Larger doses of nondepolarizing steroidal neuromuscular blocking drugs such as pancuronium, vecuronium, and rocuronium (1–3) are required to maintain paralysis during surgical procedures in patients receiving chronic anticonvulsant therapy (CAT) (carbamazepine or phenytoin). These patients recover quicker from neuromuscular blockade, which could put them at risk if movement occurred during a delicate neurosurgical procedure. The mechanism behind this faster recovery is not well understood. Pharmacokinetic (PK) as well as pharmacodynamic (PD) factors could be implicated, including hepatic enzyme induction, greater plasma protein binding, increased number or decreased sensitivity of the acetylcholine receptors, or direct competition for binding sites (4–6).
With the exception of one study (7), recovery from the neuromuscular blockade induced by atracurium, a neuromuscular blocking drug of the benzylisoquinolinium class, does not seem to be affected by anticonvulsants (8,9). In contrast to atracurium, plasma esterases are not involved in the in vitro degradation of cisatracurium and Hofmann hydrolysis is the rate-limiting step (10). In a retrospective analysis of three PK studies in patients (11–13), Kisor et al. (14) have estimated that only 23% of cisatracurium would be eliminated in an organ-dependent manner, 16% of which being through renal clearance. Consequently, it should be less susceptible to enzymatic induction or to other factors that influence neuromuscular blockade recovery. Cisatracurium could therefore be an interesting alternative to steroidal drugs to maintain neuromuscular blockade in patients receiving anticonvulsant therapy. Furthermore, its cardiovascular stability and the absence of histamine liberation are assets that differentiate it from atracurium, making it more suitable for use in neurosurgery. In fact, Koenig and Edwards (15) have found that the duration of action of a bolus of cisatracurium was not affected by anticonvulsants. However, they observed that the 10%–25% recovery index was shorter in the group chronically treated with carbamazepine or phenytoin than in a control group. This study could not conclude what caused this faster recovery because neither PD nor PK profiles were characterized.
We hypothesized that the faster recovery of the neuromuscular blockade in patients receiving CAT originates from PD factors and, concurrently, from smaller effect compartment concentrations. In addition, no data are available in the literature to predict if the speed of infusion needed to maintain a neuromuscular block at 5% of T1 will be different between an anticonvulsant-treated group and a control group.
This study was therefore designed to compare the PK/PD of cisatracurium after a prolonged infusion in patients treated with and without anticonvulsant therapy.
After IRB approval and written informed consent, 30 ASA physical status I–III patients, aged 18–70 yr, needing general anesthesia of at least 5 h of duration for elective neurosurgery in the dorsal decubitus position were enrolled in the study. Patients in the anticonvulsant group (n = 15) had been taking anticonvulsants (carbamazepine or phenytoin) for at least 4 wk. The blood concentrations of the anticonvulsants used were within the therapeutic range before surgery. Patients in the control group (n = 15) had never taken any anticonvulsant medication. Exclusion criteria included patients with cardiac, renal, hepatic, or neuromuscular impairment, and history of alcohol or drug abuse. Patients with a body mass index >30, and patients that had taken medications known to interfere with neuromuscular blockade (aminoglycosides, tetracyclines, clindamycine, procainamide, quinidine, magnesium sulfate, or calcium channel blockers) were also excluded.
This study was conducted in compliance with the good clinical research practice in PK/PD studies of neuromuscular blocking drugs (16,17).
Anesthesia was standardized for all patients. The attending anesthesiologist was not blinded to the study groups. One hour before surgery, the patients were premedicated with midazolam 0.07 mg/kg (maximum 5 mg) IM. Upon arrival in the operating room, standard monitoring equipment was applied and two IV lines were inserted, one of which was set at a constant flow rate and exclusively reserved for cisatracurium infusion. An arterial canula was inserted for direct measurement of arterial blood pressure and blood sampling. Rectal temperature was maintained above 35.5°C using a warming blanket. Ventilation was adjusted to maintain end-tidal CO2 between 30 and 35 mm Hg at the beginning of the surgery and no further change of ventilation was allowed during the study period. No IV or arterial catheters were inserted on the arm used for the neuromuscular monitoring.
Anesthesia was induced with propofol (0.5–3.0 mg/kg) and sufentanil (0.25–1.0 μg/kg). Supramaximal stimulation was then obtained and a 5-min equilibration period was allowed while maintaining anesthesia with 0.5%–1.0% end-tidal isoflurane with mask ventilation. A single bolus of cisatracurium (0.1 mg/kg) was then injected in the dedicated IV line over 5 s. After tracheal intubation, anesthesia consisted of 50% N2O in oxygen, 0.5%–1.0% end-tidal isoflurane, and sufentanil infusion (0.1–0.5 μg · kg−1 · h−1) to maintain mean arterial blood pressure within 20% of the baseline preinduction value.
Neuromuscular function was monitored using a Datex-Engstrom Neuromuscular Transmission Module (M-NMT) (Datex-Engstrom, Helsinki, Finland). The electromyogram of the adductor pollicis was recorded using five disposable Ag/AgCl electrodes placed as follows: two stimulating electrodes along the ulnar nerve at the wrist, two recording electrodes, one on the adductor pollicis and the second on the lateral surface of the index finger, and one ground electrode between the stimulating and the recording electrodes. The skin was cleaned with alcohol before electrodes were applied. The outstretched arm was enveloped in a cotton blanket to minimize heat loss and stabilized in a protective frame to prevent movement and interference during anesthesia. The M-NMT system was calibrated after induction but before administration of cisatracurium with pulses of 200 μs, at a rate of 2 Hz, starting from 10 mA with increments of 5 mA. The maximal current obtained was then increased by 15%, yielding the supramaximal stimulation. The system was set to deliver supramaximal train-of-four (TOF) stimulations (200 μs, at 2 Hz) every 12 s for an equilibration period of 5 min (18). After complete suppression of T1, the TOF stimulations were executed every minute. On the return of the first twitch (T1% > 0), an infusion of cisatracurium was started at 1.4 μg · kg−1 · min−1 (Harvard pump; Harvard Apparatus, Natick, MA) and was adjusted by 0.1 μg · kg−1 · min−1 increments or decrements every 3 min to maintain the T1% at 5% of baseline (between 3% and 7%). After a stable infusion was obtained for a period of 2 h, the perfusion was stopped to allow spontaneous recovery. The steady-state was defined as no change of the T1% value with a stable infusion rate (IR). During this period of 2 h (steady-state) as well as during the recovery period (these two periods corresponded to the low-stimulation intracranial period), the concentration of isoflurane as well as the ventilatory variables were kept constant. TOF stimulations were once again executed every 12 s to monitor the recovery. Neostigmine (0.04 mg/kg) along with glycopyrrolate (0.01 mg/kg) were given IV to every patient after maximal recovery (no change of the T1 value for 10 min) to obtain the final T1 value. Data were recorded using the Datex-Ohmeda AS/3 PC Data Collection Software (Datex-Ohmeda, Helsinki, Finland).
The time from the beginning of the cisatracurium injection to the first noticeable decrease of the T1 height was recorded as the lag time. The time until 95% and 100% T1 depression was considered as the onset time and the time to maximal depression, respectively. The recovery time was defined as the time until return of the twitch after maximal depression. At the end of the infusion, the time to spontaneous recovery of T1 to 10%, 25%, and 75% was recorded. The 10%–25% and the 25%–75% recovery indices were calculated. The data taken during the recovery period were normalized to the final T1 value.
Arterial blood samples (7 mL) were taken to measure plasma concentrations of cisatracurium. A sample was taken before induction of anesthesia to serve as a baseline for protein binding measurements. Two samples were taken 20 and 10 min before the end of the cisatracurium infusion (steady-state). Five other samples were taken at various times after the end of the infusion, during the recovery of T1 (at 10%, 20%, 30%, 40%, and 50% of the initial T1 value). The arterial blood was withdrawn in a chilled sodium heparin tube, and immediately centrifuged at high speed for 1 min. The plasma was separated and transferred to a precooled polypropylene tube containing 62.5 μL of sulfuric acid 2 M. The plasma samples were then immediately frozen on dry ice to be stored at −80°C until high-performance liquid chromatography (HPLC) analysis that was performed in a single batch for every patient at the end of the study. The whole manipulation process, from withdrawal to freezing, took <2 min for each sample.
Determination of cisatracurium concentration in plasma was performed by using HPLC. Bond-Elut® phenyl solid-phase extraction cartridges (Varian, Harbor City, CA) were used for extraction of cisatracurium. N-methyl laudanosine (500 ng/mL plasma) was used as the internal standard. After several purification steps, the eluent was half evaporated using a Speed-Vac concentrator (model SC210A; Savant Instruments, Farmingdale, NY). An aliquot was injected directly into the HPLC system by using an autosampler SIL-9A (Shimadzu, Kyoto, Japan). A Phenomenex Spherisorb SCX column (150 × 4.6 mm, inner diameter 5 μm; Phenomenex, Torrance, CA) was used for HPLC separation using a stepwise gradient (Thermo Separation Products, Riviera Beach, FL). The mobile phase changed from a first phase (14 mM Na2SO4 in 0.5 mM H2SO4/acetonitrile 40:60) during 5 min to a second phase (70 mM Na2SO4 in 0.5 mM H2SO4/acetonitrile 40:60) during 6 min. The solvent flow rate was 2.0 mL/min and the column was maintained at 50°C. The excitation and emission wavelengths of the fluorescence detector (Hewlett Packard, Waldbronn, Germany) were set at 280 and 320 ηm, respectively. This method, published by Bryant et al. (19) for urine samples, was slightly modified and fully validated in our laboratory. The assay is specific for cisatracurium and its metabolites, but the latter were not quantified. The coefficient of variation for between-run precision was <7% at a concentration of 6.5–2500 ng/mL. The percentage of accuracy of the assay was 100.9% ± 3%.
Determination of cisatracurium protein binding was performed for two subgroups of five patients. The selection was made to ensure that no bias was introduced by anticonvulsant therapy. For in vitro protein binding determination, blood samples were obtained before induction of anesthesia using ethylenediaminetetraacetic acid as anticlotting agent. Plasma was separated, flash frozen, and stored at −70°C. A method similar to that described by Cameron et al. (20) was used to determine the free fraction of cisatracurium on thawed samples.
PK variables of cisatracurium were calculated using a noncompartmental approach. For each patient, total body clearance (CL) was obtained by dividing the mean IR maintained during the last 2 h by the mean steady-state plasma concentration of cisatracurium required to maintain 95% block (Cpss95). The terminal elimination rate (Kel) was obtained by performing linear regression on the plasma concentrations measured during the recovery period (5 data points). The apparent volume of distribution (Vdβ) was then deduced by dividing CL by Kel.
Based on the available literature (21,22) for an estimated IR of 1.4 μg · kg−1 · min−1 (standard deviation of 0.7 μg · kg−1 · min−1), assuming a β of 0.2 and an α of 0.05, it was calculated that 15 patients in each group were needed to detect a 50% increase of the speed of infusion to maintain T1 at 5% of the control in the group taking anticonvulsants. In Koenig and Hoffman’s (3) study, 2 groups of 14 patients were sufficient to demonstrate a 50% difference in the 10%–25% recovery index between patients treated with or without anticonvulsants.
Nonparametric variables were analyzed using the Fisher’s exact test. Parametric variables were analyzed using the t-test and analysis of variance. The between-group difference between the onset time and the various recovery times and indices were analyzed with two-factor analysis of variance (two-way analysis of variance). PK parameters were compared using the Student’s t-test for unpaired data. A P value < 0.05 was considered significant.
Demographic and intraoperative data are presented in Table 1. The two groups were comparable regarding age, gender distribution, body mass index, and ASA physical status. No patients had preoperative motor deficit. Intraoperative data were also comparable between the two groups. Both groups had a similar and negative intraoperative fluid balance and the doses of mannitol given to provide brain relaxation were not different between the two groups. The average body temperature during the surgery was also identical in both groups and well within the normal range.
Table 2 presents the neuromuscular transmission data. CAT did not result in prolonged onset times (lag time, onset, and time to obtain maximal suppression of T1). However, when compared with the control group, CAT resulted in faster recovery times. The recovery, after the initial bolus, as well as time to obtain T1 25% and T1 75% after the termination of the cisatracurium infusion were significantly shorter in the anticonvulsant group. The recovery indices T1 0%–25% and T1 25%–75% were also significantly decreased in the anticonvulsant group. The speed of infusion of cisatracurium needed to maintain a 95% depression of T1 at steady-state was significantly more rapid in the anticonvulsant group.
Individual plasma concentrations were normalized by dividing them by their corresponding steady-state IR (Fig. 1). This normalization disclosed the increased interpatient variability in the CAT group. For illustration purposes, plasma concentration was then sorted by sampling time, and subsets of 13–15 data points were averaged. The midtime of the corresponding period was used for sampling time.
Table 3 presents the PK variables derived from the noncompartmental analysis. The protein binding of cisatracurium did not differ between both subsets of five patients. The CL of cisatracurium for the last 2 h of perfusion (steady-state) was significantly faster in the anticonvulsant group when compared with the control group. However, no differences were observed for the apparent Vdβ and Kel. The Cpss95 was significantly larger in the anticonvulsant group.
Subgroup analysis revealed that patients receiving phenytoin alone had similar PK/PD profiles to those receiving carbamazepine alone.
This study shows that treatment with phenytoin or carbamazepine affects the PK/PD of cisatracurium in patients undergoing lengthy intracranial procedures. The major findings of the study are that patients treated with anticonvulsants need a more rapid IR of cisatracurium to steadily maintain a given twitch depression during a neurosurgical procedure and that they have a faster recovery of cisatracurium-induced neuromuscular blockade after an infusion of the drug. The increase in CL, although mostly responsible for the faster recovery, cannot account for the larger dose required in patients receiving CAT. The larger Cpss95 observed in patients treated with anticonvulsants suggests a resistance to the effect of cisatracurium, indicating that almost 50% of the changes would be of PD origin.
These results confirm the findings obtained by Koenig and Edwards (15) in a similar population of neurosurgical patients treated with phenytoin or carbamazepine either chronically (>2 weeks) or acutely (<2 weeks). Those authors administered a single bolus (4 × 95% effective dose) of cisatracurium and compared the onset and recovery profiles with a control group not taking any anticonvulsant. They found that the T1 0%–25% recovery index was halved in the anticonvulsant-treated group when compared with the controls, with no difference between the acutely and chronically treated patients. There was no difference between the two groups in the times to obtain a T1 10% and a T1 25%, but the authors did not look at late recovery variables such as times to obtain a T1 50% and T1 75% recovery. In our study, the patients were allowed to completely recover from the neuromuscular blockade while under a stable plane of general anesthesia and the recovery variables were studied after a lengthy infusion of cisatracurium and not only after a single bolus (with the exception of recovery). The time to obtain a T1 10% was not altered in the anticonvulsant group, but the time to recovery and the times to obtain a T1 25% and 75% were shortened by 25%, 20%, and 26%, respectively. We showed proportional reduction for the T1 0%–25% and 25%–75% recovery indices.
Koenig and Edwards (15) did not conduct PD profiles in their study; therefore, they could not provide an explanation as to why this increased speed of recovery occurred.
In the present study, the patients in the anticonvulsant group showed a significant increase in the CL of cisatracurium. The CL obtained for our control group, 5.7 mL · kg−1 · min−1, is within the range already published (4.6–5.7 mL · kg−1 · min−1) (9,11,12). Thus, the CL observed for the anticonvulsant pretreated group (7.1 mL · kg−1 · min−1) represents a 25% increase. Similar protein binding and intraoperative fluid balance in both groups are in agreement with the lack of difference in cisatracurium Vdβ. The increased CL would be mostly attributable to an enhanced contribution of biliary excretion to the overall Kel of cisatracurium.
Kisor et al. (14) have estimated that 23% of the Kel of cisatracurium would result from biliary excretion. In patients chronically taking anticonvulsants, enzymatic induction could possibly increase both the speed of the hepatic metabolism/biliary excretion and the percentage of the drug eliminated via these pathways. Maximal induction has been shown to occur after 1–3 weeks of treatment for both drugs (23). In fact, Pirttiaho et al. (24,25) have observed that liver weight was increased by 30% in patients treated with anticonvulsants when compared with normal patients, which was accompanied by a proportional increase of hepatic blood flow. The 25% increase in CL observed in our study would be compatible with such an increase in blood flow and, in turn, biliary excretion. In comparison, Alloul et al. (6) have found that the CL of vecuronium was more than doubled (increase of 135%) in neurosurgical patients treated with carbamazepine. This much larger increase of CL correlates well with the predominantly hepatic uptake and Kel of vecuronium in humans.
In contrast, however, De Wolf et al. (11) have shown that patients undergoing liver transplantation for end-stage liver disease had a slightly larger CL of cisatracurium. This was attributed to an increase in the Vdβ, when compared with normal control. The urinary CL of the drug was not increased in this group of patients.
The Cpss95 observed in our control group is within the range (135–205 ng/mL) of the 95% effective concentration (EC95) calculated from data published by other groups of investigators (11,26,27) after correction for the potentiating effect of isoflurane. Correction for the free fraction was not deemed necessary because the results for the two subsets of patients were not different.
In our study, patients in the anticonvulsant pretreated group needed an IR 44% faster than the control group to maintain a 95% twitch depression under steady-state conditions. This, together with a 20% increase of the Cpss95, indicates a clear resistance to the effect of cisatracurium. Our study design did not provide the usual variables required to describe PK/PD relationships (EC50, γ, and keo values) and perform simulations. Our approach was meant to reproduce the clinical setting. In PK/PD studies, conclusions regarding sensitivity are often drawn from EC50 values only. This may not be appropriate because all three variables contribute to the concentration-effect relationship. Similar EC50 values do not necessarily mean similar sensitivity because changes in γ can alter the shape of the sigmoid and lead to different EC90 or EC95 values. These values are even more clinically relevant because endotracheal intubation, maintenance of adequate surgical relaxation, or pharmacological reversal of blockade take place at the extremes of the sigmoid. This is why, in our opinion, the Cpss95 is a valid indicator of the patient’s sensitivity to cisatracurium.
Alderdice and Trommer (28) have shown, using an in vitro frog sciatic nerve sartorius muscle preparation, that carbamazepine depresses postjunctional sensitivity to released acetylcholine producing a parallel decrease of amplitude of miniature end-plate potentials. They could not elucidate the mechanism of this increase of resistance. In a similar in vitro model using phenytoin in mouse, Gage et al. (29) have demonstrated that the reduction of amplitude was caused by a reduction both in the quantal content of end-plate potentials (presynaptic effect) and in the amplitude of the voltage response to quanta of acetylcholine (postsynaptic effect). This up-regulation of the acetylcholine receptor could be explained by the known neuromuscular blockade enhancing effect of phenytoin and carbamazepine (30,31). A prolonged antagonism of acetylcholine receptor by the anticonvulsant would likely cause this increased resistance at the receptor level.
In conclusion, patients chronically treated with anticonvulsants are resistant to the effect of cisatracurium and have an increased CL of the drug. These patients need a more rapid speed of infusion to steadily maintain a given twitch depression and show a faster recovery from neuromuscular blockade after a prolonged infusion. This increase is, however, not as rapid as that required for aminosteroidal drugs. Cisatracurium could then be a viable choice to maintain neuromuscular blockade in patients taking anticonvulsants.
1. Spacek A, Neiger FX, Krenn CG, et al. Rocuronium-induced neuromuscular block is affected by chronic carbamazepine therapy. Anesthesiology 1999;90:109–12.
2. Hickey DR, Sangwan S, Bevan JC. Phenytoin-induced resistance to pancuronium. Anesthesia 1988;43:757–9.
3. Koenig HM, Hoffman WE. The effect of anticonvulsant therapy on two doses of rocuronium-induced neuromuscular blockade. J Neurosurg Anesthesiol 1999;2:86–9.
4. Roth S, Ebrahim ZY. Resistance to pancuronium in patients receiving carbamazepine. Anesthesiology 1987;66:691–3.
5. Kim CS, Ardnold FJ, Iani MS, et al. Decreased sensitivity to metocurine during long-term phenytoin therapy may be attributable to protein binding and acetylcholine receptor changes. Anesthesiology 1992;77:500–6.
6. Alloul K, Whalley DG, Shutway, et al. Pharmacokinetic origin of carbamazepine-induced resistance to vecuronium neuromuscular blockade in anesthetized patients. Anesthesiology 1996;84:330–9.
7. Tempolhoff R, Modica PA, Jellish WS, Spitznagel EL. Resistance to atracurium-induced neuromuscular blockade in patients with intractable seizure disorders treated with anticonvulsants. Anesth Analg 1990;71:665–9.
8. Spacek A, Neiger FX, Spiss CK, Kress HG. Atracurium-induced neuromuscular block is not affected by chronic anticonvulsant therapy with carbamazepine. Acta Anesthesiol Scand 1997;41:1308–11.
9. Ornstein E, Atteo RS, Schwartz AE, et al. The effect of phenytoin on the magnitude and duration of neuromuscular block following atracurium or vecuronium. Anesthesiology 1987;67:191–6.
10. Welch RM, Brown A, Ravitch J, Dahl R. The in vitro degradation of cisatracurium, the R, cis
-isomer of atracurium, in human and rat plasma. Clin Pharmacol Ther 1995;58:132–42.
11. De Wolf AM, Freeman JA, Scott VL, et al. Pharmacokinetics and pharmacodynamics of cisatracurium in patients with end-stage liver disease undergoing liver transplantation. Br J Anaesth 1996;76:624–8.
12. Lien CA, Schmith VD, Belmont MR, et al. Pharmacokinetics of cisatracurium in patients receiving nitrous oxide/opioid/barbiturate anesthesia. Anesthesiology 1996;84:300–8.
13. Ornstein E, Lien CA, Matteo RS, et al. Pharmacodynamics and pharmacokinetics of cisatracurium in geriatric surgical patients. Anesthesiology 1996;84:520–5.
14. Kisor DF, Schimth VD, Wargin WA, et al. Importance of the organ-independent elimination of cis-atracurium. Anesth Analg 1996;83:1065–71.
15. Koenig HM, Edwards TL. Cis-atracurium-induced neuromuscular blockade in anticonvulsant treated neurosurgical patients. J Neurosurg Anesthesiol 2000;12:314–8.
16. Viby-Mogensen J, Englbaek J, Eriksson LI, et al. Good clinical research practice (GCRP) in pharmacodynamic studies of neuromuscular blocking agents. Acta Anaesthesiol Scand 1996;40:59–74.
17. Viby-Mogensen J, Ostergaard D, Donati F, et al. Pharmacokinetic studies of neuromuscular blocking agents: good clinical research practice (GCRP). Acta Anaesthesiol Scand 2000;44:1169–90.
18. Lee GC, Lyengar S, Szenohradszky J, el at. Improving the design of muscle relaxant studies. Anesthesiology 1997;86:48–54.
19. Bryant BJ, James CD, Cook DR, Harrelson JC. High performance liquid chromatography assay for cisatracurium and its metabolites in human urine. J Liq Chromatogr Relat Technol 1997;20:2041–51.
20. Cameron M, Donati F, Varin F. In vitro plasma protein binding of neuromuscular blocking agents in different subpopulations of patients. Anesth Analg 1995;81:1019–25.
21. Belmont MR, Lien CA, Quessy S, et al. The clinical neuromuscular pharmacology of 51W89 in patients receiving nitrous oxide/opioid/barbiturate anesthesia. Anesthesiology 1995;82:1139–45.
22. Eastwood NB, Boyd AH, Parker CJR, Hunter JM. Pharmacokinetics of 1R-cis 1′R-cis atracurium besylate (51W89) and plasma laudanosine concentrations in health and chronic renal failure. Br J Anaesth 1995;75:431–5.
23. Anderson GD. A mechanistic approach to antiepileptic drug interactions. Ann Pharmacother 1998;32:554–63.
24. Pirttiaho HI, Sotaniemi EA, Pelkonen RO, Pitkanen U. Hepatic blood flow and drug metabolism in patients on enzyme-inducing anticonvulsants. Eur J Clin Pharmacol 1982;22:441–5.
25. Pirttiaho HI, Sotaniemi EA, Ahokas JT, Pitkanen U. Liver size and indices of drug metabolism in epileptics. Br J Clin Pharmacol 1978;6:273–8.
26. Tran VT, Fiset P, Varin F. Pharmacokinetics and pharmacodynamics of cisatracurium after a short infusion in patients under propofol anesthesia. Anesth Analg 1998;87:1158–63.
27. Bergeron L, Bevan DR, Berrill A, et al. Concentration-effect relationship of cisatracurium at three different dose levels in the anesthetized patient. Anesthesiology 2001;95:314–23.
28. Alderdice MT, Trommer BA. Differential effects of the anticonvulsants phenobarbital, ethosuximide and carbamazepine on neuromuscular transmission. J Pharmacol Exp Ther 1980;215:92–6.
29. Gage PW, Lonergan M, Torda TA. Presynaptic and postsynaptic depressant effects of phenytoin sodium at the neuromuscular junction. Br J Pharmacol 1980;69:119–21.
30. Nguyen A, Ramzan I. Acute in vitro neuromuscular effects of carbamazepine and carbamazepine-10,11-epoxide. Anesth Analg 1997;84:886–90.
31. Cammu G. Interactions of neuromuscular blocking drugs. Acta Anaesthesiol Belg 2001;52:357–63.