Anesth Analg 2004;98:377-381
© 2004 International Anesthesia Research Society
doi: 10.1213/01.ANE.0000096003.64059.97
TECHNOLOGY, COMPUTING, AND SIMULATION
Phonomyography and Mechanomyography Can Be Used Interchangeably to Measure Neuromuscular Block at the Adductor Pollicis Muscle
Thomas M. Hemmerling, MD DEAA,
Guillaume Michaud,
Guillaume Trager, DESS,
Stéphane Deschamps, DESS,
Denis Babin, MSc, and
François Donati, MD PhD, FRCPC
From the Neuromuscular Research Group (NRG), Department of Anesthesiology, Centre Hospitalier de l’Université de Montréal (CHUM) Hôtel-Dieu, Université de Montréal, Montréal, Canada
Address correspondence and reprint requests to T. M. Hemmerling, MD, DEAA, Department of Anesthesiology, Université de Montréal, Hôtel-Dieu, 3580 Rue St. Urbain, Montréal (Québec) H2W 1T8, Canada. Address email to thomashemmerling{at}hotmail.com
 |
Abstract
|
|---|
The standard of neuromuscular monitoring is the measurement of the force of contraction (mechanomyography, MMG). Phonomyography (PMG) consists of recording low-frequency sounds created during muscle contraction. In this study, we compared and used both methods to determine neuromuscular blockade (NMB) at the adductor pollicis muscle. In 14 patients, PMG was recorded via a small condenser microphone taped to the thenar mass, and a standard mechanomyographic device was applied to the same arm. In another group of 14 patients, only PMG was measured. After induction of anesthesia, the ulnar nerve was stimulated supramaximally using single twitch stimulation (0.1 Hz) for onset and train-of-four (TOF) stimulation every 12 s during offset of NMB produced by mivacurium 0.1 mg/kg. Onset and recovery indices measured by the 2 methods were compared using Student’s t-test (P < 0.05). Similar comparisons were made between the two PMG groups (with or without special board). Agreement between PMG and MMG was examined using a Bland-Altman test. Onset was 165 (68) s versus 172 (67) s [mean (SD)], and maximum blockade was 89 (10)% versus 90 (11)%, for PMG and MMG respectively (NS). Time to 25%, 75%, and 90% recovery was 16.5 (4.2) min, 22.1 (6.9) min, and 24.5 (8.2) min, respectively for PMG, not different from 16.7 (4) min, 22.8 (8.1) min, and 24.8 (8.8) min for MMG. Mean bias was 0% with limits of agreement of -10 and + 10% of twitch height for all signals (MMG minus PMG). Time to TOF of 0.5, 0.7, 0.8, and 0.9, was 1 min faster with PMG than with MMG, with limits of agreement of -1.5 to 3.5 min. Pharmacodynamic data derived without or with special arm fixation were not significantly different. MMG and PMG can be used interchangeably to determine NMB at the adductor pollicis muscle. PMG is easier to apply, does not need a special monitoring board and could be a reliable monitor to determine NMB in daily routine.
IMPLICATIONS: Mechanomyography and phonomyography (PMG), a novel method of monitoring neuromuscular blockade (NMB) by recording low-frequency sounds emitted by muscle contraction, can be used interchangeably to determine NMB at the adductor pollicis muscle. PMG is easier to apply, does not need a special monitoring board and could be a reliable monitor to determine NMB in daily routine.
|
Introduction
|
|---|
Mechanomyography (MMG) measures the actual force created by muscle contraction and is considered to be the gold standard of neuromuscular monitoring. However, difficult set-up procedures requiring special arm boards prevent its use in clinical routine. Furthermore, MMG cannot be used at muscles other than the adductor pollicis, such as the corrugator supercilii or the orbicularis oculi muscle.
Recently, we described phonomyography (PMG) as a new method of neuromuscular monitoring (1). The method is based on the fact that muscle contraction evokes low-frequency sounds (2,3) that can be recorded using special microphones. Initially, this method was called "acoustic myography" and used an air-chamber interface between the skin and the microphone (4,5). There was good correlation with force measurement, although limits of agreement were as wide as ±40%. This might be attributable to the limited frequency response of these microphones, which could not detect frequencies <20 Hz. In two recent studies, (1,6) we evaluated the usefulness of small condenser microphones that could detect frequencies as low as 2.5 Hz. We determined signal characteristics and frequency response of neuromuscular blockade at the corrugator supercilii (1), a small muscle around the eye (7), and we showed that peak frequencies were in the 3–5 Hz range, with most of the signal power being under 20 Hz. We called this method of measuring sounds created by muscle contraction using small condenser microphones PMG. We hypothesized that microphones capable of recording low-frequency sounds would give better agreement with MMG than previous systems. This study used PMG to measure neuromuscular blockade at the adductor pollicis muscle and determine its agreement with MMG.
|
Methods
|
|---|
After approval of the local ethics committee and obtaining informed consent, 28 patients undergoing general surgery were included in the study. Patients with coexisting neuromuscular disease or patients on medication known to interact with neuromuscular transmission were excluded.
After arrival in the operating room, routine monitoring (noninvasive blood pressure, pulse oximetry, 5-lead electrocardiography) was applied. Anesthesia was induced with remifentanil 0.25–0.5 µg · kg-1 · min-1; 2 min later, propofol 2–3 mg/kg was injected. After loss of consciousness and ventilation via face mask for 2 min with 100% oxygen, a laryngeal mask airway (LMA) (size 4 for women, size 5 for men; LMA company, Southampton, UK) was inserted and controlled ventilation commenced with minute ventilation set to maintain an end-tidal PETCO2 of 3.5–4.5 kPa. Anesthesia was maintained with 1–1.5 MAC of sevoflurane in a breathing gas mixture of 30% oxygen in air to maintain a bispectral index (BIS) (A-2000 monitoring system; Aspect Medical Company, Newton, MA) of 50. Analgesia was provided by remifentanil 0.05–0.25 µg · kg-1 · min-1 throughout surgery. In 14 patients, the force of contraction of the adductor pollicis was measured using a force transducer. A specially molded cast was used to stabilize the arm in position. A small condenser microphone (diameter: 1.6 cm, Model 1010; Grass Instruments, Astro-Med, Inc., West Warwick, RI; frequency response: 2.5 Hz to 5 kHz, signal output: 20–40 mV into 1 M
) was attached to the middle of the thenar mass using a routine gluing tape to record the acoustic signals produced by the contraction of the adductor pollicis muscle (Fig. 1). The microphone signal was amplified and band pass filtered between 0.5 Hz and 1000 Hz using an AC/DC amplifier. The signals were continuously sampled at 100 Hz using the Polyview® software package (Polybytes, Cedar Rapids, IA), digitized and stored on a portable microcomputer. The single-twitch PMG signal was measured peak-to-peak (Fig. 1). Neuromuscular block was measured by MMG and PMG simultaneously at the same arm.
View larger version (80K):
[in this window]
[in a new window]
|
Figure 1. A, illustration of location of force transducer (including the molded mechanomyographic cast device) and the microphone. B, M1 and M2 indicate the points in between the amplitude of the signal was measured with both methods (n = 14 patients).
|
|
In another group of 14 patients, only PMG was used to measure neuromuscular blockade. The small condenser microphone was attached as described above, and stimulation and recording of signals was performed as described above. However, there was no specially molded cast and the hand was simply attached to a routine armboard with gluing tape (Fig. 2). This second group of patients was created to exclude the influence of positioning the arm in a specially molded cast—as necessary for MMG—for PMG measurements. In all patients, the ulnar nerve was stimulated at the forearm with single supramaximal twitches via surface electrodes using a constant current stimulator (Innervator®; Fisher and Paykel Healthcare, Auckland, New Zealand) that generated single-twitch square pulses of 0.2 ms with a current intensity between 0 and 70 mA (single twitch 0.1 Hz for onset, train-of-four (TOF) stimulation every 12 s for offset of neuromuscular blockade). Supramaximal stimulation was determined using single-twitch stimulation at 0.1 Hz. After at least 5 min of supramaximal stimulation and stable baselines for both recordings, mivacurium 0.1 mg/kg was injected within 5 s into a fast-flowing solution of Ringer’s lactate solution. Onset, maximum effect, and offset of neuromuscular blockade after mivacurium 0.1 mg/kg were determined.
The first twitch response was used to analyze onset time (time to reach maximum decrease of twitch response), T25%, T75%, and T90% (time to reach 25%, 75%, and 90% of control twitch response). The maximum effect was determined as the maximum decrease of the twitch response and recorded. Time to reach a TOF ratio of 0.5, 0.7, 0.8, and 0.9 was calculated for MMG and PMG simultaneously.
Data were compared between PMG and MMG using a paired Student’s t-test, P < 0.05 was considered as showing a significant difference. Group size was determined to reach a power of more than 0.95; it was calculated based on an estimated difference of 20% of mean onset time and T25% between the two methods. The mean difference of all signals determined using both methods and the limits of agreement between the two methods were analyzed using the Bland-Altman test (8). Correlation between the two methods was calculated using Pearson’s test. Data are presented as mean (SD). Onset, offset and maximum effect as determined by PMG were also compared between patients having the arm placed in a special cast and patients in which only PMG was applied using unpaired Student’s t-test; P < 0.05 was considered as showing a significant difference.
|
Results
|
|---|
In all patients, pharmacodynamic results with both methods could be obtained. Recordings of signals were continued until TOF was more than 0.9 in all patients. Hand temperature was more than 35.5°C in all patients. A typical PMG signal is shown in Figure 1b.
Onset, maximum effect and time to reach 25%, 75%, and 90% of control twitch were similar for MMG and PMG witha correlation of r = 0.95 for all signals (Table 1). Mean bias was 0% with limits of agreement of -10 and + 10% for all signals (MMG minus PMG) (Fig. 3). Recovery of TOF ratios were not different between the two methods and are shown in Figure 4 for both PMG and MMG in 14 patients. For the time to a TOF of 0.5, 0.7, 0.8, and 0.9, there was a correlation of r = 0.99 and a mean bias of 1 min with limits of agreement of -1.5 to 3.5 min (MMG minus PMG).
View larger version (22K):
[in this window]
[in a new window]
|
Figure 3. Bland-Altman test for all evoked signals; Mean bias was 0% with limits of agreement of -10 and + 10% for all signals (mechanomyography minus phonomyography).
|
|
View larger version (11K):
[in this window]
[in a new window]
|
Figure 4. Time to reach train-of-four ratios of 0.5, 0.7, 0.8, 0.9 measured using phonomyography (black) and mechanomyography (gray). Data presented as means and standard deviation.
|
|
Onset, offset, and peak effect of neuromuscular blockade as determined by PMG were not different between patients having the hand placed in a special cast or having the hand simply attached to the arm rest (Table 1).
There was only minimal drift (T1% higher at full recovery than the initial value) as monitored by MMG (T1% final, 107.5 ± 3.8%) or PMG (105.4 ± 2.9%). Reverse fade was observed in 6 of 28 patients with PMG and in 3 of 14 patients with MMG after the start of TOF stabilization, which subsided in all but 2 patients (PMG) after the stabilization period of 5 min. In those two patients, the fourth (1 patient) or third (1 patient) twitch showed a higher amplitude than the first twitch, but none was higher than 105% of the first twitch.
|
Discussion
|
|---|
Onset and offset of neuromuscular blockade measured using PMG were very close to those obtained using MMG. Agreement between MMG and PMG was very good with minimal bias. Both methods can be used interchangeably to measure onset, offset, and maximum effect of neuromuscular blockade after mivacurium at the adductor pollicis muscle. The clinically important recovery to TOF-ratios of more than 0.5 was very similar for both methods with PMG generally reaching the equivalent TOF-ratios 1–2 min earlier than MMG. This does not seem to be clinically important.
The agreement between PMG and MMG in this study was considerably better than that reported in previous studies (4,5). Virtually all measured values were within 10% of each other in the present study and bias was 0%. In those two studies where comparisons were made at the adductor pollicis, some measurements differed by more than ±40% in both studies and bias was 10% in one (5). The better performance of the microphone used in the present study is probably attributable to the absence of an air chamber between the skin and the microphone and, especially, to a better frequency response. The contribution of low frequencies (2–20 Hz) to the PMG signal is important, and the use of a microphone that can pick up signals below the hearing range (20 Hz) is essential.
Reverse fade, i.e., TOF ratios more than 0.1 in the absence of neuromuscular blocking drugs, was seen in only a few patients with both methods and did not affect the measurements. There was no significant difference between pharmacodynamic data derived using PMG with the arm fixed in a rigid position using the molded arm cast used for MMG and data derived from PMG with the arm glued to a regular arm board without additional fixation. The advantage of PMG lies in the fact that it is less difficult to apply and does not need special set-up procedures, such as a special cast or board to put the hand in a defined position or a preload. This method only needs a condenser microphone small (special in its ability to record very low frequency sounds), which is glued to the thenar region, and simple attachment of the arm on a routine armrest. It is therefore a simple, noninvasive, easy-to-apply method for routine monitoring. The fact that PMG can be applied at various muscles of interest in neuromuscular research (corrugator supercilii muscle, adductor pollicis muscle, adducting laryngeal muscles, diaphragm) and the good-to-very-good agreement with methods measuring the actual force of muscle contraction make it an interesting method for simultaneous monitoring of several muscles for which MMG is difficult or impossible to apply.
There are still some problems with this method. As PMG measures low-frequency sounds, artifacts are possible (1). Vessel pulsations can cause rhythmic sounds, which are detected as small waves of the baseline (Fig. 5). However, these artifacts are easily recognizable and do not disturb the evoked measurements. Interferences with electrocautery were noted with both MMG and PMG; during these interferences, no signals can be obtained. If at one stage, PMG is used in neuromuscular monitoring devices for clinical use, technical solutions need to be found to diminish the interferences as it has been done with most other devices used in the operating room.
Because the microphone is simply fixed to the skin, care needs to be taken to avoid detaching the tape and therefore losing contact between microphone and skin surface. If this happens, the initial signal and the signal after recovery from neuromuscular blockade are not necessarily of the same height or shape.
We have evaluated several aspects of PMG: a frequency domain analysis and determination of peak frequency to identify the technical properties of the method have shown that suitable microphones need to be able to record sounds between 2 and 50 Hz. The duration of control stimulation does not change the measured onset, maximum effect, or offset of neuromuscular blockade (9). What makes PMG an interesting method for research is the fact that it can easily be applied at all muscles of interest (diaphragm, larynx, adductor pollicis muscles, and eye muscles), unlike current methods. What makes PMG an attractive method for clinical routine is the ease of application and absence of elaborate devices for installation. We believe that the fundamental tests have been so promising that the next logical step should be to incorporate PMG in a monitoring prototype.
|
Acknowledgments
|
|---|
Supported, in part, by departmental internal funds.
|
References
|
|---|
- Hemmerling TM, Donati F, Beaulieu P, Babin D. PMG: signal characteristics, best recording site and comparison with acceleromyography. Br J Anaesth 2002; 88: 389–93.[Abstract/Free Full Text]
- Barry DT. Muscle sounds from evoked twitches in the hand. Arch Phys Med Rehabil 1991; 72: 573–5.[ISI][Medline]
- Barry DT, Cole NM. Muscle sounds are emitted at the resonant frequencies of skeletal muscle. IEEE Trans Biomed Engl 1990; 37: 525–31.
- Dascalu A, Geller E, Moalem Y, et al. Acoustic monitoring of intraoperative neuromuscular block. Br J Anaesth 1999; 83: 405–9.[Abstract/Free Full Text]
- Bellemare F, Couture J, Donati F, Plaud B. Temporal relation between acoustic and force responses at the adductor pollicis during nondepolarizing neuromuscular block. Anesthesiology 2000; 93: 646–52.[Medline]
- Hemmerling TM, Babin D, Donati F. PMG as a novel method to determine neuromuscular blockade at the laryngeal adductor muscles: comparison with the cuff pressure method. Anesthesiology 2003; 98: 359–63.[ISI][Medline]
- Plaud B, Debaene B, Donati F. The corrugator supercilii, not the orbicularis oculi, reflects rocuronium neuromuscular blockade at the laryngeal adductor muscles. Anesthesiology 2001; 95: 96–101.[ISI][Medline]
- Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clincial measurement. Lancet 1986; i: 307–10.
- Hemmerling TM, Donati F, Babin D, Beaulieu P. Duration of control stimulation does not affect onset and offset of neuromuscular blockade at the corrugator supercilii muscle measured with PMG or acceleromyography. Can J Anaesth 2002; 49: 913–7.[Abstract/Free Full Text]
Accepted for publication August 21, 2003.
Top
Abstract
Introduction
