Mivacurium, a potent short-acting non-depolarizing neuromuscular blocking agent, is metabolized by plasma cholinesterase. The short duration of action of mivacurium is in large part related to its hydrolysis by plasma cholinesterase [1,2]. Patients with liver failure exhibited markedly reduced plasma cholinesterase activity, which produced prolonged recovery of neuromuscular blockade [3,4]. Several earlier studies demonstrated a negative correlation between plasma cholinesterase activity and duration of neuromuscular block in healthy patients [5,6] and those with hepatic failure [3,4], and also reported conflicting results in healthy subjects [1,2]. However, the small number of patients included in human studies with hepatic failure does not permit us to draw conclusions about mivacurium.
The most common method of producing experimental liver injury is by administering multiple doses of carbon tetrachloride (CCl4) [7,8]. Hepatic fibrosis is successfully induced in rabbits by subcutaneous injection of CCl4 . To test the pharmacodynamic effects of mivacurium following the severity of liver dysfunction in an experimental model, we examined dose-response relationships and the duration of neuromuscular blockade and confirmed the relationship between recovery indices and plasma cholinesterase.
After approval by our Institutional Animal Care and Use Committee, we used 66 male adult Korean white rabbits, weighing 2.1-3.4 kg. The animals were allowed to adapt to the animal care facility for 1 week, with access to standard rabbit chow (Baehab, Ahnsan, Korea). The rabbits were randomly allocated into three groups. Hepatic injury was induced by 99.5% CCl4 (Showa Chemical Ind., Tokyo, Japan), which was mixed with corn oil (1 : 1 vol/vol) and injected subcutaneously on the left thigh site at a dose of 0.3 mL kg−1 (Group M; n = 21) and 0.6 mL kg−1 (Group S; n = 24), respectively, three times a week for 11 weeks. Three rabbits from S group suddenly died of causes related to CCl4 and were excluded from the study. For the control group (Group C), corn oil was injected subcutaneously at a dose of 0.3 mL kg−1 in the same way (n = 21). The doses and duration for CCl4 were decided after a preliminary study. Rabbits were kept in individual cages and blood samples were obtained every week from the ear vein 24 h after CCl4 administration. Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were measured by a Hitachi 7600 chemical autoanalyzer (Hitachi Co., Tokyo, Japan). Plasma cholinesterase activity was measured using a Hitachi 747 chemical autoanalyzer (Hitachi Co., Tokyo, Japan) according to the method of Garry .
Three rabbits in each group were randomly assigned to the dose-response study, the time-course study and the cumulative study, respectively. Randomization was based on computer-generated codes. At the end of the period of injection of CCl4, each animal was anaesthetized with propofol 1.5 mg kg−1 intravenously (i.v.) using a marginal vein of the left ear, and anaesthesia was maintained with propofol 0.2 mg kg−1 min−1. A tracheostomy was performed and the rabbit was ventilated with an animal respirator (Shinano Co., Tokyo, Japan). End-tidal carbon dioxide (CO2SMO; Novametrix Inc., Wallingford, CT, USA) was monitored and maintained at 4.2-5.2 kPa. Rectal temperature was controlled at approximately 38°C with a Blanketrol II (Cincinnati Sub-Zero Inc., Cincinnati, OH, USA) and heat lamps. The right ear vein was cannulated for mivacurium administration. A common carotid artery was cannulated for monitoring arterial pressure and intermittent analysis of arterial blood gases. A four-limb electrocardiogram was used for heart rate monitoring. I.v. fluid administration using a syringe pump was maintained at 6 mL kg−1 h−1 (0.9% NaCl) during the experiment.
The common peroneal nerve was stimulated supramaximally at the posterolateral aspect of the knee with 0.2 ms pulses derived from a peripheral nerve stimulator (DualStim; Life-Tech Inc., Stafford, TX, USA), which was opposite to the side used for CCl4 injection. Train-of-four (TOF) stimuli (2 Hz) were applied once every 10 s. The tibialis anterior muscle was detached from its insertion and tied to a force transducer (45196A, San-ei Co., Tokyo, Japan). The twitch response was quantified mechanomyographically with preload tension and the mechanomyogram was recorded on a multichannel recorder. Neuromuscular block was quantified by the first twitch (T1) of the TOF.
After stable recording of neuromuscular transmission had been established for 30 min, each of the following predetermined doses of mivacurium was used to establish single dose-response curves and was administered by random allocation: 10, 20, 30 or 40 μg kg−1 in C and M groups; 20, 30 or 40 μg kg−1 in S group, respectively (each n = 7). The doses of mivacurium were chosen following a pilot study. Each dose of mivacurium was withheld until the muscle twitch had recovered from the preceding dose and had remained at baseline value for at least twice the duration of the block of the preceding dose. The neuromuscular response was recorded as the maximum depression of twitch tension, expressed as a percentage of the control value. The percentage values for twitch depression in each group were transformed to probits and plotted against the logarithm of the dose. Regression lines were compared using the analysis of covariance. The effective dose (ED)50 and ED95 values were calculated from the log-probit regression lines for each group.
The twitch recordings after the administration of mivacurium 60 μg kg−1 in each group (each n = 7), respectively, were evaluated for the following variables: time from the end of the injection of mivacurium to maximal twitch suppression (onset); time from the end of the injection of the initial dose to recovery of T1 in the TOF to a value of 1%, 25%, 75% and 95% of control twitch tension (T1(1, 25, 75, 95)); time from 25% to 75% twitch recovery (recovery index (RI)); time from the end of the injection of the initial dose to a TOF ratio (T4/T1) of 70% (TOF (70)).
The subsequent administration of mivacurium 20 μg kg−1 in each group (each n = 7), respectively, was injected repeatedly 10 times until 95% recovery of T1 in the TOF following the initial dose of mivacurium 60 μg kg−1. The intervals after repeated doses were evaluated in each group, respectively.
The rabbits were killed at the end of the experimental period with an overdose of thiopental sodium. Three fragments of each liver lobe were collected and processed for light microscopy. The tissues were fixed in 10% formalin and were then embedded in paraffin. Serial sections were cut and stained with haematoxylin-eosin and Masson's trichrome. An observer blinded to the treatment the rabbits had received performed the histological examination. Liver biopsies were evaluated using Knodell's histological activity index (HAI) scoring system, which is most widely used . Knodell's HAI scoring was scored as portal inflammation (0-4), intralobular degeneration and focal necrosis (0-4), and periportal necrosis (0-10) for necro-inflammation, and the degree of fibrosis from 0 to 4 (0: none; 1: fibrous portal expansion; 3: bridging fibrosis; 4: cirrhosis). Mild hepatitis (1-5), moderate hepatitis (6-10) and severe hepatitis (11-22)  were classified in this study, respectively.
We chose to analyse our liver data with a non-parametric Kruskal-Wallis test, followed, where significant, by the U-test. For the neuromuscular data, slopes, ED50 and ED95 were compared between the three groups by linear regression analysis and analysis of covariance. Comparison of the pharmacodynamic variables between groups was made using the Kruskal-Wallis test, and correlation of 95% recovery time and cholinesterase activity with Pearson's correlation coefficient. To isolate the group that differed from the others, Dunnett's multiple comparison procedure was applied. Values are reported as mean ± SD or median (interquartile range (25-75% percentiles)). Statistical analysis was performed using SPSS statistical software, version 11.0 (SPSS Inc., Chicago, IL, USA). P < 0.05 was considered statistically significant.
After CCl4 treatment for 11 weeks, both M and S groups were found to have increased AST, ALT and bilirubin, decreased plasma cholinesterase, and no gain of weight, respectively, compared with placebo (P < 0.0001; Table 1). Serum albumin and total protein levels in S group were lower than those in Groups C and M (P < 0.0001), but there were no differences between C and M groups. Table 2 represents the histological evaluation of each groups as severe hepatitis (n = 20) and moderate hepatitis (n = 1) in Group S, and mild hepatitis (n = 21) in Group M. Knodell's HAI scores in Group S were significantly increased compared with those in the other groups (P < 0.0001), and there was a significant difference between Groups C and M (P < 0.001). One rabbit with moderate hepatitis from Group S was excluded from cumulative measurements. Ascites was found in Group S (11.3 ± 4.8 mL) (n = 9), but not in the other two groups.
CCl4 treatment was associated with a rightward shift of the mivacurium dose-response curves (Fig. 1). The slopes for Groups C, M and S were 3.6 ± 0.5, 3.3 ± 0.4 and 4.1 ± 0.7, respectively. Although the slopes are not significantly different, the elevations of the dose-response curves for Group S lie significantly to the right of the other groups (P < 0.0001). As a result, the calculated ED50 values for C, M and S groups were 17.1 ± 2.6, 18.2 ± 2.7 and 31.8 ± 3.2 μg kg−1, respectively. Corresponding ED95 values were 30.5 ± 3.6, 34.2 ± 3.9 and 46.2 ± 4.4 μg kg−1, respectively. It was significantly higher in Group S compared with Groups C and M (P < 0.0001).
The response of mivacurium 60 μg kg−1 is shown in Table 3. The onset in Group S was more rapid than those in the other groups, although this was not statistically significant. The rabbits in Group S receiving mivacurium had significantly prolonged recovery indices compared with Groups C (P < 0.001) and M (P < 0.05), respectively. One rabbit in Group S had a markedly prolonged recovery (T1(95): 62 min).
In the repeated bolus injection study with mivacurium 20 μg kg−1, the T1(95) and RI were significantly prolonged in Group S (18.8 ± 1.4 min and 5.6 ± 0.8 min, respectively) compared with Group C (11.6 ± 1.0 min and 2.1 ± 0.4 min) and Group M (12.6 ± 0.9 min and 2.4 ± 0.5 min) (P < 0.0001), whereas constant intervals during the repeated injections were found in both hepatic injuries (Fig. 2).
Linear regression analysis revealed a significant negative correlation between plasma cholinesterase and the duration until 95% recovery: Group C (r2 = 0.49; P = 0.006), Group M (r2 = 0.42; P = 021) and Group S (r2 = 0.49; P = 0.006), respectively (Fig. 3).
We found that severe hepatitis was associated with a rightward shift of the dose-response curve of mivacurium, and a prolonged recovery from neuromuscular blockade compared with placebo and mild hepatitis, whereas constant intervals during the repeated injections and a negative correlation with the activity of plasma cholinesterase were found even in severe hepatic injury.
CCl4 has long been known to produce liver cirrhosis in mice ; however, the continuous monitoring of biochemical changes during the progression of liver injury to cirrhosis and the monitoring of neuromuscular blockade were not always possible due to the small size of the animals . CCl4-induced liver injury in rabbits was described by Ugazio and colleagues  with microsomal induction by phenobarbital of long duration (6 months), and by Brandao and colleagues  with intragastric CCl4 and phenobarbital in drinking water within 16 weeks in 87.5%, respectively. In the present study, we were able to obtain severe hepatitis (93%) within 11 weeks without phenobarbital induction with similar mortality . Because of the variable sensitivity of rabbits to CCl4, wide variation was found in plasma AST and ALT levels at the end of treatment as experimental toxic hepatitis  and liver cirrhosis , and we also found corresponding results.
The ED50 obtained from the dose-response curves in healthy subjects corresponds with what is usually calculated for mivacurium . In this study, the twofold rightward shift in the dose-response curves of mivacurium in rabbits with severe hepatitis might be related to the larger initial volume of distribution, up-regulation of perijunctional immature acetylcholine receptors or an increase in serum protein binding, similar to changes seen in sepsis, burn injury or muscle disuse atrophy, to be the cause for the resistance . This mechanism has been proposed to explain the so-called ‘resistance’ to mivacurium and pancuronium in cirrhotic patients and hence to a less intense block [4,18]. Mild hepatitis had no influence on the potency of mivacurium, which might be related to the large hepatic metabolic activity or a less change of plasma protein binding.
In severe hepatitis, the duration of mivacurium appeared to be prolonged and unpredictable (T1(95): 62 min): the time to 95% recovery of T1/T0 of 34.5 ± 7.8 min contrasts with 21.1 ± 3.2 min in healthy subjects. No difference in the duration between mild hepatitis and healthy subjects is also of interest. The duration of action of a bolus dose of a muscle relaxant is a complicated function of neuromuscular junction sensitivity, distribution volumes and plasma clearance of the drug such that even twofold changes in plasma clearance of a muscle relaxant might not influence recovery from single bolus doses . An accelerated tendency of the onset of neuromuscular blockade in severe hepatitis compared with both placebo and mild hepatitis was found in this study, although this was not statistically significant (Table 3). Patients with hepatic failure have altered haemodynamics, in particular increased cardiac output , which may enhance delivery of the drug to the site of action  and hasten the onset of block.
Several studies reported that the patient's plasma cholinesterase activity has a significant negative correlation with recovery indices after mivacurium in healthy patients or hepatic cirrhosis [3-6,21], but not in earlier studies of the neuromuscular blockade given to healthy subjects [1,2]. This discrepancy may result from differences in the methods used to assay plasma cholinesterase. We also found a significant negative correlation between plasma cholinesterase activity and the duration of neuromuscular blockade even in severe hepatic injury (Fig. 3). This study would suggest, however, that neuromuscular block produced by mivacurium was not potentiated in spite of the reduced cholinesterase activity to about 70% in mild hepatitis (251.2 ± 22.1 IU L−1). Plasma cholinesterase activity in rabbits is reported as approximately 20% of that in human beings [22,23]. Plasma cholinesterase activity in liver cirrhosis was reduced to 1/4 of that in the normal human being . In this study it was markedly reduced to 1/7 of placebo as 53.2 ± 9.1 IU L−1, which might produce the potentiation of neuromuscular blockade.
In the clinical study, repeated doses should be given at a recovery level of 25% control twitch tension due to clinical relaxation . However, we obtained a recovery level of 95% control twitch tension and RI after the repeated doses, which can make more accurate variables of the recovery indices after neuromuscular blockade. Severe hepatitis induced a prolongation of recovery after repeated doses, but maintained the constant intervals (Fig. 2). This suggests that repeated low doses of mivacurium might not be cumulative even at the markedly low activity level of plasma cholinesterase. This implies that mivacurium may be a suitable neuromuscular drug by repeated doses to maintain target twitch depression even in patients with hepatic injury if neuromuscular monitoring is available. We gave up to 10 repeated doses for the cumulative measurement to compare with previous communications .
Ascites was found in approximately 50% of rabbits in an experimental model of extensive sinusoidal fibrosis , which was similar to our study. Ascites resulted from the deleterious combination of cirrhosis, high portal pressure and exudation, which was associated with splanchnic arterial vasodilatation and systemic (including renal) vasoconstriction . On the other hand, others have reported the absence of ascites, despite the occurrence of experimental well-defined cirrhosis [14,15].
One limitation of our study is the lack of information regarding the pharmacokinetics of mivacurium in CCl4-treated rabbits. This limitation precludes us from making any comments on the pharmacokinetic alteration of our results.
We conclude that, in severe hepatitis, mivacurium had decreased potency and a prolonged duration of action. In mild hepatitis, the potency and duration of mivacurium were not different compared with placebo. Both severe and mild hepatitis required no adjustments of different doses for repeated injection after desired depth of neuromuscular block, and had a negative correlation between plasma cholinesterase activity and the indices of recovery after mivacurium.
Supported by Research Fund HY-2003-I-043 of the Hanyang University.
1. Savarese JJ, Ali HH, Basta SJ et al.
The clinical neuromuscular pharmacology of mivacurium chloride BW (B1090U). A short-acting nondepolarizing ester neuromuscular blocking drug. Anesthesiology
2. Rigg JD, Wilson AC, Pollard BJ. Mivacurium or vecuronium for muscular relaxation in day-case surgery. Eur J Anaesthesiol
3. Cook DR, Freeman JA, Lai AA et al.
Pharmacokinetics of mivacurium in normal patients and in those with hepatic or renal failure. Br J Anaesth
4. Devlin JC, Head-Rapson AG, Parker CJ, Hunter JM. Pharmacodynamics of mivacurium chloride in patients with hepatic cirrhosis. Br J Anaesth
5. Ostergaard D, Jensen FS, Jensen E, Skovgaard LT, Viby-Mogensen J. Influence of plasma cholinesterase activity on recovery from mivacurium-induced neuromuscular blockade in phenotypically normal patients. Acta Anaesthesiol Scand
6. Phillips BJ, Hunter JM. Use of mivacurium chloride by constant infusion in the anephric patient. Br J Anaesth
7. Muriel P, Moreno MG, Hernandez MC, Chavez E, Alcantar LK. Resolution of liver fibrosis in chronic CCl4 administration in the rat after discontinuation of treatment: effect of silymarin, silibinin, colchicine and trimethylcolchicinic acid. Basic Clin Pharmacol Toxicol
8. Kamalakkannan N, Rukkumani R, Varma PS, Viswanathan P, Rajasekharan KN, Menon VP. Comparative effects of curcumin and an analogue of curcumin in carbon tetrachloride-induced hepatotoxicity in rats. Basic Clin Pharmacol Toxicol
9. Jin B, Alter HJ, Zhang ZC et al.
Reversibility of experimental rabbit liver cirrhosis by portal collagenase administration. Lab Invest
10. Garry PJ. A manual and automated procedure for measuring serum cholinesterase activity and identifying enzyme variants. Differentiation by means of Tris and phosphate buffers. Clin Chem
11. Gronbaek K, Christensen PB, Hamilton-Dutoit S et al.
Interobserver variation in interpretation of serial liver biopsies from patients with chronic hepatitis C. J Viral Hepat
12. Lee YS, Yoon SK, Chung ES et al.
The relationship of histologic activity to serum ALT, HCV genotype and HCV RNA titers in chronic hepatitis C. J Korean Med Sci
13. Stowell RE, Lee CS, Tsuboi KK, Villasana A. Histochemical and microchemical changes in experimental cirrhosis and hepatoma formation in mice by carbon tetrachloride. Cancer Res
14. Brandao CG, Ferreira HH, Piovesana H et al.
Development of an experimental model of liver cirrhosis in rabbits
. Clin Exp Pharmacol Physiol
15. Ugazio G, Bosia S, Cornaglia E. Experimental model of cirrhosis in rabbits
exposed to carbon tetrachloride by inhalation. Exp Commun Mol Pathol Pharmacol
16. Kim KS, Jeon JW, Koh MS, Shim JH, Cho SY, Suh JK. The duration of immobilization causes the changing pharmacodynamics of mivacurium and rocuronium in rabbits
. Anesth Analg
17. Mayer B, Fink H, Bogdanski R, Stadler J, Blobner M. Inflammatory liver disease shortens atracurium-induced neuromuscular blockade in rats. Eur J Anaesthesiol
18. Duvaldestin P, Agoston S, Henzel D, Kersten UW, Desmonts JM. Pancuronium pharmacokinetics in patients with liver cirrhosis. Br J Anaesth
19. Liu H, Gaskari SA, Lee SS. Cardiac and vascular changes in cirrhosis: pathogenic mechanisms. World J Gastroenterol
20. Audibert G, Donati F. The onset of rocuronium, but not of vecuronium or mivacurium, is modified by tourniquet inflation. Anesth Analg
21. Green DW, Fisher M, Sockalingham I. Mivacurium compared with succinylcholine in children with liver disease. Br J Anaesth
22. Ostergaard D, Jensen FS, Skovgaard LT, Viby-Mogensen J. Dose-response relationship for mivacurium in patients with phenotypically abnormal plasma cholinesterase activity. Acta Anaesthesiol Scand
23. Kim KS, Kim KH, Shin WJ, Yoo HK. Neuromuscular interaction between mivacurium and esmolol in rabbits
24. Blanloeil Y, Delaroche O. Decrease in plasmatic cholinesterase activity in severe bacterial infections: comparison with the decrease observed in severe liver cirrhosis. Ann Fr Anesth Reanim
25. Wanless IR, Belgiorno J, Huet PM. Hepatic sinusoidal fibrosis induced by cholesterol and stilbestrol in the rabbit: 1. Morphology and inhibition of fibrogenesis by dipyridamole. Hepatology
26. Arroyo V, Jimenez W. Complications of cirrhosis. II. Renal and circulatory dysfunction. Lights and shadows in an important clinical problem. J Hepatol