Share this article on:

Intramuscular Versus Surface Electromyography of the Diaphragm for Determining Neuromuscular Blockade

Hemmerling, Thomas M. MD, DEAA; Schmidt, Joachim MD; Wolf, Tobias; Hanusa, Christian; Siebzehnruebl, Ernst MD; Schmitt, Hubert MD

doi: 10.1097/00000539-200101000-00021
Anesthetic Pharmacology: Research Report

We determined the neuromuscular blockade of 0.2 mg · kg−1 mivacurium at the diaphragm by using two new methods of electromyographic (EMG) monitoring and compared it with acceleromyography of the orbicularis oculi (OO) and the corrugator supercilii (CS) muscle. After the induction of anesthesia in 15 patients undergoing gynecologic laparoscopic surgery, evoked EMG responses at the diaphragm were obtained by using skin electrodes at the back of the patient, placed lateral to T12/L1 or L1/L2, and a laparoscopically applied wire electrode inserted into the dorsolateral portion of the diaphragm. Acceleromyography at the right OO and the left CS was performed. The facial and phrenic nerves were stimulated transcutaneously (onset: every 10 s, offset: every 15 s, single twitch stimulation). Lag and onset time, peak effect, and clinical duration (time to reach 75% of control value and time to reach 90% of control value) were measured and the results were compared by using analysis of variance;P < 0.05 showed significant difference. Pearson’s correlation test and the Bland-Altman test were used to compare the two diaphragmatic monitoring methods. Mean peak effects of >98% were reached at all sites. Onset times at diaphragm (skin, IM) were significantly (P < 0.005) shorter than at the CS or OO (100 ± 14 s and 98 ± 16 s vs 147 ± 39 s, 185 ± 38 s) without being statistically different between OO and CS. There was a good correlation of lag, onset time, time to reach 75% of control value, and time to reach 90% of control value (r = 0.8, 0.9, 0.8, and 0.75;P < 0.01) between the two diaphragmatic methods. Mean difference and limits of agreements are −2 ± 15 s, 1 ± 21 s, −1 ± 2.3 min, and −2 ± 3.4 min. We showed a shorter onset and clinical duration at the diaphragm in comparison with CS and OO. Two methods of EMG of the diaphragm correlated well and showed good comparability. The novel method of surface diaphragmatic EMG at the patient’s back may be useful during routine clinical anesthesia.

Implications: The novel method of monitoring the diaphragmatic neuromuscular blockade (NMB) at the patient’s back showed good correlation and good comparability with the IM NMB derived from an endoscopically inserted wire electrode and might be clinically used. The simultaneous determination of the NMB at the orbicularis oculi and the corrugator supercilii muscle did not show that either of these muscles was a good indicator of the diaphragmatic response.

Departments of Anesthesiology and Gynecology, University Erlangen-Nuremberg, Germany

August 17, 2000.

Address correspondence and reprint requests to T. M. Hemmerling, CHUM, Hôtel-Dieu, 3840 Rue St. Urbain, Montréal, Québec H2W 1T8, Canada. Address e-mail to

Clinically, intraoperative onset, magnitude, and duration of neuromuscular blockade (NMB) is mainly determined by monitoring the neuromuscular response at peripheral muscles such as the orbicularis oculi (OO) or the adductor pollicis muscle. For abdominal surgery, however, in which constant control and maintenance of a certain degree of NMB is important, especially in laparoscopic procedures, the diaphragm and its NMB greatly influence abdominal pressure.

Studies over the last 15 years have enhanced our knowledge about how the diaphragm reacts to neuromuscular blocking drugs and how its pharmacodynamic profile differs from more peripheral muscles (1–5). Monitoring of the NMB at the diaphragm, whether performed via measurements of the transdiaphragmatic pressure or via surface electromyography, has been well established in research, but not introduced into clinical practice. The main reasons were that the former method was too invasive and impracticable to use in routine clinical anesthesia. The latter has always required skin electrodes placed at the 7th or 8th intercostal space lateral to the midclavicular line. This area cannot be used in abdominal surgery because it is part of the sterile area.

We therefore looked for an alternative site to monitor the diaphragmatic neuromuscular response. To validate this new area lateral of the vertebra T12/L1 or L1/L2, we compared electromyographic (EMG) responses obtained via skin electrodes with abdominal EMG responses of the diaphragm obtained via a laparoscopically inserted bipolar wire electrode. Neither method has been presented previously. Another purpose of the study was to investigate the simultaneous acceleromyographic (AMG) measurements at the OO and the corrugator supercilii muscle (CS). This muscle has been the focus of a recent study (6) that concluded that the response to neuromuscular blocking drugs corresponds well with the response of the larynx, but the authors did not measure the diaphragmatic response.

Back to Top | Article Outline


After approval of the local ethics committee and informed consent, 15 women undergoing gynecologic surgery were included in the study. Pregnant women, patients with neuromuscular, hepatic, renal disease or patients receiving medications known to interact with neuromuscular blocking drugs were excluded.

Anesthesia was induced by using 20 μg · kg1 alfentanil, followed by a target-controlled infusion of propofol (target concentration: 4 μg · mL1) programmed to reach the target concentration within 30 s. After the induction of anesthesia, the patients were orotracheally intubated. Anesthesia was maintained with target-controlled infusion of propofol (target concentration: 3 μg · mL1) and alfentanil at increments of 10 μg · kg1 at the discretion of the anesthesiologist. Mechanical ventilation was adjusted to achieve ETco2 pressure of 26–35 mm Hg.

Two Ag/AgCl-skin-electrodes were attached lateral to vertebra T12/L1 or L1/L2 (wherever the maximum response was possible) on the left paravertebral side of the back, inferior to the 12th rib, placed 2 cm apart for EMG-monitoring of the response of the left diaphragmatic muscular crux to the phrenic nerve stimulation (Fig. 1). All patients were positioned on a soft foam operating table to avoid pressure necrosis as a result of the electrodes.

The left phrenic nerve was transcutaneously stimulated with an external bipolar nerve stimulator (Multiliner®; Tönnies company, Wuerzburg, Germany) on the inferolateral edge of the sternocleidomastoid muscle. The probe of the external nerve stimulator was attached at the neck with an elastic band. It delivers a current between 0 and 70 mA. Single twitch-stimulation (0.1 Hz, pulse width: 0.2 ms) was performed on the left neck to determine the supramaximal stimulation and recorded by using Multiliner® (Toennies) software. The current was increased from 0 to the current with the maximal EMG response, and then increased by 10 mA to assure supramaximal stimulation. The medial part of the right superciliary arch was equipped with an AMG probe to record the neuromuscular response of the CS. The left upper eyelid was equipped with an AMG probe to record evoked responses of the OO. Stimulation of the upper branches of the facial nerves was performed on both sides by using two Ag/AgCl-electrodes attached to the skin 2 cm anterior to the ear lobe. The automatic calibration set-up of the TOF-guardINMT (Organon Helsinki, Finland) was used to determine supramaximal stimulation on both sides (single twitch, 0.1 Hz).

Laparoscopic surgery was then commenced. Through the infraumbilical incision the first trocar was inserted and the endoscopic camera introduced. A second trocar was inserted into left lower abdominal position. Through this trocar, a bipolar wire electrode (Dr. Osypka, Grenzach, Germany) was inserted into the left dorsolateral part of the diaphragm (Fig. 2a–d). Care was taken to place the recording part in optimal contact to the diaphragm. The wire electrode was then pulled through the routine laparoscopic trocar and connected to the Multiliner®. Phrenic stimulation was restarted; the amplitudes of both diaphragmatic compound action potentials (peak-to-peak) were measured and recorded.

After no change in the neuromuscular response could be detected on all four sites for 10 min, the patients received 0.2 mg · kg1 mivacurium, injected within 15 s into a fast-flowing infusion of lactated Ringer’s solution. No further dose of any muscle relaxant was applied. Body temperature was measured at the forehead and kept >35.6°C by using a heating blanket (Bair Hugger; Augustine Medical, Eden Prairie, MN).

The time from the end of injection of the muscle relaxant to the first twitch depression, the maximum twitch depression (Lag time, Onset time) as well as the peak effect (%-reduction of the maximal neuromuscular response) of the NMB were measured. To determine the clinical duration of the NMB, single twitch responses every 15 s were used. Time to reach 75% of control value (TH75) and time to reach 90% of control value (TH90) were measured.

The results were expressed as mean ± sd and range; analysis of variance (paired analysis of variance) was used to compare the pharmacodynamic variables between diaphragm, the OO, and the CS, corrected for the number of comparisons;P < 0.05 was regarded as showing a significant difference. Correlation analysis between the times measured at the diaphragm was done by using Pearson’s test, with P < 0.05 showing significant difference. The Bland-Altman test was used to compare the results of the two diaphragmatic recording sites; mean difference and limits of agreement were calculated.

Back to Top | Article Outline


In all 15 female patients with a mean age of 44 ± 16 yr (range 18–78 yr), a mean weight of 64 ± 10 kg (range 46–81 kg), determination of the supramaximal stimulation at all four monitoring sites was successful. No side effects resulting from the transcutaneous stimulation of the recurrent and phrenic nerve with a mean of 45 ± 6 mA (range 30–55 mA), such as arrhythmias or skin irritation, were noted. In all patients, laparoscopic placement of the wire electrode was performed in <10 min. During monitoring of the onset and clinical duration of the NMB, the electrode stayed in place. No skin irritation was found on the patients’ backs because of the monitoring electrodes.

Maximum neuromuscular blockades were more than a mean of 98% at all sites (Table 1). The lag and onset time were significantly more rapid at the diaphragm as compared with either the OO or the CS. Lag and onset time were not statistically different between the CS and the OO (Table 1).

There was a good correlation between the two monitoring sites of the diaphragm (r = 0.8, 0.9 for lag and onset time, P < 0.01) with narrow limits of agreement of −2 ± 15 s for onset time, 1 ± 21 s for lag time, skin versus wire electrodes, respectively (Fig. 3). The clinical duration of the NMB as measured by TH75 and TH90 was shorter at the diaphragm than at the OO or the CS (P < 0.01, Table 1). The clinical duration of NMB at the diaphragm correlated well at r = 0.8, 0.75 for TH75 and TH90 and show limits of agreement of −1 ± 2.3 min for TH75, −1 ± 3.4 min for TH90, skin versus wire electrodes, respectively.

Back to Top | Article Outline


We present two new methods of EMG monitoring of the diaphragmatic neuromuscular response. One method modifies the usual monitoring position of the diaphragmatic NMB (7th or 8th intercostal space, midclavicular to axillary line) and introduces a position at the back of the patient, thus enabling to monitor the NMB during abdominal surgery. The other method uses a bipolar wire electrode laparoscopically inserted into the diaphragm providing reference IM EMG measurements for validation of the novel site of surface skin recording.

We found that 0.2 mg · kg1 mivacurium caused a much faster onset and shorter clinical duration of the NMB at the diaphragm than at the OO and even faster than at the CS, a muscle that has been the focus of recent neuromuscular monitoring research (6).

Donati et al. (1) introduced surface EMG of the diaphragm to monitor the NMB; it was an easy and objective way to measure onset and offset of the NMB at this respiratory muscle. To introduce diaphragmatic monitoring into clinical practice, however, skin electrodes attached at the 7th or 8th intercostal space between the midclavicular and the anterior axillary line cannot be used. They would be inside the sterile surgical field where monitoring of the diaphragmatic relaxation would be most necessary in open and laparoscopic abdominal surgery.

We have therefore monitored the diaphragmatic NMB at a site on the back of the patient, lateral to the vertebra T12/L1 or L1/L2, where the electrodes do not interfere with surgery. Because the lumbar diaphragm inserts with its muscular crura into the first two to three lumbar vertebrae, it seemed possible to monitor its response to phrenic stimulation at that site (7) (Fig 1). No patient suffered any skin irritation because all patients were placed on a soft foam operating table; in this setting, we routinely apply skin electrodes for electrocardiographic monitoring on the patient’s back in cardiac anesthesia without any postoperative skin irritation. To avoid the influence of abdominal distension resulting from pneumoperitoneum, removal of the gas before the measurements is important. No surgical intervention took place during the measurements of the clinical duration to avoid displacement of the wire electrode. In a pilot study, we have used this technique in open abdominal surgery before the start of the current study to test its ability to monitor the NMB of the diaphragm. The opening of the abdomen itself, however, might influence the tension of the diaphragm, thus giving false measurements of onset and offset of the NMB for a given neuromuscular blocking drug.

The comparability of the two monitoring sites was very good; correlation analysis and Bland-Altman test showed good correlation and narrow limits of agreements.

In contrast to a recent study of Dhonneur et al. (5), we did not use invasive transcutaneous needle stimulation of the phrenic nerve. There was only minimal activation of the muscles supplied by the brachial plexus when we stimulated the phrenic nerve via superficial stimulation probe that did not influence the recording of the diaphragm. If at any point, phrenic nerve stimulation and subsequent monitoring of the NMB at the diaphragm would come into clinical practice, percutaneous stimulation via skin probes would be advantageous.

The CS is a small muscle originating from the medial superciliary arch that inserts into the skin of the medial forehead. Its contraction wrinkles the forehead. It is innervated by the temporal branch of the facial nerve (VII). AMG monitoring of the CS was first presented by Plaud and Donati (6). They stated that when 0.6 mg · kg −1 rocuronium is applied, the CS is a better indicator of laryngeal muscle relaxation than the OO. They measured the maximum block at the CS and the larynx and determined a comparable maximum effect and the time to reach 25% of control value and concluded that the neuromuscular response measured at the CS is a good reflection of the laryngeal muscle response. There is, however, no description of the onset time in that abstract.

The difference between the OO and the CS has been the focus of a recent anatomical study. Goodmurphy and Ovalle (8) studied both muscles morphologically and found that the OO muscle fibers are small, rounded, and 89% fast-twitch type-II-muscle fibers, whereas the CS muscle fibers are larger, polygonal, and 49% of fast-twitch type II. The capillary index of the CS, however, was 2.4 times the capillary index of the OO. Despite those structural differences between the CS and the OO, this study did not show any statistically significant difference in onset or duration between the two muscles.

In the current study, the onset and offset of the NMB of 0.2 mg · kg1 mivacurium at the diaphragm was much shorter and did not correlate with either the OO nor the CS. More studies are needed with other muscle relaxants to define the position of the CS as an indicator for diaphragm or laryngeal relaxation and its use in routine clinical anesthesia.

In conclusion, this study shows that surface EMG of the diaphragm can be traced at the back of the patient with reliable, reproducible monitoring that correlates well with IM EMG. This might be a first step toward the use of EMG monitoring of the diaphragm in clinical routine because its monitoring site does not interfere with abdominal surgery.

Back to Top | Article Outline


1. Donati F, Antzaka C, Bevan DR. Potency pancuronium at the diaphragm and the adductor pollicis muscle in humans. Anesthesiology 1986; 65: 1–5.
2. Chauvin M, Lebrault C, Duvaldestin P. The neuromuscular blocking effect of vecuronium on the human diaphragm. Anesth Analg 1987; 66: 117–22.
3. Donati F, Meistelman C, Plaud B. Vecuronium neuromuscular blockade at the diaphragm, the orbicularis oculi, and the adductor pollicis muscles. Anesthesiology 1990; 73: 870–5.
4. Cantineau JP, Porte F, d’Honneur G, Duvaldestin P. Neuromuscular effects of rocuronium on the diaphragm and adductor pollicis muscles in anesthetized patients. Anesthesiology 1994; 81: 585–90.
5. Dhonneur G, Kirov K, Slavov V, Duvaldestin P. Effects of an intubating dose of succinylcholine and rocuronium on the larynx and diaphragm. Anesthesiology 1999; 90: 951–5.
6. Plaud B, Donati F. The corrugator supercilii, not the orbicularis oculi, reflects rocuronium neuromuscular blockade of the adducting laryngeal muscles [abstract]. Anesthesiology 1999; 91: A1032.
7. Moore KL, Dalley AF. Clinically oriented anatomy. Baltimore: Lippincott Williams & Wilkins, 1999: 291.
8. Goodmurphy CW, Ovalle WK. Morphological study of two human facial muscles: orbicularis oculi and corrugator supercilii. Clin Anat 1999; 12: 1–11.
© 2001 International Anesthesia Research Society