Segmental study of the median nerve versus comparative tests in the diagnosis of mild carpal tunnel syndrome.

Abstract

The aims of this study were to analyze normative data of nerve conduction studies (NCS) by optimal transformations, and compare the utility of electrodiagnostic tests in detecting mild carpal tunnel syndrome (CTS).
In 131 hands of patients with mild CTS and 136 hands of controls, the segmental study of the median nerve between the digit-palm and palm-wrist segments, and the median-to-ulnar and median-to-radial comparative tests were performed. Normal limits were derived by calculating the mean+/-2 standard deviations of the optimally transformed data of the controls. The specificity, sensitivity, and misclassification rate were calculated to evaluate the utility of each test.
All tests had high specificities, ranging from 98.5 to 100%. The distoproximal latency ratio (DPLR) of the median nerve showed the highest sensitivity and the difference between the median and radial sensory latencies (D1M-D1R) the second highest, but there was no statistical difference between them. The difference between the median and ulnar mixed nerve latencies in the palm-to-wrist segment (PM-PU) showed the lowest sensitivity. Misclassification rates of the DPLR, D1M-D1R, and PM-PU were 6.9, 3.8, and 6.1%, respectively.
Optimal transformation of NCS data is mandatory to diminish the effect of skewness and enhance the diagnostic accuracy. As compared to the comparative tests, the segmental study of the median nerve is more easily applied and yields higher sensitivity in detecting mild CTS.
With a high diagnostic yield and easy application, the segmental study of the median nerve may routinely be used to evaluate patients with mild CTS.

Full text

Page 1

Segmental study of the median nerve versus comparative tests
in the diagnosis of mild carpal tunnel syndrome
Jau-Jiuan Sheua,b,*, Rey-Yue Yuana,b, Hung-Yi Chiouc, Chaur-Jong Hua,b, Wei-Ta Chena,b
aDepartment of Neurology, Taipei Medical University Hospital, 252 Wu-Hsing Street, Taipei, Taiwan, ROC
bDepartment of Neurology, College of Medicine, Taipei Medical University, Taipei, Taiwan, ROC
cSchool of Public Health, Taipei Medical University, Taipei, Taiwan, ROC
Accepted 4 February 2006
Available online 4 April 2006
Abstract
Objective: The aims of this study were to analyze normative data of nerve conduction studies (NCS) by optimal transformations, and
compare the utility of electrodiagnostic tests in detecting mild carpal tunnel syndrome (CTS).
Methods: In 131 hands of patients with mild CTS and 136 hands of controls, the segmental study of the median nerve between the digit–palm
and palm–wrist segments, and the median-to-ulnar and median-to-radial comparative tests were performed. Normal limits were derived by
calculating the meanG2 standard deviations of the optimally transformed data of the controls. The specificity, sensitivity, and
misclassification rate were calculated to evaluate the utility of each test.
Results: All tests had high specificities, ranging from 98.5 to 100%. The distoproximal latency ratio (DPLR) of the median nerve showed the
highest sensitivity and the difference between the median and radial sensory latencies (D1MKD1R) the second highest, but there was no
statistical difference between them. The difference between the median and ulnar mixed nerve latencies in the palm-to-wrist segment (PMK
PU) showed the lowest sensitivity. Misclassification rates of the DPLR, D1MKD1R, and PMKPU were 6.9, 3.8, and 6.1%, respectively.
Conclusions: Optimal transformation of NCS data is mandatory to diminish the effect of skewness and enhance the diagnostic accuracy. As
compared to the comparative tests, the segmental study of the median nerve is more easily applied and yields higher sensitivity in detecting
mild CTS.
Significance: With a high diagnostic yield and easy application, the segmental study of the median nerve may routinely be used to evaluate
patients with mild CTS.
q 2006 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
Keywords: Carpal tunnel syndrome; Median nerve; Electrodiagnosis; Segmental study; Comparative test; Transformation
1. Introduction
The diagnosis of carpal tunnel syndrome (CTS) requires
confirmation of the symptoms and signs with objective
electrodiagnostic tests which identify and localize dysfunc-
tion of the median nerve in the carpal tunnel (Johnson,
1993). With increased knowledge of CTS, many patients
with typical CTS symptoms are referred earlier and fail to
show abnormalities using diagnostic criteria created by
conventional electrodiagnostic methods (Jablecki et al.,
1993; Jackson and Clifford, 1989). In order to improve the
electrodiagnostic yield, a number of nerve conduction
studies (NCS) have been developed, which include: (1)
segmental study of the median nerve with stimulation
proximal and distal to the carpal tunnel (Andary et al., 1996;
Buchthal and Rosenfalck, 1971; Cruz Martinez et al., 1978;
Kimura, 1978, 1979; Kuntzer, 1994; Lew et al., 2005;
Monga et al., 1985; Padua et al., 1996; Sharma et al., 2001;
Wongsam et al., 1983); (2) sensory latency or conduction
velocity (CV) difference between the median and ulnar
nerves (Charles et al., 1990; Foresti et al., 1996; Jackson and
Clifford, 1989; Johnson et al., 1981; Lauritzen et al., 1991;
Uncini et al., 1989, 1993); (3) sensory latency or CV
Clinical Neurophysiology 117 (2006) 1249–1255
www.elsevier.com/locate/clinph
1388-2457/$30.00 q 2006 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.clinph.2006.02.004
*Corresponding author. Address: Department of Neurology, Taipei
Medical University Hospital, 252 Wu-Hsing Street, Taipei, Taiwan, ROC.
Tel.: C886 2 27372181x3308; fax: C886 2 27372758.
E-mail address: jjs0556@mail.tmch.org.tw (J.-J. Sheu).

Page 2

difference between the median and radial nerves (Carroll,
1987; Cassvan et al., 1988; Ghavanini et al., 1996; Jackson
and Clifford, 1989; Johnson et al., 1987; Pease et al., 1989);
and (4) median and ulnar latency difference after palmar
stimulation (Jackson and Clifford, 1989; Mills, 1985;
Sander et al., 1999). There are few studies of comparisons
of the segmental study of the median nerve with
comparative tests of the median and ulnar or radial nerves,
and there is no agreement on which of these tests is more
sensitive or specific in diagnosis of mild CTS (Andary et al.,
1996; Demirci and Sonel, 2004; Kuntzer, 1994; Padua et al.,
1996; Pease et al., 1989).
Normal limits for NCS parameters are usually calculated
as the meanG2 standard deviations (SD) of controls
assuming the data are normally distributed. However,
many variables in electrodiagnostic medicine are skewed
in one direction, and do not follow a normal distribution in
the normal population. Optimal transformation is applied to
reduce the degree of skewness and to convert data into a
distribution more closely approximating a Gaussian curve.
Hence, normative data are best derived from the meanG2
SD of the optimally transformed data (Campbell and
Robinson, 1993; Dorfman and Robinson, 1997; Robinson
et al., 1991). To the best of our knowledge, there is no study
of comparisons of different electrodiagnostic tests of CTS,
which apply transformation methods for normal limits. The
aims of this study were to analyze normative data of NCS by
optimal transformations of the control data, and compare the
utility of the segmental study of the median nerve with the
median-to-ulnar and median-to-radial comparative tests in
detecting mild CTS.
2. Patients and methods
2.1. Control subjects
One hundred and thirty-six healthy subjects (mean age
49.67 years; range 21–79 years; 82.4% females) were
evaluated with unilateral NCS. A screening history and
physical examination were carried out for all subjects to
exclude any obvious peripheral neuropathy, neuromuscular
disease, as well as relevant systemic conditions such as
diabetes, uremia, excessive alcohol intake, or toxin
exposure.
2.2. Patients
Patients referred to the electrodiagnostic laboratory were
prospectively evaluated. The diagnosis of CTS was made
clinically based upon paresthesia in the median nerve
territory with histories and physical examination suggestive
of CTS, including at least one of the following: (1) nocturnal
paresthesia exacerbation; (2) symptoms precipitated by
manual activities such as using hand tools or driving a car;
(3) a positive Phalen’s sign or Tinel’s sign; (4) weakness or
atrophy of the thenar muscles. Subjects with a history or
physical examination suggestive of a neuromuscular
disorder other than CTS or with abnormal ulnar distal
motor latency (DML) or sensory latency were excluded.
Totally, 235 hands in 153 patients were included. Eighty-
five percent of the hands showed nocturnal exacerbation of
paresthesia. In 70% of the hands, symptoms were
precipitated by manual activities. A positive Phalen’s sign
or Tinel’s sign was present in 65% of the hands. Weakness
or atrophy of the thenar muscles was only present in 10% of
the hands. One hundred and four symptomatic hands had
abnormal median DML or sensory latency (see results for
the criteria of abnormality) making obvious the electro-
diagnosis of CTS. Of these 104 hands, 13 had an absent
median sensory response and 3 had an absent median motor
response. One hundred and thirty-one mild CTS hands of
104 patients (mean age 49.25 years; range 28–74 years;
84.6% female) had normal median DML and sensory
latency.
2.3. Electrodiagnostic methods
A Medelec Synergy electromyograph (Medelec, Surrey,
England) was used in the study. Filters were set at 2 Hz and
10 kHz for motor studies and at 20 Hz and 2 kHz for sensory
studies. The sweep speed was set at 1 ms/division. One
centimeter disc recording electrodes were used for motor
studies and the median and ulnar mixed nerve studies, and
ring recording electrodes were used for sensory studies.
Supramaximal stimuli of 0.05–0.1 ms were delivered by a
hand-held bipolar stimulator. For sensory studies, the
potentials were recorded by averaging 16 responses, and a
gain setting of 10 mV/division was used to determine
latencies measured from the stimulus artifact to the negative
peak. Latencies were measured to the nearest 0.05 ms using
a cursor and digital display. Hand skin temperatures were
continuously monitored and maintained at 32–34 8C during
the procedures.
2.3.1. Segmental study of the median nerve
The segmental study of the median nerve was performed
in control and CTS hands. The antidromic sensory latency
of the median nerve was recorded from digit 3 with the
active recording electrode placed at the proximal inter-
phalangeal (PIP) joint. The reference electrode was placed
4 cm distal to the active electrode or placed distally with a
maximal possible interelectrode spacing (of at least 3 cm) in
the small hands (2 of 136 control hands and 5 of 235 CTS
hands). The median nerve was stimulated 7 cm proximally
in the palm and 14 cm proximally at the wrist by a bipolar
stimulator. Distances were measured in fully extended
hands with a flexible tape measure. The sensory conduction
time across the carpal tunnel (palm–wrist latency, PWL)
was determined by subtracting the latency obtained in the
palm (digit–palm latency, DPL) from the latency at the wrist
(digit–wrist latency, DWL). Two indices, for segmental
J.-J. Sheu et al. / Clinical Neurophysiology 117 (2006) 1249–12551250

Page 3

comparison of conduction between digit 3 and the palm and
between the palm and the wrist, were calculated as follows:
Distoproximal latency ratioðDPLRÞ ZDPL=PWL;
and
Distoproximal latency differenceðDPLDÞ ZDPLKPWL:
2.3.2. Comparative tests
We performed three comparative tests in control and
mild CTS hands. The first comparative test compared the
median and ulnar antidromic sensory latencies at digit 4.
The active recording electrode was placed at the PIP joint,
and the reference electrode was placed 4 cm distally or
placed distally with a maximal possible interelectrode
spacing (of at least 3 cm) in the small hands. Stimulation
was delivered at a distance of 14 cm over the median and
ulnar nerves at the wrist. The difference between the median
and ulnar latencies was calculated as D4MKD4U.
The second test compared the median and radial
antidromic sensory latencies at the thumb. The active
recording electrode was placed at the metacarpophalangeal
joint, and the reference electrode was placed distally with a
maximal possible interelectrode spacing (of at least 3 cm).
Stimulation was delivered at a distance of 10 cm over the
median nerve at the wrist and the radial nerve on the
dorsolateral surface of the wrist. The difference between the
median and radial latencies was calculated as D1MKD1R.
The third test compared the median and ulnar mixed
nerve latencies in the palm-to-wrist segment. The median
nerve was stimulated in the palm, between the second and
third metacarpal bones, at a distance of 8 cm distal to the
median recording site at the wrist. The ulnar palmar latency
was made similarly, but with the recording over the ulnar
nerve at the wrist and the stimulation between the fourth and
fifth metacarpal bones. The difference between the median
and ulnar palmar latencies was calculated as PMKPU.
2.4. Statistical analysis
For the NCS parameters of the controls, transformations
were performed to bring the coefficient of skewness closer
to zero, and to convert data to a normal distribution. The
ideal normal limits of the controls were derived from the
meanG2 SD of the optimally transformed data and by
converting these endpoints back to original units, or the
meanG2 SD of the raw data if they followed a normal
distribution.
The specificity of each test was calculated as: (number of
control hands with a normal test result/number of control
hands)!100. The sensitivity of each test was calculated as:
(number of mild CTS hands with an abnormal test
result/number of mild CTS hands)!100. The misclassifi-
cation rate was determined by counting the percentage of
mild CTS hands that were called normal by the criteria of
abnormality using the raw data, but would have been
abnormal using that of the optimally transformed data, or
the percentage of mild CTS hands that were called abnormal
by the criteria of abnormality using the raw data, but would
have been normal using that of the optimally transformed
data. The McNemar chi-square statistic was used to test the
statistical significance of comparisons between the sensi-
tivities of the segmental study of the median nerve and the
comparative tests. A P-value of !0.05 was considered
statistically significant.
3. Results
3.1. Control subjects
The results of electrodiagnosis in the controls are
summarized in Table 1. The raw data of the DPLD and
D4MKD4U closely followed a normal distribution and no
method of transformation brought the coefficient of
skewness closer to zero. The lower normal limit of
Table 1
Summary of normal values in 136 control hands
Variable MeanGSDRange Uncorrected
coefficient of
skewness
Criteria of
abnormality
by raw data
Best trans-
formation
method
Corrected
coefficient of
skewness
Criteria of
abnormality
by trans-
formed data
Median DML (ms)
DWL (ms)
DPLR
DPLD (ms)
D4MKD4U (ms)
D1MKD1Ra(ms)
PMKPUa(ms)
3.22G0.30
3.19G0.24
1.33G0.15
0.44G0.18
0.09G0.15
0.21G0.13
0.08G0.13
2.60–3.85
2.70–3.75
1.03–1.83
0.05–0.95
K0.25–0.40
K0.10–0.45
K0.20–0.40
0.26
0.23
0.27
O3.85
O3.70
!1.02
!0.05
O0.40
O0.50
O0.35
Log
Log
Log
None
None
Square
Square root
0.07
0.06
O3.90
O3.70
!1.05
K0.02
K0.06
0.05
K0.23
0.10
0.01
0.00
O0.45
O0.40
Median DML, median distal motor latency from wrist to abductor pollicis brevis; DWL, median sensory latency from wrist to digit 3; DPLR, distoproximal
latency ratio; DPLD, distoproximal latency difference; D4MKD4U, difference between median and ulnar sensory latencies at digit 4; D1MKD1R, difference
between median and radial sensory latencies at thumb; PMKPU, difference between median and ulnar mixed nerve latencies from palmar stimulation; SD,
standard deviation.
aTransformation of (raw valueC1) due to the presence of negative values, such as square root of (raw value of PMKPUC1).
J.-J. Sheu et al. / Clinical Neurophysiology 117 (2006) 1249–12551251

Page 4

the DPLD was 0.05 ms and the upper normal limit of the
D4MKD4U was 0.40 ms.
The median DML, DWL, DPLR and PMKPU values
showed a positive skew, and log or square root transform-
ations brought the coefficient of skewness closer to zero.
The D1MKD1R showed a negative skew, and square
transformation brought the coefficient of skewness closer to
zero. After an optimal transformation, the upper normal
limit of the median DML was 3.90 ms and the upper normal
limit of the DWL was 3.70 ms. The lower normal limit of
the DPLR was 1.05, the upper normal limit of the D1MK
D1R were 0.45 ms, and the upper normal limit of the PMK
PU was 0.40 ms.
3.2. Specificities, sensitivities, and misclassification rates of
the segmental study and comparative tests
The specificities, sensitivities, and misclassification
rates of the segmental study of the median nerve and the
median-to-ulnar and median-to-radial comparative tests
are shown in Table 2. All tests had high specificities,
ranging from 98.5 to 100%, whether raw data or
transformed data were used. The highest diagnostic
yield was obtained when all tests were combined, and
the combined sensitivity was 84.7% using the criteria of
abnormality with an ideal normal limit. Using the criteria
of abnormality with an ideal normal limit, the DPLR
showed the highest sensitivity being !1.05 in 77.9% of
mild CTS hands, and the D1MKD1R showed the second
highest sensitivity being O0.45 ms in 74.0% of hands.
The DPLD was !0.05 ms in 71.8% of hands, and the
D4MKD4U was O0.40 ms in 70.2% of hands. The
PMKPU showed the lowest sensitivity being O0.40 ms
in only 53.4% of hands.
Using the meanK2 SD of the raw data of the DPLR as a
normal limit misclassified 6.9% of mild CTS hands as
normal, when compared with the meanK2 SD of the log-
transformed data. Using the meanC2 SD of the raw data of
the D1MKD1R as a normal limit misclassified 3.8% of
mild CTS hands as normal, when compared with the
meanC2 SD of the square-transformed data. Using the
meanC2 SD of the raw data of the PMKPU as a normal
limit misclassified 6.1% of mild CTS hands as abnormal,
when compared with the meanC2 SD of the square root-
transformed data.
3.3. Comparisons of sensitivities of the segmental study and
comparative tests
The results of comparisons between the sensitivities of
the segmental study of the median nerve and the
comparative tests are shown in Table 3. Both DPLR and
DPLD had much greater sensitivity compared with the
PMKPU (PZ0.00). Although the DPLR had the highest
sensitivity, there were no significant differences compared
to the D4MKD4U or D1MKD1R.
Table 2
Sensitivities, specificities, and misclassification rates of electrodiagnostic tests
TestCriteria of abnormality by raw data Criteria of abnormality by ideal normal limitsa
Sensitivity (%)Specificity (%) Sensitivity (%) Specificity (%)Misclassification rate
(%)
DPLR
DPLD
D4MKD4U
D1MKD1R
PMKPU
71.0
71.8
70.2
70.2
59.5
100
100
100
100
99.3
77.9
71.8
70.2
74.0
53.4
98.5
100
100
100
100
6.9
3.8
6.1
aIdeal normal limits: derived from meanG2 SD of the raw data of DPLD and D4MKD4U, and from meanG2 SD of the optimally transformed data of
DPLR, D1MKD1R, and PMKPU. All abbreviations as in Table 1.
Table 3
Comparisons of sensitivitiesaof segmental study and comparative tests in 131 mild CTS hands
Test D4MKD4UD1MKD1RPMKPU
Positive
handsb
Negative
handsc
P-valuePositive
hands
Negative
hands
P-valuePositive
hands
Negative
hands
P-value
DPLR
Positive hands
Negative hands
DPLD
Positive hands
Negative hands
84
8
18
21
0.08 90
7
12
22
0.3669
1
33
28
0.00
80
12
14
25
0.85 84
13
10
24
0.6864
6
30
31
0.00
aSensitivity using criteria of abnormality by ideal normal limits.
bPositive hands: hands with an abnormal test result.
cNegative hands: hands with a normal test result. All abbreviations as in Table 1.
J.-J. Sheu et al. / Clinical Neurophysiology 117 (2006) 1249–12551252

Page 5

4. Discussion
This prospective study meets all 6 criteria recommended
by the American Association of Electrodiagnostic Medicine
(AAEM) Quality Assurance Committee, and our results are
in accordance with the AAEM’s statement that the
segmental study of the median nerve is the most sensitive
test in the electrodiagnosis of mild CTS (Jablecki et al.,
2002). Our results are also in agreement with those of
Andary et al. (1996) and Uncini et al. (1993) that the
difference between median and ulnar nerve latencies from
palmar stimulation has relatively low sensitivity because
both sensory and motor fibers rather than only sensory fibers
are tested. Although there was no significant difference
between the sensitivity of the segmental study of the median
nerve and that of the D1MKD1R, there were 12 cases with
an abnormal test result of the DPLR but a normal test result
of the D1MKD1R, and another 7 cases with an abnormal
test result of the D1MKD1R but a normal test result of the
DPLR (Table 3). These electrophysiologic findings suggest
that the effect of compression of branches of the median
nerve under the carpal tunnel is not uniform but affects
certain branches more than others among different patients.
The sensitivities of the tests in our study were lower than
those of Demirci and Sonel (2004), Padua et al. (1996) and
Pease et al. (1989). The difference is primarily due to
various degrees of severity of CTS among these studies.
First, CTS patients studied by Padua et al. (1996) were not
limited to those with mild disease, and hence higher
sensitivities were reported. Second, the upper normal limit
of the median sensory latency of the controls in our study
(3.7 ms) was lower than the criterion used by Pease et al.
(4.0 ms) (1989). Demirci and Sonel (2004) also reported a
higher median sensory latency among the CTS group. Our
lower normal limit for the median sensory latency results in
more cases of mild CTS with the normal comparative tests
and segmental study of the median nerve. Pease et al. (1989)
reported that the sensitivity of the ratio of sensory
conduction of the median nerve between the wrist–palm
segment and wrist–digit segment was lower than that of the
median-to-ulnar or median-to-radial comparative tests in
detecting mild CTS. Andary et al. (1996) had similar
observations and concluded that the ratio of median sensory
latency across the wrist to the latency from the wrist to the
digit added no more yield to the diagnosis of CTS. In
contrast, Demirci and Sonel (2004) and Padua et al. (1996)
reported a higher sensitivity of the segmental study of the
median nerve than that of the comparative tests. Our results
showed that the DPLR had the highest sensitivity and the
D1MKD1R the second highest sensitivity, but no signifi-
cant difference existed between them. The disparity in the
sensitivities among different studies is probably a conse-
quence of selection biases in the choice of the study
population, differences in methodology, the use of different
cutoff points to define an abnormal value (AAEM, 2002),
and different statistical methodologies (Robinson et al.,
1991).
Our results are in accordance with those of Robinson
et al. (1991) that motor and sensory latencies were
positively skewed. We also demonstrated that sensory
latency ratio of the median nerve was positively skewed, but
latency differences of the segmental study and comparative
tests may follow a normal distribution, or be positively or
negatively skewed. A boundary to the left (on the side of
shorter latency) is a possible explanation for positive skews
of latency measurements (Robinson et al., 1991). However,
the reasons for positive skews of latency ratio of the median
nerve and positive or negative skews in latency differences
of the comparative tests are not known, and further studies
based on a larger number of subjects may be needed.
Although transformation seemed to produce only a small
absolute change in the normal limits (0.03 for the DPLR,
and 0.05 ms for the D1MKD1R and PMKPU), this is
meaningful in terms of the percentage of hands misclassi-
fied. For example, the distribution of the DPLR was
positively skewed in the controls, and 6.9% of mild CTS
hands would be misclassified as normal if the abnormal
cutoff value was based on the meanK2 SD of the raw data,
leading to diagnostic underestimation (Table 2). Thus,
application of optimal transformations to generate norma-
tive values in the segmental study of the median nerve and
the comparative tests is important for diminishing the effect
of skewness and enhancing diagnostic accuracy.
The rationale for the segmental study of the median
nerve is that the slow-conducting segment of the nerve
within the carpal tunnel is quite short, and the segment of
the nerve distal to the carpal tunnel is little impaired in early
CTS. If this normally conducting segment distal to the
carpal tunnel is included in latency measurements as in the
conventional techniques, the abnormality may be diluted
and the overall conduction time may remain within normal
limits (Jackson and Clifford, 1989). The shorter the nerve
segment enclosing the abnormality, the more prominent the
slowing of conduction will be. The disadvantages of the
orthodromic sensory conduction study performed by Padua
et al. (1996) include the small amplitude of the sensory
nerve action potential (SNAP), interference by the stimulus
artifact, and hence difficulty in accurately identifying the
onset latency for velocity calculation, especially when
testing the short nerve segment (Wilbourn, 1994). Thus,
Padua et al. (1996) mentioned that all measurements used to
calculate the distoproximal ratio of the CV must be made
with extreme care because of the relatively short distances
of the digit to the palm and the palm to the wrist. To tackling
such problems, we performed antidromic stimulation over
the already premeasured, marked points in the palm and at
the wrist, and measured peak latencies rather than onset
latencies for both DPLR and DPLD calculations. In most of
the hands, the distance between the PIP joint and the tip of
the finger exceeded 4 cm, and we placed the reference
electrode 4 cm distal to the active electrode. In extremely
J.-J. Sheu et al. / Clinical Neurophysiology 117 (2006) 1249–12551253

Page 6

small hands (2 of 136 control hands and 5 of 235 CTS
hands), we placed the reference electrode distally with a
maximal possible interelectrode spacing (of at least 3 cm) as
done at the thumb. An interelectrode spacing of the
recording electrodes exceeding 3 cm can avoid distortion
of the SNAP waveform (Wilbourn, 1994). The amplitude of
the SNAP obtained by this method is larger than that by
orthodromic stimulation, and measurement of the peak
latency is easier and more accurate than that of the onset
latency. The formulas of the DPLR and DPLD are simpler
than that of the distoproximal ratio of the CV (digit-to-palm
CV, palm-to-wrist CV, and their ratio), and hence diminish
the inherent error of calculation (Sharma et al., 2001). In
addition, similar formulas for calculation, and the agree-
ment of the antidromic techniques and the peak latency
measurements in our study increase the comparability
among the DPLD, and median-to-ulnar and median-to-
radial comparative tests. The DPLR and DPLD also take
advantage of the comparative approach in which each
patient serves as his own control, and intersubject variability
in the electrodiagnosis is eliminated (Carroll, 1987; Padua
et al., 1996; Uncini et al., 1993).
The AAEM recommended comparison of median
sensory or mixed nerve conduction through the carpal
tunnel to NCS of proximal or distal segments of the median
nerve if normal median sensory NCS across the wrist with a
distance of 13–14 cm. (AAEM, 2002; Jablecki et al., 2002).
Our study shows that the segmental study of the median
nerve, requiring only an additional stimulus in the palm, is
easily applied in electrodiagnostic laboratories. As
compared to the median-to-ulnar and median-to-radial
comparative tests, the segmental study of the median
nerve is simple and timesaving, and produces lesser
discomfort to patients. In addition, the DPLR had the
highest diagnostic yield. We recommend that the segmental
study of the median nerve may routinely be used in the
evaluation of patients with mild CTS. Optimal transform-
ation of NCS data is mandatory to diminish the effect of
skewness and to enhance the diagnostic accuracy. Each
laboratory should establish its own normal limits of NCS
parameters in the electrodiagnosis of CTS based on
optimally transformed data.
References
American Association of Electrodiagnostic Medicine, American Academy
of Neurology, American Academy of Physical Medicine and
Rehabilitation. Practice parameter for electrodiagnostic studies in
carpal tunnel syndrome: summary statement. Muscle Nerve 2002;25:
918–22.
Andary MT, Fankhauser MJ, Ritson JL, Spiegel N, Hulce V, Yosef M,
Stanton DF. Comparison of sensory mid-palm studies to other
techniques in carpal tunnel syndrome. Electromyogr Clin Neurophysiol
1996;36:279–85.
Buchthal F, Rosenfalck A. Sensory conduction from digit to palm and from
palm to wrist in the carpal tunnel syndrome. J Neurol Neurosurg
Psychiatry 1971;34:243–52.
Campbell WW, Robinson LR. Deriving reference values in electro-
diagnostic medicine. Muscle Nerve 1993;16:424–8.
Carroll GJ. Comparison of median and radial nerve sensory latencies in the
electrophysiological diagnosis of carpal tunnel syndrome. Electro-
encephalogr Clin Neurophysiol 1987;68:101–6.
Cassvan A, Ralescu S, Shapiro E, Moshkovski FG, Weiss J. Median and
radial sensory latencies to digit I as compared with other screening tests
in carpal tunnel syndrome. Am J Phys Med Rehabil 1988;67:221–4.
Charles N, Vial C, Chauplannaz G, Bady B. Clinical validation of
antidromic stimulation of the ring finger in early electrodiagnosis of
mild carpal tunnel syndrome. Electroencephalogr Clin Neurophysiol
1990;76:142–7.
Cruz Martinez A, Barrio M, Perez Conde MC, Gutierrez AM. Electro-
physiological aspects of sensory conduction velocity in healthy adults.
1. Conduction velocity from digit to palm, from palm to wrist, and
across the elbow, as a function of age. J Neurol Neurosurg Psychiatry
1978;41:1092–6.
Demirci S, Sonel B. Comparison of sensory conduction techniques in the
diagnosis of mild idiopathic carpal tunnel syndrome: which finger,
which test? Rheumatol Int 2004;24:217–20.
Dorfman LJ, Robinson LR. AAEM minimonograph #47: normative data in
electrodiagnostic medicine. Muscle Nerve 1997;20:4–14.
Foresti C, Quadri S, Rasella M, Tironi F, Viscardi M, Ubiali E. Carpal
tunnel syndrome: which electrodiagnostic path should we follow? A
prospective study of 100 consecutive patients. Electromyogr Clin
Neurophysiol 1996;36:377–84.
Ghavanini MR, Kazemi B, Jazayeri M, Khosrawi S. Median–radial sensory
latencies comparison as a new test in carpal tunnel syndrome.
Electromyogr Clin Neurophysiol 1996;36:171–3.
Jablecki CK, Andary MT, So YT, Wilkins DE, Williams FH. Literature
review of the usefulness of nerve conduction studies and electro-
myography for the evaluation of patients with carpal tunnel syndrome.
AAEM Quality Assurance Committee. Muscle Nerve 1993;16:
1392–414.
Jablecki CK, Andary MT, Floeter MK, Miller RG, Quartly CA, Vennix MJ,
Wilson JR. Second literature review of the usefulness of nerve
conduction studies and electromyography for the evaluation of patients
with carpal tunnel syndrome. Muscle Nerve 2002;June 1. doi:
10.1002/mus.10215.
Jackson DA, Clifford JC. Electrodiagnosis of mild carpal tunnel syndrome.
Arch Phys Med Rehabil 1989;70:199–204.
Johnson EW. Diagnosis of carpal tunnel syndrome. The gold standard. Am
J Phy Med Rehabil 1993;72:1.
Johnson EW, Kukla RD, Wongsam PE, Piedmont A. Sensory latencies to
the ring finger: normal values and relation to carpal tunnel syndrome.
Arch Physl Med Rehabil 1981;62:206–8.
Johnson EW, Sipski M, Lammertse T. Median and radial sensory latencies
to digit I: normal values and usefulness in carpal tunnel syndrome. Arch
Phys Med Rehabil 1987;68:140–1.
Kimura J. A method for determining median nerve conduction velocity
across the carpal tunnel. J Neurol Sci 1978;38:1–10.
Kimura J. The carpal tunnel syndrome: localization of conduction
abnormalities within the distal segment of the median nerve. Brain
1979;102:619–35.
Kuntzer T. Carpal tunnel syndrome in 100 patients: sensitivity, specificity
of multi-neurophysiological procedures and estimationof axonalloss of
motor, sensory and sympathetic median nerve fibers. J Neurol Sci 1994;
127:221–9.
Lauritzen M,LiguoriR,Trojaborg
conduction along the ring finger in normal subjects and in patients
with a carpal tunnel syndrome. Electroencephalogr Clin Neuro-
physiol 1991;81:18–23.
W. Orthodromicsensory
J.-J. Sheu et al. / Clinical Neurophysiology 117 (2006) 1249–1255 1254

Page 7

Lew HL, Date ES, Pan SS, Wu P, Ware PF, Kingery WS. Sensitivity,
specificity, and variability of nerve conduction velocity measurements
in carpal tunnel syndrome. Arch Phys Med Rehabil 2005;86:12–16.
Mills KR. Orthodromic sensory action potentials from palmar stimulation
in the diagnosis of carpal tunnel syndrome. J Neurol Neurosurg
Psychiatry 1985;48:250–5.
MongaTN,ShanksGL,PooleBJ.Sensorypalmarstimulationinthediagnosis
of carpal tunnel syndrome. Arch Phys Med Rehabil 1985;66:598–600.
Padua L, Lo Monaco M, Valente EM, Tonali PA. A useful electro-
physiologic parameter for diagnosis of carpal tunnel syndrome. Muscle
Nerve 1996;19:48–53.
Pease WS, Cannell CD, Johnson EW. Median to radial latency
difference test in mild carpal tunnel syndrome. Muscle Nerve
1989;12:905–9.
Robinson LR, Temkin NR, Fujimoto WY, Stolov WC. Effect of statistical
methodology on normal limits in nerve conduction studies. Muscle
Nerve 1991;14:1084–90.
Sander HW, Quinto C, Saadeh PB, Chokroverty S. Median and ulnar palm–
wrist studies. Clin Neurophysiol 1999;110:1462–5.
Sharma KR, Rotta F, RomanoJ, AyyarDR. Early diagnosisof carpal tunnel
syndrome: comparison of digit 1 with wrist and distoproximal ratio.
Neurol Clin Neurophysiol 2001;2:2–10.
Uncini A, Lange DJ, Solomon M, Soliven B, Meer J, Lovelace RE. Ring
finger testing in carpal tunnel syndrome: a comparative study of
diagnostic utility. Muscle Nerve 1989;12:735–41.
Uncini A, Di Muzio A, Awad J, Manente G, Tafuro M, Gambi D.
Sensitivityofthree median-to-ulnar
diagnosis of mild carpal tunnel syndrome. Muscle Nerve 1993;
16:1366–73.
Wilbourn AJ. Sensory nerve conduction studies. J Clin Neurophysiol 1994;
11:584–601.
Wongsam PE, Johnson EW, Weinerman JD. Carpal tunnel syndrome: use
of palmar stimulation of sensory fibers. Arch Phys Med Rehabil 1983;
64:16–19.
comparativetests in
J.-J. Sheu et al. / Clinical Neurophysiology 117 (2006) 1249–12551255