- Email: [email protected]
Relation between needle electromyography and nerve conduction studies in patients with carpal tunnel syndrome
246
Relation Between Needle Electromyography and Nerve Conduction Studies in Patients With Carpal Tunnel Syndrome Robert A. Werner, MD, James W. Albers, MD, PhD ABSTRACT. Werner RA, Albers JW. Relation between needle electromyography and nerve conduction studies in patients with carpal tunnel syndrome. Arch Phys Med Rehabii 1995;76:246-9. • Four hundred eighty cases of electrodiagnostically confirmed carpal tunnel syndrome were reviewed to determine if the findings on nerve conduction studies could predict the presence or absence of fibrillation potentials or motor unit changes on the needle examination of the abductor pollicis brevis (APB). The needle examination is more uncomfortable and the ability to predict the findings in this setting from standard nerve conduction studies (NCS) would make the test more acceptable to patients. All patients had median and ulnar nerves (both sensory and motor) tested, as well as the needle evaluation of the APB. Two hundred thirty-one patients had an abnormal needle evaluation as defined by presence of one of the following conditions: abnormal spontaneous activity, increased motor unit action potential (MUAP) amplitude, or increased M U A P polyphasia. One hundred five patients had fibrillation potentials. The mean median motor and sensory amplitudes and latencies, as well as age, did differ in the normal and abnormal needle examination groups, but the sensitivity for predicting an abnormality ranged from 57% to 68%. The ratio of the median to the ulnar amplitudes did not improve the sensitivity of predicting the abnormal needle findings. Motor and sensory evoked potential latencies were the most important predictors of an abnormal needle examination.
© 1995 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation The standard electrodiagnostic screening examination for carpal tunnel syndrome (CTS) includes nerve conduction studies (NCS) but the needle electromyography (EMG) of the median innervated muscles in the hand is considered optional by some electromyographers. ~ The NCS provide information about the conduction velocity of the nerve fibers and an estimate of the number of nerve fibers activated. Because CTS is primarily a focal demyelinating process the NCS are very sensitive to relative changes in the conduction velocity across the carpal canal. Axonal loss is found in more severe cases and is reflected in a decreased amplitude of the median motor and sensory evoked responses, and presence of abnormal spontaneous activity or motor unit action potential (MUAP) changes (large amplitude, polyphasia, decreased recruitment, or increased duration). The NCS can establish the diagnosis but does not clearly define the severity. The purpose of this study was to determine if the findings of axonal loss noted on the needle examination could be predicted from the standard NCS evaluation. The needle examination in the hand is particularly uncomfortable and usually is not a critical component in establishing the diagnosis of CTS. ~ It is important in determining the presence of From the Physical Medicine and Rehabilitation VA Medical Center (Dr. Werner); Department of Physical Medicine and Rehabilitation (Drs. Wemer, Albers), Department of Neurology (Dr. Albers), University of Michigan Medical Center, Ann Arbor, MI. Submitted for publication May 24, 1994. Accepted in revised form August 24, 1994. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organizations with which the authors are associated. Reprint requests to Robert A. Wemer, MD, Physical Medicine and Rehabilitation Service (117), Ann Arbor VA Medical Center, 2215 Fuller Rd, Ann Arbor, MI 48105. © 1995 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 0003-9993/95/7603-308553.00/0
Arch Phys Med Rehabil Vol 76, March 1995
axonal loss and ongoing denervation in the abductor pollicis brevis or opponens muscle. This information frequently is used to make a clinical treatment decision, specifically if surgery should be pursued) If the information about axonal loss could be predicted from the data collected during the NCS, then the patient would be spared the unpleasant needle evaluation with little sacrifice in information. The needle examination remains important to identify proximal lesions such as cervical radiculopathy or more proximal median nerve entrapment. The relationship of sensory nerve action potential (SNAP) and compound muscle action potential (CMAP) with EMG findings is poorly defined. Several authors 36 suggested that a 50% drop in CMAP was associated with axonal loss but this has not been rigorously investigated. The association between the presence of denervation potentials and measures of SNAP and CMAP have not been reported. This study attempts to model the electrophysiologic measures, as well as age, height, and weight, to determine if the results of the needle examination can be predicted. METHODS
This study was a retrospective review of all cases of median mononeuropathy from the EMG laboratory database that were diagnosed in the last 3 years. Cases were selected if they included evaluation of the right arm with median and ulnar motor and sensory evoked responses and needle evaluation of the right abductor pollicis brevis. Any case that also had a diagnosis of peripheral polyneuropathy, cervical radiculopathy, brachial plexopathy, or more proximal median mononeuropathy was excluded from the evaluation. Diagnosis of a median mononeuropathy was based on a relative prolongation of the peak median sensory response of 0.5msec compared with the ulnar sensory response (14cm,
NEEDLE ELECTROMYOGRAPHY AND NCS IN CTS, Werner
antidromic) or a 0.3msec prolongation of the median compared with the ulnar midpalmar peak sensory evoked response. 7-9 An abnormal EMG examination was defined as any study that found abnormal spontaneous activity, large MUAPs or polyphasic MUAPs. All the needle examinations were performed with disposable concentric needles and the evaluation was of the low threshold motor units. The low and high frequency filters were set at 20Hz and 10kHz, respectively. The degree of increase in the MUAP amplitude was subjectively graded on a scale of 0 to 4. Zero represented a normal amplitude (<2mV) and 4+ represented units over 8mV in amplitude. Every attempt was made to optimize the rise time to report the maximum amplitude accurately. Polyphasia was graded subjectively on a scale of 0 to 4 with greater than 10% being defined as abnormal. Trigger and delay line technique was used for all evaluations of polyphasia. Fibrillation potentials also were graded on a 0 to 4 scale as define as follows: 0 no abnormal spontaneous activity; + equivocal findings like a single run or train of persistent fibrillation or very slow trains o f positive waves; 1+ unequivocal positive waves or fibrilla~ons in two areas of muscle that persist greater than 2 seconds; 2+ spontaneous potentials at three or more sites; 3+ spQntaneous activity in nearly all sites; 4 + abundant spontaneous activity nearly filling the screen and obscuring the baseline (adapted from Kimural°). The analysis consisted of dividing the population based on the presence or absence of abnormalities on the EMG examination and identifying the differences in the median sensory or motor amplitude or latency between the two subgroups. Any abnormality, such as an increased amplitude, abnormal polyphasia, or abnormal spontaneous activity, was classified as an abnormal EMG examination. The population also was grouped according to the presence or absence of fibrillation potentials, the most objective needle examination measurement. The ratio of the median motor to ulnar motor amplitude also was analyzed by the normal versus abnormal EMG evaluation; the distribution of these variables was evaluated and risk ratios determined for various electrophysiologic measures. The sensitivity and specificity of several threshold values for the SNAP and CMAP for predicting the presence of abnormalities on the EMG examination were determined. The electrophysiologic measures and demographic data were combined and, using logistic regression, a model was created to predict the presence or absence of EMG abnormalities.
RESULTS Four hundred eighty patients met the entrance criteria and were used as the basis of analysis. The;mean age was 47.5 years (range 17 to 87) with 65% being women. Two hundred thirty-one patients had an abnormal EMG examination, but only 105 had abnormal spontaneous activity. The mean amplitude and latency of the median and ulnar sensory and motor evoked responses, as well as the ratio of median to ulnar amplitude for patients with normal and abnormal EMG examination, are presented in table 1. There was a significant difference between the two group in regard to the median sensory and motor evoked amplitudes and latencies but not
247
Table 1: Comparison of Electrophysiologic Measures in Patients With and Without Abnormalities on EMG Examination Needle Examination
Median sensory amplitude (uV) Median motor amplitude (mV) Median sensory latency (msec) Median motor latency (msec) Ratio: median/ulnar motor amplitude Age (yrs)
Normal EMG (n = 249)
Abnormal EMG (n = 231)
T Test
p Value
21.4 (10.2)
15.1 (9.0)
7.06
<0.001
9.5 (4.0)
7.8 (3.8)
4.52
<0.001
4.1 (0.7)
4.5 (0.8)
6.39
<0.001
4.3 (0.8)
4.9 (1.1)
6.78
<0.001
0.9 (0.4)
0.9 (0.6)
0.70
0.49
45.3 (14.2)
53.0 (14.5)
6.23
<0.001
Results are presented as a mean (ISD).
for the ratio of median motor to ulnar motor amplitude. The patients with an abnormal EMG examination had a more prolonged latency and a smaller amplitude for the motor and sensory evoked responses. The patients with an abnormal EMG examination also tended to be older (53 compared with 45 years). The same electrophysiologic measures are presented in table 2, dichotomized by the presence or absence of abnormal spontaneous activity. The 105 patients with abnormal spontaneous activity had smaller amplitudes and more prolonged latencies compared with the group with any abnormal findings on their needle examination. Again this group was significantly worse than the group without evidence of spontaneous activity. Note that the ratio of the median motor to the ulnar motor amplitude was not different when comparing the groups with and without evidence of abnormal spontaneous activity. Table 3 shows the relationship between the graded MUAP amplitude found on EMG and the amplitudes and latencies of the evoked responses in the hand. There is a clear trend noted for both evoked amplitude and latency. The larger the MUAP amplitude, the smaller the evoked response and the more prolonged the latency. This relationship held for both sensory and motor responses. There was less variability of the latencies compared with the amplitude measures as represented by the relatively large standard deviation of the amplitude measures. A similar but less robust relationship was noted with the graded spontaneous activity and the electrophysiologic measures on NCS (see table 3). The trend was strongest with the median sensory amplitude and weakest with the median motor and sensory latencies. The relative risk of finding abnormal spontaneous activity on EMG examination, if the median sensory amplitude is less than 12#V, is 2.23 (95% CI: 1.6, 3.1). This means that the finding of abnormal spontaneous activity is twice as likely compared with someone with a larger amplitude. If the median motor latency is more than 5. lmsec, the relative risk for abnormal spontaneous activity is 2.44 (95% CI: 1.8, 3.4). If the median sensory latency is greater than 4.8msec, the relative risk is 2.43 (95% CI: 1.7, 3.4). Arch Phys Med Rehabil Vo| 76, March 1995
248
NEEDLE ELECTROMYOGRAPHY AND NCS IN CTS, Werner
Table 2: Comparison of Electrophysiologic Measures in Patients With and Without Spontaneous Activity on EMG Examination Spontaneous Activity
Median sensory amplitude (uV) Median motor amplitude (mV) Median sensory latency (msec) Median motor latency (msec) Ratio: median/ulnar motor amplitude Age (yrs)
Normal (n = 375)
Abnormal
(n = 105)
T Test
p
19.6 (10.2)
13.9 (8.9)
5.30
<0.001
9.0 (3.9)
7.5 (4.0)
2.74
0.006
4.2 (0.7)
4.6 (0.9)
5.31
<0.001
4.4 (0.8)
5.1 (1.3)
6.32
<0.001
0.9 (0.5)
0.9 (0.6)
0.12
0.91
48.1 (14.5)
52.5 (14.4)
2.81
0.005
Results are presented as a mean (ISD).
In an attempt to control for other factors when estimating the relative risk, a logistic regression analysis was performed with the dependent variable being presence or absence of any abnormality on the EMG examination. Any estimate of relative risk with a 95% confidence interval that contains the integer 1.0 is not considered significant and thus could have occurred by chance. This analysis produces an odds ratio that is an estimate of the relative risk for this population. The odds ratio of 1.55 (95% CI: 1.1, 2.2) for the median motor latency signifies a 55% increased risk for each increase of lmsec in the motor latency. When comparing an individual with a latency of 5.5msec to someone with a latency of 3.5msec, the increased latency of 2msec would translate into a 110% increased risk of finding EMG abnormalities. Similarly, there is a decreased risk of finding an abnormal EMG examination if the evoked amplitude is larger. For each lmV increase in the motor evoked response, the risk of finding an abnormality on the EMG examination is reduced 9% (odds ratio 0.91, 95% CI: 0.84, 0.99). For each additional year of life, there is a 2% increase in the risk for an abnormal EMG examination (odds ratio 1.02, 95% CI: 1.01, 1.04). Neither the median sensory amplitude and latency nor the ratio of median to ulnar motor amplitude were significant independent risk factors in this analysis. This model only explains 12% of the variance. DISCUSSION This study shows that NCS are poorly predictive of abnormal EMG findings on an individual basis. The best model to predict the presence of spontaneous activity or motor unit changes in the abductor pollicis brevis (APB) could only explain 12% of the variance. We documented that patients with an abnormal EMG have smaller amplitude median evoked responses and more prolonged latencies, and that older patients with CTS are more likely to have an abnormal EMG examination. There was no influence of gender or the ratio of the median motor to ulnar motor amplitude on EMG findings. The percentage of patients with CTS who had EMG abnormalities was 48%. This is in line with several other studies reporting 40% and 41% abnormal EMG in patients of Arch Phys Med Rehabil Vol 76, March 1995
Studies by Buchthal e t al, 3 and Marinacci ~2 noted a much higher percentage (91% and 96% respectively), but most likely described a population of more advanced or severe cases. The percentage of cases with fibrillation potentials also reflected that of Stevens 9 and Kimura. 11 We found 22% of patients with CTS had spontaneous activity compared with their findings of 18% and 22% respectively. The linear relationship between the degree of MUAP amplitude changes on needle examination and NCS abnormalities was suspected, but this is the first time that it has been analyzed statistically. Despite this relatively linear relationship, we still are unable to use this information to accurately predict the EMG findings. The sensitivity of a single parameter, with a defined threshold, in predicting an abnormal EMG examination ranged from 57% to 68%, but the specificity was poor, ranging from 58% to 61%. Combining parameters did not improve these figures. Other thresholds were tested, but all sacrificed sensitivity for specificity. This study does support the contention that a decrease in SNAP amplitude is more sensitive than a decreased CMAP amplitude in documenting axonal lOSS. 5'7'9"11 There was a drop in amplitude in both measures in the cases with EMG abnormalities in the APB, but this was more pronounced in the sensory amplitude. Also the CMAP amplitude was still within the normal range for most cases whereas the SNAP amplitudes were generally below the normal cutoff. The median motor and sensory latencies were strongly correlated with the presence of axonal loss. The lack of value in comparing the median motor with the ulnar motor amplitude was C T S . 9'11
Table 3: The Mean Median Motor and Sensory Evoked Amplitude and Latency of the Patients Stratified by EMGGraded Amplitude or Graded Spontaneous Activity EMG Amplitude Grade Evoked Responses
Median motor amplitude (mV) Median sensory amplitude (uV) Median motor latency (msec) Median sensory latency (msec)
(n = 324)
+/(n = 44)
1+ (n = 87)
2+ (n = 28)
3+ (n = 8)
9.5 (4.7)
8.4 (3,8)
7.7 (3.7)
6.4 (3.4)
6.4 (3.6)
20.7 (10.4)
17.8 (11.3)
13.9 (8.1)
12.5 (8.9)
10,7 (7.3)
4.4 (0.9)
4.9 (1.0)
5.0 (1.2)
5.1 (1.0)
4.9 (0.8)
4.2 (0.7)
4.5 (0,9)
4.6 (1.0)
4.8 (0.9)
4,8 (0.6)
Normal
EMG: Spontaneous Activity Grade
Median motor amplitude (mV) Median sensory amplitude (uV) Median motor latency (msec) Median sensory latency (msec)
Normal (n = 391)
+/-
1+
2+
3+
(n = 14)
(n = 45)
(n = 24)
(n = 4)
9.0 (4.4)
9.6 (4.0)
8.6 (5.5)
7.9 (4.0)
2.5 (2.0)
19.5 (10.4)
16.3 (11.8)
15.2 (9.4)
11.5 (7.4)
11.2 (7.1)
4.4 (0.9)
5.0 (1.2)
5.0 (1.5)
5.2 (1.0)
6.8 (2.6)
4.3 (0.7)
4.8 (0.9)
4.5 (0.9)
4.8 (1.0)
4.8 (0.9)
Results are presented as a mean (1SD).
NEEDLE ELECTROMYOGRAPHY AND NCS IN CTS, Werner
surprising but seems to reflect the extreme variability of the motor amplitudes and relative resistance of the motor axons in CTS patients. We feel that the EMG evaluation is an important component of the screening of patients with CTS. It helps rule out a more proximal nerve entrapment and may show the extent of thenar muscle denervation. The information from this evaluation may be important in making clinical management decisions related to C T S , 2 and this information cannot be predicted from the NCS. References 1. American Association of Electrodiagnostic Medicine, American Academy of Neurology, American Academy of Physical Medicine and Rehabilitation: Practice parameters for electrodiagnostic studies in Carpal tunnel syndrome: summary statement. Muscle Nerve 1993;16(12): 1390-1. 2. Hatter BT. Indications for surgery in work-related compression neuropathies of the upper extremity. Occup Med: State Art Rev 1989;4(3):48595. 3. Buchthal F, Rosenfalck A. Sensory conduction from digit to palm and
249
from palm to wrist in the carpal tunnel syndrome. J Neurol Neurosurg • Psychiatry 1971;34:243-52. 4. Eisen A. Electrodiagnosis of radiculopathy. Neurol Clin 1985;3(3):495510. 5. Felsenthal G. An overview of the clinical application of electromyography and nerve conduction techniques. Part III. prognostic and treatment implications. MD State Med J 1982;31(11):60-2. 6. Wilbourn AJ. Electrodiagnosis of plexopathies. Neurol Clin 1985;3 (3):511-29. 7. Jackson D, Clifford JC. Electrodiagnosis of mild carpal tunnel syndrome. Arch Phys Med Rehabil 1989;70(3):199-204. 8. Redmond MD, Rivner MH. False positive electrodiagnostic tests in carpal tunnel syndrome. Muscle Nerve 1988; 11(5):511-8. 9. Stevens JC. AAEM mini monograph #26: The electrodiagnosis of carpal tunnel syndrome. Muscle Nerve 1987; 10(2):99-113. 10. Kimura J. Electrodiagnosis in disease of nerve and muscle: principles and practice. Philadelphia: Davis, 1983:267. 11. Kimura I, Ayyar DR. The carpal tunnel syndrome: electrophysiologic aspects of 639 symptomatic extremities. Electromyogr Clin Neurophysiol 1985;25:151-64. 12. Marinacci AA. Comparative value of conduction velocity and electromyography in the diagnosis of carpal tunnel syndrome• Arch Phys Med Rehabil 1964;45:548-54.
Arch Phys Med Rehabil Vol 76, March 1995
Relation Between Needle Electromyography and Nerve Conduction Studies in Patients With Carpal Tunnel Syndrome Robert A. Werner, MD, James W. Albers, MD, PhD ABSTRACT. Werner RA, Albers JW. Relation between needle electromyography and nerve conduction studies in patients with carpal tunnel syndrome. Arch Phys Med Rehabii 1995;76:246-9. • Four hundred eighty cases of electrodiagnostically confirmed carpal tunnel syndrome were reviewed to determine if the findings on nerve conduction studies could predict the presence or absence of fibrillation potentials or motor unit changes on the needle examination of the abductor pollicis brevis (APB). The needle examination is more uncomfortable and the ability to predict the findings in this setting from standard nerve conduction studies (NCS) would make the test more acceptable to patients. All patients had median and ulnar nerves (both sensory and motor) tested, as well as the needle evaluation of the APB. Two hundred thirty-one patients had an abnormal needle evaluation as defined by presence of one of the following conditions: abnormal spontaneous activity, increased motor unit action potential (MUAP) amplitude, or increased M U A P polyphasia. One hundred five patients had fibrillation potentials. The mean median motor and sensory amplitudes and latencies, as well as age, did differ in the normal and abnormal needle examination groups, but the sensitivity for predicting an abnormality ranged from 57% to 68%. The ratio of the median to the ulnar amplitudes did not improve the sensitivity of predicting the abnormal needle findings. Motor and sensory evoked potential latencies were the most important predictors of an abnormal needle examination.
© 1995 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and
Rehabilitation The standard electrodiagnostic screening examination for carpal tunnel syndrome (CTS) includes nerve conduction studies (NCS) but the needle electromyography (EMG) of the median innervated muscles in the hand is considered optional by some electromyographers. ~ The NCS provide information about the conduction velocity of the nerve fibers and an estimate of the number of nerve fibers activated. Because CTS is primarily a focal demyelinating process the NCS are very sensitive to relative changes in the conduction velocity across the carpal canal. Axonal loss is found in more severe cases and is reflected in a decreased amplitude of the median motor and sensory evoked responses, and presence of abnormal spontaneous activity or motor unit action potential (MUAP) changes (large amplitude, polyphasia, decreased recruitment, or increased duration). The NCS can establish the diagnosis but does not clearly define the severity. The purpose of this study was to determine if the findings of axonal loss noted on the needle examination could be predicted from the standard NCS evaluation. The needle examination in the hand is particularly uncomfortable and usually is not a critical component in establishing the diagnosis of CTS. ~ It is important in determining the presence of From the Physical Medicine and Rehabilitation VA Medical Center (Dr. Werner); Department of Physical Medicine and Rehabilitation (Drs. Wemer, Albers), Department of Neurology (Dr. Albers), University of Michigan Medical Center, Ann Arbor, MI. Submitted for publication May 24, 1994. Accepted in revised form August 24, 1994. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organizations with which the authors are associated. Reprint requests to Robert A. Wemer, MD, Physical Medicine and Rehabilitation Service (117), Ann Arbor VA Medical Center, 2215 Fuller Rd, Ann Arbor, MI 48105. © 1995 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 0003-9993/95/7603-308553.00/0
Arch Phys Med Rehabil Vol 76, March 1995
axonal loss and ongoing denervation in the abductor pollicis brevis or opponens muscle. This information frequently is used to make a clinical treatment decision, specifically if surgery should be pursued) If the information about axonal loss could be predicted from the data collected during the NCS, then the patient would be spared the unpleasant needle evaluation with little sacrifice in information. The needle examination remains important to identify proximal lesions such as cervical radiculopathy or more proximal median nerve entrapment. The relationship of sensory nerve action potential (SNAP) and compound muscle action potential (CMAP) with EMG findings is poorly defined. Several authors 36 suggested that a 50% drop in CMAP was associated with axonal loss but this has not been rigorously investigated. The association between the presence of denervation potentials and measures of SNAP and CMAP have not been reported. This study attempts to model the electrophysiologic measures, as well as age, height, and weight, to determine if the results of the needle examination can be predicted. METHODS
This study was a retrospective review of all cases of median mononeuropathy from the EMG laboratory database that were diagnosed in the last 3 years. Cases were selected if they included evaluation of the right arm with median and ulnar motor and sensory evoked responses and needle evaluation of the right abductor pollicis brevis. Any case that also had a diagnosis of peripheral polyneuropathy, cervical radiculopathy, brachial plexopathy, or more proximal median mononeuropathy was excluded from the evaluation. Diagnosis of a median mononeuropathy was based on a relative prolongation of the peak median sensory response of 0.5msec compared with the ulnar sensory response (14cm,
NEEDLE ELECTROMYOGRAPHY AND NCS IN CTS, Werner
antidromic) or a 0.3msec prolongation of the median compared with the ulnar midpalmar peak sensory evoked response. 7-9 An abnormal EMG examination was defined as any study that found abnormal spontaneous activity, large MUAPs or polyphasic MUAPs. All the needle examinations were performed with disposable concentric needles and the evaluation was of the low threshold motor units. The low and high frequency filters were set at 20Hz and 10kHz, respectively. The degree of increase in the MUAP amplitude was subjectively graded on a scale of 0 to 4. Zero represented a normal amplitude (<2mV) and 4+ represented units over 8mV in amplitude. Every attempt was made to optimize the rise time to report the maximum amplitude accurately. Polyphasia was graded subjectively on a scale of 0 to 4 with greater than 10% being defined as abnormal. Trigger and delay line technique was used for all evaluations of polyphasia. Fibrillation potentials also were graded on a 0 to 4 scale as define as follows: 0 no abnormal spontaneous activity; + equivocal findings like a single run or train of persistent fibrillation or very slow trains o f positive waves; 1+ unequivocal positive waves or fibrilla~ons in two areas of muscle that persist greater than 2 seconds; 2+ spontaneous potentials at three or more sites; 3+ spQntaneous activity in nearly all sites; 4 + abundant spontaneous activity nearly filling the screen and obscuring the baseline (adapted from Kimural°). The analysis consisted of dividing the population based on the presence or absence of abnormalities on the EMG examination and identifying the differences in the median sensory or motor amplitude or latency between the two subgroups. Any abnormality, such as an increased amplitude, abnormal polyphasia, or abnormal spontaneous activity, was classified as an abnormal EMG examination. The population also was grouped according to the presence or absence of fibrillation potentials, the most objective needle examination measurement. The ratio of the median motor to ulnar motor amplitude also was analyzed by the normal versus abnormal EMG evaluation; the distribution of these variables was evaluated and risk ratios determined for various electrophysiologic measures. The sensitivity and specificity of several threshold values for the SNAP and CMAP for predicting the presence of abnormalities on the EMG examination were determined. The electrophysiologic measures and demographic data were combined and, using logistic regression, a model was created to predict the presence or absence of EMG abnormalities.
RESULTS Four hundred eighty patients met the entrance criteria and were used as the basis of analysis. The;mean age was 47.5 years (range 17 to 87) with 65% being women. Two hundred thirty-one patients had an abnormal EMG examination, but only 105 had abnormal spontaneous activity. The mean amplitude and latency of the median and ulnar sensory and motor evoked responses, as well as the ratio of median to ulnar amplitude for patients with normal and abnormal EMG examination, are presented in table 1. There was a significant difference between the two group in regard to the median sensory and motor evoked amplitudes and latencies but not
247
Table 1: Comparison of Electrophysiologic Measures in Patients With and Without Abnormalities on EMG Examination Needle Examination
Median sensory amplitude (uV) Median motor amplitude (mV) Median sensory latency (msec) Median motor latency (msec) Ratio: median/ulnar motor amplitude Age (yrs)
Normal EMG (n = 249)
Abnormal EMG (n = 231)
T Test
p Value
21.4 (10.2)
15.1 (9.0)
7.06
<0.001
9.5 (4.0)
7.8 (3.8)
4.52
<0.001
4.1 (0.7)
4.5 (0.8)
6.39
<0.001
4.3 (0.8)
4.9 (1.1)
6.78
<0.001
0.9 (0.4)
0.9 (0.6)
0.70
0.49
45.3 (14.2)
53.0 (14.5)
6.23
<0.001
Results are presented as a mean (ISD).
for the ratio of median motor to ulnar motor amplitude. The patients with an abnormal EMG examination had a more prolonged latency and a smaller amplitude for the motor and sensory evoked responses. The patients with an abnormal EMG examination also tended to be older (53 compared with 45 years). The same electrophysiologic measures are presented in table 2, dichotomized by the presence or absence of abnormal spontaneous activity. The 105 patients with abnormal spontaneous activity had smaller amplitudes and more prolonged latencies compared with the group with any abnormal findings on their needle examination. Again this group was significantly worse than the group without evidence of spontaneous activity. Note that the ratio of the median motor to the ulnar motor amplitude was not different when comparing the groups with and without evidence of abnormal spontaneous activity. Table 3 shows the relationship between the graded MUAP amplitude found on EMG and the amplitudes and latencies of the evoked responses in the hand. There is a clear trend noted for both evoked amplitude and latency. The larger the MUAP amplitude, the smaller the evoked response and the more prolonged the latency. This relationship held for both sensory and motor responses. There was less variability of the latencies compared with the amplitude measures as represented by the relatively large standard deviation of the amplitude measures. A similar but less robust relationship was noted with the graded spontaneous activity and the electrophysiologic measures on NCS (see table 3). The trend was strongest with the median sensory amplitude and weakest with the median motor and sensory latencies. The relative risk of finding abnormal spontaneous activity on EMG examination, if the median sensory amplitude is less than 12#V, is 2.23 (95% CI: 1.6, 3.1). This means that the finding of abnormal spontaneous activity is twice as likely compared with someone with a larger amplitude. If the median motor latency is more than 5. lmsec, the relative risk for abnormal spontaneous activity is 2.44 (95% CI: 1.8, 3.4). If the median sensory latency is greater than 4.8msec, the relative risk is 2.43 (95% CI: 1.7, 3.4). Arch Phys Med Rehabil Vo| 76, March 1995
248
NEEDLE ELECTROMYOGRAPHY AND NCS IN CTS, Werner
Table 2: Comparison of Electrophysiologic Measures in Patients With and Without Spontaneous Activity on EMG Examination Spontaneous Activity
Median sensory amplitude (uV) Median motor amplitude (mV) Median sensory latency (msec) Median motor latency (msec) Ratio: median/ulnar motor amplitude Age (yrs)
Normal (n = 375)
Abnormal
(n = 105)
T Test
p
19.6 (10.2)
13.9 (8.9)
5.30
<0.001
9.0 (3.9)
7.5 (4.0)
2.74
0.006
4.2 (0.7)
4.6 (0.9)
5.31
<0.001
4.4 (0.8)
5.1 (1.3)
6.32
<0.001
0.9 (0.5)
0.9 (0.6)
0.12
0.91
48.1 (14.5)
52.5 (14.4)
2.81
0.005
Results are presented as a mean (ISD).
In an attempt to control for other factors when estimating the relative risk, a logistic regression analysis was performed with the dependent variable being presence or absence of any abnormality on the EMG examination. Any estimate of relative risk with a 95% confidence interval that contains the integer 1.0 is not considered significant and thus could have occurred by chance. This analysis produces an odds ratio that is an estimate of the relative risk for this population. The odds ratio of 1.55 (95% CI: 1.1, 2.2) for the median motor latency signifies a 55% increased risk for each increase of lmsec in the motor latency. When comparing an individual with a latency of 5.5msec to someone with a latency of 3.5msec, the increased latency of 2msec would translate into a 110% increased risk of finding EMG abnormalities. Similarly, there is a decreased risk of finding an abnormal EMG examination if the evoked amplitude is larger. For each lmV increase in the motor evoked response, the risk of finding an abnormality on the EMG examination is reduced 9% (odds ratio 0.91, 95% CI: 0.84, 0.99). For each additional year of life, there is a 2% increase in the risk for an abnormal EMG examination (odds ratio 1.02, 95% CI: 1.01, 1.04). Neither the median sensory amplitude and latency nor the ratio of median to ulnar motor amplitude were significant independent risk factors in this analysis. This model only explains 12% of the variance. DISCUSSION This study shows that NCS are poorly predictive of abnormal EMG findings on an individual basis. The best model to predict the presence of spontaneous activity or motor unit changes in the abductor pollicis brevis (APB) could only explain 12% of the variance. We documented that patients with an abnormal EMG have smaller amplitude median evoked responses and more prolonged latencies, and that older patients with CTS are more likely to have an abnormal EMG examination. There was no influence of gender or the ratio of the median motor to ulnar motor amplitude on EMG findings. The percentage of patients with CTS who had EMG abnormalities was 48%. This is in line with several other studies reporting 40% and 41% abnormal EMG in patients of Arch Phys Med Rehabil Vol 76, March 1995
Studies by Buchthal e t al, 3 and Marinacci ~2 noted a much higher percentage (91% and 96% respectively), but most likely described a population of more advanced or severe cases. The percentage of cases with fibrillation potentials also reflected that of Stevens 9 and Kimura. 11 We found 22% of patients with CTS had spontaneous activity compared with their findings of 18% and 22% respectively. The linear relationship between the degree of MUAP amplitude changes on needle examination and NCS abnormalities was suspected, but this is the first time that it has been analyzed statistically. Despite this relatively linear relationship, we still are unable to use this information to accurately predict the EMG findings. The sensitivity of a single parameter, with a defined threshold, in predicting an abnormal EMG examination ranged from 57% to 68%, but the specificity was poor, ranging from 58% to 61%. Combining parameters did not improve these figures. Other thresholds were tested, but all sacrificed sensitivity for specificity. This study does support the contention that a decrease in SNAP amplitude is more sensitive than a decreased CMAP amplitude in documenting axonal lOSS. 5'7'9"11 There was a drop in amplitude in both measures in the cases with EMG abnormalities in the APB, but this was more pronounced in the sensory amplitude. Also the CMAP amplitude was still within the normal range for most cases whereas the SNAP amplitudes were generally below the normal cutoff. The median motor and sensory latencies were strongly correlated with the presence of axonal loss. The lack of value in comparing the median motor with the ulnar motor amplitude was C T S . 9'11
Table 3: The Mean Median Motor and Sensory Evoked Amplitude and Latency of the Patients Stratified by EMGGraded Amplitude or Graded Spontaneous Activity EMG Amplitude Grade Evoked Responses
Median motor amplitude (mV) Median sensory amplitude (uV) Median motor latency (msec) Median sensory latency (msec)
(n = 324)
+/(n = 44)
1+ (n = 87)
2+ (n = 28)
3+ (n = 8)
9.5 (4.7)
8.4 (3,8)
7.7 (3.7)
6.4 (3.4)
6.4 (3.6)
20.7 (10.4)
17.8 (11.3)
13.9 (8.1)
12.5 (8.9)
10,7 (7.3)
4.4 (0.9)
4.9 (1.0)
5.0 (1.2)
5.1 (1.0)
4.9 (0.8)
4.2 (0.7)
4.5 (0,9)
4.6 (1.0)
4.8 (0.9)
4,8 (0.6)
Normal
EMG: Spontaneous Activity Grade
Median motor amplitude (mV) Median sensory amplitude (uV) Median motor latency (msec) Median sensory latency (msec)
Normal (n = 391)
+/-
1+
2+
3+
(n = 14)
(n = 45)
(n = 24)
(n = 4)
9.0 (4.4)
9.6 (4.0)
8.6 (5.5)
7.9 (4.0)
2.5 (2.0)
19.5 (10.4)
16.3 (11.8)
15.2 (9.4)
11.5 (7.4)
11.2 (7.1)
4.4 (0.9)
5.0 (1.2)
5.0 (1.5)
5.2 (1.0)
6.8 (2.6)
4.3 (0.7)
4.8 (0.9)
4.5 (0.9)
4.8 (1.0)
4.8 (0.9)
Results are presented as a mean (1SD).
NEEDLE ELECTROMYOGRAPHY AND NCS IN CTS, Werner
surprising but seems to reflect the extreme variability of the motor amplitudes and relative resistance of the motor axons in CTS patients. We feel that the EMG evaluation is an important component of the screening of patients with CTS. It helps rule out a more proximal nerve entrapment and may show the extent of thenar muscle denervation. The information from this evaluation may be important in making clinical management decisions related to C T S , 2 and this information cannot be predicted from the NCS. References 1. American Association of Electrodiagnostic Medicine, American Academy of Neurology, American Academy of Physical Medicine and Rehabilitation: Practice parameters for electrodiagnostic studies in Carpal tunnel syndrome: summary statement. Muscle Nerve 1993;16(12): 1390-1. 2. Hatter BT. Indications for surgery in work-related compression neuropathies of the upper extremity. Occup Med: State Art Rev 1989;4(3):48595. 3. Buchthal F, Rosenfalck A. Sensory conduction from digit to palm and
249
from palm to wrist in the carpal tunnel syndrome. J Neurol Neurosurg • Psychiatry 1971;34:243-52. 4. Eisen A. Electrodiagnosis of radiculopathy. Neurol Clin 1985;3(3):495510. 5. Felsenthal G. An overview of the clinical application of electromyography and nerve conduction techniques. Part III. prognostic and treatment implications. MD State Med J 1982;31(11):60-2. 6. Wilbourn AJ. Electrodiagnosis of plexopathies. Neurol Clin 1985;3 (3):511-29. 7. Jackson D, Clifford JC. Electrodiagnosis of mild carpal tunnel syndrome. Arch Phys Med Rehabil 1989;70(3):199-204. 8. Redmond MD, Rivner MH. False positive electrodiagnostic tests in carpal tunnel syndrome. Muscle Nerve 1988; 11(5):511-8. 9. Stevens JC. AAEM mini monograph #26: The electrodiagnosis of carpal tunnel syndrome. Muscle Nerve 1987; 10(2):99-113. 10. Kimura J. Electrodiagnosis in disease of nerve and muscle: principles and practice. Philadelphia: Davis, 1983:267. 11. Kimura I, Ayyar DR. The carpal tunnel syndrome: electrophysiologic aspects of 639 symptomatic extremities. Electromyogr Clin Neurophysiol 1985;25:151-64. 12. Marinacci AA. Comparative value of conduction velocity and electromyography in the diagnosis of carpal tunnel syndrome• Arch Phys Med Rehabil 1964;45:548-54.
Arch Phys Med Rehabil Vol 76, March 1995