Journal of the Neurological Sciences, 1978, 38: 1-10 © Elsevier/North-HollandBiomedicalPress

1

A METHOD FOR DETERMINING MEDIAN NERVE CONDUCTION VELOCITY ACROSS THE CARPAL TUNNEL

JUN KIMURA* Division of Clinical Electrophysiology, Department of Neurology, College of Medicine, University of Iowa, Iowa City, Iowa 52242 (U.S.A.)

(Received26 January, 1978) (Accepted 7 February, 1978)

SUMMARY Palmar stimulation was used to assess median nerve conduction across the carpal tunnel. In 50 hands from 25 control subjects, motor and sensory lateneies in the wristto-palm segment (mean 4- SD: 1.15 4- 0.21 msec and 1.12 :k 0.21 msec respectively) were less than half the conventional terminal latencies in the wrist-to-muscle and wristto-digit segment (3.01 -4- 0.44 msec and 2.47 -4- 0.39 msec). Motor and sensory conduction velocities (MNCV and SNCV) in the wrist-to-palm segment (56.0 -4- 7.6 m/sec and 58.7 + 7.5 m/see respectively) were comparable to those in the elbow-to-wrist segment (57.0 q- 4.5 m/see and 62.4 + 5.7 m/see). In 20 symptomatic hands from 13 patients with mild carpal tunnel syndrome, delay in motor and sensory terminal latencies (3.91 + 0.67 msec and 2.90 q- 0.57 msec) was primarily attributable to increased conduction time in the wrist-to-palm segment (1.96 -4- 0.59 msec and 1.58 -4- 0.49 msec) and not in the remaining more distal portions. Consequently, MNCV and SNCV were significantly(P < 0.001) slowed when calculated in the segment across the carpal tunnel (36.6 4- 11.2 m/see and 44.9 4- 11.8 m/see), even though the conventional terminal lateneies from the stimulus site at the wrist were often within normal limits.

INTRODUCTION The clinical value of nerve conduction studies has long been established and the technique for evaluating the median nerve is now well standardized (Hodes, Larrabee and German 1948; Lambert 1962; Kaeser 1970; Daube 1977). In the conventional method, electrical stimulation is applied at the wrist and elbow, and less commonly at * Correspondingauthor: Jtm Kimura, M.D., Professorof Neurology,Chief,Divisionof Clinical Electrophysiology,UniversityHospitalsand Clinics,Iowa City,Iowa 52242, U.S.A.

the axilla, where selective activation of the median nerve is often difficult because of its proximity to the ulnar nerve (Kimura 1976). Additionally, although apparently not widely known, median motor fibers are accessible with electrodes placed at midpalm where the nerve branches to innervate thenar muscles. With activation of the nerve at the palm, sensory nerve action potentials can also be recorded antidromically from an electrode around the second or third digits. Palmar stimulation, therefore, allows calculation of latency and conduction velocity of motor and sensory impulses across the carpal tunnel in the wrist-to-palm segment. The purpose of this communication is to describe this technique, to establish normal values and variations, and to discuss abnormalities found in a few patients with the carpal tunnel syndrome. METHODS The median nerve was stimulated using surface electrodes at 3 points along its course and sensory nerve and muscle action potentials were recorded in the conventional manner. At the elbow, shocks were applied on the anterior surface of the upper arm over the brachial pulse. For stimulation at the wrist, the cathode was one cm proximal to the distal crease on the volar surface. The terminal portion of the median nerve was stimulated at midpalm over the second carpometacarpal space (Fig. 1). In a few normal subjects, additional stimuli were given at several points across the wrist to demonstrate

Fig. 1. Stimulation of the median nerve with the cathode at midpalm and anode 2 cm proximally. Sensory nerve and muscle action potentials were recorded from the second digit and abductor pollicis brevis respectively. Although simultaneous recordings o f these two potentials were possible using a two-channel recorder, they were ordinarily assessed separately with shocks of just maximal intensity.

Sensory Nerve Potential

Site of Stimulation

.[

0

......

.

Muscle Action Potential

, ',,,.,

,

, ~.~

1 2 3

;

~''.

/

;"

%

• ..~,..,,.,,~s

4

f

"i~"

f ~ ' "

~..~ .-.

5 6

t' ,:

',N ~ ",.."'.'x~.~, - ~

7 ! ::

8 9

. %.-" k ",

/'

,

.

10 11

. p - ' ~ . .~. ~", ~-- _ _ ~

~

.

- •

12 Stimulus

0.5 msec

0.5 msec

Fig. 2. Sensory nerve and muscle action potentials in a normal subject recorded after stimulation of the median nerve at multiple points across the wrist. The site of each stimulus is indicated by a number on the left representing the distance in cm from the most distal stimulus point at midpalm. Both motor and sensory latencies increased progressively as the stimulus site was moved proximally along the course of the nerve.

progressive increase of' motor and sensory latencies as the stimulus site was moved proximally (Fig. 2). The anode was 2 cm proximal to the cathode and the ground electrode was placed around the forearm. Based on the previously described anatomical course (Robbins 1963), palmar stimulation was aimed at the thenar nerve, a recurrent branch of the median nerve innervating the thenar muscles. In each hand tested, the palm was stimulated at different sites and the most distal point of stimulation was chosen which produced an appropriate thumb twitch indicating contraction of median innervated thenar muscles. Johnson and Shrewbury (1970) suggested that a line drawn from the scaphoid to the radial cleft of the ring finger would intersect the thenar flexion crease at about the location of the thenar nerve. The site of palmar stimulation in our study was usually slightly distal to tbis point. Although motor response was used for selection of the stimulus site, anti° dromic sensory nerve potentials could usually be elicited without adjusting the location of the stimulating electrodes. To overcome the problem of stimulus artifact a specially designed amplifier with short blocking time (1.0 msec) and a low noise (0.5 #V RMS at bandwidth of 2,000

c/sec) was developed. Full details of this amplifier will be described elsewhere (Walker and Kimura, in preparation). In brief, the input and interstage coupling capacitators are eliminated so that the feed-forward path from electrodes to output is strictly a DC amplifier. The amplifier thus recovers virtually immediately because it contains no significant capacitators to store charge. To establish a low frequency cutoff and to eliminate DC electrode offset, an integrator feedback loop subtracts an offset signal from the stage 1 input. To achieve fast recovery, a circuit detects if the stage 1 output exceeds limits near saturation in either direction and, via control circuit, opens the feedback loop. Frequency response used was 20-2,000 c/sec for sensory nerve and 20-32,000 c/sec for muscle action potentials. Recording electrodes for muscle potentials were placed over the belly in the center of the abductor pollicis brevis (G1) and tendon just distal to the metacarpophalangeal joint (G2). The sensory nerve action potentials were recorded antidromically with ring electrodes placed around the proximal (GD and distal (G2) interphalangeal joints of the second digit. At least 4 trials were given at each stimulus site to confirm the consistency of the recorded response. Shock intensity was gradually advanced until further increase no longer altered the size of sensory nerve or muscle action potentials. In general, less intense stimulation was required to elicit maximal sensory than maximal motor response. The voltage required to activate maximal response was generally similar at different sites of stimulation. In some subjects greater intensity was necessary at the palm, which, combined with the proximity of the stimulus point to the recording electrodes, tended to cause a larger stimulus artifact. This problem was circumvented by the use of a fast recovery amplifier. Both motor and sensory latencies were measured from the stimulus artifact to the onset of the initial negative peak (upward deflection). With stimulation at the palm or wrist, the latencies were measured using a sweep speed of 0.2 msec, 0.5 msec or 1.0 msec per division. To determine the latency, a marker was positioned to the desired spot of the waveform on the oscilloscope, and the value was automatically displayed digitally. The amplitude was measured from the baseline to the negative peak for both sensory nerve and muscle potentials using a sensitivity of 20/~V and 2 mV per division respectively. The surface distance between the two stimulus sites at the wrist and palm was carefully measured in mm with a ruler. Motor and sensory nerve conduction velocities (MNCV and SNCV) in the wrist-to-palm segment was determined according to the conventional formula, dividing the distance by the latency difference between the two evoked potentials. RESULTS

(I) Normal values The normal range of latency and amplitude were established in 25 subjects, 8 healthy volunteers and 17 patients with unrelated disorders. There were 13 males and 12 females ranging in age from 15 to 58 years (average age, 36). All subjects were studied on both sides. In each hand tested, sensory nerve and muscle action potentials were easily elicited with palmar stimulation. Normal values and variations among different

Muscle Action PotenUal

/

Normal

\

/

I

/\

\,

/

CTS 1

-

!

/\

/ CTS 2

I

/ i

i

CTS 3

/"

_J

\

I

/

\ ~\

J

\

SUmulus at palm

I

I

2•V

Stimulus at wrist

1 msec

Fig. 3. Muscle action potential recorded from abductor pollicis brevis after stimulation of the median nerve at midpalm and at wrist. Compared to a normal hand (top tracing), motor latencies were increased not only over the wrist-to-muscle but also palm-to-muscle segment in moderately advanced carpal tunnel syndrome (CTS 1 and 2). In contrast, the palm-to-muscle latency was normal whereas the wrist-to-muscle latency was considerably increased in a mild case (CTS 3).

Sensory Nerve Potential

/ Normal

CTS 1

t~,

. , - !,

i

I

!

I

)

I

"~" i '., t

CTS 2 i

I

I

\, CTS 3 ~

",,..,,:

,w.l~,d.~-,,,%, ,~,%.. " )

Stimulus at palm

I

Stimulus at wrist

-,,."~"~ ..... i

120 I~V 1 msec

Fig. 4. Sensory nerve potential antidromically recorded from the second digit after stimulation of the median nerve at midpalm and at wrist in the same hands shown in figure 3. Like motor latancies, wristto-digit sensory latencies were markedly increased in all three symptomatic hands (CTS 1, 2 and 3) when compared to a normal hand (top tracing). The palm-to-digit latency was normal in a mild case (CTS 3) but slightly increased in more severely affected hands (CTS 1 and 2).

TABLE 1 NORMAL VALUES (MEAN A: SD) Latency difference (msec)

Conduction velocity (m/sec)

Nerve Amplitude segments (mY for motor, tested /~V for sensory)

Latency (msec)

Motor fibers 50 hands from 25 subjects

palm wrist elbow

7.6 4- 3.6 7.2 4- 3.5 7.1 4- 3.4

1.86 4- 0.36 3.01 4- 0.44 1.15 i 0.21 56.0 4- 7.6 7.34 4- 0.78 4.33 4- 0.47 57.0 4- 4.5

Sensory fibers 50 hands from 25 subjects

digit palm wrist elbow

41.8 4- 21.7 38.6 4- 18.7 23.4 4- 9.2

1.34 4- 0.31 59.9 4- 9.2 2.47 4- 0.39 1.21 ± 0.21 58.7 4- 7.5 6.40 4- 0.77 3.90 4- 0.50 62.4 4- 5.7

subjects are summarized in Table 1. There was no difference between responses on the right and left sides. Therefore, the two sides were combined for calculation of the mean and standard deviation. In normal subjects, m o t o r and sensory latencies in the wrist-to-palm segment were significantly (P < 0.05) smaller than the conduction time in the remaining distal segments to the muscle and digit respectively. There was no significant difference between M N C V in the elbow-to-wrist segment and that in the wrist-to-palm segment (P > 0.1). The SNCV was the same in the wrist-to-palm and palm-to-digit segments (P > 0. l) but significantly faster in the elbow-to-wrist segment (P < 0.01). The amplitude of muscle potential was the same whether the nerve was stimulated at the palm, wrist or elbow (P > 0.5). The antidromically activated sensory nerve potential became progressively smaller as the stimulating electrodes were moved proximally although the difference was statistically significant (P < 0.001) only between wrist and elbow and not between palm and wlist (P > 0.1). For the purpose of this study, the upper and lower limits of normal were defined as the mean 4- 2 SD for latencies and conduction velocities. Therefore, motor latencies in the wrist-to muscle and wrist-to-palm segment were considered delayed if they exceeded 3.9 and 1.6 msec resl:ectively. The M N C V was considered slow if it was less than 48 m/sec in the elbow-to-wrist segment and 41 m/sec in the wrist-to-palm segment. Similarly, sensory latencies in the wrist-to-digit and wrist-to-palm segment were regarded as abnormal if they were greater than 3.3 and 1.5 msec respectively. The SNCV was considered slow if it was less than 51 m/sec in the elbow-to-wrist segment and 44 m/sec in the wrist-to-palm segment.

(2) Carpal tunnelsyndrome The patient group consisted o f 3 males and 10 females 38-59 years of age (average age 47), all with a clinically typical but mild carpal tunnel syndrome. All patients were seen within a few months of the onset of clinical symptoms. Advanced cases with a longer duration of illness or those showing significant muscle atrophy were excluded. Although each patient was tested on both sides, only clinically affected hands were in-

TABLE 2 PATIENTS WITH MILD CARPAL TUNNEL SYNDROME (MEAN -4- SD) Nerve Amplitude Latency segments (mY for motor, ( m s e c ) tested /zV for sensory)

Latency Conduction difference velocity (m/sec) (msec)

Motor fibers palm 20 hands from 13 subjects wrist elbow

5.8 4- 2.4 4.9 -4- 2.2 4.5 4- 1.9

1.93 -4- 0.30 3.91 -4- 0.67 1.96 -4- 0.59 36.6 4- 11.2 8.17 -4- 0.90 4.35 -4- 0.39 54.7 -4- 4.5

Sensory fibers digit 20 hands from 13 subjects palm wrist elbow

30.3 -4- 16.4 28.5 4- 13.2 18.0 -4- 6.6

1.32 -4- 0.23 59.6 ± 8.6 2.90 4- 0.57 1.58 -4- 0.49 44.9 -4- 11.8 6.79 -4- 0.68 3.90 -4- 0.37 58.6 4- 7.4

cluded for analysis (Table 2). There were 20 symptomatic hands, 14 from 7 patients affected on both sides, and 6 from the remaining patients with unilateral disease, 3 on the right and 3 on the left. In each hand, sensory nerve and muscle action potentials were easily elicited with stimulation at the palm, wrist and elbow. Based solely on the conventional wrist-tomuscle latency, abnormalities of motor fibers were apparent in 10 (50 ~o) but not in the remaining 10 (50 ~). Palmar stimulation revealed slowed MNCV in the wrist-to-palm segment in 4 additional hands (20 ~), but the remaining 6 symptomatic hands (30 ~o) were undetected. Sensory nerve conduction was slowed on the basis of wrist-to-digit latency only in 4 ( 2 0 ~ ) and was normal in 16 hands (80~). An additional 5 ( 2 0 ~ ) hands were considered abnormal after palmar stimulation, but no slowing was detected in the remaining 11 (55 ~ ) symptomatic hands. Taking both motor and sensory conduction studies into consideration, electrophysiological abnormalities were found in all but 5 symptomatic hands (25 ~). If it had not been for palmar stimulation, however, 5 additional hands (25 ~ ) with clinical signs would have been regarded as normal. The motor and sensory latencies were significantly greater in patients than in normals over the wrist-to-palm segment (P < 0.001) but not over the remaining distal segment to the muscle and digit respectively (P > 0.1). Consequently, both MNCV and SNCV across the carpal tunnel were significantly reduced in patients when compared to normals (P < 0.001). In patients, unlike normals, the latency across the carpal tunnel was greater than that in the palm-to-muscle or palm-to-digit segments although the difference was statistically significant only for the sensory fibers (P < 0.05). In patients the MNCV and SNCV were both significantly slower in the wrist-to-palm segment than in the more distal or proximal segments (P < 9.001). Both sensory nerve and muscle action potentials were significantly (P < 0.05) smaller in amplitude than the corresponding normal values regardless of the site of stimulation. There was no significant difference in amplitude of muscle potentials elicited by stimulation at palm, wrist and elbow (P > 0.1). There was no significant difference in sensory nerve action potentials between stimulation at palm and wrist (P > 0.5), but the potentials elicited with stimulation at the elbow were significantly (P < 0.01) smaller.

COMMENTS Buchthal and Rosenfalck (1971) described a technique of determining SNCV in the median nerve in the palm-to-wrist segment. In their study, sensory fibers of the digits were stimulated either with surface or needle electrodes and orthodromic nerve action potentials were recorded using steel needles placed near the nerve at the palm and wrist. Electronic averaging of 500-2000 traces were employed when necessary. A sim ilar method was reported by Casey and Le Quesne (1972) who recorded digital nerve action potentials for determination of SNCV. More recently Eklund (1975) stimulated sensory fibers percutaneously at the palm and recorded nerve action potentials from the wrist using surface electrodes. With this technique, however, the recorded potentials may not be solely those of sensory fibers; antidromic motor potentials are registered at the wrist if motor fibers are simultaneously activated at the palm. In the present study, sensory nerve potentials were recorded from the second digit after stimulation of the median nerve at the palm, wrist and elbow. The potentials thus recorded antidromically were in general greater in amplitude than the previously reported orthodromic potentials (Wiederholt 1970; Buchthal and Rosenfalck 1971). The wrist-to-palm and palm-to-digit latencies were the same in normals whether tested antidromically as in the present study or orthodromically as was done by Buchthal, Rosenfalck and Trojaborg (1974). In agreement with these authors, we found slowing along the median nerve sensory fibers from wrist to palm in patients with clinical signs and symptoms of the carpal tunnel syndrome when conduction time from wrist to digit was normal. As a diagnostic test, therefore, the sensory fibers may be studied antidromically or orthodromically in assessing the wrist-to-palm and palm-to-digit segments of the median nerve. In contrast to the sensory fibers, conduction of the motor fibers has not previously been assessed in the wrist-to-palm segment. Instead, it has been the common practice to resort to the wrist-to-muscle conduction time, ordinarily referred to as the terminal latency. A delay in terminal latency, however, may be caused by increased conduction time in the wrist-to-palm segment or in the palm-to-muscle segment or both. Using the technique ofpalmar stimulation described here, it is possible to assess motor conduction in these two segments separately. Palmar stimulation also allows determination of MNCV in the segment across the carpal tunnel, the most common site of median nerve compression. The length of the nerve between the two stimulus points at the wrist and palm is sholter (mean ± SD: 64 ± 7 mm in 50 normal hands) than is ordinarily recommended for calculation of MNCV. Fortunately, however, the surface distance can be measured very accurately along a straight line in the wrist-to-palm segment. Further, using a fast sweep speed, the latency can be determined with accuracy considerably better than 0.1 msec. Therefore, despite the short length of the nerve segment in question, MNCV obtained by this technique should be reasonably accurate. Our findings support this contention since variation of MNCV was only slightly greater in this segment than in more proximal segments. Determination of nerve conduction velocity is important in clinical assessment of the carpal tunnel syndrome (Simpson 1956; Gilliatt and Sears 1958; Kaeser 1963;

9 Phalen 1966; Thomas, Lambert and Cseuz 1967; Duensing, Lowitzsch, Thornwirth and Vogel 1974; Daube 1977). In some of the patients included in this study, the palmto-muscle and palm-to-digit latencies were appreciably slower than normal, perhaps resulting from secondary degeneration of distal motor and sensory fibers respectively. A more consistent finding in the mild carpal tunnel syndrome, however, was selective slowing of MNCV and SNCV in the wrist-to-palm segment, indicating conduction abnormalities limited to the carpal tunnel. Since the time required for the passage of impulse in the wrist-to-palm segment is normally less than half that of the wrist-tomuscle or wrist-to-digit latency, exclusion of the normal conduction time in the segment distal to the palm should make it possible to demonstrate mild conduction delay across the involved segment. Indeed, in 5 (25 ~o) of 20 hands with clinical signs of the carpal tunnel syndrome, the conventional terminal latencies were normal, yet MNCV or SNCV in the wrist-to-palm segment was distinctly slow. The incidence of conduction abnormalities in our study was less than previously described by others probably because we selected patients with mild symptoms of recent onset (Buchthal and Rosenfalck 1971 ; Buchthal et al. 1974; Kaeser 1963; Phalen 1966; Thomas et al. 1967; Duensing et al. 1974). In these previous studies, sensory nerve conduction studies were generally considered the most sensitive indication of a lesion. This was not so in our subjects but a further study with a larger number of patients is necessary to clarify this discrepancy. It seems justified to conclude, however, that palmar stimulation is a useful addition to the conventional method in assessing median motor and sensory fibers for a lesion across the carpal tunnel. To allow a greater distance between the wrist and palm for improved accuracy of MNCV and SNCV, it is possible in most subjects to stimulate the median nerve at the wrist 3-5 cm proximal to the distal crease (Fig. 2). A problem inherent in determining nerve conduction time over short distances is the control of stimulus artifact. A fast recovery amplifier circumvents this difficulty and further improves accuracy of measurement of nerve conduction velocity in this region (Walker and Kimura, in preparation). ACKNOWLEDGEMENT The authors wish to thank Mr. D. David Walker, MSEE for electrical engineering assistance, and Ms. Sheila Mennen, Joanne Colter and Karen Thompson for technical assistance.

REFERENCES Buchthal, F. and A. Rosenfalck (1971) Sensory conduction from digit to palm and from palm to wrist in the carpal tunnel syndrome, J. Neurol. Neurosurg. Psychiat., 34: 243-252. Buchthal, F., A. Rosenfalck and W. Trojaborg (1974) Electrophysiologicalfindings in entrapment of the median nerve at wrist and elbow, J. Neurol. Neurosurg. Psychiat., 37" 340-360. Casey, E. B. and P. M. Le Quesne (1972) Digital nerve action potentials in healthy subjects, and in carpal tunnel and diabetic patient, J. Neurol. Neurosurg. Psychiat., 35: 612-623. Daube, J. R. (1977) Special Course in Clinical Electromyography, American Academy of Neurology, pp. 13-22.

I0 Duensing, F., K. Lowitzsch, V. Thorwirth and P. Vogel (1974) Neurophysiologische Befunde beim Karpaltunnelsyndrom, Z. Neurol., 206: 267-284. Eklund, G. (1975) A new electrodiagnostic procedure for measuring sensory nerve conduction across the carpal tunnel, Uppsala J. Med. Sci., 80: 63-64. Gilliatt, R. W. and T. A. Sears (1958) Sensory nerve action potentials in patients with peripheral nerve lesions, J. Neurol. Neurosurg. P6ychiat., 21 : 109-118. Hodes, R., M. G. Larrabee and W. German (1948) The human electromyogram in response to nerve stimulation and the conduction velocity of motor axons, Arch. Neurol. Psychiat., 60: 341)-365. Johnson, R. K. and M. M. Shrewsbury (1970) Anatomical course of the thenar branch of the median nerve - - Usually in a separate tunnel through the transverse carpal ligament, J. Bone Jt Surg., 52-A : 269-273. Kaeser, H. E. (1963) Diagnostische Probleme beim Karpaltunnelsyndrom, Dtsch. Z. Nervenheilk., 185: 453-470. Kaeser, H. E. (1970) Nerve conduction velocity measurements. In: P. J. Vinken and G. W. Bruyn (Eds.), Handbook of Clinical Neurology, Vol. 7 (Diseases of Nerves, Part I), North-Holland Publishing Comp., Amsterdam, pp. 116-196. Kimura, J. (1976) Collision technique - - Physiologic block of nerve impulses in studies of motor nerve conduction velocity, Neurology (Minneap.), 26: 680-682. Lambert, E. H. (1962) Diagnostic value of electrical stimulation of motor nerves, Electroeneeph. olin. NeurophysioL, Suppl., 22: 9-16. Phalen, G. S. (1966) The carpal tunnel syndrome, seventeen years' experience in diagnosis and treatment of six hundred fifty-four hands, J. Bone Jt Surg., 48-A: 211-228. Robbins, H. (1963) Anatomical study of the median nerve in the carpal tunnel and etiologies of the carpal-tunnel syndrome, J. Bone Jt Surg., 45-A: 953-966. Simpson, J. (1956) Electrical signs in the diagnosis of carpal tunnel and related syndromes, J. NeuroL Neurosurg. Psychiat., 19: 275-280. Thomas, J. E., E. H. Lambert and K. A. Cseuz (1967) Electrodiagnostic aspects of the carpal tunnel syndrome, Arch. NeuroL (Chic.), 16: 635-641. Walker, D. D. and J. Kimura, A fast-recovery electrode amplifier for electrophysiology, In preparation. Wiederholt, W. C., (1970) Median nerve conduction velocity in sensory fibers through carpal tunnel, Arch. phys. Med. Rehab., 51 : 328-330.