Advances in Clinical Neurophysiology
(Supplements to Clinical Neurophysiology, Vol. 57)
Editors: M. Hallett, L.H. Phillips, II, D.L. Schomer, J.M. Massey
© 2004 Elsevier B.Y. All rights reserved
Chapter 19
Traumatic injury to peripheral nerves**
Lawrence R. Robinson *
Department of Rehabilitation Medicine, University of Washington, Seattle, WA 98195 (USA)
Abstract
This chapter reviews the epidemiology and classifi-
cation of traumatic peripheral nerve injuries, the
effects of these injuries on nerve and muscle, and
how electrodiagnosis is used to help classify the
injury. Mechanisms of recovery are also reviewed.
Motor and sensory nerve conduction studies, needle
EMG, and other electrophysiologic methods are
particularly useful for localization of peripheral
nerve injuries. They are used to detect and quantify
the degree of axon loss. Data from electrophysio-
logic studies can guide treatment decisions and assist
in the assessment of prognosis.
1. Epidemiology of peripheral nerve trauma
Traumatic injury to peripheral nerves results in
considerable disability across the world. In peace-
time, peripheral nerve injuries result commonly from
motor vehicle accident trauma, and less commonly
* Correspondence to: Professor Lawrence R. Robinson,
Box 359740, Rehabilitation Medicine, Harborview Medi-
cal Center, 325 Ninth Avenue, Seattle, WA 98104, USA.
Tel: +12067313167; Fax: +12067316554;
E-mail: lorenzo@u.washington.edu
** Used by permission, copyright 2000 AAEM. This
chapter is based on the monograph "Traumatic Injury to
Peripheral Nerves" published by the AAEM in Muscle and
Nerve, 2000, 23: 863-873. The monograph can be
purchased from AAEM at + 1(507)-288-0100.
173
from penetrating trauma, i.e. falls and industrial
accidents. Out of all patients admitted to Level I
trauma centers, an estimated 2-3% have peripheral
nerve injuries (Selecki et aI., 1977; Noble et al.,
1998). If plexus and root injuries are also included,
the incidence is about 5% (Noble et aI., 1998).
In the upper limb, the nerve injured most com-
monly is the radial nerve, followed by the ulnar and
median nerves (Selecki et aI., 1977; Noble et al.,
1998). Lower limb peripheral nerve injuries are less
common. The sciatic nerve is injured most fre-
quently, followed by the peroneal and rarely the
tibial or femoral nerves. Fractures of bones and
adjacent nerve injuries aften occur together. An
example is humeral fracture with associated radial
neuropathy.
In wartime, peripheral nerve trauma is much more
common, and much of our knowledge about periph-
eral nerve injury, repair and recovery comes from the
medical experience in World Wars I and II, and
subsequent conflicts (Haymaker and Woodhall,
1953; Seddon, 1975; Sunderland, 1978).
Peripheral nerve injuries may be seen in isolation,
but they also often accompany CNS trauma. This not
only compounds the disability, but recognition of the
peripheral nerve lesion is often problematic. Sixty
percent of patients with peripheral nerve injuries
have a traumatic brain injury (Noble et aI., 1998).
Conversely, 10-34% of patients with traumatic brain
174

injury admitted to rehabilitation units have asso- 2. Classification of nerve injuries

ciated peripheral nerve injuries (Garland, 1981;
Stone and Keenan, 1988; Cosgrove et al., 1989). It is
often easy to miss peripheral nerve injuries in the
setting of CNS trauma. Since the neurologic history
and examination is limited, early hints to a super-
imposed peripheral nerve lesion might be only
flaccidity, areflexia, and reduced limb movement.
Peripheral nerve injuries have significant impact,
as they impede recovery of function and return to
work. There is also a risk of secondary disability
from falls, fractures, or other injuries. An under-
standing of the classification, pathophysiology and
electrodiagnosis of these lesions is critical for
appropriate diagnosis, localization and management
of peripheral nerve trauma.
The classifications of Seddon (1975) and Sunderland
(1978) are the most commonly used schemes for
description (Table 1). The former is more commonly
used in the literature. The terms neurapraxia,
axonotmesis, and neurotmesis are used to describe
peripheral nerve injuries (Seddon, 1975). Neu-
rapraxia is a comparatively mild injury. There is
motor and sensory loss but no Wallerian degenera-
tion. The distal nerve conducts normally. Focal
demyelination and/or ischemia are thought to be the
etiologies for conduction block. Recovery may occur
within hours, days, weeks, or a few months.
Axonotmesis occurs commonly in crush injuries. The
axons and their myelin sheaths are broken, yet the

TABLE 1

CLASSIFICATION SYSTEMS FOR NERVE INJURY

Seddon Sunderland Pathology Prognosis

classification classification

Neurapraxia First degree Myelin injury or ischemia
Axonotmesis Axons disrupted
Variable stromal disruption
Second degree Axons disrupted
Endoneuria! tubes intact
Perineurium intact
Epineurium intact
Third degree Axons disrupted
Endoneurial tubes disrupted
Perineurium intact
Epineurium intact
Fourth degree Axons disrupted
Endoneurial tubes disrupted
Perineurium disrupted
Epineurium intact
Neurotmesis Fifth degree Axon disrupted
Endoneurial tubes disrupted
Perineurium disrupted
Epineurium disrupted
Excellent recovery in weeks to months
Good to poor, depending upon integrity of
supporting structures and distance to muscle
Good, depending upon distance to muscle
Poor
Axonalmisdrrection
Surgery may be required
Poor
Axonal misdirection
Surgery usually required
No spontaneous recovery
Surgery required
Prognosis after surgery guarded

Adapted from Dillingham, 1998.


surrounding stroma (i.e. the endoneurium, perineu-
rium, and epineurium) remains partially or fully
intact. Wallerian degeneration occurs, but sub-
sequent axonal regrowth may proceed along the
intact endoneurial tubes. Recovery ultimately
depends upon the degree of internal disorganization
in the nerve as well as the distance to the end organ.
Sunderland's classification (below) further divides
this category. Neurotmesis describes a nerve that has
been either completely severed or is so markedly
disorganized by scar tissue that axonal regrowth is
impossible. Examples are sharp injury, some traction
injuries or injection of toxic drugs. Prognosis for
spontaneous recovery is extremely poor without
surgical intervention.
The Sunderland classification (Sunderland, 1978)
is a more subdivided scheme for description of
peripheral nerve injuries. There are five groups
instead of three. In first degree injury, conduction
block occurs, but the stroma is completely intact. It
corresponds to Seddon's classification of neura-
praxia, and the prognosis is good. In second degree
injury, the axon is transected, but stroma is intact.
Recovery can occur by axonal regrowth along
endoneurial tubes. In third degree injury, the axon
and endoneurial tubes are transected, but the sur-
rounding perineurium is intact. Recovery depends
upon how well the axons can cross the site of the
lesion to find endoneurial tubes. In fourth degree
injury, there is loss of continuity ofaxons, endo-
neurial tubes and perineurium. Individual nerve
fascicles are transected, and the continuity of the
nerve trunk is maintained only by the surrounding
epineurium. Traction injuries commonly produce
this type of lesion. The prognosis is usually poor
without surgical intervention, due to the marked
internal disorganization of guiding connective tissue
elements and associated scarring. In fifth degree
injury, the entire nerve trunk is transected, and it and
is similar to Seddon's neurotmesis.
Some authors have described another "degree" of
injury, known as 6th degree injury (Mackinnon and
Dellon, 1988). This is a mixed lesion, since both
axon loss and conduction block occur in different
fibers. This type lesion is probably quite common,
175
and it requires skillful electrodiagnostic data collec-
tion and analysis to separate it from pure axon loss
lesions.
3. Effects of neurapraxia on nerve and muscle
As noted above, neurapraxic injuries to peripheral
nerves may be due to ischemia or focal demyelina-
tion. When ischemia for a brief period (i.e. up to
6 h) is the underlying cause, there is usually no
structural change in the nerve (Gilliatt, 1980),
though there may be edema in other nearby tissues.
On the other hand, in neurapraxic lesions due to
focal demyelination, there are anatomic changes of
the myelin sheath that spare the axon. Tourniquet
paralysis has been used to produce an animal model
of a neurapraxic lesion (Ochoa et aI., 1972), though
it is recognized that acute crush injuries may differ
in mechanism from tourniquet injury. In this model,
anatomic changes along the nerve are most marked
at the edge of the tourniquet, where a significant
pressure gradient exists between the tourniquet and
non-tourniquet areas. The pressure gradient essen-
tially "squeezes" out the myelin. This produces
invagination of one paranodal region into the next.
As a result, there is an area of focal demyelination at
the edge of the tourniquet (Ochoa et aI., 1972).
Larger fibers are more severely affected than smaller
fibers.
In the area of focal demyelination, impulse
conduction from one node of Ranvier to the next is
slowed, as current leakage occurs. This prolongs the
time for impulses to reach threshold at successive
nodes of Ranvier. Slowing of conduction velocity
along this nerve segment ensues. More severe
demyelination results in complete conduction block.
This has been reported to occur when internodal
conduction times exceed 500-600!J..s (Rasminsky
and Sears, 1972). Since there are very few sodium
channels in internodal segments of myelinated
nerves, conduction in demyelinated nerves cannot
simply proceed slowly as it would for normally
unmyelinated nerves. Thus sufficient demyelination
will produce conduction block rather than more
severe slowing.
176
There are relatively few changes in muscle as a
result of neurapraxic lesions. Disuse atrophy can
occur when neurapraxia is more than transient.
There remains debate as to whether muscle fibril-
lates after a purely neurapraxic lesion (see below).
4. Effects ofaxonotmesis on nerve and muscle
Soon after an axonal lesion, the process of Wallerian
degeneration begins in nerve fibers. This process is
described well elsewhere (Miller, 1987; Dumitru,
1995), and it will be reviewed only briefly here.
There are changes in both the axon and the nerve cell
body. In the axon, a number of changes occur in the
first 2 days, including leakage of axoplasm from the
severed nerve, swelling of the distal nerve segment,
and subsequent disappearance of neurofibrils in the
distal segment. By day 3, there is fragmentation of
both axon and myelin beginning digestion of myelin
components. By day 8, the axon has been digested,
and Schwann cells are attempting to bridge the gap
between the two nerve segments. Nerve fibers may
also degenerate for a variable distance proximally;
depending upon the severity of the lesion, this
retrograde degeneration may extend for several
centimeters.
If the lesion is sufficiently proximal, there are also
a number of changes that occur in the nerve cell
body after nerve trauma. Within the first 48 h, the
Nissl bodies (the cell's rough endoplasmic retic-
ulum) break apart into fine particles. By 2-3 weeks
after injury, the cell nucleus becomes displaced
eccentrically, and the nucleolus is becomes eccen-
trically placed within the nucleus. These changes
may reverse during recovery.
5. Electrodiagnosis: timing of changes and
determining degree of injury
5.1. The compound muscle action potential
(eMAP)
5.1.1. Neurapraxia
In purely neurapraxic lesions, the compound muscle
action potential will change immediately after injury,
assuming one can stimulate both above and below
the site of the lesion (Fig. 1). When one records from
distal muscles and stimulates distal to the site of the
lesion, the CMAP should always be normal, since
axonal loss and Wallerian degeneration have not
occurred. Stimulation proximal to the lesion will
produce a smaller or absent CMAP, as conduction in
some or all fibers is blocked. It should be remem-
bered that amplitudes normally fall with increasing
distance between stimulation and recording sites;
hence there is some debate about how much of a
drop in amplitude is sufficient to demonstrate
conduction block. Amplitude drops exceeding 20%
over a 25 em distance or less are clearly abnormal.
Smaller changes over shorter distances may also
suggest conduction block. In addition to conduction
block, there may also be conduction slowing across
the site of a partial lesion. This slowing may be due
to either loss of faster conducting fibers or demyeli-
nation of surviving fibers. All these changes in the
CMAP will generally persist until recovery takes
place, typically by no more than a few months post
injury. Most importantly, the distal CMAP will never
drop in amplitude in purely neurapraxic injuries,
since no axon loss or Wallerian degeneration occur,
and the distal nerve segment remains normally
excitable.
5.1.2. Axonotmesis and neurotmesis
Electrodiagnostically, complete axonotmesis (equiv-
alent to Sunderland grades 2, 3 and 4) and complete
neurotmesis look the same, since the difference
between these types of lesions is the integrity of the
supporting structures that lack electrophysiologic
function. These lesions can be grouped together as
axonotmesis for the purpose of this discussion.
Immediately after axonotmesis and for a "few
days" thereafter, the CMAP and motor conduction
studies look the same as those of a neurapraxic
lesion. Nerve segments distal to the lesion remain
excitable. They demonstrate normal conduction,
while proximal stimulation results in an absent or
small response from distal muscles. Early on, this
picture looks the same as conduction block, and it
can be confused with neurapraxia. Hence neu-
rapraxia and axontomeis cannot be distinguished
 177

Fig. 1. Diagrammatic representation of changes in the compound muscle action potential (CMAP) after (A) neurapraxia,
(B) axonotmesis or neurotmesis, and (C) mixed lesions.
178
until sufficient time for Wallerian degeneration in all
motor fibers has occurred. This is typically about 9
days post injury (Chaudry and Comblath, 1992).
As Wallerian degeneration occurs, the amplitude
of the CMAP elicited with distal stimulation will
fall. It starts at about day 3 and is complete by about
day 9 (Chaudry and Comblath, 1992). Neuromus-
cular transmission fails before nerve excitability
does (Gilliatt and Taylor, 1959; Gilliatt and Hjorth,
1972). Thus, in complete axonotmesis at day 9, one
has a very different picture from neurapraxia. There
are absent responses from stimulation above and
below the lesion. Partial axon loss lesions will
produce small amplitude motor responses, and the
amplitude of the CMAP is roughly proportional to
the number of surviving axons. One can compare
side-to-side CMAP amplitudes to estimate the
degree of axon loss, though inherent side-to-side
variability of up to 30-50% limits the accuracy of
the estimate. The use of the CMAP amplitude to
estimate the degree of surviving axons is only
reliable early after injury, before axonal sprouting
has occurred. Use of this technique at a later time
after injury will lead to underestimation of the
degree of axon loss.
5.1.3. Mixed lesions
Lesions which have a mixture of axon loss and
conduction block provide a unique challenge. These
can usually be sorted out by carefully examining
amplitudes of the CMAP elicited from stimulation
above and below the lesion and by comparison of the
amplitude with distal stimulation to that obtained
from the other side. The percentage of axon loss is
best estimated by comparing the CMAP amplitude
from distal stimulation with that obtained con-
tralaterally. The percentage conduction block in the
remaining axons is best estimated by comparing
amplitudes or areas obtained with stimulation distal
and proximal to the lesion. Thus, if a LrnV response
is obtained with proximal stimulation, a 2 mV
response is obtained distally, and a 10 mV response
is obtained with distal stimulation contralaterally,
one can deduce that probably about 80% of the
axons are lost. Of the remaining 20%, half are
blocked (neurapraxic) at the lesion site. As men-
tioned above, this analysis is most useful only in the
acute phase, before reinnervation by axonal sprout-
ing occurs.
5.2. F-waves
F-waves may change immediately after the onset of
a neurapraxic lesion. In complete block, F-waves
will be absent. In partial lesions, changes can be
more subtle, since F-waves are elicited from only
3-5% of the axon population at one time (Fisher,
1992). Partial lesions may have normal miminal and
mean F-wave latencies, with reduced or possibly
normal penetrance. While F-waves are conceptually
appealing for detection of proximal lesions (e.g.
brachial plexopathies) they provide useful additional
or unique information in only a few instances. They
are sometimes useful in very early proximal lesions
when conventional studies are normal since stimula-
tion does not occur proximal to the lesion. They are
not very good at distinguishing axon loss lesions
from conduction block.
5.3. Compound or sensory nerve action potentials
5.3.1. Neurapraxia
The sensory nerve action potential (SNAP) and
compound nerve action potential (CNAP) will show
changes similar to the CMAP after focal nerve
injury. In the setting of neurapraxia, there is a focal
conduction block at the site of the lesion, and
amplitude with distal stimulation is preserved. The
criteria for establishing conduction block in sensory
nerve fibers are substantially different than that for
the CMAP. When one records nerve action poten-
tials, there is normally a greater drop in amplitude
over increasing distance between stimulating and
recording electrodes due to temporal dispersion and
phase cancellation (Kimura et aI., 1986). Amplitude
drops of 50-70% over a 25 em distance are not
unexpected, and it is less clear just what change in
amplitude is abnormal. A large focal change over a
small distance is probably significant. Slowing may
also accompany partial conduction blocks, as it does
for the CMAP. Responses elicited with stimulation

and recording distal to the lesion are normal in pure
neurapraxic injuries.
5.3.2. Axonotmesis and neurotmesis
Immediately after axonotmesis, the SNAP looks the
same as it does in a neurapraxic lesion. Nerve
segments distal to the lesion remain excitable and
demonstrate normal conduction, while proximal
stimulation results in an absent or small response.
Hence neurapraxia and axontomeis cannot be distin-
guished until Wallerian degeneration has occurred in
all sensory fibers, typically about 11 days post injury
(Chaudry and Cornblath, 1992). It takes slightly
longer for sensory nerve studies to demonstrate loss
of amplitude than for motor studies, i.e. 11 days vs.
9 days, due to the earlier failure of neuromuscular
transmission compared to nerve conduction.
5.4. Needle electromyography
5.4.1. Neurapraxia
The needle EMG examination in purely neurapraxic
lesions will show neurogenic changes in recruitment,
but there is debatable about abnormalities in sponta-
neous activity. As mentioned earlier, opinions differ
as to whether fibrillation potentials are recorded after
a purely neurapraxic lesion. One study of peripheral
nerve lesions in baboons did not demonstrate
fibrillation potentials in purely neurapraxic lesions
(Gilliatt et aI., 1978). On the other hand, in a study
of purely neurapraxic lesions in rats (Cangiano and
Lutzemberger, 1977), fibrillation potentials were
said to occur in blocked, but not denervated, muscle
fibers. There are limited reports of fibrillation
potentials in humans with apparent predominantly
neurapraxic nerve lesions (Trojaborg, 1978; Yuska
and Wilbourn, 1998), but it is impossible to know
whether or not any axon loss had occurred in these
patients. Needle EMG is more sensitive for detecting
motor axon loss than nerve conduction studies,
and hence it is easy to imagine situations in which
nerve conduction studies are within normal limits,
but needle EMG detects minimal or mild axon
loss.
179
Independent of whether or not the needle EMG
demonstrates fibrillation potentials in neurapraxia,
the most apparent change on needle EMG will be a
change in recruitment. These occur immediately
after injury. In complete lesions (i.e. complete
conduction block) there will be no motor unit action
potentials (MUAPs). In incomplete neurapraxic
lesions, there will be reduced numbers of MUAPs
that firie more rapidly than normal (i.e. reduced or
discrete recruitment). Recruitment changes alone are
not specific for neurapraxia or axon loss.
Since no axon loss occurs in neurapraxic injuries,
there will be no axonal sprouting and no changes in
MUAP morphology (e.g. duration, amplitude or
phases) anytime after injury.
5.4.2. Axonotmesis and neurotmesis
A number of days after an axon loss lesion, the
needle EMG will demonstrate fibrillation potentials
and positive sharp waves. The time between injury
and onset of fibrillation potentials will depend in part
upon the length of the distal nerve stump. When the
lesion is distal and the distal stump is short, it takes
only 10-14 days for fibrillation potentials to
develop. With a proximal lesion and a longer distal
stump (e.g. ulnar innervated hand muscles in a
brachial plexopathy), fibrillation potentials and pos-
itive sharp waves develop fully in 21-30 days
(Thesleff, 1974). The electromyographer must be
acutely aware of the time since injury to avoid
underestimation of severity when a study is per-
formed early after injury. This is also important to
prevent minsinterpretation of increasing fibrillation
potentials over time as a sign of worsening nerve
injury.
Fibrillation potential and positive sharp wave
density are usually graded on a 1--4 scale. This is an
ordinal scale that indicates greater severity with
higher numbers. It is not an interval or ratio scale,
i.e. 4+ is not twice as bad as 2+ or 4 times as bad
as 1 +. Moreover, a grade of 4 + fibrillation poten-
tials does not reflect complete axon loss. In fact, it
may occur when only a minority ofaxons have been
lost (Buchthal, 1982; Dorfman, 1990). Evaluation of
180
recruitment and the distally elicited CMAP ampli-
tude are necessary before one can decide whether or
not complete axon loss has occurred.
Fibrillation potential size decreases with time
following injury. Kraft (1990) has demonstrated that
fibrillation potentials amplitudes are several hundred
microvolts in the first few months after injury.
However, when lesions are more than 1 year old,
fibrillation potential amplitudes are unlikely to
exceed 100 I.L V. Fibrillation potentials will also
decrease in number as reinnervation occurs, but this
finding is not usually useful clinically for two
reasons. First, since a qualitative or ordinal scale of
fibrillation density is used typically in the absences
of an accurate quantitative measurement of fibrilla-
tion density, comparison of fibrillation potential
numbers from one examination to the other is not
reliable (Dorfman, 1990). Second, even in complete
lesions, fibrillation density will eventually reduce,
since the muscle becomes fibrotic and the number of
viable muscle fibers falls. In this case, a reduction in
fibrillation potential numbers does not predict recov-
ery, but rather muscle fibrosis.
Fibrillation potentials may also occur after direct
muscle injury. Partanen and Danner (1982) have
demonstrated that patients who undergo muscle
biopsy have persistent fibrillation potentials that
begin after 6-7 days and continue for up to 11
months. In patients who have undergone multiple
trauma, coexisting direct muscle injury is common,
and it can be potentially misleading when one
attempts to localize a lesion.
When there are surviving axons after an incom-
plete axonal injury, the remaining MUAPs are
initially normal in morphology, but they demonstrate
reduced or discrete recruitment. Axonal sprouting
will be manifested by changes in the morphology of
existing MUAPs. The amplitude, duration, and the
percentage of polyphasic MUAPs will increase as
the motor unit territory increases (Erminio et al.,
1959; Dorfman, 1990). This process begins soon
after injury. Microscopic studies demonstrate out-
growth of nerve sprouts that begins 4 days after
partial denervation (Hoffman, 1950; Miller, 1987).
Single fiber EMG studies have demonstrated an
increase in fiber density 3 weeks after nerve injury
(Massey and Sanders, 1991).
In complete lesions, the only possible mechanism
of recovery is axonal regrowth. The earliest needle
EMG finding in this case is small, polyphasic, often
unstable, MUAPs. These have been previously
described as "nascent potentials". Observation of
these potentials is dependent upon the establishment
of axon regrowth as well as new neuromuscular
junctions. This observation represents the earliest
evidence of reinnervation, and it usually precedes
the onset of clinically evident voluntary movement
(Dorfman, 1990). These potentials represent the
earliest definitive evidence of axonal reinnervation in
complete lesions. When one performs the examina-
tion in a search for new motor unit potentials, it is
important to accept only "crisp", nearby motor unit
potentials with a short rise-time, since distant
potentials recorded from other muscles can be
mistaken for evidence of intact innervation.
5.4.3. Mixed lesions
When there is a lesion with axon loss and conduction
block, needle EMG examination can be potentially
misleading if it is interpreted in isolation. If, for
example, a lesion results in destruction of 50% of the
original axons and conduction block of the other
50%, needle EMG will demonstrate abundant (e.g.
4+) fibrillation potentials and no voluntary MUAPs.
The electromyographer should not then conclude
that there is a complete axonal lesion, but should
instead carefully evaluate the motor nerve conduc-
tion studies to determine whether there is neuro-
praxia, axonotmesis, or both. The important point
here is to not take the presence of abundant
fibrillation potentials and absent voluntary MUAPs
as evidence of complete denervation.
6. Localization of traumatic nerve injuries
The localization of peripheral nerve injuries is
sometimes straightforward, but it is potentially
complicated by a variety of pitfalls. Localization is
usually performed by two methods: (1) detection of

focal slowing or conduction block on nerve conduc-
tion studies; and (2) assessment of a pattern of
denervation on needle EMG.
Localization of peripheral nerve lesions by nerve
conduction studies usually requires focal slowing or
conduction block as one stimulates above and below
the lesion. To see such a change there must either be
focal demyelination or ischemia, or the lesion should
be so acute that degeneration of the distal stump
has not yet occurred. Thus, lesions with partial
or complete neurapraxia (due to either demyelina-
tion or ischemia) can be well localized with motor
nerve conduction studies, as can very acute axonal
injuries.
In pure axonotmetic or neurotmetic lesions, it is
more difficult, if not impossible, to localize the
lesion using nerve conduction studies. In such a
case, there will be mild and diffuse slowing in the
entire nerve due to loss of the fastest fibers, or there
will be no response at all. Conduction across the
lesion site will be no slower than in other nerve
segments. In addition, if enough time for Wallerian
degeneration has elapsed (i.e. at least 9 days for
motor fibers or 11 days for sensory fibers), there will
be no change in amplitude as one traverses the site
of the lesion. Thus, pure axon loss lesions are not
well localized along a nerve by nerve conduction
studies.
There are some cases in which indirect inferences
can be made about the location of purely axonal
lesions. For instance, if the ulnar motor response is
very small or absent and the median motor response
is normal, this implies an ulnar neuropathy rather
than a lower brachial plexus lesion. However, in such
an instance, the site of pathology along the ulnar
nerve may not be well defined.
Another indirect inference that can be made from
sensory nerve conduction studies is the placement of
the lesion at a pre- versus post-ganglionic location.
Lesions that are proximal to the dorsal root ganglion,
i.e, at the pre-ganglionic level (proximal root, cauda
equina, spinal cord) tend to have normal sensory
nerve action potential amplitudes, even in the setting
of reduced or absent sensation (Brandstater, 1983;
Tackman and Radu, 1983). This is a particularly bad
181
prognostic sign, since it indicates possible root
avulsion. On the other hand, lesions that occur distal
to the dorsal root ganglion have small or absent
SNAPs (when these are recorded in the appropriate
distribution). Thus, SNAPs may be useful to differ-
entiate root vs. plexus or other pre- vs. post-
ganglionic locations. A limitation, particularly in
partial lesions, is the wide variability in SNAP
amplitudes in normal individuals. Mixed pre- and
post-ganglionic lesions are also potentially difficult
to interpret.
The other major electrodiagnostic method for
determining the site of nerve injury is the needle
EMG. Conceptually, if one knows the branching
order to various muscles under study, one can
determine that the nerve injury is between the
branches to the most distal normal muscle and the
most proximal abnormal muscle. There are, how-
ever, a number of potential problems with this
approach. First, the branching and innervation
pattern for muscles is not necessarily consistent from
one person to the next. Sunderland (1978) has
demonstrated the great deal variability in branching
order to muscles in the limbs, as well as the the
number of branches going to each muscle and the
order in which nerve or nerves supply each muscle.
Thus, the typical branching scheme may not apply to
the patient being studied, and consequently the
lesion site can be misidentified.
Second, the problem of muscle trauma and
associated needle EMG findings can be misleading.
As mentioned earlier, direct muscle trauma can
produce positive sharp waves and fibrillation poten-
tials that persist for months after injury (Partanen
and Danner, 1982). Practically speaking, this can
result in erroneous localization or misdiagnosis of
more than one lesion. For example, in the setting of
humeral fracture with associated radial neuropathy,
the triceps muscle often demonstrates fibrillation
potentials due to direct muscle trauma. However,
one could be misled to localize the radial nerve
lesion to the axilla or higher rather than spiral
groove, if the triceps findings are not recognized to
be the result of direct muscle trauma rather than
nerve injury.
182
Third, the problem of partial lesions can make for
misdiagnosis at more distal sites. In partial ulnar
nerve lesions at the elbow, for example, the forearm
ulnar-innervated muscles are often spared (Campbell
et al., 1989). This is thought to be due at least
partially to sparing of the fascicles in the nerve that
are preparing to branch to the flexor digitorum
profundus and the flexor carpi ulnaris. They may be
in a relatively protected position. This finding could
lead one to inadvertently localize the lesion to the
distal forearm or wrist. Similarly, a lesion involving
the median nerve in the arm (above the elbow) has
been reported to cause findings only in the anterior
interosseous nerve distribution (Wertsch et aI.,
1985). Intraneural topography must be considered
when one makes a diagnosis based on nerve
branching (Wertsch et aI., 1994).
7. Mechanisms of recovery
There are several possible mechanisms of recovery
after traumatic nerve injury. Knowledge of these
mechanisms and the type of nerve injury will allow
estimation of the probable course of recovery.
For motor fibers, resolution of conduction block
(in neurapraxic lesions), muscle fiber hypertrophy
(in partial lesions), distal sprouting of spared axons,
and axonal regeneration from the site of injury, may
contribute to the recovery of strength.
Resolution of conduction block, whether due
to ischemia or demyelination, is the earliest mecha-
nism for recovery of strength after nerve injury.
Improvement after a solely ischemic lesion is
relatively quick. Demyelinating injuries take longer,
as remyelination over an injured segment may
take as long as several months (Fowler et aI., 1972).
The rate of recovery depends upon the severity of
demyelination and the length of the demyelinated
segment.
In normal adults who perform strengthening
exercises, there are generally two mechanisms for an
increase in force production. The initial neural
mechanisms are followed by later muscle fiber
hypertrophy. The initial neural mechanisms are
thought to involve improved synchronization of
motor unit firing (Milner-Brown et aI., 1975;
Moritani and De Vries, 1979), and they result in
increased efficiency (defined as muscle force per unit
of electrical activity) in the absence of muscle fiber
changes. After several weeks, muscle fiber hyper-
trophy begins, and it results in further increases in
strength. In patients with partial nerve lesions, it is
unclear to what extent neural changes alone (i.e,
increased efficiency of firing) can contribute to
increased strength, since there is loss of nerve fibers.
However, it is likely that working the existing
muscle fibers to fatigue in the setting of partial nerve
injuries produces enlargement of muscle fibers and
consequent increases in force production.
Partial axonotmesis of motor nerves also produces
distal sprouting of motor fibers from intact axons. It
has been observed that, within 4 days after nerve
injury, sprouts start to form from intact axons,
typically from distal nodes of Ranvier (nodal
sprouts) or from nerve terminals (terminal sprouts)
near denervated muscle fibers (Hoffman, 1950).
Partial recovery in twitch tension has been reported
as early as 7-10 days post injury (Brown et al.,
1981), though electrophysiologic correlates (e.g.
polyphasic, long duration MUAPs) usually take
longer. Sometimes, when axonal regeneration
occurs, those muscle fibers reinnervated by distal
sprouting become dually innervated, i.e. by both the
sprout and the newly regenerated fiber (Hoffman,
1950; Guth, 1962). It is not well understood how
multiple synapses are reduced.
Axonal regeneration contributes to recovery in
both partial and complete axonotmesis and, with
surgical approximation, neurotmesis. In complete
axon loss lesions, this is the only mechanism for
muscle recovery. It is noted that in the 24--36 h after
injury, the proximal nerve stump begins to sprout
regenerating axons, and they begin to penetrate the
area of injury. The recovery that results from this
process depends upon the degree of injury, presence
of scar formation, approximation of the two nerve
ends, and the age of the subject.
In relatively minor axonotmetic lesions, in which
the endoneurial tubes are preserved (i.e, Sunderland
2nd degree injuries), the axons can traverse the

segment of injury in 8-15 days. They then regenerate
along the distal nerve segment at a rate of 1-5 mm/
day (Sunderland, 1978). The rate is slightly faster for
crush injuries than for sharp laceration. It is also
slightly faster for proximal injuries, and it is faster in
younger individuals.
In more severe axonotmetic lesions in which there
is distortion of endoneurial tubes with or without
perineurial disruption (Sunderland 3rd and 4th
degrees), the prognosis for spontaneous regrowth is
worse. Extensive scarring reduces the speed at which
regenerating axons can traverse the lesion. More
importantly, it reduces the likelihood that they will
ever reach their end organs. When regrowth occurs,
it may also be misdirected to the wrong end organ. In
some cases, particularly when a large neuroma is
present, surgical intervention is required.
In complete neurotmesis (Sunderland 5th degree),
axonal regrowth will not occur unless the nerve ends
are freed from scar and surgically re-approximated.
After surgical intervention, using either direct
approximation or cable grafting, nerve growth will
often occur along the endoneurial tubes of the distal
segments. The use of cable grafts (e.g. sural nerve
graft) does not provide axons directly since, they die
after harvesting. The graft simply provides a path-
way for axonal regrowth (Seddon, 1963; Wood,
1998).
In complete lesions, recovery of motor function
will also depend upon integrity of the muscle when
the axon reaches it. Muscles remain viable for
reinnervation for 18-24 months post-injury. Past this
time, motor axon regrowth makes little difference,
since muscle fibers are no longer viable due to
fibrosis and atrophy. For example, in complete lower
trunk brachial plexus lesions, recovery of hand
function is usually not expected, no matter how good
the surgical grafting might be; it simply takes too
long for axons to reach the muscle.
Recovery of sensory function depends upon
different mechanisms than motor recovery. There
may be redistribution of sensory territory after an
axonal injury. Intact fibers often provide cutaneous
sensation to a larger area than they did prior to the
injury (Weddell and Glees, 1941; Speidel, 1942).
183
The mechanisms of axonal regeneration are similar
to those mentioned above for motor axons. An
important difference, however, is that one does not
have end organs that may degenerate after 18-24
months as muscle does; hence sensory recovery may
continue for a longer period of time than motor
recovery does.
8. Electrodiagnostic evaluation of prognosis
Determination of the pathophysiology of a periph-
eral nerve traumatic injury can help with the
estimation of prognosis. Those injuries that are
completely or largely neurapraxic have a good
prognosis for recovery within a few months (usually
up to three months post injury). Resolution of
ischemia and remyelination should be complete by
this time.
Mixed injuries typically have two or more phases
of recovery (Fig. 2). The neurapraxic component
resolves quickly, and muscle fiber hypertrophy can
provide additional recovery. The axonal component
is slower, since it depends upon distal axonal
sprouting and axonal regeneration from the site of
the lesion. Thus patients usually experience a
relatively rapid partial, but incomplete, recovery
followed by a slower phase of additional recovery.
Sensory recovery may proceed for a longer time than
motor (Fig. 3).
Partial axon loss lesions usually represent axo-
notmesis, though a partial neurotmesis (e.g. a
laceration through part of the nerve) cannot always
be excluded in such cases. In axonotmesis, recovery
will depend upon axonal sprouting and regeneration.
Thus there will be some early recovery followed by
a later phase of recovery if or when regenerating
axons reach their end organs. The amplitude of the
CMAP provides some guide to prognosis. In facial
nerve lesions, it has been demonstrated that patients
with CMAP amplitudes 30% or more of the normal
side have an excellent outcome. Those whose
amplitudes are 10-30% of the normal side have good
but not always complete recovery, and those with
amplitudes less than 10% of the normal side have a
poor outcome (Sillman et al., 1992).
184

Conceptual Model of Strength Increases After a Mixed Lesion

Injury
normal IE
to- Axonal Regeneration
Strength
+- MuscleFiber Hypertrophy
+-Olstal Axon Sprouting

+- Resolution or Conduction Block

I
time 18 months

Fig. 2. Conceptual representation of mechanisms for increases in strength after a mixed lesion of a peripheral nerve.
The processes represented are not temporally distinct, but they may merge. Maximal recovery is usually achieved by
18-24 months.

Complete axonotmesis and neurotmesis have the
worst prognosis. Recovery depends solely upon
axonal regeneration which mayor may not occur,
depending upon the degree of injury to the nerve. In
many cases of complete axon loss it is not possible
to know the degree of nerve injury except by surgical
exploration. The only alternative is to look for
evidence of early reinnervation after the lesion. As a
consequence, it is often recommended to wait 2-4
months and look for evidence of reinnervation in
previously completely denervated muscles near the
site of the lesion (Kline, 1990; Wood, 1998). Those
lesions that have some spontaneous recovery are
usually treated conservatively, since operative repair
is unlikely to improve upon natural recovery. Those
with no evidence of axonal regrowth usually have
operative exploration with possible nerve grafting.
Acknowledgment
Paula Micklesen is thanked for production of figures
for this chapter.

Conceptual Model of Sensory Improvement After a Mixed Lesion

Injury
normal - It'
Sensation
"Axonal Regeneration

+- Redistribution of SensofY Function

+- Resolution of Conduction Block
I
time 18 months
Fig. 3. Conceptual representation of mechanisms for improvement in sensation after a mixed lesion of a peripheral nerve.
The processes represented are not necessarily temporally distinct, but they may merge. Recovery may continue for longer
than 18 months since it is not dependent upon muscle viability.

References
Brandstater, M.E. and Fullerton, M. Sensory nerve conduction
studies in cervical root lesions. Can. 1. Neurol. Sci., 1983, 10:
152.
Brown, M.e., Holland, RL. and Hopkins, A.G. Motor nerve
sprouting. Ann. Rev. Neurosci., 1981, 4: 17-42.
Buchthal, E Fibrillations: c1nical electrophysiology. In: W.J. Culp
and J. Ochoa (Eds.), Abnormal Nerves and Muscle Generators.
Oxford University Press, New York, 1982,632-662.
Campbell, w.w., Pridgeon, RM. and Riaz, G. et al: Sparing of
the flexor carpi ulnaris in ulnar neuropathy at the elbow.
Muscle Nerve, 1989, 12: 965-967.
Cangiano, A., Lutzemberger, L. and Nicotra, L. Non-equivalence
of impulse blockade and denervation in the production of
membrance changes in rat skeletal muscle. J. Physiol. 1977,
273: 691-706.
Chaudry, V. and Cornblath, D.R. Wallerian degeneration in
human nerves: Serial electrophysiological studies. Muscle
Nerve, 1992, 15(6): 687-693.
Cosgrove, J.L., Vargo, M. and Reidy, M.E. A prospective study of
peripheral nerve lesions occuring in traumatic brain-injured
patients. Am. J. Phys. Med. Rehabil., 1989,68: 15-17.
Dillingham, T.R Approach to Trauma of Peripheral Nerves. 1998
AAEM Course C: Electrodiagnosis in Traumatic Conditions -
Course Handout, 7-12. AAEM, Rochester, MN.
Dorfman, L.J. Quantitative clinical electrophysiology in the
evaluation of nerve injury and regeneration. Muscle Nerve,
1990, 13: 822-828.
Durnitru, D. Electrodiagnostic medicine. Hanley and Belfus,
Philadelphia, PA, 1995,341-384.
Erminio, E, Buchthal, E and Rosenfalck, P. Motor unit territory
and muscle fiber concentration in paresis due to peripheral
nerve injury and anterior hom cell involvement. Neurology,
1959,9: 657-671.
Fisher, M.A. Minimonograph no. 13: H reflex and F wave:
physiology and clinical indications. Muscle Nerve, 1992, 15:
1223-1233.
Fowler, T.J., Danta, G. and Gilliatt, RW. Recovery of nerve
conduction after a pneumatic tourniquet: observations on the
hind-limb of the baboon. 1. Neurol. Neurosurg. Psychiatry,
1972,35: 638-647.
Garland, D.E. and Bailey, S. Undetected injuries in head-injured
adults. Clin. Orthop. Relat. Res., 1981, 155: 162-165.
Gilliatt, RW. Acute compression block. In: A.J. Sumner (Ed.),
The Physiology of Peripheral Nerve Disease. Philadelphia,
W.B. Saunders, 1980,287-315.
Gilliatt, R.W. and Hjorth, R.J. Nerve conduction during Wallerian
degeneration in the baboon. 1. Neurol. Neurosurg. Psychiatry,
1972,35,335-341.
Gilliatt, RW. and Taylor, J.e. Electrical chagnes following
section of the facial nerve. Proc. R. Soc. Med., 1959, 52:
1080-1083.
185
Gilliatt, RW., Westgaard, R.H. and Williams, LR. Extrajunctional
acetylcholine sensitivity of inactive muscle fibres in the
baboon during prolonged nerve pressure block. 1. Physiol.,
1978,280:499-514.
Guth, L. Neuromuscular function after regeneration of inter-
rupted nerve fibers into partially denervated muscle. Exp.
Neurol., 1962,6: 129-141.
Haymaker, W. and Woodhall, B. Peripheral nerve injuries. W.B.
Saunders, Philadelphia, PA, 1953.
Hoffman, H. Local reinnervation in partially denervated muscle:
a histophysiological study. Aust. J. Exp. Bioi. Med. Sci., 1950,
28: 383.
Kimura, J., Machida, M. and Ishida, T. et al. Relation between
size of compound sensory or muscle action potential, and
length of nerve segment. Neurology, 1986, 36: 647-652.
Kline, D.G. Surgical repair of peripheral nerve injury. Muscle
Nerve, 1990, 13(9): 843-852.
Kraft, G.H. Fibrillation amplitude and muscle atrophy following
peripheral nerve injury. Muscle Nerve, 1990, 13: 814-821.
MacKinnon, S.E. and Dellon, A.L. Surgery of the peripheral
nerve. Thieme Medical Publishers Inc., New York, 1988.
Massey, 1.M. and Sanders, D.B. Single-fiber EMG demonstrates
reinnervation dynamics after nerve injury. Neurology, 1991,
41(7), 1150-1151.
Miller, R.G. AAEE rninimonograph no. 28: Injury to Peripheral
Motor Nerves. Muscle Nerve, 1987, 10: 698-710.
Milner-Brown, H.S., Stein, R.B. and Lee, RG. Synchrnonization
of human motor units: Possible role of exercise and supraspi-
nal reflexes. EEG Clin. Neurophysiol., 1975,38: 245-254.
Moritani, T. and DeVries, H.K. Neural factors vs. hypertrophy in
the time course of muscle strength gain. Am. J. Phys. Med.,
1979,58: 115-130.
Noble, 1., Munro, e.A., Prasad, V.S.S.v. and Midha, R Analysis
of Upper and Lower Extremity Peripheral Nerve Injuries in a
Population of Patients with Multiple Injuries. J. Trauma, 1998,
45(1), 116-122.
Ochoa, J., Fowler, T.J. and Gilliatt, RW. Anatomical changes in
peripheral nerves compressed by a pneumatic tourniquet. J.
Anat, 1972, 113:433-455.
Partanen, J.V. and Danner, R. Fibrillation potentials after muscle
injury in humans. Muscle Nerve, 1982, 5(9S): S70-S73.
Rasrninsky, M. and Sears, T.A. Internodal conduction in undis-
sected demyelinated nerve fibres. J. Physiol., 1972, 227:
323-350.
Seddon, H.J. Nerve grafting. J. Bone Joint Surg. (Br.), 1963,45:
447-455.
Seddon, H.J. Surgical disorders of the peripheral nerves, 2nd Ed.
Churchill Livingstone, New York, 1975,336, 21-23.
Selecki, B.R, Ring, LT., Simpson, Da.A., Vanderfield, G.K. and
Sewell, M.E Trauma to the central and peripheral nervous
systems: Part II, a statistical profile of surgical treatment in
New South Wales. 1977 Aust. n.z: J. Surg., 1982, 52:
111-116.
186
Sillman, J.S., Niparko, 1.K., Lee, S.S. and Kileny, P.R. Prognostic
value of evoked and standard electromyography in acute facial
paralysis. Otolaryngol. Head Neck Surg., 1992, 107(3):
377-381.
Speidel, C.c. Studies of living nerves: growth adjustments of
cutaneous terminal arborization. J. Camp. Neurol., 1942, 76:
57-73.
Stone, L. and Keenan, M.A. Peripheral nerve injuries in the adult
with traumatic brain injury. Clin. Orthop. Relat. Res., 1988,
233: 136-144.
Sunderland, S. Nerves and nerve injuries, 2nd Ed. Churchill
Livingstone, New York, 1978. 1046 pp., 133-138.
Tackman, W. and Radu, E.W. Observations of the application of
electrophysiological methods in the diagnosis of cervical root
compressions. Eur. Neurol., 1983, 22: 397-404.
Thesleff, S. Physiological effects of denervation of muscle. Ann.
N.Y. Acad. Sci., 1974,228: 89-103.
Trojaborg, W. Early electrophysiological changes in conduction
block. Muscle Nerve, 1978, I: 400-403.
Weddell, G. and Glees, P. The early stages in the degeneration of
cutaneous nerve fibers. J. Anat., 1941,76: 65-93.
Wertsch, 1.1., Sanger, 1.R. and Matloub, H.S. Pseudo-anterior
interosseous nerve syndrome. Muscle Nerve, 1985,8: 69.
Wertsch, 1.1., Oswald, T.A. and Roberts, M.M. Role of intra-
neural topography in diagnosis and localization in
electrodiagnostic medicine. Physical Medicine and Rehabilita-
tion Clinics ofNorth America, 1994,5(3): 465-475.
Wood, M.B. Surgical Approach to Peripheral Nervous System
Trauma. 1998 AAEM Course C: Electrodiagnosis in Traumatic
Conditions - Course Handout. AAEM, Rochester, MN,
27-36.
Yuska, M.A. and Wilbourn, A.1. Incidence of fibrillation
potentials in "pure" conduction block mononeuropathies.
Muscle Nerve, 1998,21: 1572 (abs).