Characterization of the pharmacokinetic-pharmacodynamic (PK-PD) relationship of muscle relaxants in clinical anaesthesia practice is important because it enables the design of infusion regimens and closed loop infusion systems. To obtain these data, the model proposed by Sheiner and colleagues  is frequently applied. This model is simple and suffices in many cases of reversible, receptor-mediated drug effects . In these cases the disappearance of drug effect parallels the decline of the plasma concentration of the drug.
Mivacurium has some special features interfering with the characterization of its PK-PD relationship by the model proposed by Sheiner and colleagues. For example, Laurin and colleagues  found that an effect compartment linked to the peripheral compartment was necessary to successfully model their PK-PD data of mivacurium. In contrast to other muscle relaxants, the rapid clearance and decay of the plasma concentration do not parallel the speed of recovery from mivacurium-induced neuromuscular block [4-7]. Usually a high plasma clearance is also accompanied by a rapid onset of effect [8,9]. However, mivacurium onset is longer than for several other muscle relaxants with a less rapid disappearance from the plasma, like rocuronium and rapacuronium [10,11].
Another challenge for successfully modelling the PK-PD data is that mivacurium is a mixture of three isomers: the trans-trans (57%), cis-trans (37%) and cis-cis (6%) isomers . The potency of these isomers is only known from animal studies (pers. comm. Maehr et al. Anesthesiology 1991; 75: A772; Belmont et al. Anesth Analg 1993; 76: S18) and showed an equipotency for the trans-trans and cis-trans isomers and a more than 10 times less potent cis-cis isomer.
The aim of our study was to offer more insight in the PK-PD relationship of the mivacurium isomers using different PK-PD models and the currently available information about mivacurium.
The study protocol was approved by the institutional Medical Ethics Committee. Written informed consent was obtained from each patient. Fourteen patients, American Society of Anesthesiologists (ASA) Grade I-II, aged between 25 and 55 yr participated in the study. Minor to intermediate, predominantly ear, nose and maxillofacial procedures were performed causing minimal blood loss. Patients with known liver or kidney disease, cardiovascular impairment or with neuromuscular disorders were excluded from this study. Also patients receiving medication known to interact with neuromuscular blocking agents (e.g. corticosteroids, magnesium, antibiotics, anticonvulsive medication, etc.) were not admitted to the protocol.
Anaesthesia and mivacurium administration
Premedication consisted of midazolam 7.5 mg orally on the morning of the surgery. Routine monitoring was applied according to the ASA standards for basic intraoperative monitoring. Anaesthesia was induced with thiopentone, 4-6 mg kg−1, and fentanyl, 1-2 μg kg−1. Anaesthesia was maintained with a mixture of nearly 100% oxygen and isoflurane in such a manner that an end-tidal concentration of 0.8% was achieved rapidly. After calibration of the neuromuscular monitor equipment following the induction of anaesthesia, mivacurium was administered at a rate of 30 μg min−1 kg−1, to obtain sufficient data during the onset phase of neuromuscular block . To achieve approximately 90% twitch depression, the infusion was stopped at 70% twitch depression. Endotracheal intubation was performed at maximum relaxation. Controlled ventilation was started to maintain the end-tidal partial pressure of carbon dioxide between 4.0 and 4.6 kPa. Anaesthesia was maintained with nitrous oxide/oxygen in a ratio 2/1 and isoflurane, aiming at an end-tidal concentration of 0.4%.
Neuromuscular function was monitored throughout the experiment by measuring the twitch tension of the adductor pollicis muscle until at least 90% twitch recovery from the mivacurium-induced neuromuscular block was obtained. The arm was wrapped up to keep the temperature of the hand above 32.5°C. Calibration of the measuring equipment was performed by supramaximal stimulation of the ulnar nerve at the wrist with single stimuli of 0.2 ms at a frequency of 0.1 Hz during 1 min. The resultant contraction was measured by a force displacement transducer with a preload between 250 and 400 g. The mean of the last six twitches before the onset of neuromuscular block was used to reset the twitch height at 100%.
Blood sampling procedure
The radial artery of the extremity that was not used for neuromuscular monitoring was cannulated. Arterial blood samples of 5 mL were drawn according to the following scheme: a blank sample before administration of mivacurium; four samples during onset of block; two samples at maximum block; four samples during recovery of block and one sample at 30, 45, 60, 75, 90, 120 and 150 min after the start of the mivacurium infusion. After sampling, the blood was immediately transferred into 10 mL EDTA tubes containing ecothiopate 0.25% 0.4 mL (phospholine iodide), an irreversible cholinesterase inhibitor, to prevent in vitro breakdown of mivacurium. Within 2 h of sampling the samples were centrifuged and the plasma was pipetted into clean tubes and stored at −18°C until analysis.
Analysis of plasma samples
High performance liquid chromatography (HPLC) was used for the quantitative determination of the plasma concentrations of the three stereoisomers: trans-trans, cis-trans and cis-cis mivacurium. The detection procedure was based on the methods described by Lacroix and colleagues  and Kleef and colleagues . A detailed description is given in the Appendix.
Plasma concentration-time data of the trans-trans, the cis-trans and the cis-cis isomer were analysed separately using the program MultiFit (written by J.H.P.). The analysis was performed with a dose ratio of 0.57:0.37:0.06 (trans-trans:cis-trans:cis-cis) for the isomers . The mean values and variance-covariance matrix of the pharmacokinetic population parameters of a one-compartment model (CL, V1), a two-compartment model (CL, V1; CL12, V2) and a three-compartment model (CL, V1; CL12, V2; CL13, V3), and the residual variance were estimated by an iterative Bayesian procedure as described in the literature [15,16], using the plasma concentration-time data, actual dose and duration of infusion of each patient participating in the study. Each parameter was normalized for body weight. A lognormal distribution for both the pharmacokinetic population parameters and the plasma concentration measurement errors was assumed. The correctness of the latter assumption was tested by visual inspection of the graphs of the residuals plotted against time and against concentration. Goodness-of-fit was evaluated from visual inspection of the measured and calculated data points and of the residuals plotted against time and against concentration. The choice between a one-, two- and three-compartment model was based on Akaike's Information Criterion . The lowest value of this statistical criterion indicates the best fitting model.
Two linked models were investigated. In the first model, the actual plasma concentration data were linked to effect data using an effect compartment of negligible volume connected to the plasma compartment and characterized by the rate constant of transport ke0[1,18]. The second model consisted of an interstitial space compartment that was interposed between the plasma and effect compartment in order to link plasma concentration and effect. The interstitial space compartment is connected to the plasma compartment by first-order kinetics, characterized by the rate constant kip, according to the following equation:
where Cis and C are the concentrations of the muscle relaxant in the interstitial and plasma compartment, respectively.
Similarly, the effect compartment is connected to the interstitial compartment by first-order kinetics, characterized by the rate constant kei, according to the following equation:
where Ce is the concentration of the muscle relaxant in the effect compartment.
If it is assumed that the volume of the interstitial compartment is negligible compared to the volume of the plasma compartment, and that the volume of the effect compartment is negligible compared to the volume of the interstitial compartment, both equations can be solved numerically, using the actual plasma concentration data for C.
Neuromuscular block was calculated using the sigmoid Emax model (Hill equation) with parameters EC50 (concentration in the effect compartment at 50% of maximum effect) and γ (slope of the concentration-effect relationship).
Assuming that the trans-trans and cis-trans isomers are equally potent, the sum of the measured plasma concentrations of the trans-trans and cis-trans isomers was used to drive the link model in the PK-PD analysis, using the method with a non-parametric kinetic model as described by Unadkat and colleagues . The cis-cis isomer was considered not to exert neuromuscular blocking activity in this study design, delivering a single dose of mivacurium as a short-term infusion to reach submaximal block .
Subsequently, four different scenarios were used to challenge both link models.
Scenario A: A range of different potency ratios of cis-trans and trans-trans isomers were tested, using the following formula:
The potency adjustments at each time point were incorporated in the analysis as follows:
where C is the effective mivacurium concentration used to drive the link model in the PK-PD analysis.
Scenario B: The possibility of contribution of the cis-cis isomer to neuromuscular block was investigated by assuming an EC-ratio (EC50(trans-trans)/EC50(cis-cis)) of 0.1, similar to the procedure described for the cis-trans and trans-trans isomers. This one-tenth potency of the cis-cis isomer was used according to existing literature .
Scenario C: The possibility of peripheral breakdown of mivacurium was taken into account using an extension of our interstitial space model (see above) as is visualized in Figure 1.
Scenario D: The model described by Donati and Meistelman , assuming 85% receptor occupation to obtain 50% neuromuscular block was tested with and without an interposed interstitial space compartment.
The twitch height-time data were analysed by an iterative Bayesian two-stage analysis, using the program PkPdFit (written by Proost JH). The goodness-of-fit was evaluated by visual inspection of the measured and calculated data, by the degree of minimization of the hysteresis loop, by the residual standard deviation and by the Akaike's Information Criterion value.
The patient characteristics data, infusion data and actual pharmacodynamic data are presented in Table 1.
The individual plasma concentration-time data are presented in Figure 2. The results of the pharmacokinetic analysis of the three mivacurium isomers are summarized in Table 2. For the trans-trans isomer, a two-compartment model fitted the data significantly better than a one- or a three-compartment model (Akaike's Information Criterion values 141, 32 and 119 for one-, two- and three-compartment models, respectively). For the cis-trans isomer, the performance of a one- or a two-compartment model was almost similar, but significantly better than a three-compartment model (Akaike's Information Criterion values 112, 114 and 270 for one-, two- and three-compartment models, respectively). For the cis-cis isomer, a two-compartment model fitted the data significantly better than a one- or a three-compartment model (Akaike's Information Criterion values 264, 92 and 153 for one-, two- and three-compartment models, respectively).
For each of the two-compartment models, the model with weight-normalized parameters, as presented in Table 2, had a lower Akaike's Information Criterion than that with non-normalized parameters. The residual coefficient of variation for the two-compartment model with weight-normalized parameters (Table 2) was 24%, 29% and 27% for the trans-trans, cis-trans and cis-cis isomer, respectively.
The results of the two applied link models are presented in Table 3. The systematic deviation between the actual and calculated time course of action with the plasma compartment link model is most obvious during the onset of block (Fig. 3a). The superiority of the interstitial space link model (Fig. 3b) is confirmed by the difference in Akaike's Information Criterion value, 8977 (interstitial space) and 11 559 (plasma compartment) and the residual standard deviations, 2.7% (interstitial space) and 5.4% (plasma compartment). Figure 4 shows the mivacurium plasma concentration profile in a typical patient and the calculated concentration profiles in the effect compartment and interstitial compartment.
Next we tested the four scenarios described in the Methods section.
Scenario A: The tested potency (EC50) ratios ranged from 1 (equally potent) to 0 (no activity from cis-trans isomer). The data for the interstitial space link model are shown in Table 3. Likewise, the interstitial space link model performed better than the plasma compartment link model using different potency ratios. According to the Akaike's Information Criterion value, the performance of the model was slightly better with a potency ratio of 0. The values of kip, kei and γ are hardly influenced by the variation in the potency ratio.
Scenario B: Assuming a cis-cis/trans-trans potency ratio of 0.1, the plasma compartment model performs slightly better (Akaike's Information Criterion value 11 531) and the interstitial space model performs worse (Akaike's Information Criterion value 9016) compared to the analysis assuming that the cis-cis isomer did not contribute to the neuromuscular block.
Scenario C: Breakdown of mivacurium in the interstitial compartment and/or the effect compartment was taken into account using an extension of the model as is illustrated in Figure 1. The parameters kis(m) and ke(m) were found to be unidentifiable, in which m stands for metabolism. Using fixed values for these parameters, these extended models resulted in an almost identical fit: the values for the sum of kip and kis(m), for the sum of kei and ke(m), and for γ were independent of the fixed values for kis(m) and ke(m). The values for EC50 decreased when kis(m) or ke(m) increased (data not shown).
Scenario D: Accounting for the 'buffered diffusion' hypothesis, using the model of Donati and Meistelman , did not result in a more satisfactory PK-PD model. It resulted in Akaike's Information Criterion values of 9159 (interstitial space) and 11 721 (plasma compartment) and parameter values approximating those obtained by the corresponding model that did not take into account receptor-binding sites (data not shown).
The concentration-effect relationship of mivacurium in this study is better described by an interstitial space link model than by the conventional plasma compartment link model.
Also Laurin and colleagues  failed to describe the concentration-effect relationship of mivacurium by a standard plasma compartment link model. To improve the fitting, they performed a parametric modelling analysis of mivacurium, linking an effect compartment to the peripheral compartment. Their peripheral compartment link model has several features in common with our interstitial space model, and their results are broadly comparable to our results. However, the interstitial model, interposing a pharmacokinetically indistinguishable compartment (i.e. the amount of drug in this compartment is assumed to be negligible) between the central and effect compartment (Fig. 1) can be performed non-parametrically, i.e. using the plasma concentration measurements directly, thus avoiding any influence of the pharmacokinetic modelling on the PK-PD modelling. Therefore, this model might be a more elegant and flexible solution for the description of the fate of mivacurium.
In comparison with other non-depolarizing muscle relaxants mivacurium has an apparently aberrant pharmacological behaviour. Firstly, a rapid clearance can contribute to a short onset time, like for example succinylcholine [8,9]. However, the time to peak effect of mivacurium is not short compared to other muscle relaxants [10,11]. Secondly, the metabolism by plasma cholinesterase is so rapid (between 70% and 88% of the rate of succinylcholine ) that the two active isomers, i.e. trans-trans and cis-trans, have virtually disappeared from plasma before complete recovery is obtained (Figs 3 and 4). Previously, the rate of recovery from mivacurium-induced neuromuscular block was found to be independent of the rate of its plasma concentration decay in patients with normal plasma cholinesterase activity . The delay between clearance and recovery must lie within the sequence of events of the rate of binding and unbinding to receptor-binding sites, diffusion to and from the receptor site and diffusion of mivacurium from the interstitium into the plasma.
Based on the results of their isolated forearm experiments, Feldman and colleagues speculated about this apparent discordance between rapid plasma clearance and relatively slow recovery [20,21]. They concluded that mivacurium might be retained in the biophase and/or at redistribution sites. In additional isolated arm experiments with a different approach, they tried to demonstrate that recovery of block does not follow immediately, in spite of rapidly declining plasma concentrations of non-depolarizing muscle relaxants. According to Feldman biophase-binding sites and the dissociation rate from these sites may be responsible for the delayed time course of action of non-depolarizing muscle relaxants . However, redistribution phenomena within the recirculated arm may also play a role in the ongoing onset and relatively slow recovery. Earlier, Hull recognized that there may be one ore more first-order lag-factors in the relationship between plasma concentration and effect and suggested already the role of 'buffered diffusion' . In a model proposed and described by Donati and Meistelman , the overcapacity of receptor-binding sites leading to the so-called 'buffered diffusion' may explain the inverse relationship between potency and onset. For mivacurium, a relatively potent non-depolarizing muscle relaxant with an EC50 of 98 μgL−1 (9.10−8M), this effect of buffering of the drug molecules would be pronounced and cause a delay in the onset of block. However, in the present study PK-PD modelling that allows 'buffering' of drug molecules, did not result in a better fit (Scenario D). The parameter values approximated the results obtained by the application of the traditional plasma compartment link model. This finding does not support the 'buffered diffusion' hypothesis.
In an elegant study under steady state infusion conditions Ezzine and colleagues  demonstrated the existence of peripheral elimination of mivacurium isomers. This peripheral metabolism of mivacurium is also a factor to be considered in our search for a proper PK-PD model. However, modelling taking into account the influence of possible peripheral metabolism of mivacurium (Scenario C) yielded no improvement of the fittings and resulted in smaller values for the kip, kei and EC50 than shown in Table 3.
Nigrovic and colleagues described a more anatomical and physiological model with an interposed, interstitial space compartment and used it to explain the inverse correlation between potency and speed of onset of muscle relaxants . This model was tested by Beaufort and colleagues , who found that the plasma concentration data of various muscle relaxants fitted satisfactorily with this adapted model after rapid and early arterial sampling. Also, in our study the interstitial space link model, based on the model of Nigrovic and colleagues , was able to model the concentration-effect relationship of mivacurium satisfactorily.
The plasma compartment link model was not appropriate in our study and in the study of Beaufort and colleagues , whereas it has been used satisfactorily in many studies with various muscle relaxants. It is likely that the plasma compartment link model allows an adequate description of the effect compartment concentration in cases where the plasma concentration is not changing rapidly. In that case the capillary and receptor site can be modelled by a single step. However, if plasma concentration changes rapidly, as for mivacurium or during the 1st minute in the study of Beaufort and colleagues , the plasma compartment link model fails and inclusion of an interstitial compartment may allow a better description of the concentration profile in the effect compartment (Fig. 4).
The separation of the rate constants kei and kip is somewhat artificial because the values of both rate constants can be interchanged without influencing the outcome. In the flow-volume model, Nigrovic reasoned and calculated a flow/volume ratio of 0.153 min−1 with the receptors as an integral part of the interstitial space [27,28]. The simulated data following this approach correlate well with the observed onset times for vecuronium. One could argue that this rate constant of 0.153 min−1 must be the same for all muscle relaxants, assuming that they do not influence their own blood flow. If this is true, our values for the rate constants kip and kei in this study should be interchanged. This problem of identifiability of rate constants makes the interpretation of the data hazardous. However, it still can be concluded that the transport of mivacurium between plasma and effect site and plasma cannot be described adequately by a single rate constant, suggesting that recovery is retarded by buffering of mivacurium between plasma and effect site.
To date, no information is available regarding the potency ratios of the three mivacurium isomers in human beings. Animal research revealed that the trans-trans and cis-trans isomers are equipotent and that the cis-cis isomer is approximately 10 times less potent. In addition, Lien and colleagues  demonstrated that the cis-cis isomer contributes for less than 5% to the neuromuscular blocking effect after a mivacurium infusion, despite the markedly lower clearance of this isomer. This finding has been confirmed in other studies [29,30]. In our study, PK-PD analysis taking into account a cis-cis/trans-trans potency ratio of 0.1 hardly influences the results (Scenario B) and did not suggest a significant effect from the cis-cis isomer. Therefore, we evaluated a range of potency ratios of trans-trans and cis-trans isomers, neglecting neuromuscular blocking activity of the cis-cis isomer. A potency of 0 for the cis-trans isomer slightly improved the performance of the interstitial space compartment link model (Scenario A). However, it is very unlikely that the potency of the cis-trans isomer is 0 while animal research revealed an equipotency for both isomers. More importantly, using those different potency ratios seemed to leave the PK-PD characteristics (kip, kei, γ) relatively unchanged. Still, it might be interesting to investigate the concentration-response relationship and potency ratios of the single isomers administered separately in human beings.
In summary, the PK-PD relationship of mivacurium is well described by interposing an interstitial space compartment between the plasma and effect compartment. Based on earlier research and on our results with alternative models, we conclude that the delay in the concentration-effect relationship of mivacurium may be explained by the relatively slow transport between plasma and effect compartment via the interstitium.
Reference compounds were kindly donated by GlaxoSmithKline BV, Zeist, The Netherlands. J. Roggeveld, who performed the analytical work, was funded by Organon International BV, PO Box 20, 5240 BH Oss, The Netherlands. The authors like to thank Ursula Kleef BSc, analytical chemist; Douglas Eleveld MEng, research fellow; Ann de Haes MD, research fellow; Marjon Dijkema MD, resident in anaesthesiology; Tessa Dijkstra, student in pharmacy; Ria Carpay, office manager and Ton Beaufort MD PhD, staff anaesthesiologist for their support during the conduct of this study.
Quantitative determination of the plasma concentrations of the three isomers of mivacurium
The three stereoisomers of mivacurium (GlaxoSmithKline BV, Zeist, The Netherlands) were isolated from the plasma by a liquid-liquid extraction with potassium iodide as a counter-ion. To 1000 μL plasma, 250 ng of the internal standard Org 22175 (Organon Laboratories Ltd., Newhouse, Scotland) and 100 μL of a 5% potassium iodide solution in water were added. The mixture was shaken for 30 s on a vortex mixer with 6 mL of dichloromethane. After centrifugation for 10 min at 740g, the upper phase was discarded and the organic layer was evaporated to dryness at room temperature under a stream of nitrogen. The residue was dissolved in 125 μL of mobile phase and 100 μL was introduced onto the HPLC system. Separation of the isomers was performed on a Hypersyl BDS C 18 column (Alltech, Emmen, The Netherlands) combined with a Nova-Pak® C18 Guard-Pak (Waters Corp, Milford, MA, UK). The column was maintained at 26°C. The mobile phase was acetonitrile-water (1:2 v/v) containing 0.005 M octane sulphonic acid. The flow rate was 0.7 mL min−1. After separation, the eluent was extracted with dichloroethane, and the organic phase was led to a fluorometric detector operating at a wavelength of 220 nm (excitation) and 320 nm (emission). Calibration graphs were constructed by applying linear regression on the logarithmically transformed data of peak area ratio analyte/internal standard (response ratio) and the amount of analyte, and were linear in a range of at least 5-250 ng for each isomer. The intra-day variability was 9% (mean value) for each compound, as indicated by the coefficients of variation in the range of 5-250 ng. The mean absolute deviation of the accuracy samples 15, 100 and 1000 ng mL−1, was 2.4%, 2.9% and 1.9% for the trans-trans, cis-trans and cis-cis isomers, respectively. The lower limit of quantification, defined as the minimum concentration that could be detected with an accuracy and precision better than 15%, was 5 ng mL−1 for each isomer.
1. Sheiner LB, Stanski DR, Vozeh S, Miller RD, Ham J. Simultaneous modeling of pharmacokinetics and pharmacodynamics: application to d
-tubocurarine. Clin Pharmacol Ther
2. Venitz J. Pharmacokinetic-pharmacodynamic modeling of reversible drug effects. In: Derendorf H, Hochhaus G, eds. Handbook of Pharmacokinetic/Pharmacodynamic Correlation.
Boca Raton, FL: CRC Press Inc, 1995: 1-34.
3. Laurin J, Donati F, Nekka F, Varin F. Peripheral link model as an alternative for pharmacokinetic-pharmacodynamic modeling of drugs having a very short elimination half-life. J Pharmacokinet Pharmacodyn
4. Savarese JJ, Ali HH, Basta SJ, et al.
The clinical neuromuscular pharmacology of mivacurium chloride (BW B1090U). Anesthesiology
5. Lien CA, Schmith VD, Embree PB, Belmont MR, Wargin WA, Savarese JJ. The pharmacokinetics and pharmacodynamics of the stereoisomers of mivacurium in patients receiving nitrous oxide/opioid/barbiturate anesthesia. Anesthesiology
6. Østergaard D, Rasmussen SN, Viby-Mogensen J, Pedersen NA, Boysen R. The influence of drug-induced low plasma cholinesterase activity on the pharmacokinetics and pharmacodynamics of mivacurium. Anesthesiology
7. Nijs N, Duvaldestin P, Slavov V, Dhonneur G. Is the recovery profile of mivacurium independent of the rate of decay of its plasma concentration in patients with normal plasma cholinesterase activity? Acta Anaesthesiol Scand
8. Proost JH, Wierda JMKH, Meijer DKF. An extended pharmacokinetic/pharmacodynamic model describing quantitatively the influence of plasma protein binding, tissue binding, and receptor binding on the potency and time course of action of drugs. J Pharmacokinet Biopharm
9. Beaufort TM, Nigrovic V, Proost JH, Houwertjes MC, Wierda JMKH. Inhibition of the enzymic degradation of suxamethonium and mivacurium increases the onset time of submaximal neuromuscular block. Anesthesiology
10. Van den Broek L, Wierda JMKH, Smeulers NJ, van Santen GJ, Leclerq MGL, Hennis PJ. Clinical pharmacology of rocuronium (Org 9426): study of the time course of action, dose requirement, reversibility, and pharmacokinetics. J Clin Anesth
11. Schiere S, Proost JH, Schuringa M, Wierda JMKH. Pharmacokinetics and pharmacokinetic-dynamic relationship between rapacuronium (Org 9487) and its 3-desacetyl metabolite (Org 9488). Anesth Analg
12. Lacroix M, Tu TM, Donati F, Varin F. High-performance liquid chromatographic assays with fluorometric detection for mivacurium isomers and their metabolites in human plasma. J Chromatogr B
13. Viby-Mogensen J, Ostergaard D, Donati F, et al.
Pharmocokinetic studies of neuromuscular blocking agents: good clinical research practice (GCRP). Acta Anaesthesiol Scand
14. Kleef UW, Proost JH, Roggeveld J, Wierda JMKH. Determination of rocuronium and its putative metabolites in body fluids and tissue homogenates. J Chromatogr B
15. Mentre F, Gomeni R. A two-step iterative algorithm for estimation in nonlinear mixed-effect models with an evaluation in population pharmacokinetics. J Biopharm Stat
16. Bennett JE, Wakefield JC. A comparison of a Bayesian population method with two methods as implemented in commercially available software. J Pharmacokinet Biopharm
17. Akaike H. An information criterion. Math Sci
18. Unadkat JD, Bartha F, Sheiner LB. Simultaneous modeling of pharmacokinetics and pharmacodynamics with non-parametric kinetic and dynamic models. Clin Pharmacol Ther
19. Donati F, Meistelman C. A kinetic-dynamic model to explain the relationship between high potency and slow onset time for neuromuscular blocking drugs. J Pharmacokinet Biopharm
20. Campkin NTA, Hood JR, Feldman SA. Recovery of mivacurium and doxacurium versus vecuronium in the isolated forearm. Anaesthesia
21. Feldman SA, Hood JR, Campkin NTA, Rehm S. Sensitivity to second dose of mivacurium. Anaesthesia
22. Feldman S. Biophase binding: its effect on recovery from non-depolarising neuromuscular block. Anaesth Pharmacol Rev
23. Hull CJ. Pharmacodynamics of non-depolarizing neuromuscular blocking agents. Br J Anaesth
24. Ezzine S, Donati F, Varin F. Mivacurium arteriovenous gradient during steady state infusion in anesthetized patients. Anesthesiology
25. Nigrovic V, Banoub A, Diefenbach C, Mellinghof H, Buzello W. Onset of the neuromuscular block simulated in an anatomical model. Br J Clin Pharmacol
26. Beaufort TM, Nigrovic V, Proost JH, Houwertjes MC, Kleef UW, Wierda JMKH. Do plasma concentrations obtained from early arterial blood sampling improve pharmacokinetic/pharmacodynamic modeling? J Pharmacokinet Biopharm
27. Nigrovic V, Banoub M. Onset of the nondepolarizing neuromuscular block in humans: quantitative aspects. Anesth Analg
28. Nigrovic V. Neuromuscular block by vecuronium: simulation with a flow-volume model. Eur J Anaesthesiol
29. Østergaard D, Viby-Mogensen J, Pedersen NA, Holm H, Skovgaard LT. Pharmacokinetics and pharmacodynamics of mivacurium in young adult and elderly patients. Acta Anaesthesiol Scand
30. Ledowski T, Wulf H, Ahrens K, et al.
Neuromuscular block and relative concentrations of mivacurium isomers under isoflurane versus propofol anaesthesia. Eur J Anaesthesiol
Keywords:© 2004 European Academy of Anaesthesiology
NEUROMUSCULAR NON-DEPOLARIZING AGENTS, mivacurium; PHARMACOLOGY, pharmacokinetics, pharmacodynamics; CHOLINESTERASES, pharmacokinetics