This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Te c h n i c a l Is s u e s
a n d Po t e n t i a l
Complications of
Nerve Conduction
Studies and Needle
E l e c t ro m y o g r a p h y
Devon I. Rubin, MD

KEYWORDS
 Nerve conduction studies  Needle electromyography
 Complications  Pitfalls  Anomalous anatomy
 Anticoagulation  Pacemaker  Pain

Nerve conduction studies (NCS) and needle electromyography (EMG) provide impor-
tant and complementary information as part of an electrodiagnostic study for evalu-
ating patients with suspected neuromuscular disorders. The information obtained
from these techniques helps to define the underlying pathologic changes that may
involve nerves, neuromuscular junctions, or muscle fibers. Although the techniques
are standardized and straightforward in experienced hands, potential technical prob-
lems that are encountered during the studies may interfere with accurate and reliable
acquisition of information and interpretation of the data. Recognition, identification,
and correction of various technical problems are critical to the reliable interpretation
of any electrodiagnostic study.
Each of the techniques is safe and generally associated with only mild, transient
discomfort when performed by experienced physicians. However, NCS involve the
administration of electric current, thereby posing some potential risks in certain
clinical circumstances. Similarly, because needle EMG involves inserting a needle
percutaneously into muscle tissue, potential risks and complications may occur in
rare instances. This article reviews technical aspects that should be considered
to assist in accurate interpretation and safe performance of an electrodiagnostic
study.

Department of Neurology, Mayo Clinic, 4500 San Pablo Road, Jacksonville, FL 32224, USA
E-mail address: rubin.devon@mayo.edu

Neurol Clin 30 (2012) 685–710

doi:10.1016/j.ncl.2011.12.008 neurologic.theclinics.com
0733-8619/12/$ – see front matter Ó 2012 Elsevier Inc. All rights reserved.
 Author's personal copy
686 Rubin

TECHNICAL PROBLEMS DURING NCS

An NCS is performed by (1) placing an electrical stimulator on the skin directly over
a nerve being tested, (2) placing recording electrodes at a distant site along the
same nerve or over a muscle innervated by that nerve, and (3) applying an electrical
stimulus sufficiently strong to depolarize all of the axons within the nerve. Technical
problems may be encountered at each of these steps and, if not identified and cor-
rected, may result in false-positive or false-negative interpretations of the study.
When this occurs, the results may not truly reflect the integrity of the underlying neuro-
muscular anatomy being studied. Identifying technical problems requires a high
degree of acumen and compulsiveness during the performance of the studies as
well as close scrutiny of the waveforms, because the interpreting physician cannot
identify these types of errors simply by reviewing the numerical data. Several types
of technical, physiologic, and anatomic issues that may be encountered during an
NCS are reviewed in this article.

Nerve Stimulation Problems

In order to adequately assess the integrity of a nerve during an NCS, all of the axons
within the nerve being tested must be sufficiently depolarized. Insufficient or submax-
imal stimulation of the axons may occur when a nerve cannot be easily identified or
when technical issues prevent the current administered on the surface of the skin
from penetrating deeply enough through the subcutaneous tissues to depolarize the
nerve in entirety. In contrast, administering excessive electrical current can produce
abnormal responses as a result of current spread and depolarization of neighboring
nerves. Both of these stimulation problems can alter the normal recorded response.

Imprecise nerve localization

During an NCS, the stimulator should be placed as close to the nerve as possible to
ensure maximal stimulation with the lowest possible stimulus intensity and to minimize
current spread to nearby nerves. In most of the common NCS performed, particularly
at distal sites of stimulation such as the wrist or ankle, the nerves are superficial and
readily identified. However, individual anatomic variation, soft tissue edema, limb
deformities, postsurgical changes or scarring, and obesity can impair nerve identifica-
tion. For example, in a patient with a subluxed ulnar nerve or who has had surgical
transposition of the ulnar nerve, the nerve may lie anterior, rather than posterior, to
the medial epicondyle, which may not be readily apparent when performing the study.
The optimal method to identify the location of a nerve is through the sliding tech-
nique, in which the stimulator is placed over the most likely site of the nerve being
studied and the stimulus intensity is increased in small increments until a threshold,
submaximal response is obtained. The stimulator is then moved a few millimeters in
one direction parallel to the nerve without changing the stimulus intensity. A higher
amplitude response indicates closer proximity to the nerve, whereas a lower ampli-
tude response indicates that the stimulator has moved farther away from the nerve.
The stimulator is moved several times in both directions until the site of the maximum
amplitude waveform is identified. At that site, the stimulus intensity is then increased
until a supramaximal response is obtained (Fig. 1).

Reversal of cathode-anode orientation

During routine NCS using a bipolar electrical stimulator, the cathode (the site of nerve
depolarization) is placed directly over the nerve, pointed toward the recording elec-
trodes, with the anode approximately 2 cm distant from the cathode. It is at these
cathode stimulation sites that the conduction velocity (CV) distant measurements or
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 687

Fig. 1. Sliding technique on ulnar nerve conduction at the elbow. Note the different ampli-
tudes with the same stimulation intensity at different sites at the elbow. The highest ampli-
tude (at the fourth trace) indicates the site where the stimulator is closest to the nerve.

measurements for distal latency calculation are made. If the cathode and anode are
inadvertently reversed, the nerve is depolarized approximately 2 cm farther away
from the recording electrode, producing several technical abnormalities:
 Inaccurate CV. Because the site of nerve depolarization occurs approximately 2
cm farther than the suspected cathode site, reversal of the orientation at 1 of the
2 sites during assessment of CV along a nerve segment can produce an errone-
ously calculated CV (falsely slower or faster than normal, depending on whether
the reversal occurred at the distal or proximal stimulation site).
 Falsely prolonged distal latency. When the cathode-anode reversal occurs at
a distal stimulation site, the distance between the site of nerve depolarization
and the recording electrode is 2 cm longer than expected; therefore, the distal
latency may be prolonged by approximately 0.4milliseconds (Fig. 2).
 Anode block. A theoretic block of conduction of some of the fibers may occur as
a result of hyperpolarization of axons at the anode. Anode block can lead to
a higher stimulus intensity, which is necessary to depolarize the axons, and the
excess current may spread to adjacent nerves, as described later.
Methods to eliminate inadvertent reversal of cathode and anode are listed in Box 1.

Submaximal nerve stimulation

The compound muscle action potential (CMAP) or sensory nerve action potential
(SNAP) amplitude reflects the number and integrity of functioning axons. Appropriate
interpretation and quantitation of an NCS requires stimulation of all axons within
a nerve. If the nerve is not maximally stimulated, the number of conducting fibers is
underestimated and the recorded amplitude may be falsely low, mimicking disease.
Several physiologic factors, such as limb edema, fibrosis, obesity, or deep nerves
at proximal sites may contribute to understimulation. In addition, high stimulator
 Author's personal copy
688 Rubin

Fig. 2. Median sensory NCS with reversal of the stimulator cathode and anode causing a pro-
longed distal latency and inaccurate CV (top). Normal stimulus orientation (bottom).
(Normal distal latency is <3.6 milliseconds.) DL, distal latency.

impedance also leads to submaximal stimulation. Submaximal stimulation may lead to

the following findings (Fig. 3):
 Falsely low CMAP or SNAP amplitude. This finding may be misinterpreted as
indicating a neuromuscular disorder. The occurrence at a proximal but not distal
stimulation site may mimic a focal conduction block or anomalous innervation.
 Falsely prolonged distal latency or slowed CV. In a submaximally stimulated
nerve, some of the large-diameter, faster conducting fibers are not depolarized.
As a result, the maximal recorded CV is slower than normal and could mimic
a focal mononeuropathy (eg, ulnar neuropathy at the elbow or median neurop-
athy at the wrist).

Box 2 lists methods that can be used to minimize or correct for submaximal
stimulation.

Box 1
Methods to eliminate reversal of cathode and anode

 Ensure that the cathode is clearly noted on the stimulator (usually with a mark or polarity
sign)
 Observe the stimulator orientation before stimulation with every NCS
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 689

Fig. 3. Tibial motor (abductor hallucis recording) NCS. (A) Understimulation at the knee
showing a low CMAP amplitude and slower CV. (B) Supramaximal stimulation at the knee.

Overstimulation of the nerve producing current spread to other nerves

An electrical stimulus applied to the skin overlying a nerve produces current flow
through the extracellular tissue surrounding the nerve. Because these tissues have
the ability to conduct electrical charge, increasing stimulus intensity increases the
spread of the charge over a wider area (volume conduction). Excess stimulus intensity
may result in spread of current to, and inadvertent depolarization of, neighboring
nerves. Common sites of occurrence include areas where different nerves are in close
proximity to each other and where a higher amount of current may be needed because
of a deeper nerve location, such as the tibial nerve at the knee with spread to the pero-
neal nerve (and vice versa), median and ulnar nerves at the elbow, facial nerve stimula-
tion (with spread to trigeminal), and at the Erb point during brachial plexus stimulation.
Overstimulation and current spread leads to several potential problems (Fig. 4):

 Erroneously higher amplitudes. This problem is caused by the summation of

responses from more than 1 nerve and muscles supplied by the additional nerve

Box 2
Methods to eliminate submaximal nerve stimulation

 The sliding technique to localize the nerve.

 Reduction of stimulator impedance with conducting paste.
 Closely observing the waveform following each stimulus for increasing amplitude with
increasing stimulus intensity. If the amplitude increases without a significant change in
waveform morphology, additional axons have been depolarized.
 Supramaximal stimulation is ensured when a 10% to 20% (usually 5–10 mA) increase in
stimulus intensity produces no further increase in amplitude.
 Incrementally increasing the duration of the stimulus (eg, from 0.2 milliseconds to 1
millisecond) if a supramaximal response is not obtained with the maximal stimulus intensity
(eg, 100 mA).
 Using a stimulator with a larger cathode and anode separation, a monopolar stimulator, or,
in some cases, a monopolar needle stimulator for deeper nerves (eg, the tibial at the knee).
 Author's personal copy
690 Rubin

Fig. 4. The effect of overstimulation of the median nerve at the wrist during a median
motor NCS. (Top) Submaximal nerve stimulation producing depolarization of less than
100% of muscle fibers from the abductor pollicis brevis (APB). (Middle) Maximal stimulation
of the median nerve resulting in complete depolarization of 100% of APB muscle fibers.
(Bottom) Overstimulation of the median nerve with spread to the ulnar nerve resulting in
depolarization of 100% of median fibers and the addition of some ulnar-innervated muscle
fibers in the region of the thenar eminence.

that is depolarized. The higher-than-normal amplitude may lead to a false-

negative interpretation (interpreting the study as normal, when the actual
response would be abnormally low). There is also typically a change in waveform
morphology with overstimulation.
 Inaccurate onset latency and CV.
 False identification of anomalous anatomy. The findings may mimic anomalous
innervation, such as a median-to-ulnar anastomosis (from overstimulation at
the elbow) or accessory peroneal nerve (from overstimulation at the knee).
 False-negative repetitive stimulation. The current may spread directly to the
muscle, bypassing the neuromuscular junction, and lead to a false-negative
repetitive stimulation study in a patient with a neuromuscular junction disorder.
Clues to identify, and methods used to reduce, overstimulation are listed in Box 3.

Nerve Recording Problems

Appropriate application of the recording electrodes during an NCS is important to
obtain an accurate and maximal response from the nerve or muscle. Problems related
to the recording electrode size and placement can lead to abnormally recorded
responses.
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 691

Box 3
Clues to identify, and methods to reduce, overstimulation

Clues
 Higher CMAP amplitude and area at proximal site of stimulation compared with distal site.
 Different waveform morphology between the proximal and distal sites of stimulation.
 Initial positive deflection (indicating volume conduction from a distant source).
Methods to reduce
 Reduce impedance with appropriate skin preparation (abrade the skin and use of electrode
paste).
 Use the sliding technique to localize the nerve.
 Increase the stimulus intensity by small increments (ie, 5–10 mA) with each shock.
 Observe the waveform morphology following each stimulus increment. The point at which
the morphology does not change with 1 or 2 small increases in intensity is the supramaximal
point.
 Observe the muscle contraction to determine correct nerve stimulation (eg, dorsiflexion with
peroneal nerve stimulation, plantar flexion with tibial nerve stimulation).

Electrode size
The recording electrodes should be large enough to record the action potential from all
of the muscle fibers or nerves that are depolarized. Electrodes that are too small rela-
tive to the muscle may produce a lower response than normal because they record
from a small portion of the muscle. For most NCS, electrodes 2 to 10 mm in diameter
are sufficient. For large muscles or muscles such as the anterior tibialis or biceps,
larger recording electrodes (2 cm or greater in diameter) should be used.

Electrode placement during motor NCS

Incorrect placement of either the active or reference electrode can alter the
morphology of the waveforms and affect the recorded parameters (distal latency,
amplitude, and CV). When recording over a muscle during a motor NCS, the active
(G1) electrode should be placed directly over the muscle end-plate zone, whereas
the reference (G2) electrode is placed at a variable but defined site, usually along
the tendon of the muscle. The motor end-plate region is the site of origin of action
potential generation along the muscle fibers following nerve stimulation. Once initi-
ated, the action potentials travel along the fibers away from the end-plate zone.
When G1 electrode placement is over the end-plate region, the initial waveform
deflection is upward (negative). An initial downward (positive) deflection, or positive
dip, indicates that the recorded muscle fiber action potentials have originated at
a distance from the G1 electrode and are traveling toward, rather than away from,
the electrode through the extracellular tissue (volume conduction). This positive initial
deflection is present with all sites of stimulation (Fig. 5).
Although a positive dip is the most characteristic feature of G1 electrode placement
off the end-plate region, a low CMAP amplitude, sometimes with an atypical waveform
morphology, can also occur as a result of incorrect G1 placement. Therefore,
a problem with G1 electrode placement should always be considered when an unex-
pectedly low CMAP amplitude is recorded (Fig. 6). Clues to identify, and methods to
correct for, improper G1 electrode placement are listed in Box 4.
The placement of the G2 electrode also has an effect on the recorded CMAP.1–3
Because the recording is performed using a differential amplifier, the difference in
 Author's personal copy
692 Rubin

Fig. 5. Median motor NCS (abductor pollicis brevis recording) with incorrect G1 placement.
(A) CMAP amplitude is reduced and a positive initial deflection is present at the wrist and
elbow. (B) The normal response after the G1 electrode was moved.

Fig. 6. A 40-year-old patient with leg pain and a normal examination. (A) The tibial CMAP
amplitude is low (normal amplitude at knee >4.0 mV). (B) After moving the G1 electrode,
the amplitude nearly doubles.
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 693

Box 4
Clues to identify, and methods to reduce, improper G1 electrode placement

Clues
 Initial positive deflection from baseline at all stimulation sites (proximal and distal)
 Unexpectedly low CMAP amplitudes
 Unusual CMAP waveform morphology
Methods to correct
 Know nerve conduction setup based on laboratory technique and reference values
 Carefully identify the muscle to be recorded before applying electrodes
 If initial positivity or unexpectedly low amplitude is observed, systematically move G1 until
initial positivity disappears or amplitude becomes maximal
 Ensure that G1 and G2 inputs are correctly placed in the preamplifier

the electrical fields between the G1 and G2 electrodes is amplified. Although, in most
standard motor NCS, the G2 electrode is placed over the muscle tendon, the G2 elec-
trode is not electrically silent because it records a volume-conducted response trav-
eling along the fibers toward the electrode. Therefore, the responses that are recorded
and interpreted consist of the electrical fields recorded from each electrode. If the G1
and G2 electrodes are placed too close together, the responses obtained with each
electrode will be similar, thereby reducing the recorded CMAP amplitude.
The normal effect of the G2 electrode on the CMAP morphology, even with place-
ment over the muscle tendon, is greater with ulnar and tibial waveform morphologies
than with other nerves. This difference results from the close proximity of ulnar or tibial
innervated muscles in the hand or foot to the G2 electrode. As a result, the G2
recorded response contributes more to the final resulting CMAP than with other
nerves where there are fewer muscles near the G2 electrode.1–3 This contribution of
the recorded potentials from intrinsic foot muscles recorded by the G2 electrode is
the likely cause of the normal reduction in CMAP amplitude between the ankle and
knee of up to 50% during the tibial motor NCS.3 When interpreting the studies accord-
ing to reference values that use standard recording montages, the effect of the G2
electrode placement is inherent in the normative data. Care must be taken to ensure
that the same setup is used during each NCS that is used for the reference values.

Limb movement during the study

Once the electrodes have been applied over the muscle during motor NCS, any move-
ment of the muscle or limb can shift the position of G1 and the relationship with the
end-plate zone, thereby altering the morphology of the waveform. Careful attention
to ensure that the limb position remains stable through each conduction study will
reduce this problem.

Electrode placement during sensory NCS

Sensory NCS record a traveling volley of the summated sensory fiber action potentials
traveling toward the recording electrodes along the nerve. During sensory NCS, the
recording electrodes are placed as close as possible to the nerve to obtain the
maximal response. The farther away the G1 recording electrode is from the action
potential generator, the lower the amplitude of the response. This finding is particularly
important in sensory studies, in which the amplitudes are much lower than motor
 Author's personal copy
694 Rubin

responses. The G1 and G2 electrodes should be placed 3.5 to 4 cm apart to maximize

the summation of the potential as recorded with a differential amplifier. If the 2 elec-
trodes are less than 3 cm apart, the SNAP amplitude decreases (Fig. 7).
Motor response interference with SNAP
Another source of error in sensory studies is mistaking a volume-conducted motor
response for a sensory response. This error is most commonly seen in antidromic
studies (particularly the ulnar) in which a mixed motor-sensory nerve is stimulated
while recording over the sensory branches in a digit (Fig. 8). The volume-conducted
motor response may interfere with or distort the sensory potential. When it is unclear
whether a recorded waveform is a true sensory response or a volume-conducted
motor response, moving the G1 recording electrode a short distance away from the
stimulator will slightly prolong the latency of the sensory response, whereas the
latency of the volume-conducted motor response will not change. Several methods
can be used to correct for motor interference of the sensory response (Box 5).4

Distance Measurement Errors

During most NCS, the latency from the distal site of stimulation to the recording elec-
trodes (distal latency) and the CV in a nerve segment are assessed. Both of these
parameters rely on accurately measuring the length of the nerve between the stimula-
tion and recording sites (Fig. 9). The measuring tape should always follow the course
of the nerve with the limb remaining in the position tested. This rule applies especially
in ulnar NCS when the study is performed with the arm in a flexed position. If the
distance is measured with the arm extended after the stimulation is performed with
the arm flexed, a falsely shorter distance than the actual length of the nerve will be
measured, resulting in an apparently slower CV.
The result of a measurement error is proportional to the distance between the 2
stimulation sites, with a greater error the shorter the distance. For example, in a nerve
with a CV of 50 m/s, a 5-mm error in distance measurement over a distance of 20 cm
will produce a 2.4-m/s error in velocity. The same measurement error over a distance
of 10 cm will produce a 4.8-m/s error in velocity. Methods to ensure correct distance
measurements are listed in Box 6.

Fig. 7. Median antidromic sensory NCS with reduced interelectrode distance between G1
and G2 electrodes.
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 695

Fig. 8. (A) Ulnar sensory antidromic NCS showing prominent motor artifact (arrows) that
mimics a sensory response. (B) Reduction in the motor artifact with movement of electrodes
more distal on the digit. M-P, metacarpal-phalangeal; PIP, proximal interphalangeal joint.

Cool Limb Temperature

One of the more common pitfalls in the performance of NCS is not measuring limb
temperature and performing a study on a cool limb. Cool limb temperatures can signif-
icant alter most NCS parameters and mimic focal or diffuse disease of the nerves.
Lower temperatures affect the channel kinetics along the nerve by prolonging the
duration of opening of the sodium and potassium channels and slowing electrotonic
spread of the potential along the nerve.5,6 These physiologic changes lead to (Fig. 10):
 Higher amplitude and prolonged durations of CMAP and SNAP responses
 Prolonged distal latencies (by approximately 0.2 milliseconds per  C decrease)
 Slowed CV (by 1.8–2.0 m/s per  C decrease)

Box 5
Methods to correct for motor interference of SNAP

 Use the smallest current necessary to produce a supramaximal response

 Move recording electrodes as far away as possible from the muscle groups generating the
motor response (eg, place G1 near proximal interphalangeal joint rather than metacarpal-
phalangeal joint on the ring finger during ulnar antidromic NCS)
 Shield the recording electrodes with gauze
 Author's personal copy
696 Rubin

Fig. 9. Ulnar motor study showing differences in CV with different distance measurements
between the wrist and elbow.

 Reduction in degree of conduction block (when present)

 Repair of neuromuscular transmission defect (lessens degree of decrement on
repetitive nerve stimulation [RNS]).
The slowing of CVs or prolongation of the distal latencies that may occur with cool
limb temperature can often lead to the erroneous interpretation of a study as indicating
peripheral neuropathy, carpal tunnel syndrome, or ulnar neuropathy. Factors such as
muscle atrophy, exposure to cold, and metabolic disorders such as hypothyroidism

Box 6
Methods to ensure correct distance measurement

 Mark site of stimulator cathode with a pen immediately after nerve stimulation (before the
stimulator is removed from the skin)
 Hold the tape measure along the course of the nerve during measurement
 Use calipers for proximal nerve segments (eg, distance around shoulder or neck for brachial
plexus or root stimulation)
 To ensure reliability, repeat measurement if an unexpected CV is obtained
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 697

Fig. 10. Median antidromic sensory study. (Top) Performed with cool limb temperature
causing a prolonged distal latency, mimicking carpal tunnel syndrome. (Bottom) After
warming (normal distal latency <3.6 milliseconds).

may predispose a limb to cooling. In our laboratory, limb surface temperature should
be greater than 32 C on the hand and greater than 30 C on the foot. Although the
degree of slowing of CV per  C decrease in temperature is known, the relationship
is not linear and therefore using a simple correction factor is not reliable, so maintain-
ing warm limb temperature should be performed in all studies (Box 7).
Cool limb temperature can also affect repetitive stimulation studies, producing
a false-negative study. Because cooler temperature prolongs the open time of the
acetylcholine receptors and lowers the activity of acetylcholinesterase at the neuro-
muscular junction, thereby increasing the safety margin of neuromuscular transmis-
sion, the degree of decrement on repetitive NCS may be less in a cooler limb
(comparable with the clinical ice-cube test).7 Warming the limb has been shown to
increase the yield of repetitive NCS.8

Technical Problems During RNS

The technical problems that occur during RNS include all of the issues related to
routine NCS described earlier, but include another potential problem related to the
need to ensure stability of the responses obtained following each of multiple, rapidly

Box 7
Methods to correct for a reduction in limb temperature

 Accurately measure limb temperature over the dorsum of the hand or foot during all studies.
 Warm a cool limb with a heat lamp, warm bath (10 minutes at 40 C), or heating pads. This
procedure often improves the velocity and latencies and normalizes the study.
 Author's personal copy
698 Rubin

administered stimuli. With RNS technique, multiple sequential supramaximal stimuli

are given at rates of 2 to 50 Hz. With each stimulus, the stimulator and recording elec-
trodes must remain in a fixed position over the nerve and muscle, respectively. Move-
ment of the stimulator off the nerve, change in the position of the recording electrodes
over the muscle, or contraction of the limb all may alter the CMAP waveform between
stimuli and produce false decrement. In most RNS studies, any decrement observed
should first be considered the result of technical factors unless the typical physiologic
features are shown. Characteristic features of true, physiologic decrement from
a defect of neuromuscular transmission and patterns indicating false decrement
caused by technical factors are listed in Table 1 and shown in Fig. 11. Several
methods can be helpful to minimize false decrement during RNS (Box 8).

Anatomic Nerve Variations with Anomalous Innervation

Variations in peripheral nerve anatomy are common and failure to recognize anoma-
lous anatomy may lead to erroneous diagnoses when a study may be normal. The
most commonly encountered anomalous variations in the arm and leg are the
median-to-ulnar anastomoses (Martin-Gruber anastomosis [MGA], median-to-ulnar
crossover) and accessory peroneal nerve, respectively. Although a detailed review
of anomalous innervation is beyond the scope of this article, the main findings of
MGA and accessory peroneal nerve are briefly reviewed.

MGA
The MGA is a common variation that is present in 15% to 31% of individuals and is bilat-
eral in up to 68% of individuals.9,10 This anatomic variation consists of a communication
between the median and ulnar nerve fibers in the forearm, whereby fibers that are
destined to supply ulnar-innervated muscles course through the median nerve in the
upper arm and proximal forearm, and cross over to the ulnar nerve in the forearm before
innervating the destined muscles. The fibers may branch off from the median nerve
proper or the anterior interosseus branch.11 Sensory fibers are not involved. In rare
instances, the origin of the crossing over fibers may be more proximal and located
above the elbow, thereby mimicking an ulnar neuropathy at the elbow on NCS.12
The muscles supplied by the crossing over fibers vary among individuals and
include 1 or more of the following: (1) first dorsal interosseus (FDI), (2) abductor digiti

Table 1
Characteristic features of true and false decrement on RNS

Features of True Decrement Features of False Decrement

Stable baseline without baseline noise Fluctuating or unstable baseline
Tapering pattern of amplitude reduction Higher amplitude with second or subsequent
stimuli than with first stimulus or
increasing amplitude from second to
fourth stimuli
Largest decrease in amplitude/area between Largest decrease in amplitude/area between
first and second stimuli stimuli other than first and second
Repair of degree of decrement immediately Worsening of decrement immediately
following 10 s of exercise following 10 s of exercise
Reproducible degree of decrement with Markedly different degrees of decrement
repeated trials with repeated baseline trials at short
intervals (eg, 20% decrement at trial 1, 3%
decrement 60 s later)
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 699

Fig. 11. (A) False decrement during repetitive stimulation at 2 Hz caused by limb movement.
(B) True decrement in a patient with myasthenia gravis.

minimi (ADM), (3) adductor pollicis, or (4) flexor pollicis brevis. In approximately half of
individuals with MGA, only 1 muscle is innervated by the crossing over fibers (FDI>ad-
ductor pollicis>ADM).10,13–15 The patterns of findings seen on routine median and/or
ulnar motor NCS in the presence of a MGA depend on which muscle(s) the crossing
over fibers innervate.

Fibers supply the ADM

This type of MGA should be suspected during a routine ulnar motor NCS recording
from the ADM, when the amplitude recorded with wrist stimulation is more than
20% higher than the response recorded with elbow stimulation. This CMAP amplitude
reduction may simulate an ulnar neuropathy with a partial focal conduction block.
Because the MGA fibers cross over to the ulnar nerve in the midforearm, stimulation
of the ulnar nerve at the below-elbow site 5 cm distal to the medial epicondyle (prox-
imal to the crossover fibers) yields an amplitude that is similar to the above-elbow
stimulation site, in contrast with ulnar neuropathies at the elbow where the reduction
occurs between the below-elbow and above-elbow stimulation sites. This type of
MGA can be confirmed by stimulating the median nerve at the wrist and elbow while
recording over the ADM (Fig. 12, Table 2).

Fibers supply FDI or flexor pollicis brevis

In this type of MGA, the crossing over fibers supply ulnar muscles that are located in
the hand adjacent to the thenar region (FDI, adductor pollicis, or abductor pollicis bre-
vis). This pattern of MGA is identified on a routine median motor NCS recording from

Box 8
Methods to minimize falsely abnormal decrement during RNS

 Perform 2 to 3 baseline studies at rest before exercise to ensure consistent results and
technical reliability
 Stabilize the limb with wooden boards or Velcro straps to prevent movement of the limb
during stimulation
 Ensure supramaximal nerve stimulation
 Author's personal copy
700 Rubin

Fig. 12. MGA to the ADM (anastomosing fibers denoted by dashed line).

the thenar eminence. When the median nerve is stimulated at the elbow, in addition to
recording from the usual median-innervated thenar muscles, the electrodes over the
abductor pollicis brevis (APB) also record a volume-conducted response from the
nearby ulnar-innervated muscles that are supplied through the crossing over fibers.
This situation results in a higher-than-normal CMAP amplitude caused by the summa-
tion of the additional muscle fiber action potentials of the adductor pollicis, flexor pol-
licis brevis, and/or FDI. When the median nerve is stimulated at the wrist, distal to the
crossover, the recorded response is a pure median response (Fig. 13, Table 3).
A distinctive pattern is seen in up to 20% of individuals with a type II MGA and
a median neuropathy at the wrist (carpal tunnel syndrome).16,17 With this combination,
an initial positive deflection may be seen in the CMAP waveform with median nerve
stimulation at the elbow, whereas, with wrist stimulation, no initial positivity is present
(Fig. 14), resulting from focal slowing of the median fibers at the wrist, which delays the
conduction of the true median fibers innervating the median muscles (eg, APB). As
a result of this median fiber slowing, the crossing over fibers that supply the ulnar
muscles in the thenar region conduct faster than the slowed median fibers and the
response recorded from the crossing over fibers precedes the response from the
true median fibers. The finding of a positive deflection with elbow stimulation has rarely
been described in patients with carpal tunnel syndrome with otherwise normal

Table 2
Findings of MGA to ADM

Nerve Stimulated Recording Site Findings

Ulnar ADM CMAP with above-elbow stimulation >20% lower than
wrist stimulation
CMAP with below-elbow stimulation >20% lower than
wrist stimulation (below-elbow and above-elbow
responses are similar)
Median APB Normal results
Median ADM Performed to confirm MGA
Initial negative CMAP obtained with elbow stimulation
(normal individuals have no response)
No response with wrist stimulation

Abbreviation: APB, abductor pollicis brevis.

 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 701

Fig. 13. MGA to the ulnar muscles in the thenar region (anastomosing fibers denoted by
dashed line).

conduction studies.17 In patients with more severe carpal tunnel syndrome and
a higher degree of slowing in the median nerve, the response recorded from the
crossing over fibers may be seen as a completely separate waveform from the true
median waveform.18

Accessory peroneal nerve

In most individuals, the extensor digitorum brevis (EDB) is supplied by a branch of the
deep peroneal nerve. In up to 28% of the population, the axons supplying the EDB
travel within the superficial peroneal nerve and course around the lateral malleolus
before innervating the EDB (termed the accessory peroneal nerve).9,19 The accessory
peroneal nerve innervates the fibers in the EDB muscle to a variable degree. In most
cases, only a percentage of fibers (usually the lateral fibers) are supplied by the acces-
sory branch, with the medial fibers supplied directly by the deep peroneal nerve,
although in some cases the accessory peroneal nerve supplies the entire muscle.

Table 3
Findings of a MGA to the FDI/adductor pollicis/flexor pollicis brevis (thenar region muscles)

Nerve Stimulated Recording Site Findings

Median APB Higher CMAP amplitude with above-elbow than
wrist stimulation
Change in CMAP morphology between the
above-elbow and wrist stimulation sites
Ulnar ADM Normal results. CMAP amplitude of above-elbow
stimulation slightly lower than wrist stimulation
Ulnar APB CMAP recorded at wrist and elbow stimulation,
with larger amplitude at wrist stimulation (this
is also seen in individuals without MGA)
 Author's personal copy
702 Rubin

Fig. 14. MGA with superimposed carpal tunnel syndrome. The positive deflection with
elbow stimulation reflects CMAP originating from the adductor pollicis and deep head of
flexor pollicis brevis that are supplied through the crossing fibers.

An accessory peroneal nerve is identified by a higher CMAP amplitude (recorded from

the EDB) with knee stimulation compared with ankle stimulation, and is confirmed by
the presence of a recorded response from the EDB with stimulation behind the lateral
malleolus (Box 9, Fig. 15). A higher CMAP amplitude at the knee than the ankle can
also occur from overstimulation at the knee, and observation of the appropriate
muscle twitch to ensure that current spread to the tibial nerve is not occurring is
important.
In most situations, this finding is of no clinical significance. However, in lesions of the
deep peroneal nerve, clinical and electrophysiologic sparing of the toe extensors
could occur in the presence of an accessory peroneal nerve. Also, in laboratories
where the peroneal nerve is stimulated at the ankle before the knee, caution should
be used when a low CMAP amplitude is obtained, which could be misinterpreted as
the result of underlying abnormality, and stimulation at the knee should always be
performed.

Other anomalous anatomy

Other anatomic variations can produce confusing findings on NCS. One rare anomaly
in the hand is the Riche-Cannieu anastomosis (all ulnar hand), in which all of the thenar
muscles are supplied by the ulnar nerve.20,21 This anomaly produces an absent

Box 9
NCS findings of an accessory peroneal nerve

 Higher CMAP amplitude with stimulation at the knee and fibular head than with stimulation
at the ankle
 Stimulation behind the lateral malleolus produces an initially negative CMAP recorded from
the EDB
 CMAP amplitude recorded from lateral malleolus stimulation typically equals the difference
between the knee and ankle CMAP amplitudes
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 703

Fig. 15. Accessory peroneal nerve. Higher peroneal CMAP amplitude with knee stimulation
compared with ankle stimulation. Stimulation behind the lateral malleolus produces
a CMAP.

median motor response from the APB with median nerve stimulation at the elbow and
wrist. This anastomosis should be suspected when a patient has an absent median
motor CMAP response despite normal thenar muscle bulk and strength and normal
findings on needle examination in the thenar muscles. The Riche-Cannieu anasto-
mosis is confirmed by the presence of a normal CMAP response, with an initially nega-
tive deflection, following stimulation of the ulnar nerve and recording from the APB.
Another uncommon anomalous sensory innervation involving the hand is superficial
radial sensory innervation to the dorsal, ulnar portion of the hand, which is normally
supplied by the dorsal ulnar cutaneous (DUC) branch of the ulnar nerve. This anomaly
has been identified in approximately 16% of normal individuals.22 Consideration and
testing for this anomaly is important in situations in which electrodiagnostic testing is
performed to localize the site of an ulnar neuropathy and a low or absent DUC
response is obtained. In this situation, checking for anomalous radial innervation
can be performed by moving the stimulator to the radial nerve while recording over
the dorsal ulnar hand.

RISKS OF NCS
Performance in Patients with Pacemakers or Cardiac Defibrillators
Because the performance of NCS requires the administration of a variable amount of
electric current across the skin and into neighboring tissue, there is a theoretic risk
when performing the studies in patients with pacemakers or cardiac defibrillators. In
these patients, if the electric current administered reached the cardiac device, it could
possibly be interpreted as a cardiac conduction signal and inhibit or trigger the device,
leading to abnormal pacing or reprogramming of the device.23–25 In some instances,
repeated stimuli could potentially induce cardiac arrhythmias. Despite these theoretic
concerns, studies of routine NCS in patients with implanted cardiac devices during
 Author's personal copy
704 Rubin

continuous monitoring of the electrocardiogram and interrogation of the devices found

that the electrical signals from NCS (with stimulation up to 100 mA and 0.5 millisec-
onds stimulus duration) of the leg, arm, and, in some patients, at the Erb point,
were not sensed by, and never affected, the programming of the devices.25 These
studies suggested that NCS are safe in this patient population. The safety of NCS in
patients with temporary transvenous cardiac pacemakers has not been studied and
NCS should be avoided in patients with these types of pacemakers.

Performance in Patients with Peripheral Intravenous Lines

Similar to the risk to patients with cardiac devices, there is a theoretic risk of current
flow to cardiac structures in patients with peripheral intravenous lines. With standard
NCS using surface stimulation, the skin and subcutaneous tissues have a high resis-
tance to current flow. However, patients with intravenous lines have a breach in the
high-resistance tissue, possibly allowing current to flow through the intravascular fluid
to cardiac tissue. In a study of NCS in 20 patients with peripheral intravenous lines and
implanted cardiac devices that could monitor electrical activity in the heart, surface
electrocardiograms never detected the peripherally administered current and there
was no interference on the settings or function of the cardiac devices.26 In this study,
NCS were performed in the same limb as the intravenous lines, with stimulation per-
formed at the wrist and elbow sites administering up to a 100-mA current with a dura-
tion of up to 0.5 milliseconds. Although only a small number of subjects were studied,
these findings suggest that the performance of routine NCS on patients with peripheral
intravenous lines is safe.
Studies have not been performed assessing the safety of NCS in patients with
central venous lines or devices with external wires that are in close proximity to the
heart. Although the risk of current flow through these structures may be low with
routine NCS and stimulation in the leg or the distal arm, caution should be used and
appropriate monitoring for cardiac conduction changes should be undertaken when
NCS are performed in these patients. Similar cautions should be used if stimulation
at proximal arm sites or Erb point is considered.

Needle Stimulation
In rare instances when supramaximal stimulation of a nerve located deep in a limb
cannot be achieved with surface stimulation, near-nerve stimulation using a monopolar
needle may be performed. When this technique is performed, the needle electrode is
slowly advanced until it is in close proximity to the nerve to allow for nerve stimulation
with as little current as possible. Although the risk of current administration through
a needle is low, a theoretic risk of current spread to neighboring tissues is present.
Because it is unlikely that needle stimulation would be performed on a nerve near
the heart, there is little, if any, practical risk with this technique. However, a risk of
hematoma formation or injury to structures within the pathway of the needle electrode
is more common. For example, injury to the brachial plexus or the development of
pneumothorax with puncture of the apex of the lung could occur with attempted nee-
dle stimulation of the brachial plexus at the Erb point, therefore needle stimulation at
the Erb point should be avoided.

TECHNICAL PITFALLS OF NEEDLE EMG

Filter Settings
The typical frequency range of most of the EMG waveforms recorded during needle
EMG is approximately 2 Hz to 30 kHz. To optimize the recorded signals with minimal
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 705

distortion and to minimize nonphysiologic noise, the filters are usually set at 20 to 30
Hz low-frequency filter (LFF) or high-pass filter and 10 to 20 kHz high-frequency filter
(HFF) or low-pass filter during most routine EMG studies. Alterations in the filter
settings have an effect on the motor unit potential (MUP) morphology and can lead
to an increase or decrease in the recorded duration of the MUP. Increasing the LFF
reduces the amplitude and filters out more of the slower frequency tail components
of the MUP, resulting in a reduction in the MUP duration. Higher LFF settings, typically
of 500 to 2000 Hz, are used during single-fiber EMG. Although the LFF settings are not
usually altered on the EMG equipment during routine studies, if unexpectedly short-
duration MUPs are recorded in all muscles during a study, the examiner should always
check to ensure that an inadvertent change in the filter settings has not occurred.

Needle Movement Problems

Reliable assessment of a muscle during needle EMG requires moving the needle
through the muscle to different locations. Several problems can occur related to nee-
dle movement during needle EMG.

Too few sites examined

Neuromuscular diseases may involve axons or muscle fibers unequally. As a result,
the electrical signals may be normal or minimally abnormal in one portion of a muscle
but more severely abnormal in a different region. For example, dermatomyositis fibril-
lation potentials and short-duration MUP may be more prominent in superficial layers
of the muscle, or peripheral nerve vasculitis may affect certain nerve fascicles more
than others, thereby affecting motor units innervating different portions of a muscle
unequally. Examination of only a single area of muscle may result in not identifying
patchy abnormalities and an abnormal study may be interpreted as normal. This
potential pitfall can be minimized by moving the needle through several different tracks
during examination of each muscle.

Too rapid needle movements with minimal pauses

During an EMG study, a muscle is examined at rest to evaluate for abnormal sponta-
neous activity such as fibrillation potentials. Fibrillation potentials may fire at various
frequencies with some firing as slow as 0.5 Hz, such as in some myopathies. Similarly,
fasciculation potentials fire randomly and often infrequently. If needle movements are
made quickly and without pausing for 0.5 to 1 second following each movement
during examination of a resting muscle, slow-firing spontaneous discharges may be
missed.
Too large needle movements
Adequately assessing multiple areas of the muscle with the needle electrode relies on
patient tolerance and cooperation of the study. The technique and size of each needle
movement can significantly affect study tolerance. The number of muscle fibers
damaged by the needle electrode and the pain experienced during needle EMG corre-
late with the size of needle movements. Larger needle movements (>1 cm) have been
shown to be more painful than small needle movements (<1 mm).27 Electromyogra-
phers should perform each study using needle movements that are as small as
possible to maximize patient tolerance.

Artifacts
Problems with interpretation of needle EMG findings may be encountered when elec-
trical artifacts similar to pathologic waveforms are recorded. For example, a cardiac
pacemaker artifact is recorded as a single spike, usually firing in a regular pattern at
 Author's personal copy
706 Rubin

a rate of approximately 1 Hz. The regular firing pattern may be confused with, and
interpreted as, a fibrillation potential. The pacemaker artifact is most often recorded
in muscle closer to the pacemaker, such as the paraspinal muscles or proximal upper
extremity muscles, but can be recorded from any muscle. It can often be distinguished
from a fibrillation potential because it does not change or disappear with needle move-
ment and the same spike firing at the same rate is found in more than 1 muscle. In
addition, as a result of the square wave morphology of the pacemaker artifact, a pace-
maker artifact has an artificial sound that is subtly different from a fibrillation potential.
Other artifacts can sometimes be confused with pathologic waveforms, such as 60-
cycle or fluorescent light artifact, mimicking a complex repetitive discharge or an inter-
mittently firing transcutaneous electrical nerve stimulator (TENS) or other electrical
stimulator, mimicking a myokymic discharge. These electrical artifacts typically fire
in a pattern that is more regular than even the regular patterns of physiologic wave-
forms, giving the waveforms artificial sounds.

POTENTIAL RISKS DURING NEEDLE EMG

Needle EMG is generally a safe procedure when performed by experienced physi-

cians. However, because the technique involves insertion of a needle through the
skin and up to several centimeters into a muscle, potential complications may rarely
occur.28 Risks related to the performance of needle EMG include examining patients
on antiplatelet or anticoagulant medications, examining patients with a variety of cuta-
neous issues, and examining muscles that are in close proximity to other organs or
structures. In addition, the discomfort associated with needle EMG can affect the
ability to optimally record the electrical activity throughout a muscle. Awareness of
these potential complications, and adjustment in the technique to reduce the risks,
are critical for a safe study and optimizing patient care.

EMG in Patients on Anticoagulation or with Bleeding Disorders

Needle examination can generally be performed without complications in patients on
anticoagulants, antiplatelet agents, or with bleeding complications, although adjust-
ments in technique and limitations may apply. The risk of performance of the needle
examination in these patients is excessive bleeding and hematoma formation. If this
were to occur in a closed compartment, such as the anterior compartment of the
leg, there is the potential for development of compartment syndrome and tissue
necrosis.29 The magnitude of the risk is low and there have been only a few reports
of paraspinal hematoma, calf hematoma, and calf artery pseudoaneurysm develop-
ment following needle examination in patients on anticoagulation.30–32 In a survey of
47 EMG laboratories in the United States, 9% of laboratories reported experiencing
at least 1 episode of bleeding complication requiring medical or surgical intervention
caused by needle EMG.32
Despite these concerns, in a recent study using ultrasound to assess for hematoma
formation in the anterior tibialis muscle following needle EMG in patients on warfarin
(with international normalized ratio [INR] >1.5 and <4.2) and antiplatelet agents (aspirin
or clopidogrel), only 2 of 101 patients on warfarin and 1 of 57 patients on antiplatelet
agents developed small (2–3 mm wide by 2–3 cm long), subclinical hematomas.33 At
this time, there is no standard of practice or consensus in electrodiagnostic medicine
regarding the highest level of anticoagulation at which a needle examination can safely
be performed without additional risk. Therefore, each case must be considered indi-
vidually and the necessity and benefits of the study must be weighed against the
potential risks. In the ideal situation, anticoagulants should be discontinued before
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 707

the study, although this increases the risk of potential thrombotic complications. Most
electromyographers prefer to know the level of anticoagulation (INR) before the study
to determine the level of risk, and the decision of whether to perform an entire or partial
study is individual. If the needle examination is performed on patients on anticoagula-
tion, some adjustment in the technique of the study may help to minimize potential
bleeding complications (Box 10).
Similar precautions should be taken in patients with thrombocytopenia. If the
platelet count is more than 30,000/mm2, the study can usually be performed safely.
For patients with hemophilia and uncommon bleeding disorders, the patient’s hema-
tologist should be consulted before performance of the needle examination.

EMG in Patients with Skin Problems

Several dermatologic conditions should lead to avoidance or limitation of the needle
examination. The needle electrode should not be inserted through infected skin (eg,
cellulitis) or into an area of prominent vasculature (eg, varicose veins or arteriovenous
dialysis shunt). In addition, patients with thin skin, such as those on corticosteroids,
may be more prone to bleeding or tearing of the skin and extra caution should be taken
during the examination.

EMG in Lymphedema
Examining a limb with lymphedema poses the risk of persistent leaking of serous fluid,
potentially increasing the risk of the development of cellulitis. Despite the absence of
studies assessing this risk, a position statement by the American Association of
Neuromuscular and Electrodiagnostic Medicine (AANEM) suggested that “reasonable
caution should be exercised in performing needle examinations in lymphedematous
regions.”34

Examining Peripleural Muscles

Examination of muscles adjacent to or near the lung poses a risk of puncturing the
pleura and inducing pneumothorax. This problem may occur with examination of
the diaphragm, rhomboids, serratus anterior, trapezius, supraspinatus, and cervical
and thoracic paraspinals. Experience in examining these muscles, and precise knowl-
edge of the location and anatomy of these muscles and their relationship to the pleura,
are critical to preventing this complication. Techniques used to reduce the risk of
pneumothorax when examining these muscles have been reviewed.28 In all cases,
when any of these muscles are examined, the needle electrode should be advanced
slowly and smoothly, listening for the sharp, clicky sound of the MUPs (indicating
close proximity). When the sound of the potentials becomes more dulled with slow
needle advancement, the needle is likely nearing the distant portion of the muscle

Box 10
Technique adjustments during needle examination in patients on anticoagulation or with
bleeding disorders

 Examine the least number of muscles necessary to make the diagnosis

 Avoid deep muscles (eg, paraspinals, diaphragm)
 Avoid muscles in tight fascial spaces (eg, tibialis anterior)
 Avoid muscles in close proximity to arteries (eg, iliopsoas, flexor pollicis longus)
 Place pressure on the puncture site for 1 to 2 minutes following the examination
 Author's personal copy
708 Rubin

and should be withdrawn. Listening for a respiratory pattern of MUP firing, indicating
the approach to the peripleural muscles, should prompt caution with continued
forward advancement of the needle. The increasing use of ultrasonography during
EMG studies to identify and observe needle insertion into the peripleural muscles
may further help to reduce this complication.

Performance of Needle EMG in Patients with Pacemaker or Defibrillator

There is no contraindication to performing the needle examination in patients with
a pacemaker or other automated defibrillator. Recognition of pacemaker artifact is
important to avoid misinterpretation of the artifact as a fibrillation potential.

Performance of Needle EMG in Patients with Prosthetic Cardiac Valves

The risk of needle EMG in patients with rheumatic or other types of valvular heart
disease or with prosthetic valves is similar to that of repeated venipuncture, and
prophylactic antibiotics are not necessary.

Reducing Pain Associated with Needle EMG

Most patients are able to tolerate the discomfort of the needle examination without
difficulty, but a few need a special approach. Pain minimization requires attention to
interactions with the patient, and particularly the technique of the needle examination.
Techniques such as distraction, continued reassurance, and an empathetic approach
to the patient during the study may improve the patient’s tolerance of the study. The
technique of needle movement has a significant impact in pain reduction. Studies
have shown that needle movements of less than 1 mm when using concentric needle
electrodes are significantly less painful than needle movements of approximately
1 cm.27,35

SUMMARY

NCS and needle EMG can be safely performed in most patients with appropriate tech-
nique and precautions in certain circumstances. Attention to detail of technique,
awareness of technical factors that can affect reliable interpretation of the study,
and appropriate troubleshooting are imperative for quality performance and accurate
interpretation of the study.

REFERENCES

1. Kincaid JC, Brashear A, Markand ON. The influence of the reference electrode on
CMAP configuration. Muscle Nerve 1993;16:392–6.
2. Brashear A, Kincaid JC. The influence of the reference electrode on CMAP
configuration: leg nerve observations and an alternative reference site. Muscle
Nerve 1996;19:63–7.
3. Barkhaus PE, Kincaid JC, Nandedkar SD. Tibial motor nerve conduction studies:
an investigation into the mechanism for amplitude drop of the proximal evoked
response. Muscle Nerve 2011;44:776–82.
4. Rutkove SB. Reduction of motor artifact in antidromic ulnar sensory studies.
Muscle Nerve 1999;22:520–2.
5. Rutkove SB, Kothari MJ, Shefner JM. Nerve, muscle, and neuromuscular junction
electrophysiology at high temperature. Muscle Nerve 1997;20:431–6.
6. Franssen H, Wieneke GH, Wokke JH. The influence of temperature on conduction
block. Muscle Nerve 1999;22:166–73.
 Author's personal copy
Pitfalls of Nerve Conduction Studies and Needle Electromyography 709

7. Hubbard JI, Jones SF, Landau EM. The effect of temperature change upon trans-
mitter release, facilitation, and post-tetanic potentiation. J Physiol 1971;216:
591–609.
8. Rutkove SB, Shefner JM, Wang AK, et al. High temperature repetitive nerve stim-
ulation in myasthenia gravis. Muscle Nerve 1998;21:1414–8.
9. Guttman L. AAEM minimonograph #2: important anomalous innervations of the
extremities. Muscle Nerve 1993;16:339–47.
10. Wilbourn AJ, Lambert EH. The forearm median-to-ulnar nerve communication:
electrodiagnostic aspects. Neurology 1976;26:368.
11. Srinivasan R, Rhodes J. The median-ulnar anastomosis (Martin-Gruber) in normal
and congenitally abnormal fetuses. Arch Neurol 1981;38:418–9.
12. Whitaker CH, Felice KJ. Apparent conduction block in patients with ulnar neurop-
athy at the elbow and proximal Martin-Gruber anastomosis. Muscle Nerve 2004;
30:808–11.
13. Sun SF, Streib EW. Martin-Gruber anastomosis: electromyographic studies: part
II. Electromyogr Clin Neurophysiol 1983;23:271–85.
14. Kimura I, Ayyar DR. The hand neural communication between the ulnar and
median nerves: electrophysiological detection. Electromyogr Clin Neurophysiol
1984;24:409–14.
15. Kimura J, Murphy MJ, Varda DJ. Electrophysiological study of anomalous inner-
vation of intrinsic hand muscles. Arch Neurol 1976;33:842–4.
16. Gutmann L. Median-ulnar nerve communications and carpal tunnel syndrome.
J Neurol Neurosurg Psychiatry 1977;40:982–6.
17. Gutmann L, Gutierrez A, Riggs JE. The contribution of median to ulnar communi-
cation in diagnosis of mild carpal tunnel syndrome. Muscle Nerve 1986;9:319–21.
18. Rubin DI, Dimberg EL. Martin-Gruber anastomosis and carpal tunnel syndrome:
morphologic clues to identification. Muscle Nerve 2010;42:457–8.
19. Lambert EH. The accessory deep peroneal nerve: a common variation in innerva-
tion of extensor digitorum brevis. Neurology 1969;19:1169–76.
20. Sachs GM, Raynor EM, Shefner JM. The all ulnar hand without forearm anasto-
mosis. Muscle Nerve 1995;18:309–13.
21. Dimitru D, Walsh NE, Weber CF. Electrophysiologic study of the Riche-Cannieu
anomaly. Electromyogr Clin Neurophysiol 1988;28:27–31.
22. Leis AA, Wells KJ. Radial nerve cutaneous innervation to the dorsum of the hand.
Clin Neurophysiol 2008;119:662–6.
23. Nora LM. American Association of Electrodiagnostic Medicine guidelines in elec-
trodiagnostic medicine: implanted cardioverters and defibrillators. Muscle Nerve
1996;19:1359–60.
24. LaBan MM, Petty D, Hauser AM, et al. Peripheral nerve conduction stimulation: its
effect on cardiac pacemakers. Arch Phys Med Rehabil 1988;69:358–62.
25. Schoeck AP, Mellion ML, Gilchrist JM, et al. Safety of nerve conduction studies in
patients with implanted cardiac devices. Muscle Nerve 2007;35:521–4.
26. Mellion ML, Buxton AE, Iyer V, et al. Safety of nerve conduction studies in patients
with peripheral intravenous lines. Muscle Nerve 2010;42:189–91.
27. Strommen JA, Daube JR. Determinants of pain in needle electromyography. Clin
Neurophysiol 2001;112:1414–8.
28. Al-Shekhlee A, Shapiro BE, Preston DC. Iatrogenic complications and risks of
nerve conduction studies and needle electromyography. Muscle Nerve 2003;
27:517–26.
29. Farrell CM, Rubin DI, Haidukewych GJ. Acute compartment syndrome of the leg
following diagnostic electromyography. Muscle Nerve 2003;27:374–7.
 Author's personal copy
710 Rubin

30. Rosioreanu A, Dickson A, Lypen S, et al. Pseudoaneurysm of the calf after elec-
tromyography: sonographic and CT angiographic diagnosis. Am J Roentgenol
2005;185:282–3.
31. Butler ML, Dewan RW. Subcutaneous hemorrhage in a patient receiving antico-
agulant therapy: an unusual EMG complication. Arch Phys Med Rehabil 1984;
65:733–4.
32. Gruis KL, Little AA, Zebarah VA, et al. Survey of electrodiagnostic laboratories
regarding hemorrhagic complications from needle electromyography. Muscle
Nerve 2006;34:356–8.
33. Lynch SL, Boon AJ, Smith J, et al. Complications of needle electromyography:
hematoma risk and correlation with anticoagulation and antiplatelet therapy.
Muscle Nerve 2008;38:1225–30.
34. AANEM Position Statement. Needle EMG in certain uncommon clinical contexts.
Muscle Nerve 2005;31:398–9.
35. Walker WC, Keyser-Marcus LA, Johns JS, et al. Relation of electromyography-
induced pain to type of recording electrodes. Muscle Nerve 2001;24:417–20.