Thus there will be some early recovery fol-lowed possibly by a later recovery if or when regeneratingaxons reach their end organs. The amplitude of the CMAPprovides some guide to prognosis. In facial nerve lesions, ithas been demonstrated that patients with CMAP amplitudes30% or more of the other side have an excellent outcome,those with 10-30% have good, but not always complete,recovery, and those with less than 10% have a poor out-come.37

Complete axonotmesis and neurotmesis have the worst prog-nosis. Recovery depends solely upon axonal regenerationwhich may or may not occur, depending upon the degree ofinjury to the nerve. In many cases of complete axon loss, it isnot possible to know the degree of nerve injury except by sur-gical exploration with or without intraoperative recording, orlooking for evidence of early reinnervation after the lesion.

SUMMARY

As a consequence, it is often recommended that an EDXphysician should wait 2-4 months before looking for evi-dence of reinnervation in previously completely denervatedmuscles near the site of the lesion.23,47 Those lesions that have


Figure 2 Conceptual representation of mechanisms forincreases in strength after a mixed lesion to a peripheral nerve.The processes represented are not temporally distinct, but maymerge. Maximal recovery is usually achieved by 18-24 months.

Figure 3 Conceptual representation of mechanisms forimprovement in sensation after a mixed lesion to a peripheralnerve. The processes represented are not necessarily temporallydistinct, but may merge. Recovery may continue for longer than18 months since it is not dependent upon muscle viability.

some spontaneous recovery are usually treated conservativelysince operative repair is unlikely to improve upon naturalrecovery. Those with no evidence of axonal regrowth usuallyhave operative exploration with possible grafting.

REFERENCES

1. Brandstater ME, Fullerton M. Sensory nerve conduction studies incervical root lesions. Can J Neurol Sci 1983;10:152.

2. Brown MC, Holland RL, Hopkins WG. Motor nerve sprouting.Ann Rev Neurosci 1981;4:17-42.

3. Buchthal F. Fibrillations: clinical electrophysiology. In: Culp WJ,Ochoa J, editors. Abnormal nerves and muscle generators. New York:Oxford University Press; 1982. p 632-662.

4. Campbell WW, Pridgeon RM, Riaz G, Astruc J, Leahy M, CrosticEG. Sparing of the flexor carpi ulnaris in ulnar neuropathy at theelbow. Muscle Nerve 1989;12:965-967.

5. Cangiano A, Lutzemberger L, Nicotra L. Non-equivalence of impulseblockade and denervation in the production of membrance changesin rat skeletal muscle. J Physiol 1977;273:691-706.

6. Chaudhry V, Cornblath DR. Wallerian degeneration in humannerves: serial electrophysiological studies. Muscle Nerve1992;15:687-693.

7. Cosgrove JL, Vargo M, Reidy ME. A prospective study of peripheralnerve lesions occuring in traumatic brain-injured patients. Am J PhysMed Rehabil 1989;68:15-17.

8. Dillingham TR. Approach to trauma of peripheral nerves. In: 1998AAEM Course C: Electrodiagnosis in traumatic conditions.Rochester, MN: American Association of ElectrodiagnosticMedicine. p 7-12.

9. Dorfman LJ. Quantitative clinical electrophysiology in the evaluationof nerve injury and regeneration. Muscle Nerve 1990;13:822-828.

10. Dumitru D. Electrodiagnostic medicine. Philadelphia: Hanley &Belfus; 1995. p 341-384.

11. Erminio F, Buchthal F, Rosenfalck P. Motor unit territory and mus-cle fiber concentration in paresis due to peripheral nerve injury andanterior horn cell involvement. Neurology 1959;9:657-671.

12. Fisher MA. AAEM minimonograph #13: H reflexes and F waves:physiology and clinical indications. Muscle Nerve 1992;15:1223-1233.

13. Fowler TJ, Danta G, Gilliatt RW. Recovery of nerve conduction aftera pneumatic tourniquet: observations on the hind-limb of thebaboon. J Neurol Neurosurg Psychiatry 1972;35:638-647.

14. Garland DE, Bailey S. Undetected injuries in head-injured adults.Clin Orthop Relat Res 1981;155:162-165.

15. Gilliatt R, Hjorth RJ. Nerve conduction during Wallerian degenera-tion in the baboon. J Neurol Neurosurg Psychiatry 1972;35:335-341.

16. Gilliatt R, Taylor JC. Electrical changes following section of the facialnerve. Proc R Soc Med 1959;52:1080-1083.

17. Gilliatt R, Westgaard RH, Williams IR. Extrajunctional acetylcholinesensitivity of inactive muscle fibres in the baboon during prolongednerve pressure block. J Physiol 1978;280:499-514.

18. Gilliatt RW. Acute compression block. In: Sumner AJ, editor. Thephysiology of peripheral nerve disease. Philadelphia: WB Saunders;1980. p 287-315.

19. Guth L. Neuromuscular function after regeneration of interruptednerve fibers into partially denervated muscle. Exp Neurol1962;6:129-141.

20. Haymaker W, Woodhall B. Peripheral nerve injuries. Philadelphia:WB Saunders; 1953.

21. Hoffman H. Local reinnervation in partially denervated muscle: ahistophysiological study. Aust J Exp Biol Med Sci 1950;28:383.

22. Kimura J, Machida M, Ishida T, Yamada T, Rodnitzky RL, Kudo Y,Suzuki S. Relation between size of compound sensory or muscleaction potential, and length of nerve segment. Neurology1986;36:647-652.

23. Kline DG. Surgical repair of peripheral nerve injury. Muscle Nerve1990;13:843-852.

24. Kraft GH. Fibrillation amplitude and muscle atrophy followingperipheral nerve injury. Muscle Nerve 1990;13:814-821.

25. Mackinnon SE, Dellon AL. Surgery of the peripheral nerve. NewYork: Thieme Medical Publishers; 1988.

26. Massey JM, Sanders DB. Single-fiber EMG demonstrates reinnerva-tion dynamics after nerve injury. Neurology 1991;41:1150-1151.

27. Miller RG. AAEE minimonograph #28: Injury to peripheral motornerves. Muscle Nerve 1987;10:698-710.

28. Milner-Brown HS, Stein RB, Lee RG. Synchronization of humanmotor units: possible role of exercise and supraspinal reflexes.Electroencephalogr Clin Neurophysiol 1975;38:245-254.

29. Moritani T, deVries HK. Neural factors versus hypertrophy in thetime course of muscle strength gain. Am J Phys Med 1979;58:115-130.

30. Noble J, Munro CA, Prasad VS, Midha R. Analysis of upper andlower extremity peripheral nerve injuries in a population of patientswith multiple injuries. J Trauma 1998;45:116-122.

31. Ochoa J, Fowler TJ, Gilliatt RW. Anatomical changes in peripheralnerves compressed by a pneumatic tourniquet. J Anat 1972;113:433-455.

32. Partanen JV, Danner R. Fibrillation potentials after muscle injury inhumans. Muscle Nerve 1982;5:S70-S73.

33. Rasminsky M, Sears TA. Internodal conduction in undissecteddemyelinated nerve fibres. J Physiol 1972;227:323-350.

34. Seddon HJ. Nerve grafting. J Bone Joint Surg Br 1963;45:447-461.35. Seddon HJ. Surgical disorders of the peripheral nerves, 2nd edition.

New York: Churchill-Livingstone; 1975. p 21-23.36. Selecki BR, Ring IT, Simpson DA, Vanderfield GK, Sewell MF.

Trauma to the central and peripheral nervous systems: part II, a sta-tistical profile of surgical treatment in New South Wales, 1977. AustN Z J Surg 1982;52:111-116.

37. Sillman JS, Niparko JK, Lee SS, Kileny PR. Prognostic value ofevoked and standard electromyography in acute facial paralysis.Otolaryngol Head Neck Surg 1992;107:377-381.

38. Speidel CC. Studies of living nerves: growth adjustments of cuta-neous terminal arborization. J Comp Neurol 1942;76:57-73.

39. Stone L, Keenan MA. Peripheral nerve injuries in the adult with trau-matic brain injury. Clin Orthop Relat Res 1988;233:136-144.


40. Sunderland S. Nerves and nerve injuries, 2nd edition. New York:Churchill-Livingstone; 1978. p 133-138.

41. Tackmann W, Radu EW. Observations of the application of electro-physiological methods in the diagnosis of cervical root compressions.Eur Neurol 1983;22:397-404.

42. Thesleff S. Trophic functions of the neuron. II. Denervation and reg-ulation of muscle. Physiological effects of denervation of muscle. AnnN Y Acad Sci 1974;228:89-103.

43. Trojaborg W. Early electrophysiological changes in conduction block.Muscle Nerve 1978;1:400-403.

44. Weddell G, Glees P. The early stages in the degeneration of cutaneousnerve fibers. J Anat 1941;76:65-93.

45. Wertsch JJ, Oswald TA, Roberts MM. Role of intraneural topogra-phy in diagnosis and localization in electrodiagnostic medicine. PhysMed Rehabil Clin N Am 1994;5:465-475.

46. Wertsch JJ, Sanger JR, Matloub HS. Pseudo-anterior interosseousnerve syndrome. Muscle Nerve 1985;8:68-70.

47. Wood MB. Surgical approach to peripheral nervous system trauma.In: 1998 AAEM Course C: Electrodiagnosis in traumatic conditions.American Association of Electrodiagnostic Medicine: Rochester,MN; p 27-36.

48. Yuska MA, Wilbourn AJ. Incidence of fibrillation potentials in “pure”conduction block mononeuropathies. Muscle Nerve 1998;21:1572.


12 AANEM Course

Acknowledgements

This work was supported by the National Institute of Arthritis andMusculoskeletal Skin Disease (NIAMS) (P60 AR48093) and theUniversity of Washington’s Multidisciplinary Clinical ResearchCenter.

INTRODUCTION

The first reports of using magnetic resonance imaging (MRI) toimage the median nerve in the carpal tunnel appeared in themid-1980s.40,50 Much progress has been made in the interven-ing two decades and currently, even on a standard 1.5 Tesla MRsystem, radiologists can now routinely obtain high-resolutionimages of peripheral nerves.1,7,10,14,15,19,20,26,28,36,37,51 The dis-semination of MRI systems using even higher field strength(>3 Tesla) promises further improvement in imaging.6,18,41

These advances and refinements in MRI have allowed inves-tigators to visualize several types of peripheral nerve patholo-gies. It is now possible to see morphological and signalchanges in nerve entrapment disorders, cervical radiculopa-thy, traumatically injured and surgically repaired peripheralnerves, and peripheral nerve tumors. The ability of MRI to

noninvasively visualize anatomic detail may lead to a betterunderstanding and diagnosis of nerve injury in general.

Why bother to image the carpal tunnel if nerve conductionstudies are already widely available? While electrodiagnostic(EDX) studies are useful and widely employed for patientswith suspected carpal tunnel syndrome (CTS), there is roomfor improvement. In patients with suspected CTS, there isvarying evidence as to how well they correlate with symptomseverity and response to treatment. For example, Prignac andcolleagues found no relationship between EDX severity andwell-accepted measures of clinical disease, the Katz-StirratHand Pain Diagram and the Carpal Tunnel SyndromeAssessment Questionnaire (CTSAQ).45 In contrast,Dennerlein and colleagues found that distal motor and sen-sory latencies did correspond to preoperative symptom sever-ity.16 There is also conflicting evidence regarding EDX stud-ies ability to predict outcome. Several investigators havefound a correlation between various EDX parameters andoutcomes after surgery.16,24,4,8 For example, Bland found thatpatients with middle grade EDX abnormalities did the bestafter surgery, compared to those with either very severe orvery mild findings. Dennerlein and colleagues showed thatdistal motor latency correlated with outcome.16 Others have

13

Peripheral Nerve Magnetic ResonanceImaging: the Median Nerve in Carpal

Tunnel Syndrome

Jeffrey G. Jarvik MD, MPH

ProfessorDepartments of Radiology, Neurosurgery, and Health Services

University of WashingtonSeattle, Washington

not found such a relationship.54,3,35,21 In the only largerandomized controlled trial of surgery versus conservativetherapy for CTS, Gerritsen and colleagues found that EDXwas not a significant predictor of which patients wouldimprove without surgery.22 Thus, there may be a useful rolefor MRI in further clarifying which patients would benefitfrom surgery for CTS.

MAGNETIC RESONANCE TECHNIQUE

High-resolution MRI of peripheral nerves requires the use ofsurface coils to increase signal-to-noise ratio. While thisauthor uses a custom-designed phased-array wrist coil, com-mercially available coils provide excellent images as well.Higher field strength systems (> 1.5 Tesla) also provide bettersignal-to-noise ratio, although their use in the peripheralnervous system has been limited to date.

This author’s group uses axial T1- and T2-weighted fast shorttau inversion recovery (STIR) images in an attempt to imagein a plane perpendicular to the long axis of the nerve. TheT1-weighted images are best for demonstrating normalanatomy. Fat is bright, muscle and nerves are intermediatesignal, and tendons are dark. The natural contrast providedby the fat that surrounds nerves, vessels, and muscle allowsready identification of these structures. The fast STIR imagessuppress fat signal, allowing greater conspicuity of structuresthat have long T2 relaxation. This includes pathologicalperipheral nerves, acutely denervated muscle, inflammatorytissue surrounding nerves and tendons, and mass lesions suchas ganglions. Some groups use T2-weighted sequences with afrequency-selective fat saturation pulse, but because of thedifficulty in obtaining uniform fat saturation due to magnet-ic field inhomogeneity, the author prefers to use the STIRsequence. Saturation pulses are also applied superior andinferior to the imaging plane to minimize vascular flow arti-fact.

This author’s current imaging protocol is as follows:

(1) coronal T1-weighted spin echo: (TR=600, TE=min-imum, 18 cm field of view (FOV), 256x192 matrix,4 mm slice thickness);

(2) axial T1-weighted spin echo: (TR=450, TE=mini-mum, 10 cm FOV, 256x256 matrix, 4 mm slicethickness, 1 mm skip, 5.5 minutes);

(3) axial fast STIR: (TR=3650, TE=54, TI=160, echotrain length = 6, 10 cm FOV, 256x224 matrix, 4mm slice thickness, 1 mm skip, 5.25 minutes).

In postoperative patients, the author repeats the T1-weightedaxial images with fat saturation after giving contrast to detectscar formation.

Chappell and colleagues pointed out a potential pitfall ininterpreting MRIs of peripheral nerves: the magic angleeffect.11 When highly ordered structures, such as nerves andtendons, are aligned at particular angles to the main magnet-ic field, the signal within these structures increases on T2-weighted images with moderate or short echo times (TE).This “magic angle” is achieved when the following formula issatisfied: 3cos2q-1=θ. Thus, when the nerve is at 55 degreesto the main magnetic field, the signal will be hyperintense onT2-weighted images. The authors point out that this condi-tion is routinely satisfied when imaging the brachial plexus,but may also be present on studies of the carpal tunnel.Careful attention must be given when positioning patientsfor these studies, and interpretation of signal intensity musttake into account the magic angle phenomenon.

ANATOMY

Peripheral nerves have three distinct layers: endoneurium,perineurium, and epineurium. Endoneurium consists of con-nective tissue and extracellular fluid that surrounds the axonsand their Schwann cells. Perineurium is a sheath that bundlesfibers together to form a fascicle. The epineurium surroundsthe multiple fascicles to form the nerve.23

Magnetic resonance imaging can reliably identify the mediannerve within the carpal tunnel and characterize its shape andsignal intensity.30 Frequently, MRI can also identify individ-ual fascicles within the epineurium. Cross-sectional (or axial)images best demonstrate these features.

In addition to the median nerve itself, MRI can delineate theanatomy of adjacent structures within and around the carpaltunnel. The median nerve is usually superficial or volar to theflexor digitorum superficialis tendons (usually the latter)(Figure 1) but may also be interposed between the flexor dig-itorum superficialis tendons (Figure 2a) and/or between theflexor pollicis longus and flexor digitorum superficialis ten-dons (Figure 2b).55

Guyon’s canal, located at the ulnar side of the carpal tunnel,contains the ulnar nerve, artery and vein. Ulnar entrapmentat the wrist may occur at Guyon’s canal, especially in patientswith hook of hamate fractures.

The MRI findings in patients with CTS include flattening ofthe median nerve within the carpal tunnel (Figure 3), highsignal of the nerve on T2-weighted images (Figure 4),

14 Peripheral Nerve Magnetic Resonance Imaging AANEM Course

enlargement of the nerve either proximal or distal to thepoint of maximal compression (Figure 5), bowing of the flex-or retinaculum, and thickening with increased signal of theflexor tendon sheaths (Figure 6), and deep palmar bursa(Figure 7).

Configurational changes in the nerve are probably the mostreliable imaging finding. Britz and colleagues9 described aflattening ratio that is frequently used to measure nerve con-figuration changes. It is the ratio of the major to minor axesof the median nerve at the level of the hamate bone. Theswelling ratio is also used, which is the ratio of the nervecross-sectional area at the level of the pisiform compared withthe cross-sectional area at the level of the distal radioulnarjoint.32

The median nerve may occasionally divide proximal to thecarpal tunnel and result in a bifid median nerve.27,46 This rel-atively uncommon anatomic variant (approximately 3% ofpatients) is important to recognize when using size criteria todetermine nerve abnormality. The bifid nerve may not meas-ure as large as an undivided nerve, leading to decreased sen-sitivity of imaging.

DIAGNOSTIC PERFORMANCE

Increased T2 signal appears to be a consistent finding,although as pointed out above, must be interpreted with cau-tion due to the magic angle effect as well as other factors that

can cause variability in signal. For example, shading of imagesbecause of surface coils can make superficial structures appearmore hyperintense than deep structures. Moreover, the nervewill normally appear hyperintense compared with the rela-tively black signal of the adjacent tendons.39 A useful internalstandard is the thenar muscles, which compared with themedian nerve should be iso- to slightly hyperintense (Figure8). The nerve signal frequently returns to near normal at the


Figure 1 Axial T1-weighted image at the level of the pisiform(P) demonstrates a normal position of the median nerve (MN-arrow), volar to the flexor tendons.

Figure 2a Axial T1-weighted image at the level of the hamatehook shows the median nerve (asterisk) interposed nervebetween flexor digitorum tendons.

Figure 2b Axial T1-weighted image at the level of the hamatehook in a dif ferent patient demonstrate the median nerveinterposed between the flexor pollicis longus and flexor digitorumtendons.

site of maximal compression, perhaps due to a loss of fluid.Quantification of nerve signal is still not clinically practicaland evaluation relies on the judgement of the observer.

In addition to spatial variation in nerve signal, there is alsotemporal variation. Experimental evidence by a number ofauthors have demonstrated T2 prolongation in injurednerves.2,13 with T2 signal peaking around day 14 after injuryand normalizing by 2 months. Moreover, Cudlip and col-leagues demonstrated a strong temporal correlation betweenT2 signal and functional assessment of an animal’s gait.

Swelling of the median nerve proximal to the carpal tunnelseems to be a consistent finding among several studies as animportant discriminator between patients with CTS andthose without. Monagle found that cross-sectional area of thenerve was 50% larger in patients compared to asymptomaticvolunteers.41 Keberle found that a swelling ratio greater thanor equal to 1.3 and a flattening ratio of greater than or equalto 3.4 had a 100% sensitivity and negative predictive value,with a specificity of 68% and positive predictive value of75%.32 In their series of 37 patients, Wu and colleaguesfound proximal nerve enlargement to have the strongest asso-


Figure 3 Axial T2-weighted short tau inversion recovery imageshows flattened and hyperintense median nerve (asterisk).

Figure 4 Axial T2-weighted short tau inversion recovery imageshows a median nerve that is extremely hyperintense (asterisk).

Figure 5 Axial T2-weighted short tau inversion recovery imagedemonstrates an enlarged and hyperintense median nerve. Thefascicles are prominent as well.

Figure 6 Axial T2-weighted short tau inversion recovery image.There is marked thickening and high signal of the flexor tendoninterspace and deep palmar bursa (asterisks). The median nerveis normal in signal.

ciation with recurrent CTS after surgery.52 Prominence of thefascicles frequently accompanies nerve swelling; this findingis best demonstrated on T2-weighted images.

A less common finding is bowing of the flexor retinaculum.This is presumably caused by increased pressure within thecarpal tunnel. Britz and colleagues described drawing a linefrom the tip of the hamate hook to the trapezial tubercle, andconsidered that abnormal bowing was present if the flexorretinaculum extended more than 2 mm anterior to this line.9

Other less commonly cited findings include high signal andthickening of the flexor tendon sheath interspaces as well asthe palmar bursa which may reflect an inflammatory

tenosynovitis/bursitis. Thenar muscle denervation is rareunless symptoms are severe. With acute denervation, there ishigh signal on STIR images in the thenar muscle that maypersist for months. Chronic denervation results in fatty infil-tration of the muscle as well as atrophy.1

Studies that have examined the diagnostic accuracy of MRIfor patients with CTS have generally suffered from smallsample sizes and various biases, such as spectrum bias, thattend to inflate estimates of accuracy.9,17,20,25,33,34,42,43,44,47,48,

49,51,53,55 The author’s experience, the most reliable imagingfindings were high T2-signal, configurational changes, andan overall global rating of nerve abnormality.30 It has alsobeen found that the length of signal abnormality on T2-weighted images correlates with the degree of median nerveconduction slowing.30

Several authors have failed to find a strong correlationbetween imaging findings and EDX studies. Deryani andcolleagues found no correlation between median nerve diam-eter or flexor retinaculum bowing and median sensory nerveaction potential distal peak latency or amplitude.17 Theauthor’s group also failed to find a strong correlation betweenimaging and EDX studies.30

POTENTIAL IMAGING INDICATIONS

Magnetic resonance imaging has the potential to offer newinsight into the diagnosis and management of patients withhand and wrist neurological symptoms. Unlike EDX studies,MRI directly visualizes the median nerve and can detectabnormalities of both configuration (nerve compression) aswell as signal (possibly indicating intraneural edema anddemyelination).20,29-31 Either or both of these findings havethe potential to be better predictors of patient outcomes thanEDX studies.

Can high-resolution MRI of the median nerve identifypatients for whom early surgery might be more efficaciousthan conservative therapy? There are currently no definitivestudies that can answer this question. Cudlip and colleagues,in their series of 30 patients, showed that low T2 signal in themedian nerve was associated with worse outcome,12 althoughthey only had one patient with a poor outcome followingsurgery. The author’s group is conducting a trial that willhopefully shed some light on this issue.38 As part of a treat-ment trial comparing surgery with nonsurgical therapy,patients are being recruited with mild to moderate CTS fromboth primary care as well as subspecialty offices (orthopedicsurgery, neurosurgery, physical medicine and rehabilitation,and neurology). Subjects undergo high resolution MRI of thecarpal tunnel prior to randomization, and are followed for a


Figure 7 Axial T2-weighted short tau inversion recovery image.There is high signal in the deep palmar bursa (arrows). There isalso high signal in the median nerve.

Figure 8 Axial T2-weighted short tau inversion recovery image.The median nerve (arrow) is isointense to muscle (asterisk).

year. Both subjects and physicians are blinded to the resultsof the MRI. This will allow the group to determine how wellMRI predicts change in symptoms and functional status, andmore importantly, if it can preoperatively help to determinethe benefit of surgery.

Magnetic resonance imaging may prove useful as a postoper-ative evaluative tool. Wu and colleagues found that contin-ued proximal enlargement of the median nerve and evidenceof tenosynovitis were associated with recurrent CTS.52 Whilepreliminary in nature, this potential use as a problem-solvingtool shows promise.

What does the future hold? As high field-strength MRI sys-tems proliferate, MRI will better delineate anatomic defini-tion of the median nerve and its surrounding structures.18,41

Farooki and colleagues, using a clinical 8 Tesla magnet, wereable to resolve tertiary tendon fiber bundles, achieving a res-olution of approximately 200 microns. Bilgen and colleagues,on a 9.4 Tesla system, identified structures at the subfascicu-lar level and were able to quantify diffusion characteristicsalong the nerve. The role of diffusion and perfusion charac-teristics of peripheral nerves, as well as functional activationstudies of the spinal cord and brain will also surely be inves-tigated.

Molecular imaging is a new field that uses a variety of tech-nologies, such as micro-positive emission tomography, opti-cal imaging, microcomputed tomography, and MRI to mon-itor fundamental cellular events in living subjects. Usingmolecular agents, it promises unparalleled specificity.Bendszus and Stoll imaged macrophages in vivo using super-paramagnetic iron oxide (SPIO) particles.5 Macrophagesconcentrate SPIO, and thus processes that result inmacrophage migration and activation will demonstrate signalalteration. Using a nerve crush injury in a rat model, theauthors were able to observe accumulation of local iron andhence signal loss, at the site of crushed nerves. This signalchange peaked at day 4 and then normalized by 14 days afterinjury. As the field of molecular imaging matures, other trac-ers will undoubtedly be developed that will be able to identi-fy specific characteristics of nerve injury and healing.

SUMMARY

While EDX studies will probably remain the primary diag-nostic test for patients with suspected CTS, MRI holds greatpromise and may well play an expanded role as the technol-ogy matures.

REFERENCES

1. Aagaard B, Maravilla KR, Kliot M. MR neurography. MR imagingof peripheral nerves. Magn Reson Imaging Clin N Am 1998;6:179-194.

2. Aagaard BD, Lazar DA, Lankerovich L, Andrus K, Hayes CE,Maravilla K, Kliot M. High-resolution magnetic resonance imagingis a noninvasive method of observing injury and recovery in theperipheral nervous system. Neurosurgery 2003;53:199-203.

3. al-Qattan MM, Bowen V, Manktelow RT. Factors associated withpoor outcome following primary carpal tunnel release in non-diabet-ic patients. J Hand Surg Br 1994;19:622-625.

4. Atroshi I, Johnsson R, Ornstein E. Patient satisfaction and return towork after endoscopic carpal tunnel surgery. J Hand Surg Am1998;23:58-65.

5. Bendszus M, Stoll G. Caught in the act: in vivo mapping ofmacrophage infiltration in nerve injury by magnetic resonance imag-ing. J Neurosci 2003;23:10892-10896.

6. Bilgen M, Heddings A, Al-Hafez B, Hasan W, McIff T, Toby B,Nudo R, Brooks WM. Microneurography of human median nerve.J Magn Reson Imaging 2005;21:826-830.

7. Blake LC, Robertson WD, Hayes CE. Sacral plexus: optimal imag-ing planes for MR assessment. Radiology 1996;199:767-772.

8. Bland J. Do nerve conduction studies predict the outcome of carpaltunnel decompression? Muscle Nerve 2001;24:935-940.

9. Britz GW, Haynor DR, Kuntz C, Goodkin R, Gitter A, Kliot M.Carpal tunnel syndrome: correlation of magnetic resonance imaging,clinical, electrodiagnostic, and intraoperative findings. Neurosurgery1995;37:1097-1103.

10. Britz GW, Haynor DR, Kuntz C, Goodkin R, Gitter A, Maravilla K,Kliot M. Ulnar nerve entrapment at the elbow: correlation of mag-netic resonance imaging, clinical, electrodiagnostic, and intraopera-tive findings. Neurosurgery 1996;38:458-465.

11. Chappell KE, Robson MD, Stonebridge-Foster A, Glover A, AllsopJM, Williams AD, Herlihy AH, Moss J, Gishen P, Bydder GM.Magic angle effects in MR neurography. AJNR Am J Neuroradiol2004;25:431-440.

12. Cudlip SA, Howe FA, Clifton A, Schwartz MS, Bell BA. Magneticresonance neurography studies of the median nerve before and aftercarpal tunnel decompression. J Neurosurg 2002;96:1046-1051.

13. Cudlip SA, Howe FA, Griffiths JR, Bell BA. Magnetic resonance neu-rography of peripheral nerve following experimental crush injury, andcorrelation with functional deficit. J Neurosurg 2002;96:755-759.

14. Dailey A, Tsuruda JS, Goodkin R, Haynor DR, Filler AG, Hayes CE,Maravilla KR, Kliot M. Magnetic resonance neurography for cervicalradiculopathy: a preliminary report. Neurosurgery 1996;38:488-492.

15. Dailey AT, Tsuruda JS, Filler AG, Maravilla KR, Goodkin R, KliotM. Magnetic resonance neurography of peripheral nerve degenera-tion and regeneration. Lancet 1997;350:1221-1222.

16. Dennerlein JT, Soumekh FS, Fossel AH, Amick BC 3rd, Keller RB,Katz JN. Longer distal motor latency predicts better outcomes ofcarpal tunnel release. J Occup Environ Med 2002;44:176-183.

17. Deryani E, Aki S, Muslumanoglu L, Rozanes I. MR imaging andelectrophysiological evaluation in carpal tunnel syndrome. YonseiMed J 2003;44:27-32.

18. Farooki S, Ashman CJ, Yu JS, Abduljalil A, Chakeres D. In vivo high-resolution MR imaging of the carpal tunnel at 8.0 tesla. SkeletalRadiol 2002;31:445-450.


19. Filler AG, Howe FA, Hayes CE, Kliot M, Winn HR, Bell BA,Griffiths JR, Tsuruda JS. Magnetic resonance neurography. Lancet1993;341:659-661.

20. Filler AG, Kliot M, Howe FA, Hayes CE, Saunders DE, Goodkin R,Bell BA, Winn HR, Griffiths JR, Tsuruda JS. Application of magnet-ic resonance neurography in the evaluation of patients with peripher-al nerve pathology. J Neurosurg 1996;85:299-309.

21. Gerritsen A, Korthals-de Bos I, PM L, de Vet H, Scholten R, BouterL. Splinting for carpal tunnel syndrome: prognostic indicators of suc-cess. J Neurol Neurosurg Psychiatry 2003;74:1342-1344.

22. Gerritsen AA, Scholten RJ, Assendelft WJ, Kuiper H, de Vet HC,Bouter LM. Splinting or surgery for carpal tunnel syndrome? Designof a randomized controlled trial [ISRCTN18853827]. BMC Neurol2001;1:8.

23. Grant GA, Goodkin R, Kliot M. Evaluation and surgical manage-ment of peripheral nerve problems. Neurosurgery 1999;44:825-839.

24. Higgs PE, Edwards DF, Martin DS, Weeks PM. Relation of preoper-ative nerve-conduction values to outcome in workers with surgicallytreated carpal tunnel syndrome. J Hand Surg Am 1997;22:216-221.

25. Horch RE, Allmann KH, Laubenberger J, Langer M, Stark GB.Median nerve compression can be detected by magnetic resonanceimaging of the carpal tunnel. Neurosurgery 1997;41:76-82.

26. Howe FA, Filler AG, Bill BA, Griffiths JR. Magnetic resonance neu-rography. J Magn Reson Imag 1992;28:328-338.

27. Iannicelli E, Chianta GA, Salvini V, Almberger M, Monacelli G,Passariello R. Evaluation of bifid median nerve with sonography andMR imaging. J Ultrasound Med 2000;19:481-485.

28. Jarvik J, Yuen E, Haynor DR, Bradley C, Fulton-Kehoe D, Smith-Weller T, Franzblau A, Kliot M, Franklin G. Magnetic resonanceneurography (MRN) in carpal tunnel syndrome (CTS). AmericanSociety of Neuroradiology (ASNR), 1999; San Diego.

29. Jarvik JG, Kliot M, Maravilla KR. MR nerve imaging of the wrist andhand. Hand Clin 2000;16:13-24, vii.

30. Jarvik JG, Yuen E, Haynor DR, Bradley CM, Fulton-Kehoe D,Smith-Weller T, Wu R, Kliot M, Kraft G, Wang L, Erlich V, HeagertyPJ, Franklin GM. MR nerve imaging in a prospective cohort ofpatients with suspected carpal tunnel syndrome. Neurology2002;58:1597-1602.

31. Jarvik JG, Yuen E, Kliot M. Diagnosis of carpal tunnel syndrome:electrodiagnostic and MR imaging evaluation. Neuroimaging Clin NAm 2004;14:93-102.

32. Keberle M, Jenett M, Kenn W, Reiners K, Peter M, Haerten R, HahnD. Technical advances in ultrasound and MR imaging of carpal tun-nel syndrome. Eur Radiol 2000;10:1043-1050.

33. Kleindienst A, Hamm B, Hildebrandt G, Klug N. Diagnosis andstaging of carpal tunnel syndrome: comparison of magnetic reso-nance imaging and intra-operative findings. Acta Neurochir1996;138:228-233.

34. Kleindienst A, Hamm B, Lanksch WR. Carpal tunnel syndrome:staging of median nerve compression by MR imaging. J Magn ResonImaging 1998;8:1119-1125.

35. Kouyoumdjian JA, Morita MP, Molina AF, Zanetta DM, Sato AK,Rocha CE, Fasanella CC. Long-term outcomes of symptomatic elec-trodiagnosed carpal tunnel syndrome. Arq Neuropsiquiatr2003;61:194-198.

36. Kuntz C 4th, Blake L, Britz G, Filler A, Hayes CE, Goodkin R,Tsuruda J, Maravilla K, Kliot M. Magnetic resonance neurography ofperipheral nerve lesions in the lower extremity. Neurosurgery1996;39:750-756.

37. Maravilla KR, Bowen BC. Imaging of the peripheral nervous system:evaluation of peripheral neuropathy and plexopathy. AJNR Am JNeuroradiol 1998;19:1011-1023.

38. Martin BI, Levenson LM, Hollingworth W, Kliot M, Heagerty PJ,Turner JA, Jarvik JG. Randomized clinical trial of surgery versus con-servative therapy for carpal tunnel syndrome [ISRCTN84286481].BMC Musculoskelet Disord 2005;6:2.

39. Mesgarzadeh M, Schneck CD, Bonakdarpour A. Carpal tunnel: MRimaging. Part I. Normal anatomy. Radiology 1989;171:743-748.

40. Middleton WD, Kneeland JB, Kellman GM, Cates JD, Sanger JR,Jesmanowicz A, Froncisz W, Hyde JS. MR imaging of the carpal tun-nel: normal anatomy and preliminary findings in the carpal tunnelsyndrome. AJR Am J Roentgenol 1987;148:307-316.

41. Monagle K, Dai G, Chu A, Burnham RS, Snyder RE. QuantitativeMR imaging of carpal tunnel syndrome. AJR Am J Roentgenol1999;172:1581-1586.

42. Murphy RX Jr, Chernofsky MA, Osborne MA, Wolson AH.Magnetic resonance imaging in the evaluation of persistent carpaltunnel syndrome. J Hand Surg Am 1993;18:113-120.

43. Oneson SR, Scales LM, Erickson SJ, Timins ME. MR imaging of thepainful wrist. Radiographics 1996;16:997-1008.

44. Pierre-Jerome C, Bekkelund SI, Husby G, Mellgren SI, Osteaux M,Nordstrom R. MRI of anatomical variants of the wrist in women.Surg Radiol Anat 1996;18:37-41.

45. Priganc VW, Henry SM. The relationship among five commoncarpal tunnel syndrome tests and the severity of carpal tunnel syn-drome. J Hand Ther 2003;16:225-236.

46. Propeck T, Quinn TJ, Jacobson JA, Paulino AF, Habra G, Darian VB.Sonography and MR imaging of bifid median nerve with anatomicand histologic correlation. AJR Am J Roentgenol 2000;175:1721-1725.

47. Radack DM, Schweitzer ME, Taras J. Carpal tunnel syndrome: arethe MR findings a result of population selection bias? AJR Am JRoentgenol 1997;169:1649-1653.

48. Soccetti A, Raffaelli P, Giovagnoni A, Ercolani P, Mercante O,Pelliccioni G. MR imaging in the diagnosis of carpal tunnel syn-drome. Ital J Orthop Traumatol 1992;18:123-127.

49. Timins ME, O’Connell SE, Erickson SJ, Oneson SR. MR imagingof the wrist: normal findings that may simulate disease.Radiographics 1996;16:987-995.

50. Weiss KL, Beltran J, Shamam OM, Stilla RF, Levey M. High-fieldMR surface-coil imaging of the hand and wrist. Part I. Normal anato-my. Radiology 1986;160:143-146.

51. West GA, Haynor DR, Goodkin R, Tsuruda JS, Bronstein AD, KraftG, Winter T, Kliot M. Magnetic resonance imaging signal changes indenervated muscles after peripheral nerve injury. Neurosurgery1994;35:1077-1085.

52. Wu HT, Schweitzer ME, Culp RW. Potential MR signs of recurrentcarpal tunnel syndrome: initial experience. J Comput Assist Tomogr2004;28:860-864.

53. Yoshioka S, Okuda Y, Tamai K, Hirasawa Y, Koda Y. Changes incarpal tunnel shape during wrist joint motion. MRI evaluation ofnormal volunteers. J Hand Surg Br 1993;18:620-623.

54. Yu GZ, Firrell JC, Tsai TM. Pre-operative factors and treatment out-come following carpal tunnel release. J Hand Surg Br 1992;17:646-650.

55. Zeiss J, Skie M, Ebraheim N, Jackson WT. Anatomic relationsbetween the median nerve and flexor tendons in the carpal tunnel:MR evaluation in normal volunteers. AJR Am J Roentgenol1989;153:533-536.


20 AANEM Course

INTRODUCTION

Three large categories of peripheral neuropathies may behelped by operative intervention—nerve injuries, entrap-ments, and tumors. There are a number of excellent booksavailable concerning surgical nerve lesions.2,7,8,22,27,31 Thereare, of course, also occasions where the surgeon can help theelectrodiagnostic physician in working up and managing amedical neuropathy.6 This is usually accomplished by nerveand/or muscle biopsy, and under exceptional circumstancesoperative exploration for suspected (but not proven) neuralpathology.

NERVE INJURIES AND TRANSECTIONS

There are two categories of mechanical nerve injury. Thesmallest category is transection or partial laceration of thenerve or plexus element. However, physicians most often pre-sume the cause of nerve injury to be from transaction or par-tial laceration. However, although an important category toconsider, these injuries represent approximately 30% of allserious nerve injuries. There are two categories of lacerationsand transactions—sharp, and dull or blunt, and thus contu-sive. This is far from a mechanistic stratification becausethese two categories have a different management algorithm.

SHARP TRANSECTION OR LACERATION

Sharp transection or laceration injuries usually involve aknife, glass, or a sharp metal edge. The forces necessary todivide the nerve are minimal and thus, force to the stumps isminimal. As a result, these are excellent cases for relativelyacute nerve repair when performed within the first 72 hourspostinjury.16 This approach offers many advantages becausethe nerve can be repaired with the expectation of a reasonableresult at the same time as the repair of associated injuries toadjacent structures such as vessels and tendons.2 Such earlyrepair, when possible, has the best outcome—even for thebrachial plexus (Table 1). If it is not repaired relatively acute-ly, the stumps retract and the need for grafts to bridge the gapincreases. Positive outcomes with grafting are possible butnot as good as with end-to-end repair.

BLUNT TRANSECTION

Blunt transection injuries are usually due to fan and propellerblades, auto metal, or other blunt objects. The blunt injury isbest managed secondarily, after a delay of several weeks.27,31

This is because the forces of transection are large and blunt,and there is a degree of proximal and distal injury which isunpredictable acutely. After several weeks, the extent of dam-

21

Surgical Management of Nerve Injuries

David G. Kline, MD

Boyd Professor and HeadDepartment of Neurosurgery

Louisiana State University Medical CenterNew Orleans, Louisiana

age can be both palpated and visualized, permitting resectionback to healthy tissue (coapting good to good rather than badto bad). If blunt transection is encountered acutely, it is bestto tack the stumps down with nonresorbable sutures to adja-cent fascial planes. This maintains length so that later end-to-end repair rather than graft repair can be performed.

Despite a penetrating mechanism, nerves can be displaced,contused, stretched, or sometimes partially divided, leadingto neuromas or lesions in continuity. It is best to wait 2 to 3months to evaluate whether surgery is necessary. After 2 to 3months, operative nerve action potential (NAP) recordingscan be used to guide decisions about resection and repair,partial resection and split repair, or simple neurolysis wherethe nerve is cleared 360 degrees around, over a length, andfreed up from surrounding scar or other injury.

OTHER INDICATIONS FOR ACUTE OPERATION

There are some other urgent indications for operating onnerve(s) acutely. The most obvious is a blood clot, whereacute compression of the nerve (unless ameliorated) can leadto either long-standing or irreversible neuropathy.22 Forexample, not all pelvic clots associated with anticoagulantssufficiently resolve spontaneously to permit recovery of

femoral nerve function. It is even less certain whether subpec-toral, axillary, popliteal, and subgluteal clots will resolve with-out acute surgical intervention.

Another relatively acute indication for surgery is a pseudoa-neurysm or atrioventricular fistula usually due to a penetrat-ing wound and affecting axillary or popliteal artery.31 Thepseudoaneurysm or fistula needs resection and a neurolysisperformed of the affected cord and nerve elements or tibialnerve.

Sometimes when a nerve in an extremity is badly swollen bytrauma, it needs exposure and neurolysis acutely, especially ifit is near a potential area of tightness or entrapment. In manyof these cases fasciotomies will be necessary, but under somecircumstances neurolysis of the nerves is also be needed.22 Agood example of this is Volkman’s ischemic contracture. Thisusually is due to distal humeral fracture and often due toelbow dislocation and brachial artery compromise. Here, inaddition to volar and dorsal forearm fasciotomies, it may beadvisable to expose the median, radial, and sometimes eventhe ulnar nerve.

LESIONS OR NEUROMAS IN CONTINUITY

Lesions or neuromas in continuity are the largest and mostdifficult category of injury to manage and therefore requirespecial attention. The mechanisms of injury are usually bluntand associated with contusion and stretch involving nerve orplexus elements. Classic, but by no means exclusive, injuriesinclude fractures and gunshot wounds (GSWs) involvinglimb, neck, shoulder, or pelvis. A large category which can beassociated with or without fracture(s) is stretch-contusion tothe plexus with or without avulsion of nerve roots. As seenunder the transection-laceration category, these mechanismscan occasionally lead to either partial or complete division ofthe element. More often than not, lesions in continuity resultfrom stretch/contusion either at a supraclavicular or infra-clavicular level.

An important first step is a careful clinical examination withgrading of all muscles and sensation, especially that of thehand or foot. An early needle electromyography (EMG)study may be indicated if the clinician suspects prior injury,entrapment, or disease. It may also be helpful to conductearly nerve conduction studies centered on sites commonlyinvolved by potential neurapraxic injuries.14 Usually withmore serious injuries, the initial needle EMG study is per-formed at approximately 3 weeks postinjury because the

22 Surgical Management of Nerve Injuries AANEM Course

Table 1 Surgical outcome in 71 brachial plexus lacerations

Elements in Sharp Blunt continuity Transection Transection Totals

Plexus cases 20 28 2 71

Plexus elements 57 83 61 201

Neurolysis 24/26 0/0 0/0 4/26(92%)(+NAPs)

Primary suture 0/0 25/31 0/0 25/31(81%)

Secondary suture 9/7 12/8 3/5 18/26(70%)

Secondary graft 22/17 21/40 25/56 63/118(118%)

Total elements 48/57 54/83 28/61 130/201(65%)

Results are given as number of elements recovering to grade 3 or better (LSUHSCsystem). Primary = repair within 72 hours of injury; secondary = delayed repair,usually after several weeks.

NAP = nerve action potential.

Wallerian degenerative process takes time. For most suspect-ed lesions in continuity, an approximate 3 month period ofrepetitive clinical and needle EMG studies is needed beforeresorting to surgery since some patients can have evidence ofearly clinical or electrical recovery and thus not require anoperation. When this does not occur, operative exploration,external neurolysis of the involved elements, and operativeNAP recordings should be performed. Based on informationfrom these studies, resection of the lesion in continuity maybe indicated.

OPERATIVE NERVE ACTION POTENTIAL RECORDINGS

To perform operative NAP recordings requires an EMGmachine for both stimulation and recording, either bipolar orpreferably tripolar electrodes for stimulation, and bipolarelectrodes for recording.10 If possible this author prefers tocheck electrodes and equipment either by recording from anadjacent, less involved or intact nerve, or by both stimulatingand recording above the lesion. An inter-electrode distance ofat least 3 cm is needed. The recording electrodes are thenmoved through the lesion in continuity to determine if arecorded response persists and whether the response is distalto the lesion or injury site. If found, a simple external neurol-ysis up and down the course at an epineurial level will suffice.

A recent analysis of NAP recordings is in press at the proceed-ings of the 13th World Congress of Neurological Surgeryheld in Marrakesh, Morocco. Recovery to a Louisiana StateUniversity Health Science Center (LSUHSC) grade 3 or bet-ter level occurred with neurolysis based on +NAPs in 1255 of1422 (94.7%) of injuries. A portion of the injury site some-times appeared worse than the rest. Facsicles or groups of fac-sicles were individually tested and those with flat traces wereresected and repaired. Those with positive traces were leftalone. This is termed a split repair. Outcomes in the splitrepair group of patients were uniformly good (94%). If thereis no conduction across the lesion and the trace is flat just dis-tal to the lesion, the lesion is resected and an epineurial end-to-end repair is performed for shorter gaps. Graft repair isusually performed using the sural nerve for longer deficits.Analyzing 1975 repairs performed under these circumstances(including favorable and unfavorable elements or nerves suchas the lower trunk, the medial cord, the ulnar nerve, and theperoneal nerve or peroneal division of the sciatic nerve),recovery to a grade 3 or better LSUHSC level was 56%.

Operative NAP recordings performed in the early weekspostinjury may confirm neurapraxia or document a partial orincomplete lesion. This will not be of use in differentiatingbetween the complete traumatic neuropathy due to either

axonotmesis or neurotmesis where there is gross continuity.To confirm regeneration or lack of regeneration, it is impor-tant to wait several months before recording NAPs. Also, thepresence of a NAP beyond a lesion in continuity a year ormore postinjury may not have the predictive ability of onerecorded in the early months. If a tourniquet is used, it needsto be removed or deflated for 30 minutes before recordingsare attempted. The technique will not work if local anesthet-ic has bathed the nerve being tested.

BRACHIAL PLEXUS INJURIES

There are special circumstances where NAP recordings takeon a different dimension. One of these circumstances is thefrequently seen stretch/contusive injury involving thebrachial plexus.12,28 Cases involving the supraclavicularplexus should be carefully worked up before surgery.1,23,30

Workup should include not only the involved limb but alsoneedle EMG studies of paraspinal muscles and more distalsensory recordings, searching for sensory nerve action poten-tials (SNAPs) suggesting avulsion.25 A cervical computerizedtomography (CT) myelogram is usually also a necessary pre-operative step.4,12 A CT scan is performed not only to lookfor meningoceles—suggesting although not certifying—avul-sion(s) but to determine whether both the anterior and pos-terior nerve roots can be identified within the dye column ateach C5 through T1 level. Usually, each potentially involvedlevel is compared to the same level on the uninvolved side.Magnetic resonance imaging (MRI) using coils and the prop-er sequences is a wonderful advance for medicine and in thehands of experienced neuroradiologists it can be an effectivetool for visualizing the nerve.8 The usual MRI performedwithout such steps is currently useful for nerve tumors butnot usually for nerve injury. At the present time, medicinecontinues to rely on the CT myelogram even though foreither technique there is a false positive and false negativeincidence. On the other hand, MRI of muscles provides theearliest test for denervation.29 Due to increased water con-tent, muscle distal to the nerve lesion may appear somewhathypodense on T1 and more likely hyperdense on T2 withina few days of loss of axonal input.

If possible, an operation is performed at 4-5 months postin-jury for C5 and C6 (Erb’s), or C5, C6, and C7 (Erb’s plus)stretches in adults. This timeframe is important because 40%or more of C5 and C6 and roughly 16% of C5, C6, and C7lesions may be identified within this time frame. Early evi-dence clinically or electrically of recovery will thereforenegate the need for an operation.12 Flail arms involving theC5 through T1 elements should be explored earlier if possi-ble since clinical and/or electrical-evidence of spontaneous


recovery is less common (5%).11,21 There is a great interestworldwide in the use of nerve transfers or what is incorrectlytermed neurotization (repair of nerve) for such lesions.2,9,22

This is accomplished by sectioning functional nerves such asaccessory, phrenic, C3 or C4 spinal nerves, thoracodorsalnerve, triceps branches, medial pectoral branches, intercostalnerves, or a portion of the ulnar nerve and transferring themto nonfunctional elements such as suprascapular nerve, thedivisions of the upper or middle trunk, or the axillary ormusculocutaneous or proximal median nerves.3,19,24 Even thecontralateral C7 spinal nerve or middle trunk has beenextended by sural grafts to provide input to proximal mediannerve. The preferential priorities for plexus repair for stretch-es are (1) elbow-flexion, (2) some shoulder abduction, (3)external rotation of shoulder, and (4) some wrist and fingermotion.16,21

This author continues to favor direct plexus repair for theselesions whenever possible supplemented by nerve transfers.16

As a result, the involved portion of the plexus is exposed,including a portion of the intraforaminal spinal nerves, andoperative NAP studies are performed. Thus, the spinal nerveis stimulated and recordings are taken downstream on distaltrunks and divisions, depending on the distal extent of theinjury on the cords of the plexus. If traces are flat, the lesionmay be postganglionic or pre- and postganglionic. Then, thisauthor sections back to a more proximal level on the spinalnerve and either finds healthy facsicular structure useful forlead-out of grafts or a scarred and sometimes avulsed ele-ment, in which case direct repair at that level is not possible.Such an element or its distal outflows are then candidates fornerve transfer(s). If the lesion is preganglionic, a relativelylarge amplitude and rapidly conducting NAP will be record-ed because of sensory fiber sparing. Direct repair in such adistribution will not be useful, but nerve transfer to it may beuseful. The most favorable operative NAP finding would bea smaller amplitude and slower conducting (20-40 m/s) NAPwhich provides strong evidence for regeneration even thoughmore distal clinical loss is complete. This would be an ele-ment not to repair or transfer into more distally.

In patients with C5, C6, clinical, and needle EMG loss, it isextremely important to expose operatively C7 to the middletrunk and to perform operative recordings on that element.This is because 1/7th of C5 andC6 patients also have seriousinvolvement of C7. This needs assessment so that its poten-tial repair is not neglected.12,16

In only 5% of flail arm cases will C5 be avulsed and in only5% of cases are all roots (C5 thru T1) avulsed. Thereforedirect repair of one or more elements is usually possible.12

Having said that, the results of nerve transfers are good sothey can almost always be added in. For example, if directrepair of lead out from C6 to anterior division of upper trunk

or lateral cord is possible, this author will still transfer intothe medial portion of the musculocutaneus nerve medial pec-toral branches on a C5 and C6, or C5, C6, and C7 stretch ortake ulnar flexor fasicles to the motor branch of the muscu-locutaneous nerve. Even in the flail arm, if C6 or C7 haveuseful outflow that will be taken to the lateral cord and some-times also the posterior cord while intercostal nerves aretaken to one-half of the more distal musculocutaneous nerve.Another example of direct repair coupled with nerve transferis that the C5, if useable for direct lead-out, may be extend-ed by grafts to the posterior division of the upper trunk andthe distal accessory nerve transferred into the more proximalsupraclavicular nerve.

Useful outcomes are still quite difficult to achieve in this cat-egory of plexus injury.12,21 By comparison, outcomes withrepair of plexus lacerations (Table 2) and even plexus gunshotwounds are far better (Table 3). Most operations for lacera-tions for plexus lacerations and GSWs are performed anteri-orly, but if damage is close to the spine (especially on lowerspinal nerves or roots) a posterior approach is indicated.5 Useof operative recordings is especially important with GSWswhere some 40-50% of the lesions in continuity produced bythis wounding mechanism have some potential for recoverywithout direct repair. Provided these lesions can be identifiedby operative recording techniques. Thus, in a series of 118cases involving brachial plexus elements with adequate fol-low-up, 46% of lesions with a finding of complete preopera-tive clinical and EMG loss in their distribution had NAPstransmitted across their lesions at 2-6 months post wound-ing. These plexus elements had only a neurolysis with a 91%recovery rate and did not require repair either by direct sutureor by grafts.13 By comparison recovery rates to LSUHSCgrade 3 or better in the suture repair group was 67% andwith grafts it was 54%. These figures include repairs for ele-ments known to have low recovery reates with repair such asC8, T1, lower trunk, and medial cord as well as those such asC5, C6, C7, upper and middle trunks, and lateral and poste-rior cords where repair is more likely to be successful.

Interestingly, despite clinical and EMG evidence of sparing offunction, there were nine plexus elements that did not con-duct and thus required resection. These elements histologi-cally were neurotemetic. The preoperative evaluation as toseverity of loss was probably complicated by anatomicalplexus variations especially at the cord level.

Gunshot wounds as well as stretch injuries involving infra-clavicular plexus levels i.e., cords and cord-to-nerve levels,did not necessarily recover spontaneously and required oper-ation. They did relatively well even if repair by suture or graftwas necessary. The anatomy of these lesions can be complexand the juxtaposition of the axillary artery and vein and theirbranches make dissection difficult, but the outcomes warrant



Table 2 Overall postoperative grades on 366 patients with supraclavicular plexus stretch injuries patients

Postoperative grade

Initial loss 0 1 1-2 2 2-3 3 3-4 4 4-5 Totals

C5/C6 ( C ) 1 0 0 0 6 14 18 9 7 5

C5/C6/C7 ( C ) 1 0 0 2 14 23 20 9 6 75

C5 to T1 (C ) 21 8 15 29 62 40 20 13 0 208

C5 to C8 ( C ) 0 0 0 0 0 1 1 0 0 2

C6/C7/C8/T1 (C ) 0 0 0 1 0 2 1 0 0 4

C7 to T1 (C ) 0 0 0 0 1 1 0 0 0 2

C8 to T1 ( C ) 2 3 2 2 0 2 0 0 0 11

C8 to T1 (I) 0 0 0 0 0 0 2 2 3 7

C7/C8/T1 (I) 0 0 0 0 0 1 1 0 0 2

Total 25 11 17 34 83 84 63 33 16 366

C = complete or nearly complete loss; I = incomplete loss.

Table 3 Outcomes of Surgery for 118 gunshot wound injuries toplexus

No. of NeurolysisType of Lesion Elements (+NAP) Suture Graft

Lesions w/ 202 46/42 21/14 135/73complete loss

Lesions w/ 91 82/78 6/5 3/2incomplete loss

Totals 293 128/120 27/19 138/75(94%) (70%) (54%)

Results are given as the total number of elements/ the number of elements recov-ering to grade 3 or higher.

NAP = nerve action potential.

this step. An interesting subset of such injuries are those ofthe axillary nerve (Table 4). The incidence of associatedplexus injuries was relatively high, but despite this, thepatient outcome was relatively good in the author’s series aswell as others.15,20

POSSIBLE PITFALLS

Some cautions referable to both pre- and postoperative elec-trical evaluation (especially regarding plexus lesions) are war-ranted. It is possible for the needle EMG to show a good dealof denervational change in muscle(s) that have recoveredclinical contraction especially in the early months postinjuryor repair, so examination and grading of muscle functionremains a priority. Unfortunately, on the other hand, earlyweak contraction does not always mean eventual good con-traction although it favors it. It is also possible, since needleEMG is a sampling procedure, to record nascents in one ormore areas of a muscle perhaps because a few fibers havefound their way back to muscle and yet over time there maynot be enough regrowth for useful contraction of that mus-cle. However, if serial needle EMG shows an increasing num-ber of nascents (perhaps with decreasing numbers of fibrilla-tions) these repetitive observations strongly favor useful

recovery. Parsonage-Turner syndrome can, of course, beresponsible for loss in the plexus distribution.7,30 Like thoracicoutlet syndrome, Parsonage-Turner syndrome is a diagnosis ofexclusion. Unless the patient has the classic neuropathiesinvolving multiple somewhat desperate plexus elements, and acharacteristic antecedent event such as immunization, pneu-monia, or excessive exercise or exertion, the clinician mustexclude such possibilities as cervical rib or elongation,“Parrott beaking” of the C7 transverse process, tumor, ordirect trauma due to operative manipulation or malpositionduring anesthesia. Having said that, Parsonage-Turner syn-drome remains an important differential diagnosis and thepatient’s recovery is not helped by operation.30

Other Operative Electrophysiologic Tests

The usefulness of skin level somatosensory studies in seriousplexus lesions at least in the first year or so postinjury is ques-tionable. The stimulus sites are usually well distal to theinjury site. For serious injuries such as GSWs, lacerations,and stretches, the information gained can be minimal.26

With the supraclavicular stretch injuries, the clinician shouldperform distal sensory NAP studies to look for preganglionicinjury of T1, C8, or C7. Reproducible studies for C6 pregan-glionic injuries are possible but difficult using the lateral


Table 4 Outcomes of axillary nerve repair in 99 patients (Number of Nerves/Mean Postoperative LSUHSC Grade)

Partial Loss Complete Loss& Positive & Positive NAP/ Resection Resection

Outcome NAP/Neurolysis* Neurolysis† & Suture† & Graft ‡

Solitary axillary palsy (56 cases) 4/4.3 6/3.8 3/3.8 43/3.8§

Axillary palsy w/ 1/4.5 3/4.3 0/0 7/3.22 suprascapular palsy (11 cases)

Axillary palsy w/radial loss (14 cases) 3/4.3 2/4.5 0/0 9/3.0

Axillary palsy w/ 7/4.0 4/4.2 0/0 7/3.1other plexus loss(18 cases)||

Totals** 15/4.1 ± 0.4 15/4.0 ± 1.0 3/3.8 ± 0.6 66/3.7 ± 1.1

* The mean preoperative grade was 2.2 (range 1-4)† Preoperative grade in each case was 0.‡ Preoperative grade in most cases was 0, except in three in which it was 1 (trace only).§ Includes two split repairs with a mean outcome grade of 4.|| Two additional axillary nerves were exposed, but irreparable due not only to lengthy involvement, but also to distal avulsion.

** The mean values are expressed as means ± SDs.

LSUHSC = Louisiana State University Health Science Center; NAP = nerve action potential; SD = standard deviation.

cutaneous branch of musculocutaneous territory for stimula-tion and recording. Reproducable studies are nonexistent forC5 nerve roots and spinal nerve.14,23

Direct stimulation of spinal nerve(s) on the operating tableand recording a spinal evoked potential (EP) or evoked cor-tical response (ECR) does, if it is positive, document someintegrity of the peripheral to central sensory pathway. It doesnot, however, guarantee that the motor concomitant is intact,although it favors such.10,26 Experimental studies have indi-cated that only a few hundred fibers need to be intact in theposterior root to record such a spinal EP or ECR after stim-ulation of its more distal spinal nerve.32

Recently, in the case of the plexus from the spinal nerves,there is interest in motor evoked potentials (MEPs) where thecerebral motor cortex is stimulated and more distal record-ings are made The implication is that if there is a positiveresponse then there is an intact motor root. Although likely,that is not necessarily so since the descending pathways fromthe cortical stimulus site used may be other than motor, justas they may be at a spinal cord level when used to evaluatecorticospinal tract function. Of some promise, but awaitingfurther experience and confirmation, is the use of paraspinalmotor unit action potential (MUAP) recordings after spinalnerve stimulation. If a response is positive after spinal nervestimulation then the motor root is intact, although only 100or less intact fibers may give such a response. The surgeonmust take great care that the stimulus does not spread to adja-cent intact spinal nerves since each segment of paraspinalmuscle has input from multiple cervical spinal segments.Unfortunately, the reverse hypothesis—that proximal C5injury usually provides paraspinal denervation—is not neces-sarily true. Chang, England, and Sumner studied 20 cases ofC5 injury including a number of avulsions and found thatthe C5 paraspinal levels were often spared denervation(unpublished data).

Stimulation of a spinal nerve or other more distal plexal ele-ment and needle or surface electrode recordings from muscleis of interest to physicians when diagnosing nerve injury. Formost plexus lesions or sciatic or pelvic level femoral lesions, itis many months postinjury before regenerating fibers areexpected to reach muscle, unless the lesions are partial orincomplete from the beginning. In addition, even thoughsuch growth has had a chance to occur, such a response mayonly indicate the arrival of a few fibers especially if samplingof the muscle is limited to one or two areas. The advantage ofoperative NAPs, at least in the early months postinjury, isthat their presence indicates at least 4000 to 5000 nerve

fibers greater than 5 µm in diameter at the recording site.The relatively early presence of this many fibers beyond alesion in continuity bodes well for future functional growth.

SUMMARY

It can be reassuring under some delayed circumstances toknow that at least some fibers have reached muscle and madeenough of a connection to record an evoked MUAP. Thus,grading the amplitude or latency of such a response and usingthat to decide for or against resection has been an unexpect-ed consequence of such studies. This author will leaves thereader to their own assessment of when a nerve or element isstimulated and whether recordings can be made from distalmuscle in the distribution of that nerve, no matter theiramplitude or latency, and whether resection of the involvednerve or element and the consequent repair necessitated isjustified.

The previous observations are based on adult nerve (andespecially plexus injuries) and may not be applicable to theunfortunate child with a birth injury to the plexus.

REFERENCES

1. Alnot JY. Traumatic brachial plexus palsy in the adult. Retro- andinfraclavicular lesions. Clin Orthop Relat Res 1988;237:9-16.

2. Birch R, Bonney G, Wynn Parry C. Surgical disorders of the periph-eral nerve. Philadelphia: WB Saunders; 1998.

3. Brandt KE, Mackinnon SE. A technique for maximizing bicepsrecovery in brachial plexus reconstruction J Hand Surg Am1993;18:726-733.

4. Carvalho GA, Nikkhah G, Matthies C, Penkert G, Samii M.Diagnosis of root avulsions in traumatic brachial plexus injuries: valueof computerized tomography myelography and magnetic resonanceimaging. J Neurosurg 1997;86:69-76.

5. Dubuisson AS, Kline DG, Weinshel SS. Posterior subscapularapproach to the brachial plexus-report of 102 patients J Neurosurg1993;79:319-330.

6. Dyck P, Thomas PK, editors. Peripheral neuropathy, 3rd edition.Philadelphia: WB Saunders; 1993.

7. England J, Sumner A. Neurologic amyotrophy: an increasinglydiverse entity. Muscle Nerve 1987;10:60-68.

8. Filler AG, Kliot M, Howe FA, Hayes CE, Saunders DE, Goodkin R,Bell BA, Winn HR, Griffiths JR, Tsuruda JS. Application of magnet-ic resonance neurography in the evaluation of patients with peripher-al nerve pathology. J Neurosurg 1996;85:229-309.

9. Gabriel E, Villavicencio A, Friedman A. Evaluation and surgicalrepair of brachial plexus injuries. Semin Neurosurg 2001;12:29-48.

10. Happel L, Kline D. Intraoperative neurophysiology of the peripheralnervous system. In: Deletis V, Shils J, editors. Neurophysiology in


neurosurgery. A modern interoperative approach. New York:Academic Press; 2002.

11. Hentz VR, Narakas A. The results of microneurosurgical reconstruc-tion in complete brachial plexus palsy. Assessing outcome and pre-dicting results. Orthop Clin North Am 1988;19:107-114.

12. Kim DH, Cho YJ, Tiel RL, Kline DG. Outcomes of surgery in 1019brachial plexus lesions treated at LSUHSC. J Neurosurg2003;98:1005-1016.

13. Kim DH, Murovic JA, Tiel RL, Kline DG. Penetrating injuries dueto gunshot wounds involving brachial plexus. Neurosurg Focus2004;16:1-6.

14. Kimura J. Electrodiagnosis in diseases of nerve and muscle principleand practice, 2nd edition. Philadelphia: FA Davis; 1989.

15. Kline D, Kim D. Axillary nerve repair in 99 patients with 101 stretchinjuries. J Neurosurg 2003;99:630-636.

16. Kline D, Tiel R. Direct plexus repair by grafts supplemental by nervetransfers. Hand Clin 2005;21:55-69.

17. Lundborg G. Nerve injury and repair. Edinburgh: Churchill-Livingstone; 1988.

18. Mackinnon S, Dellon A. Surgery of peripheral nerve. New York:Thieme; 1988.

19. Malessy M, Thomeer R. Evaluation of intercostal to musculocuta-neous nerve transfer in reconstructive brachial plexus surgery. JNeurosurg 1998;88:266-271.

20. Mikami Y, Nagano A, Ochiai N, Yamamoto S. Results of nerve graft-ing for injuries of the axillary and suprascapular nerves. J Bone JointSurg Br 1997;79:527-531.

21. Millesi H. Brachial plexus injuries. Nerve grafting. Clin Orthop RelatRes 1988;237:36-42.

22. Omer G, Spinner M, Van Beek AC. Management of peripheral nerveproblems. Philadelphia: WB Saunders; 1998.

23. Parry GJ. Electrodiagnostic studies in the evaluation of peripheralnerve and brachial plexus injuries. Neurol Clin 1992;10:921-934.

24. Samardzic M, Rasulic L, Grujicic D, Milicic B. Results of nerve trans-fer to the musculocutaneous and axillary nerves. Neurosurg2000;46:93-103.

25. Smith SJ. The role of neurophysiological investigation in traumaticbrachial plexus injuries in adults and children. J Hand Surg Br1996;21:145-148.

26. Sugioka H. Evoked potentials in the investigation of traumatic lesionsof peripheral nerves and brachial plexus. Clin Orthop Relat Res1984;184:85-92.

27. Sunderland S. Nerve injuries and their repair: a critical appraisal.Edinburgh: Churchill-Livingstone; 1991.

28. Terzis J, Strauch B. Microsurgery of the peripheral nerve a physiolog-ical approach. Clin Orthop Relat Res 1978;122:39-48.

29. West GA, Haynor DR, Goodkin R, Tsuruda JS, Bronstein AD, KraftG, Winter T, Kliot M. Magnetic resonance imaging and signalchanges in denervated muscle after peripheral nerve injury.Neurosurgery 1994;35:1077-1086.

30. Wilbourn AJ. Brachial plexus disorders. In: Dyck PJ, Thomas PK,editors. Peripheral neuropathy, 3rd edition. Philadelphia: WBSaunders; 1993.

31. Woodhall B, Beebe GW, editors. Peripheral nerve regeneration. A fol-low-up study of 3656 World War II injuries. VA MedicalMonograph. Washington, DC: US Government Printing Office;1956.

32. Zhao S, Kim DH, Kline DG, Beuerman RW, Thompson HW.Somatosensory evoked potentials induced by stimulating a variablenumber of nerve fibers in rat. Muscle Nerve 1993;16:1220-1227.


��421 First Avenue SW, Suite 300 East

Rochester, MN 55902(507) 288-0100 / Fax: (507) 288-1225

[email protected]

��

lawrence r.robinson,md jeffrey g.jarvik,md,mph david … · university of washington ... washington...

Documents