Previous studies report the pharmacokinetics of mivacurium isomers after an infusion using venous blood sampling. Although the extent of the mivacurium arterial-venous gradient is not known, the sampling site is likely to influence mivacurium pharmacokinetic parameters because the drug is rapidly metabolized as it traverses the circulation. The objectives of this study were (1) to determine the pharmacokinetics of mivacurium isomers in healthy persons after intravenous bolus administration using intensive arterial blood sampling, and (2) to characterize the formation and elimination of mivacurium metabolites in human plasma.
Eight persons classified as American Society of Anesthesiologists physical status 1 or 2 who were scheduled to undergo elective surgery under balanced anesthesia received 0.15 mg/kg mivacurium chloride as an intravenous bolus. Arterial blood samples were collected every 10 s during the first 2 min and at frequent intervals for 4 h thereafter. Plasma concentrations of mivacurium isomers and their metabolites were determined by two stereoselective high-performance liquid chromatographic methods coupled with fluorometric detection and noncompartmental pharmacokinetic parameters.
Mean elimination half-lives of the trans-trans, cis-trans, and cis-cis isomers were 2.4, 2, and 28.5 min, respectively, with corresponding mean plasma clearances of 29.2, 45.7, and 6.7 ml.min 1.kg-1. The volumes of distribution at steady state of the trans-trans, cis-trans, and cis-cis isomers were 0.047, 0.054, and 0.189 l/kg, respectively. Plasma concentrations of monoester and alcohol metabolites peaked 25 s (median) after mivacurium injection, with half-lives in the range of 90 min, except for the cis alcohol metabolite, which was only negligibly and transiently formed.
Substantial hydrolysis of mivacurium isomers by cholinesterases was confirmed by the rapid appearance of mivacurium metabolites in plasma. The intensive arterial sampling proved to be critical for the trans-trans and cis-trans isomers because the area under the curve between 0 and 2 min accounted for 75% and 86% of the total, respectively.
Mivacurium chloride is a new short-acting nondepolarizing neuromuscular blocking agent member of the benzylisoquinolinium family. It consists of a mixture of three stereoisomers: the two most active and equipotent are the trans-trans and cis-trans isomers (57% and 36% w/w, respectively) whereas the cis-cis isomer (6% w/w) has only one tenth the activity of the others in cats and monkeys. ,* The contribution of each isomer to the overall activity of mivacurium in humans is not yet known.
Rapid hydrolysis of the trans-trans and cis-trans isomers by plasma cholinesterases explains the short half-lives of these active isomers and the brief clinical duration of action of mivacurium chloride. [2,3]Renal excretion appears to be a minor elimination pathway; approximately 7% of an administered dose of mivacurium is recovered unchanged in human urine. Hydrolysis of mivacurium by plasma cholinesterases produces two types of metabolites, namely the quaternary amino alcohols (cis and trans) and quaternary monoesters (cis and trans), which account, respectively, for 44% and 46% of the dose in urine. It is not yet known if these metabolites are active.
The lack of a stereoselective analytical method has limited the availability of pharmacokinetic data for mivacurium isomers. Furthermore, although rapid hydrolysis of mivacurium by plasma cholinesterases is said to explain the brief duration of action of this relaxant, to our knowledge no supporting data concerning the plasma concentration versus the time profile of mivacurium metabolites have yet been published. A stereoselective high-performance liquid chromatographic assay with fluorometric detection for mivacurium isomers in human plasma was recently reported and was used in a pharmacokinetic study by Lien et al. to determine the pharmacokinetic and pharmacodynamic parameters of mivacurium isomers after a two-step infusion. Subsequently Head-Rapson et al. [5,6]reported the pharmacokinetic parameters of mivacurium isomers in healthy and cirrhotic patients after a 10-min infusion. These results, obtained from venous blood measurements, represent the only pharmacokinetic data available for mivacurium isomers. Although the extent of the arterial-venous gradient in the case of mivacurium is not known, the choice of the sampling site (arterial versus venous) should have an important influence on the determination of mivacurium pharmacokinetics because the drug is rapidly eliminated as it traverses the circulation.
Consequently, we began this study to describe the pharmacokinetics of mivacurium isomers after an intravenous bolus injection. The arterial concentration was measured because it is thought to be more representative of what is actually delivered to the site of action for many drugs. More specifically, the atracurium plasma concentration-effect relation differed substantially when arterial rather than venous blood was used for the same patient. To describe adequately the early concentration-time curve of mivacurium isomers and the rapid formation of their metabolites, frequent blood sampling during the first 2 min after injection was performed. In this study we also tried to evaluate the importance of this intensive blood sampling approach, which has been used successfully in our laboratory for vecuronium and atracurium pharmacokinetic studies, [9,10]on the noncompartmental pharmacokinetic parameters of mivacurium isomers. Finally, this study investigates the formation and elimination of mivacurium metabolites.
Materials and Methods
The study protocol was approved by the Royal Victoria Hospital Ethics Committee, and all participants gave written informed consent before entering the study. Eight persons classified as American Society of Anesthesiologists physical status 1 or 2 who were 18 to 40 yr old and scheduled to undergo elective surgery in which the insertion of an arterial cannula was indicated participated. Patients showing any evidence of clinically significant psychiatric, neurological, neuromuscular, pulmonary, or cardiovascular disease as well as any clinically significant impairment of hepatic or renal function were excluded. Similarly, persons taking medications known or suspected to affect neuromuscular function were also excluded.
Patients were premedicated, if necessary, with either lorazepam (0.05 mg/kg given intramuscularly), diazepam (0.1–0.2 mg/kg given orally), or midazolam (0.020.10 mg/kg given intramuscularly or intravenously). A radial artery was cannulated using local anesthesia before general anesthesia was given. General anesthesia was induced with 2 to 10 mg/kg thiopental and 0.5 to 10 micro gram/kg fentanyl administered intravenously. Anesthesia was maintained with 50–70% nitrous oxide (N2O) in oxygen administered using a face mask. Additional doses of fentanyl or thiopental were given, as needed, to maintain an adequate level of anesthesia. Tracheal intubation occurred at maximum blockade after injection of 0.15 mg/kg mivacurium, after completion of the intensive sampling period. The patient's lungs were then ventilated mechanically. Minute volume was adjusted to keep end-tidal carbon dioxide between 30 and 35 mmHg. Isoflurane (0.5–1.0% end-tidal) was added after recovery from neuromuscular blockade. During surgery, blood pressure, heart rate, electrocardiogram, body temperature, and respiration were monitored continuously. Patients were closely observed for signs of histamine release and other side effects.
The arm without the radial artery cannula was positioned, secured, and used to monitor neuromuscular function. The ulnar nerve was stimulated in single supramaximal twitches (0.2 ms at 0.1 Hz) via two surface electrodes. The resultant force of contraction of the adductor pollicis was measured using a force transducer (Grass FT-10; Grass Instrument Co., Quincy, MA), and the transducer output was recorded on a polygraph. After anesthesia was stabilized, each patient received an intravenous bolus dose of 0.15 mg/kg mivacurium chloride (Mivacron, Burroughs Wellcome, Research Triangle Park, NC). If and when additional muscle relaxation was needed (in six of eight patients), either vecuronium or pancuronium, neither of which would interfere with the assay, was administered. Neuromuscular block was not routinely reversed in this study, but reversal was necessary in two patients who received a mixture of edrophonium (0.25–1 mg/kg) and atropine (0.005–0.02 mg/kg). These reversal agents were given 3 to 4 h after the bolus administration of mivacurium, thereby avoiding any interference with mivacurium active isomers.
Arterial blood samples were collected in heparin-prepared Vacutainer tubes (Becton Dickinson, Canada) containing 0.1 mg echothiophate iodide (a plasma cholinesterase inhibitor). The arterial cannula was connected to a three-way stopcock via a small extension tubing, and dead space was minimized to 0.6 ml. The blood flow out of the cannula was approximately 3 ml every 10 s at normal arterial pressure. The first arterial sample (5 ml) was collected before mivacurium chloride was administered. Immediately after mivacurium was injected, the stopcock was opened and blood was allowed to flow freely into 12 Vacutainer tubes during the first 2 min. Tubes were changed every 10 s and the time assigned to these samples was the midpoint of the 10-s interval during which the sample was drawn. Samples (3 ml) were collected regularly at 3, 4, 5, 7, 10, 12, 15, 20, 30, 45, 60, 90, 120, 180, and 240 min. Blood samples were kept on ice, centrifuged at 1,600g for 5 min, and the collected plasma was frozen in dry ice and stored at -20 degrees C until high-performance liquid chromatographic analysis.
Plasma concentrations of mivacurium isomers and their metabolites were measured using two specific high-performance liquid chromatographic assays coupled with fluorometric detection developed in our laboratory. A novel solid-phase extraction procedure allowed good recovery of mivacurium isomers (mean, 98%) and their monoester metabolites (mean, 83%), whereas the alcohol metabolites were analyzed after direct precipitation of plasma proteins. For all mivacurium isomers and metabolites, these two assays proved to be sensitive (lower limit of quantification: 3.9–15.6 ng/ml), reproducible (coefficient of variation, less or equal to 15%), accurate (> 94%), and linear for concentrations in the therapeutic ranges. For concentrations greater than the standard curve upper limits, samples were diluted with blank plasma before plasma extraction.
Two sets of pharmacokinetic parameters for mivacurium isomers were derived for each patient: One included all 12 blood samples collected during the first 2 min (intensive blood sampling), whereas one included only the 1- and 2-min samples (limited blood sampling). After 2 min, the blood sampling times were the same in both sets. A noncompartmental approach, which allowed the inclusion of all plasma concentration points without any previous assumption about the pharmacokinetic model, was used to described the pharmacokinetics of mivacurium isomers and their metabolites, with the understanding that these products follow linear pharmacokinetics. The plasma area-under-the curve concentration-time curve (AUC) and the area-under-the first moment curve (AUMC) were calculated according to the trapezoidal rule for all mivacurium isomers and metabolites. The elimination rate constant (kel) was obtained from linear regression through the last four or five data points. Subsequently, the elimination half-lives of mivacurium isomers and their metabolites were derived from the Equation t1/2 = 0.693/kel. The mean residence time (MRT), which represents the time needed to eliminate 63.2% of the dose, was obtained for all mivacurium isomers and metabolites from the ratio AUMC/AUC. The plasma clearance of each mivacurium isomer was obtained by dividing the dose administered by the AUC; this could not be done for mivacurium metabolites because urine measurements were not done. The dose of each isomer administered was calculated by using the percentage of each isomer present in the mixture as described in the published product monograph, namely 57%, 36%, and 6%, respectively, for the trans-trans, cis-trans, and cis-cis isomers. The actual relative proportions of each isomer present in the administered product were calculated and found to be in accordance with the published data, with percentages of (mean +/- SD) 59.8 +/- 0.3, 34.8 +/- 0.3, and 5.4 +/- 0.3% for the trans-trans, cis-trans, and cis-cis isomer, respectively. The volume of distribution at steady state (Vdss) was derived from the product of MRT and plasma clearance, and the volume of distribution by area (Vd beta) was determined by dividing plasma clearance by kel. Also noted for all mivacurium isomers and metabolites were the maximum plasma concentration (Cmax) and the time from administration of mivacurium chloride to Cmax.
Results are presented as means +/- SEM. The pharmacokinetic parameters derived from intensive blood sampling were compared with those obtained after limited blood sampling (1 and 2 min), which constitutes the more traditional approach. Because each patient was his or her own control, pharmacokinetic parameters were compared using the Student's t test for paired data, with an alpha level of statistical significance of 0.05.
Intensive Blood Sampling Pharmacokinetics
Eight ASA physical status 1 or 2 (4:4) patients participated in this trial (six men, two women), with a mean (+/-SEM) age and weight of 33 +/- 2 yr and 70 +/- 6 kg, respectively. Surgical procedures performed included laminectomy, mandibular, and abdominal procedures. Mivacurium maximum block was (means +/- SEM) 95 +/- 2%, with an onset to maximum block of 5.6 +/- 0.4 min. Recovery of neuromuscular function to 25%, 50%, and 75% of baseline values was observed after 14 +/- 2 min, 17 +/- 2 min, and 21 +/-2 min, respectively. Figure 1shows the plasma concentration-time curves for all three mivacurium isomers. After the intravenous bolus administration of 0.15 mg/kg mivacurium chloride, Cmax of 4,486 +/- 403 ng/ml, 2,198 +/- 208 ng/ml, and 504 +/- 40 ng/ml for the trans-trans, cis-trans, and cis-cis isomers, respectively, were observed after 25 s (n = 6) and 35 s (n = 2)(Figure 1). Plasma concentrations of the trans-trans and cis-trans isomers declined rapidly and could not be detected in plasma 10–20 min after injection. However, plasma concentrations of the cis-cis isomer decreased more slowly and could still be detected 45 min after injection in seven patients (mean, 7.5 ng/ml).
(Table 1) summarizes the noncompartmental pharmacokinetic parameters of mivacurium isomers after intensive arterial sampling. The trans-trans and cis-trans isomers behaved similarly with short half-lives of less than 2.5 min and very rapid plasma clearances. Both isomers also showed small volumes of distribution (Vdss and Vd beta) consistent with an extracellular distribution of this large ionized molecule. In contrast, a much slower plasma clearance and a longer half-life were found for the cis-cis isomer, as were larger volumes of distribution.
Intensive versus Limited Sampling Pharmacokinetics
The peak arterial concentrations just noted could not be detected by standard limited blood sampling scheduled at 1 and 2 min after dosing. Mean peak arterial concentrations for mivacurium trans-trans, cis-trans, and cis-cis after limited sampling corresponded to 950 +/- 78 ng/ml, 405 +/- 39 ng/ml, and 137 +/- 10 ng/ml, which were significantly lower (P < 0.05) than those obtained with intensive sampling. Using the limited sampling data, the AUC covering the first 2 min (AUC0–2) was decreased by 51%, 55%, and 42% for the trans-trans, cis-trans, and cis-cis isomers, respectively, compared with the intensive sampling results (Table 1(B)). This led to an important underestimation of the AUC0-inf, which was decreased by 38% and 45% for the trans-trans and cis-trans isomers and by 10.5% for the cis-cis isomer. Consequently, an important overestimation of the pharmacokinetic parameters (MRT, plasma clearance, Vdss, and Vd beta) was observed for the trans-trans and cis-trans isomers, as summarized in Table 1. For the trans-trans isomer, the MRT was increased by 47%, the plasma clearance by 66%, the Vdss by 134%, and the Vd beta by 60%. In the case of the cis-trans isomer, the MRT was increased by 66%, the plasma clearance by 90%, the Vdss by 215%, and the Vd beta by 92%. Although significant, the effect of limited sampling was less pronounced on the pharmacokinetic parameters of the more slowly eliminated cis-cis isomer, in which the MRT was overestimated by 10%, the plasma clearance by 10%, the Vdss by 22%, and the Vd beta by 7%.
Pharmacokinetic Parameters of Mivacurium Metabolites
Rapid hydrolysis of the trans-trans and cis-trans isomers resulted in the formation of monoester and alcohol metabolites (Figure 2). Mean Cmax of 140 +/- 9 ng/ml and 1,027 +/- 121 ng/ml were noted for the cis and trans alcohols, respectively, at 15 s (n = 1), 25 s (n = 6), and 35 s (n = 1), whereas mean Cmax of 658 +/- 95 ng/ml and 624 +/- 85 ng/ml were observed at 25 s (n = 5) and 35 s (n = 3) for the cis and trans monoesters, respectively. Subsequently these metabolites were eliminated slowly, with one exception being the cis alcohol, which was only negligibly and transiently present, with a half-life less than 2 min (Table 2). The trans alcohol, cis monoester, and trans monoester behaved similarly, with half-lives of approximately 90 min.
Using an intensive arterial sampling method, we derived the pharmacokinetic parameters of mivacurium after an intravenous bolus dose. With such a rapidly eliminated drug as mivacurium, intensive sampling becomes essential because the first 2 min greatly contribute to the total plasma concentration-time profile.
This study was done to characterize the pharmacokinetics of mivacurium isomers after an intravenous bolus injection. The early pharmacokinetic events after a bolus dose, characterized by a short period in which no drug can be detected in plasma (5–10 s) followed by a sharp increase in observed drug concentrations (until 25 to 35 s), are believed to be governed by circulatory factors. To properly describe the early time profile of each mivacurium isomer, an intensive sampling procedure was required.
A similar method, used in our laboratory to characterize the pharmacokinetics and pharmacodynamics of an intravenous bolus dose of vecuronium, showed that plasma concentrations of vecuronium measured during the first minute after injection were critical for adequate noncompartmental pharmacokinetic-pharmacodynamic modeling. Ignoring these early concentration points led to an overestimation of vecuronium pharmacokinetic parameters (Vd) and to a significant increase in the pharmacodynamic parameter, keo. Similar findings were subsequently reported when this approach was used to model the concentration-effect relation of atracurium after the administration of two bolus doses of 0.5 mg/kg given 1 h apart. 
Before 1992, the lack of availability of stereoselective assays to determine mivacurium isomers in human plasma limited the information on the selective pharmacokinetics of this new drug. Brown et al. reported on such a stereoselective method, which was used subsequently in a study by Lien et al. to describe the pharmacokinetic parameters of mivacurium isomers using venous blood measurements and a two-step infusion. However, the analytical method used could not measure mivacurium metabolites because these were eliminated in the washing step of the extraction procedure. Also using a stereoselective high-performance liquid chromatographic method, Head-Rapson et al. derived the pharmacokinetics of mivacurium isomers after a short infusion or a multiple-step infusion. We used two stereoselective assays developed in our laboratory to determine the plasma concentration-time profiles of mivacurium isomers and their metabolites in human plasma. 
The pharmacokinetic parameters (plasma clearance, Vdss, Vd beta) of the trans-trans and cis-trans isomers derived from limited blood sampling (Table 1) correspond with those presented in previous studies in which mivacurium was given as an infusion. [2,5,6]The only exception is the Vdss of the trans-trans isomer, which appears 45% smaller than the one reported by Head-Rapson et al. [5,6]The half-lives of the trans-trans and cis-trans isomers, which are also independent of the sampling method, correspond with those reported earlier, [2,5,6]with values surrounding 2.2 min. If we consider the mean residence half-life of each isomer, that is the product of 0.693 and MRT, the following values are obtained: 1.18 min for the trans-trans and 0.83 min for the cis-trans. These estimates give a better approximation of the effective half-life of these isomers.
For the cis-cis isomer, the volumes of distribution obtained from limited sampling are consistent with those previously published, but the plasma clearance of this isomer is approximately 50% faster than that reported previously. This last observation accounts for the 45% decrease in half-life observed in this study for the cis-cis isomer.
The pharmacokinetic parameters of the trans-trans and cis-trans isomers derived from intensive sampling differ significantly, however, from previously published results. [2,5]Plasma clearance for both isomers are decreased by 50% and the volumes of distribution are reduced by 40–75%. Similarly, the volume of distribution of the cis-cis isomer is approximately 25% smaller. Interestingly, the plasma clearance of the cis-cis isomer derived from intensive sampling corresponds more closely with previously published data, with differences of approximately 38% rather than the 50% difference with limited sampling noted previously.
Compared with previous mivacurium studies [2,5,6]and with our limited sampling data, the Vdss derived from intensive sampling for the trans-trans and cis-trans mivacurium isomers are small, approaching plasma volumes. Methodologic factors, such as sampling frequency or site, could account for this phenomenon.
With a rapidly eliminated drug such as mivacurium, intensive sampling is critical. Indeed, for the trans-trans and cis-trans isomers, the AUC obtained between in the first 2 min represents 75% and 86% of the total AUC, respectively, whereas that of the cis-cis isomer, which is more slowly eliminated, represents only 22% of the total AUC. Thus by limiting sampling to 1 and 2 min after dosing, an important portion of the total AUC is omitted for the trans-trans and cis-trans isomers, which leads to an important underestimation of the total AUC and subsequently to an important overestimation of the plasma clearances and volumes of distributions, which, as noted previously, correspond closely with previously published data. Understandably, this effect is less pronounced with the cis-cis isomer. This observation reinforces the importance of the intravascular mixing phase in the overall determination of the pharmacokinetics of rapidly eliminated isomers.
Because compartmental analysis with limited sampling is mostly used in studies, it would have been interesting to compare it with noncompartmental analysis of an intense sampling data set. Such comparisons were made previously after an intravenous bolus dose of vecuronium, and pharmacokinetic parameters were found to be similar using both approaches. Similarly, using compartmental analysis (without peripheral elimination) of plasma concentrations obtained after 1 min, the AUC0-infderived for the trans-trans isomer of mivacurium (mean, 2,669 ng [center dot] min-1[center dot] ml-1) were similar to those we report here for the noncompartmental analysis of the intense sampling (mean, 2,932 ng [center dot] min-1[center dot] ml-1). However, when compartmental analysis was applied to plasma concentrations obtained after 30 s (Cmax), a significant increase in AUCs was observed (mean, 6,191 ng [center dot] min-1[center dot] ml-1).
The arteriovenous difference may also contribute to this discrepancy, especially for a drug such as mivacurium, which is eliminated as it traverses the circulation. Indeed, this would suggest that at steady state, venous concentrations of mivacurium would be significantly lower than corresponding arterial concentrations, and thus by using venous measurements in the determination of mivacurium pharmacokinetics, [2,5,6]we could expect to obtain lower AUCs and subsequently larger clearance and Vd values. In a previous study in which atracurium was given as an intravenous bolus dose, the AUC formed within the first 20 min, using a limited sampling schedule, was 25% smaller if calculated from venous rather than arterial concentrations. Consequently, it is believed that the differences observed between the pharmacokinetic parameters derived using the limited sampling approach and those presented by Lien et al. and Head-Rapson et al. [5,6]can be explained primarily by using different sampling sites.
Finally, the small Vss of mivacurium observed in our study may also be an artifact of the possible elimination of mivacurium outside the central compartment. According to Cook et al., mivacurium's major elimination pathway has been shown in vitro to occur via plasma cholinesterases (butyrylcholinesterases), with only a minimal contribution of acetycholinesterases. Because one major assumption underlying the noncompartmental analysis is that elimination occurs from the central compartment only, we must assume that this enzyme exists only in plasma and not in the extracellular fluid space. This might not be entirely true. Because mivacurium pharmacokinetic parameters reported in previous studies have also been derived based on this assumption, [2,5,6]the results of the present study suggest that this effect might be more pronounced after an intravenous bolus.
Mivacurium isomers are believed to be hydrolyzed by plasma cholinesterases to monoester and alcohol metabolites. All three isomers are not metabolized to the same extent, with the cis-cis isomer metabolized much more slowly, as seen by its longer half-life. The Cmax of metabolites is generally reached soon after injection, within 35 s, reemphasizing the rapid hydrolysis of the parent compound. However, the actual pathway of metabolite formation remains to be elucidated, and this study offers the first insights as to how this may happen. Surprisingly, the cis alcohol is minimally and transiently detected compared with the other three metabolites, which show higher AUC values and slower half-lives in the range of 90 min.
This study represents the first attempt to characterize the pharmacokinetic parameters of mivacurium isomers after an intravenous bolus injection using an intensive arterial blood sampling approach. The short half-lives and rapid clearances observed with the trans-trans and cis-trans isomers are consistent with the short clinical duration of action of mivacurium chloride. Furthermore, the rapid appearance of monoester and alcohol metabolites in plasma, soon after mivacurium injection, reemphasized the rapid hydrolysis of mivacurium isomers by plasma cholinesterases.
The intensive blood sampling schedule proved to be critical for an adequate determination of mivacurium pharmacokinetics because of the significant contribution to the overall area under the plasma concentration-time curve. In addition, the choice of the sampling site may explain the differences noted between our results and those that were published before.
*Maehr RB, Belmont MR, Wray DL, Savarese JJ, Wastilla WB: Autonomic and neuromuscular effects of mivacurium and isomers in cats (abstract). Anesthesiology 1991; 77:A772.