Secondary Logo

Journal Logo

Train-of-Four Stimulation for Adductor Pollicis Neuromuscular Monitoring Can Be Applied at the Wrist or Over the Hand

Nepveu, Marie-Eve MD; Donati, François MD, PhD, FRCPC; Fortier, Louis-Philippe MSC, MD, FRCPC

doi: 10.1213/01.ANE.0000141525.09320.C8
Technology, Computing, and Simulation: Research Report
Free
SDC

Adductor pollicis stimulation over the ulnar nerve at the wrist is the standard method of monitoring neuromuscular function. Stimulation over a muscle is believed to cause direct muscle contraction, but evidence for this is lacking. In this study we sought to determine whether direct muscle stimulation occurred during stimulation of the adductor pollicis in the hand and whether the responses were comparable to those observed with stimulation at the wrist. In 20 patients anesthetized with sevoflurane, 1 pair of stimulating electrodes was positioned over the ulnar nerve at the wrist. A second pair was placed between the first and second metacarpals on the palmar and dorsal aspects of the hand. The acceleromyographic response was monitored. Rocuronium 0.6 mg/kg was administered. Train-of-four (TOF) stimulations were applied at the wrist site until maximal blockade. Then, stimulation was applied to the hand site. During recovery, both sites were monitored alternately. After injection of rocuronium, 17 of 20 patients showed no twitch response at either site. One patient had a response at both stimulation sites, and two patients had responses only at the wrist site. With a Bland and Altman analysis, TOF ratios during recovery at the hand showed a bias of 0.5% and limits of agreement of ±11.8% as compared with the wrist. Stimulation in the hand causes no direct muscle stimulation because the response is no more than that produced by stimulation at the wrist. Both sites yield comparable TOF ratios.

IMPLICATIONS: Adductor pollicis monitoring is usually performed by applying electrodes over the ulnar nerve at the wrist. An alternative is to stimulate over the muscle itself in the hand. Train-of-four ratios at the hand and the wrist are comparable, and direct muscle stimulation does not occur.

Department of Anesthesiology, Hôpital Maisonneuve-Rosemont, Université de Montréal, Montréal, Québec, Canada

Accepted for publication July 21, 2004.

Address correspondence and reprint requests to François Donati, MD, PhD, FRCPC, Hôpital Maisonneuve-Rosemont, 5415 boulevard de l’Assomption, Montréal, QC, H1T 2M4, Canada. Address e-mail to francois.donati@umontreal.ca.

Adductor pollicis neuromuscular monitoring with stimulation of the ulnar nerve is widely used in anesthesia. It is the “gold standard” used to monitor the level of block and to compare neuromuscular blocking drugs and the response of other muscles to these drugs (1–3). The accepted method of stimulating the adductor pollicis uses electrodes placed at the wrist over the ulnar nerve. However, the ulnar nerve innervates many muscle groups in the hand. Apart from the adductor pollicis, some of the interossei, the third and fourth lumbricals, the muscles of the hypothenar eminence, and, in most cases, the flexor pollicis brevis are also innervated by the ulnar nerve (4). Therefore, stimulation of this nerve produces a movement not only of the thumb but also of every other finger. This could alter the visual and tactile interpretation of the train-of-four (TOF) by giving an impression of movement of the thumb when there is none. Moreover, if pulse oximetry is used on the same hand, artifacts can occur if repeated stimulations are used. In addition, the wrist is sometimes not available because of additional monitoring or patient positioning.

Stimulating the ulnar nerve more distally in the hand could yield a response more specific to the thumb. A site near the thenar muscles could be an alternative to stimulation at the wrist. This could circumvent many of the disadvantages of stimulation at the wrist. However, when electrodes are placed directly over a muscle, there is always a concern that direct muscle stimulation (depolarization of muscle fibers directly and not through the neuromuscular junction) could occur (5). This could lead to a false assumption that the patient is not completely paralyzed. One other question regarding this new method is whether it would be equivalent to the conventional method in regard to the TOF ratio during recovery.

The aim of this study was to determine whether direct muscle stimulation occurs when electrodes are placed in the palm of the hand as opposed to the wrist. Comparison was also made of TOF ratios and first twitch height (T1) with the two stimulating sites.

Back to Top | Article Outline

Methods

After approval from the hospital ethics committee and after obtaining informed consent, 20 adult patients (age 26–71 yr) with ASA physical status I or II undergoing elective surgery were included in the study. None had neuromuscular disease, and none was taking drugs that could interfere with neuromuscular transmission. Patients with a body mass index >30 kg/m2 were excluded. Before the procedure, electrocardiographic (ECG) monitoring, noninvasive arterial blood pressure monitoring, pulse oximetry, and a temperature probe were installed. The thickness of the thenar eminence was measured just proximal to the first metacarpophalangeal joint by using a millimetric caliper. The thickness of the wrist was measured just proximal to the styloid process of the ulna using the same instrument. Two pairs of standard ECG electrodes were placed on the same arm. The first pair was placed over the ulnar nerve, at the wrist (wrist site). The negative electrode was placed distally. The other pair was applied on both sides of the hand (hand site). The negative electrode was placed on the palm of the hand, between the base of the thumb and the second finger. It was positioned 1.5 cm from the proximal end of the first and second metacarpals, just medial and distal to the thenar eminence. The positive electrode was placed in the same position on the dorsal side of the hand (Fig. 1). Each pair of electrodes was connected alternately to a TOF-Watch SX (Organon, Swords, Dublin Co., Ireland) stimulator. An accelerometer was attached to the volar aspect of the thumb and connected to the stimulator. The arm and hand were immobilized with tape, leaving only the thumb free. No special arm board was used, and no preload was applied to the thumb. All data collected were transferred directly from the stimulator to a laptop computer by using the TOF-Watch Monitor software.

Figure 1

Figure 1

The induction of anesthesia was performed with midazolam 1–3 mg IV, fentanyl 2–4 μg/kg, and propofol 1–3 mg/kg. Anesthesia was maintained with sevoflurane 1%–3% (end-tidal) and fentanyl boluses as needed. Before any neuromuscular blocking drug was administered, the supramaximal current was obtained. Stimulation was applied at 1 Hz while the current was progressively increased to 60 mA. The point at which the response reached a plateau was considered to be the supramaximal current. It was measured at both stimulation sites. Then, 10 mA was added to the larger of the two values, and the stimulator was calibrated at the wrist site by using this current (CAL 1 function). Once calibration was completed, two TOF stimulations (four stimulations at 2 Hz) were performed at both sites. The fingers that moved at each site were noted. Rocuronium 0.6 mg/kg was injected while TOF stimulations were applied every 15 s at the wrist site until maximal blockade. Tracheal intubation was performed or a laryngeal mask airway was inserted. Ventilation was controlled to maintain an end-tidal CO2 partial pressure between 35 and 40 mm Hg. Neuromuscular monitoring was then switched to the hand site within 1 min of complete neuromuscular block, and TOFs were applied. Any presence or absence of response was noted. Afterward, TOFs were administered at the wrist site every 15 s until the beginning of recovery. Then, for every 10% recovery of T1 or TOF ratio at the wrist site, stimulations at the wrist site were stopped, and four TOF stimulations at 15-s intervals were measured at the hand site. Stimulation was then continued at the wrist site. All switches between sites were performed in <15 s to ensure continual TOF stimulations. Monitoring was discontinued only when the TOF ratio reached at least 80%. Additional rocuronium doses could be given during surgery as needed. At the end of surgery, neostigmine 50 μg/kg and glycopyrrolate 10 μg/kg were administered if necessary.

All T1 values were compared with the control (before rocuronium) T1 measured with supramaximal stimulation at the wrist. The TOF ratio was defined as T4/T1 at the same stimulation site. Direct muscle stimulation was assumed to occur if any response after TOF stimulation at the hand site was present in the absence of response at the wrist. The number of subjects (n = 20) was chosen so that if direct muscle stimulation >2% of control T1 was present in 40% of patients, it would be detected with a power of 0.80.

For each 10% of recovery of the T1 and TOF ratio at the wrist site, the mean of four TOFs was calculated for each stimulation site. The mean of four consecutive TOF and T1 values was calculated for the hand site and was compared with the mean of the two TOF and T1 values immediately before and the two immediately after these for the wrist site. This was done to ensure that the recordings were compared at equivalent degrees of recovery. The values from each stimulation site were compared by using a Bland and Altman analysis. All other values are presented as mean ± sd. Differences were considered statistically significant when P < 0.05.

Back to Top | Article Outline

Results

Twenty patients were included in the study: 18 women and 2 men (age, 45.9 ± 13.3 yr; weight, 60.8 ± 9.7 kg; height, 1.62 ± 0.08 m; body mass index, 23.8 ± 2.8 kg/m2; hand thickness, 26.4 ± 3.5 mm; wrist thickness, 20.9 ± 2.3 mm). There was no statistically significant difference between the mean supramaximal threshold at the wrist and at the hand (Table 1). No correlation was found between wrist thickness and wrist site supramaximal threshold or hand thickness and hand site supramaximal threshold. There was no difference between the mean initial TOF ratio at the wrist and at the hand or the mean initial T1 at the wrist and at the hand (Table 1). Initial TOF ratios were >100% in approximately half of the subjects at both sites. When TOF stimulation was applied at the wrist, all five digits showed movement, whereas only the first three digits moved when TOF stimulation was applied at the hand (Table 2).

Table 1

Table 1

Table 2

Table 2

After injection of rocuronium 0.6 mg/kg, 17 of 20 patients showed no twitch response at both sites. Two patients showed four equal twitches of 4%–5% of control T1 only at the wrist site, with no response at the hand site. A movement of the hypothenar eminence was noted in the first case, and a movement just below the electrodes at the wrist was noted in the second case. Finally, one patient had four twitches with a fade response at both sites despite rocuronium 0.6 mg/kg and an additional dose of 0.2 mg/kg. Four patients had a delay of 4–6.5 min between the time no twitch response was present at the wrist and the time there was no response at the hand. All other patients had no delay between complete neuromuscular blockade at the hand and at the wrist.

During recovery, TOF ratios obtained at the hand site were similar to those obtained at the wrist site (Figs. 2 and 3). There was a bias of +0.5%, and 95% of the values obtained at the hand were within ±11.8% of the values measured at the wrist. Bias and limits of agreement were not dependent on the degree of recovery. Twitch heights measured at the hand and the wrist were also comparable (Fig. 4). The bias for T1 was larger (+6.9%; values at the hand site were larger than those at the wrist site), and 95% of the values obtained at the hand were within ±40.3% of the values measured at the wrist. However, for a given patient, the difference between hand and wrist values tended to remain relatively constant throughout recovery.

Figure 2

Figure 2

Figure 3

Figure 3

Figure 4

Figure 4

Back to Top | Article Outline

Discussion

This study shows that neuromuscular monitoring at the hand does not produce direct muscle stimulation and that TOF ratios measured after stimulation at the hand are comparable to those obtained with stimulation at the wrist. Any response to TOF stimulation in the hand of more than 2% of control T1 in the absence of response at the wrist was assumed to represent direct muscle stimulation.

The electrodes on the hand were applied over the adductor pollicis muscle, which lies just medial and distal to the thenar eminence, between the thumb and the second finger. It was not placed over the thenar eminence because stimulation of muscles innervated by the median nerve—such as the flexor pollicis brevis, the opponens pollicis, or the abductor pollicis—could have occurred. However, the recurrent branch of the median nerve lies very close to the adductor pollicis before it enters the thenar eminence (4). It is therefore possible that a portion of the response we got while stimulating at the hand site came from contraction of muscle fibers innervated by the median nerve.

We decided to compare the two stimulation sites on the same arm, using the same accelerometer to minimize bias because there could have been a difference between baseline drift between both arms and between positioning of the accelerometer on each thumb (6–8). This way, calibration was done with only one accelerometer, and it allowed us to compare all T1 values with the control T1 at the wrist and thus estimate the relative force of contraction of the muscle at both sites. The disadvantage was that we could not stimulate both sites at the same time and had to switch the stimulating wires from one electrode pair to the other. To ensure that we compared TOF ratio and twitch height at the same time during recovery and to minimize variation between successive measurements, a mean of four TOFs was used for each stimulation site: four consecutive TOFs and T1 values for the hand site and the two immediately before and after these for the wrist site. This was done for every 10% increase in TOF or T1 at the wrist site in two minutes.

The wrist site and the hand site showed remarkable correlation with respect to TOF ratios, and TOF is the most commonly used modality in clinical practice. When both methods were compared, there was practically no bias (0.5%), and the 95% limits of agreement (±11.8%) could be within the accuracy limits of the accelerometer. In regard to T1, dispersion was more important, and there was a positive bias of 6.9% (the hand site showed higher T1 values than the wrist site). The greater T1 values are probably due to stimulation of muscle fibers of the thenar eminence innervated by the median nerve, thus increasing the force of contraction or changing the direction of the acceleration vector toward the accelerometer.

When neuromuscular function is monitored, it is usually thought that the stimulation site must be over a peripheral nerve to elicit the most accurate response (5). It is believed that stimulation over or near a muscle could produce direct muscle contraction. If direct muscle stimulation were to occur, interpretation of neuromuscular function could be hampered because muscle contractions would be present even with adequate muscle relaxation. However, very few data are available to substantiate this concern, because it appears that a high voltage or current is required to stimulate a muscle directly. Nerve and muscle fibers are excitable and can depolarize when submitted to an electrical stimulation. However, maximal direct stimulation of the adductor pollicis in humans may require as much as 1000 V. A lower voltage stimulates nerve fibers only (9). With IM electrodes placed in cat tibialis anterior muscle, maximal muscle contraction required a current of 2 mA for 1 ms when stimulated via a nerve. When a neuromuscular blocking drug was given, direct muscle contraction required a current of 15 mA for 1 ms (10). In another experiment, when IM electrical stimulation was applied, direct muscle contraction contributed only to 3%–7% of the total contraction; the rest was achieved by stimulation of IM nerve fibers (11). Similarly, our results support the notion that the current used in neuromuscular monitoring (10–60 mA for 0.2 ms) through surface electrodes is usually not sufficient to produce direct muscle contraction.

As was done in previous studies, we chose the presence of response to TOF stimulation at the hand site in the absence of response at the wrist site as the criterion for direct stimulation (12). In a study on the significance of electrode position for the orbicularis oculi, Gätke et al. (8) found that 80% of patients showed one to four acceleromyographic responses to TOF stimulation even though there was no visible response of the orbicularis oculi. They stated that direct stimulation was a possibility, but technical limitations of the TOF-Guard appear more likely. At high gains, the TOF-Guard might display a response in the absence of actual muscle contraction. In another study on acceleromyographic monitoring over the vastus medialis muscle, Saitoh et al. (13) found that only 4 of 30 patients did not show a disappearance of T1 response after the administration of vecuronium 0.1 mg/kg. In this last study, the electrodes were placed over a muscle and not over a nerve, and most patients showed a normal acceleromyographic response to muscle relaxation. With regard to the hand, Rodiera et al. (14) found that selective nerve stimulation over the ulnar nerve produced similar results as nonselective stimulation delivered over the forearm supinator longus muscle. Finally, in a study on excitability properties of human median axons measured at the motor point, Kuwabara et al. (15) provided evidence that direct muscle activation was unlikely because the stimulus-response curves they measured did not change abruptly, as would be expected with muscle fiber activation.

In our study, only one of the patients showed a four-twitch muscular response at the hand after rocuronium 0.6 mg/kg. This patient had four twitches with fade at both the wrist and hand sites and could have been more resistant to muscle relaxants. All other patients showed no twitch response at the hand site after injection of muscle relaxant. However, 2 other patients showed 4 equal twitches only at the wrist site and none at the hand site. This is the opposite of what one would expect to find. Direct muscle stimulation near the wrist could have occurred in these cases, because muscle movement near the wrist electrodes, but none near the thumb, was present. Therefore, in this study, direct muscle stimulation was observed with stimulation at the wrist but not at the hand.

In four patients, there was a delay between total block at the wrist site and total block at the hand site: the latter was slower than the former. These patients tended to have a higher initial T1 at the hand site than at the wrist site. This could be explained by the fact that nerves innervating muscles other than the adductor pollicis course in the area under the hand electrodes. These muscles could be slightly more resistant than the adductor pollicis to the effects of rocuronium.

We also looked at the number of fingers that moved when each site was stimulated. All five digits almost always moved with the wrist site. After repeated stimulation, this could cause artifacts on an oximeter placed on those fingers and could even potentially dislodge it. In contrast, when the hand site was stimulated, only the first three digits moved, and therefore an oximeter could be placed conveniently on the fourth or fifth finger. Whether tactile or visual interpretation of the TOF ratio could be more accurate with this method is not known because it was beyond the scope of this study.

In conclusion, monitoring the adductor pollicis by positioning the stimulating electrodes in the hand is an acceptable alternative to applying electrodes at the wrist. Direct muscle stimulation does not seem to be as common a phenomenon as previously believed, because it occurred in none of our cases. Neuromuscular monitoring near or over a muscle is possible. Additional studies on the possible advantages of this method of stimulation are required.

Back to Top | Article Outline

References

1. Larsen PB, Gatke MR, Fredensborg BB, et al. Acceleromyography of the orbicularis oculi muscle. II. Comparing the orbicularis oculi and adductor pollicis muscles. Acta Anaesthesiol Scand 2002;46:1131–6.
2. Hemmerling TM, Donati F. Neuromuscular blockade at the larynx, the diaphragm and the corrugator supercilii muscle: a review. Can J Anaesth 2003;50:779–94.
3. Kopman AF, Zhaku B, Lai KS. The “ intubating dose” of succinylcholine: the effect of decreasing doses on recovery time. Anesthesiology 2003;99:1050–4.
4. Gray H. Gray’s anatomy. 30th American ed. Philadelphia: Lea & Febiger, 1985.
5. Viby-Mogensen J. Neuromuscular monitoring. In: Miller R, ed. Anesthesia. 5th ed. Philadelphia: Churchill Livingstone, 2000:1363.
6. Loan PB, Paxton LD, Mirakhur RK, et al. The TOF-Guard neuromuscular transmission monitor: a comparison with the Myograph 2000. Anaesthesia 1995;50:699–702.
7. Harper NJ, Martlew R, Strang T, Wallace M. Monitoring neuromuscular block by acceleromyography: comparison of the Mini-Accelograph with the Myograph 2000. Br J Anaesth 1994;72:411–4.
8. Gätke MR, Larsen PB, Englbaek J, et al. Acceleromyography of the orbicularis oculi muscle. I. Significance of the electrode position. Acta Anaesthesiol Scand 2002;46:1124–30.
9. Hill DK, MCDonnel MJ, Merton PA. Direct stimulation of the adductor pollicis in man. J Physiol 1979;300:2P–3P.
10. Mortimer JT, Bhadra N. Peripheral nerve and muscle stimulation. In: Horch KW, Dhillon GS, eds. Neuroprosthetics: theory and practice. Singapore: World Scientific Publication Co Inc, 2004.
11. Crago PE, Peckham PH, Mortimer JT, Van der Meulen JP. The choice of pulse duration for chronic electrical stimulation via surface, nerve, and intramuscular electrodes. Ann Biomed Eng 1974;2:252–64.
12. Machetanz J, Bischoff C, Pichlmeier R, et al. Magnetically induced muscle contraction is caused by motor nerve stimulation and not by direct muscle activation. Muscle Nerve 1994;17:1170–5.
13. Saitoh Y, Nakajima H, Hattori H, et al. Neuromuscular blockade can be assessed accelerographically over the vastus medialis muscle in patients positioned prone. Can J Anaesth 2003;50:342–7.
14. Rodiera J, Calabuig R, Gual A. Selective versus non-selective neural stimulation in the monitoring of muscular relaxation during general anesthesia. J Clin Monit Comput 1999;15:341–5.
15. Kuwabara S, Bostock H, Ogawara K, et al. Excitability properties of human median axons measured at the motor point. Muscle Nerve 2004;29:227–33.
© 2005 International Anesthesia Research Society