Department of Anaesthesiology, Leiden University Medical Centre, PO Box 9600, NL-2300 RC Leiden, The Netherlands
Corresponding author. E-mail: m.j.mertens@lumc.nl
Accepted for publication: September 26, 2002
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Methods. Remifentanil concentrationtime data that had been collected during a previous pharmacodynamic interaction study in 30 female patients (ASA physical status I, aged 2065 yr) who received a TCI of remifentanil and propofol during lower abdominal surgery were used in this evaluation. The remifentanil concentrations predicted by the five parameter sets were calculated on the basis of the TCI device record of the infusion ratetime profile that had actually been administered to each individual. The individual and pooled bias [median performance error (MDPE)], inaccuracy [median absolute performance error (MDAPE)], divergence and wobble of the remifentanil TCI device were determined from the pooled and intrasubject performance errors.
Results. A total of 444 remifentanil blood samples were analysed. Blood propofol and remifentanil concentrations ranged from 0.5 to 11 µg ml1 and 0.1 to 19.6 ng ml1 respectively. Pooled MDPE and MDAPE of the remifentanil TCI device were 15 and 20% for the parameter set of Minto and colleagues (Anesthesiology 1997; 86: 1023), 1 and 21%, 6 and 21%, and 6 and 19% for the three parameter sets described by Egan and colleagues (Anesthesiology 1996; 84: 82133, Anesthesiology 1993; 79: 88192, Anesthesiology 1998; 89: 56273), and 24 and 30% for the parameter set described by Drover and Lemmens (Anesthesiology 1998; 89: 86977).
Conclusions. Remifentanil can be administered by TCI with acceptable bias and inaccuracy. The three pharmacokinetic parameter sets described by Egan and colleagues resulted in the least bias and best accuracy.
Br J Anaesth 2003; 90: 13241
Keywords: anaesthetics i.v., propofol; analgesics opioid, remifentanil; analgesic techniques, infusion; drug delivery, computerized; pharmacokinetics, models
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Materials
A palm-top computer was provided with three-compartment pharmacokinetic data for remifentanil1 to control a syringe pump for the infusion of remifentanil using the algorithm described by Hull7 (Fig. 1). The same computer, provided with three-compartment pharmacokinetic data for propofol,8 was used to control another syringe pump for the infusion of propofol. The control algorithm checked and adjusted the infusion rates every 5 s. Every minute the infusion ratetime data for remifentanil and propofol were stored on disk in the palm-top computer. The infusion ratetime data for each patient were entered off-line into a computer simulation program that calculated the predicted remifentanil concentrations when provided with one of the five evaluated pharmacokinetic parameter sets for remifentanil (Table 1).15 Lean body mass (a covariate in three out of the five evaluated pharmacokinetic parameter sets of remifentanil) was calculated as:1
|
|
For each measured blood remifentanil concentration, five predicted remifentanil concentrations corresponding to five pharmacokinetic parameter sets were thus obtained, and the performance of each parameter set was then determined.
Blood samples and assays
As described previously,6 arterial blood samples for determination of propofol and remifentanil concentrations in whole blood were collected at laryngoscopy, intubation, skin incision, the opening of the peritoneum, at awakening, and 6 min after a predicted target remifentanil concentration had been achieved during the intraoperative period.
Samples for the determination of blood propofol concentrations were transferred to test tubes containing potassium oxalate and stored at 4°C. Propofol concentrations in blood were measured at our laboratory at the Leiden University Medical Centre by reversed-phase high-performance liquid chromatography.9 The limit of quantitation was 110 ng ml1. The intra-assay coefficients of variation at plasma concentrations of 0.46, 2.33, 4.66 and 13.54 ng ml1 were all <5.0%. The corresponding inter-assay coefficients of variation were <3.7%. Propofol assays were carried out within 12 weeks.
Samples for the determination of the blood remifentanil concentration were collected into tubes containing sodium heparin and immediately transferred to tubes containing 50% citric acid (to inactivate esterases) before freezing at 20°C. The assay method was based on tandem mass spectrometry detection (LC-MS/MS) with a quantitation limit of 0.1 ng ml1. The intra-assay coefficients of varia tion at plasma concentrations of 0.1, 0.2, 30 and 50 ng ml1 were 10.0, 7.8, 2.5 and 2.7% respectively. The corresponding inter-assay coefficients of variation were 2.1, 9.0, 3.8 and 3.7%.10 The remifentanil analyses were performed by a commercial laboratory (Analytico, Breda, The Netherlands).
Analysis
From each remifentanil pharmacokinetic parameter set, micro-constants were determined using standard equations (Table 2). These were introduced, along with the volume of the central compartment, into a computer simulation program in an Excel spreadsheet for the off-line calculation7 of the predicted remifentanil concentration (Cp) for each time when a remifentanil blood sample was obtained from a patient. In order to correct for the fact that delivery performance of a TCI pump is never ideal,11 12 Cp was calculated on the basis of the delivery system record of the infusion ratetime profile actually used for each individual (as opposed to that which should have been given).13
|
PEij (%)=[(CmijCpij)/Cpij]x100(2)
where Cpij is the predicted blood remifentanil concentration in sample j from patient i, and Cmij is the measured blood concentration of remifentanil in that sample. Subsequently, the intrasubject bias (i.e. direction and size of deviation from target) and inaccuracy (i.e. size of the typical miss) of each system was assessed by determination of the median performance error (MDPEi),
MDPEi (%)=median{PEij, j=1,...,Ni}(3)
and the median absolute performance error (MDAPEi),
MDAPEi (%)=median{|PE|ij, j=1,...,Ni}(4)
where Ni is the number of blood samples for individual i.
Divergence,14 a measure of the expected systematic time-related changes in performance, was calculated for individual i as the slope obtained from linear regression of that individuals |PE|ij s against time:
where tij is the time (in min) at which blood sample ij was collected and PEij and Ni are as described earlier.
Wobble, a measure of the total intra-individual variability in PE,14 which is directly related to the ability to achieve stable drug concentrations with the computer-controlled infusion pump, was calculated for the ith individual as
Wobblei (%)=median{|PEijMDPEi|, j=1,...,Ni}(6)
where MDPEi is the median PE in individual i.
The influence of blood propofol concentration on remifentanil PE determined for the pharmacokinetic data set of Minto and colleagues,1 which was implemented on the TCI device in this study, was explored by linear regression analysis.
The pooled bias and inaccuracy of each system were assessed by determination of the MDPE and MDAPE over all 444 blood samples. Furthermore, the pooled divergence and wobble were calculated from all 444 PEs of all patients. When the 95% confidence interval of the pooled MDPE included zero, it was concluded that the bias was not significant.
Data are presented as mean (SD) unless stated otherwise.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The measured blood concentrations of remifentanil were lower than those predicted on the basis of the initially used pharmacokinetic parameter set of Minto and colleagues.1 Visual inspection of the predicted vs measured blood remifentanil concentration plots (Figs 2 and 3) showed significant overprediction of the blood remifentanil concentration by the pharmacokinetic parameter sets of Minto and colleagues1 and Drover and Lemmens5 compared with the three parameter sets described by Egan and colleagues.24 As a result of the relatively unbiased prediction of the blood remifentanil concentration by the three parameter sets of Egan and colleagues24, the PEs were evenly distributed around a range of 6 to 1% when plotted against the predicted concentration (Figs 2 and 3, right panels), whereas the more negatively biased predictions by the parameter sets of Minto and colleagues1 and Drover and Lemmens5 were distributed around 15 and 24%, respectively.
|
In Figure 4, the percentage PEs are plotted against time.
|
|
|
The total intra-individual variability in PE (i.e. wobble), a measure that is directly related to the ability to achieve stable drug concentrations with the TCI system, was acceptable for all five parameter sets (range 1221%).
Linear regression analysis demonstrated no influence of blood propofol concentration on the remifentanil PE (Fig. 6).
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
It should be emphasized that the true performance, i.e. the mismatch between the target remifentanil concentrations and the remifentanil concentrations actually measured, could only be determined for the data set of Minto and colleagues that had been implemented in the TCI system. Had we implemented one of the other data sets, the actually achieved (measured) concentrations would have been somewhat different. However, the true performance of that data set determined in this way would, in all likelihood, be very similar to the expected performance calculated on the basis of the volumes per unit of time actually delivered, as done in this study, because in the relatively narrow clinical concentration range the PEs of remifentanil are independent of the absolute remifentanil concentration.
The accuracies of the evaluated parameter sets (MDAPE) ranged from 19 to 30%. With the exception of the parameter set of Drover and Lemmens,5 these were comparable with the median absolute weighted residual (MDAWR),14 a measure of accuracy of a parameter set in its original study population. The MDAPE found in our study and the reported MDAWR in the original publications were 20 vs 15.3% for the set of Minto and colleagues,1 21 vs 15.9%, 21 vs 14.8% and 19 vs 23.1% for the three sets of Egan and colleagues,24 and 30 vs 9.7% for the parameter set of Drover and Lemmens.5 The observed bias (MDPE) of the evaluated parameter sets, when applied in TCI, was predominantly negative. In our study one of the parameter sets described by Egan2 achieved an unbiased performance.
The pharmacokinetics of remifentanil is linear (i.e. independent of dose or infusion rate) over a large dose range.2 15 16 Because remifentanil is cleared rapidly, most of the interindividual variability in concentration during a continuous infusion will reflect variability in metabolic clearance.1 In population pharmacokinetic models, an attempt has been made to explain and thereby reduce the interindividual variability of the pharmacokinetic parameters by the introduction of covariates into the model.17 In the pharmacokinetic parameter set of Minto and colleagues1 that was implemented in the TCI device in our study, the coefficient of variation in Cl1 in their model without covariates was 23%. The coefficient of variation for Cl1 was 14% when the covariates age and lean body mass were included in their final model. Age and lean body mass therefore accounted for 9% of the interindividual variability. The source of the remaining 14% variability remains unknown. Considering the coefficient of variation for Cl1 of 14%, the MDAPE (a measure of the typical error) of 20% calculated for the parameter set of Minto and colleagues1 in our study is in line with what may be expected. The reported coefficients of variation for Cl1 reported by Egan2 3 and Drover and Lemmens5 were 12, 10 and 23% respectively.
To achieve the most accurate prediction of blood concentrations of remifentanil, one should use a pharmacokinetic parameter set that has been determined in a population that most resembles the actual patients who receive the drug. The blood remifentanil samples evaluated in our study were obtained during abdominal surgery in mechanically ventilated female patients and in the presence of propofol. Our study population therefore differed from those described in the original publications in which the evaluated parameter sets were determined (Table 1).
The MDPEs of the evaluated pharmacokinetic parameter sets, when used for TCI, differed between the five parameter sets and were predominantly negative. As follows from equation (1), a negative prediction error will result from Cm being lower than Cp. Because remifentanil is cleared rapidly, most of the variability in concentration during a continuous infusion will reflect variability in metabolic clearance.1
The primary metabolic pathway of remifentanil is hydrolysis by non-specific blood and tissue esterases, resulting in formation of the carboxylic acid metabolite GI-90291. Approximately 90% of remifentanil can be recovered as the acid metabolite.18 Hydrolysis in blood, liver and kidneys each accounted for only 1% of the systemic clearance of remifentanil in male Beagle dogs, whereas muscle, intestine and brain had the largest tissue clearance rates of the tissues examined.19 The in vitro half-life of remifentanil in human blood at 37°C is approximately 60 min.20 The rapid metabolism of remifentanil is therefore more likely to be explained by hydrolysis of remifentanil in tissues than by hydrolysis in blood. In a study on the influence of arterial vs venous sampling on remifentanil pharmacokinetics, Hermann and colleagues21 demonstrated the importance of tissue metabolism in the pharmacokinetics of remifentanil. They used a pharmacokinetic model in which a one- or two-compartment venous model was linked to the central arterial compartment of a standard three-compartment, open, linear mammillary model to describe the time courses of arterial and venous remifentanil concentration data simultaneously. Elimination occurred from both the arterial and the venous central compartment. They suggest that differences between subjects in extraction ratios across the sampling site could lead to added variability in concentrations and estimated pharmacokinetic parameters. According to this study, increased tissue perfusion or decreased tissue transit time will thus lead to increased clearance of remifentanil from the blood. Indeed, Duthie and colleagues16 found a direct relation between remifentanil clearance and cardiac output, indicating that the tissues eliminating remifentanil clear the drug better when blood is perfused to them. In contrast, during haemorrhagic shock with low cardiac output and low mean arterial pressure, remifentanil elimination clearance and central volume were reduced in pigs.22
Tissue perfusion is the most important determinant of remifentanil tissue clearance and tissue clearance is the most important process determining remifentanil elimination clearance, which is the most important parameter determining the pharmacokinetics of remifentanil. This is influenced by sex and the conditions during surgery (i.e. mechanical ventilation, abdominal surgery, increased protein binding and/or the presence of propofol). These factors may explain the differences in PE between the five pharmacokinetic parameter sets.
When remifentanil is administered during propofol anaesthesia to ASA III adult patients using a TCI device driven by the pharmacokinetic parameter sets determined by Minto and colleagues1 or Drover and Lemmens,5 one should be aware that a considerable bias may exist. In the presence of propofol the actual measured remifentanil concentration, and thus the actual effect site concentration, may be 15% lower than predicted by the TCI device. This can be especially important when following remifentanil dosing guidelines, as mentioned in the literature on blood remifentanil concentrations associated with a particular effect. With the set of Egan and colleagues,2 on the other hand, an unbiased predictive performance can be achieved.
In conclusion, remifentanil can be administered by TCI with acceptable bias and inaccuracy. The use of the pharmacokinetic parameter sets of Egan and colleagues24 resulted in the best precision and accuracy in ASA III female patients. No correlation was found between blood propofol concentration and the PE of remifentanil. Although the parameter set of Egan and colleagues2 performed best in our population, we believe that a population pharmacokinetic parameter set like that of Minto and colleagues1 may prove beneficial in more heterogeneous groups of patients.
![]() |
Acknowledgement |
---|
|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Egan TD, Minto CF, Hermann DJ, Barr J, Muir KT, Shafer SL. Remifentanil versus alfentanil: comparative pharmacokinetics and pharmacodynamics in healthy adult male volunteers. Anesthesiology 1996; 84: 82133[CrossRef][ISI][Medline]
3 Egan TD, Lemmens HJ, Fiset P, et al. The pharmacokinetics of the new short-acting opioid remifentanil (GI87084B) in healthy adult male volunteers. Anesthesiology 1993; 79: 88192[ISI][Medline]
4 Egan TD, Huizinga B, Gupta SK, et al. Remifentanil pharmacokinetics in obese versus lean patients. Anesthesiology 1998; 89: 56273[ISI][Medline]
5 Drover DR, Lemmens HJ. Population pharmacodynamics and pharmacokinetics of remifentanil as a supplement to nitrous oxide anesthesia for elective abdominal surgery. Anesthesiology 1998; 89: 86977[CrossRef][ISI][Medline]
6 Mertens MJ, Vuyk J, Olofsen E, Bovill JG, Burm AGL. The pharmacodynamic interaction between remifentanil and propofol Eur J Anaesthesiol 2001; 18 (Suppl 21); 101
7 Hull CJ. Pharmacokinetics for Anesthesia, 1st Edn. Oxford: Butterworth-Heinemann, 1991
8 Gepts E, Camu F, Cockshott ID, Douglas EJ. Disposition of propofol administered as constant rate intravenous infusions in humans. Anesth Analg 1987; 66: 125663[Abstract]
9 Vuyk J, Engbers FH, Lemmens HJ, et al. Pharmacodynamics of propofol in female patients. Anesthesiology 1992; 77: 39[ISI][Medline]
10 Bender J, van den Elshout J, Selinger K, Broeders G, Dankers J, van der Heiden C. Determination of remifentanil in human heparinised whole blood by tandem mass spectrometry with short-column separation. J Pharm Biomed Anal 1999; 21: 55967[CrossRef][ISI][Medline]
11 Auty B. Infusion equipment for total intravenous anesthesia. In: Kay B, ed. Total intravenous anaesthesia. Amsterdam: Elsevier Science Publishers, 1991; 20523
12 Engbers FHM. Total intravenous anaesthesia: the equipment. In: Vuyk J, Engbers FHM, Groen-Mulder S, eds. On the Study and Practice of Intravenous Anaesthesia. Dordrecht: Kluwer Academic Publishers, 2000; 7187
13 Oei-Lim VL, White M, Kalkman CJ, Engbers FH, Makkes PC, Ooms WG. Pharmacokinetics of propofol during conscious sedation using target-controlled infusion in anxious patients undergoing dental treatment. Br J Anaesth 1998; 80: 32431
14 Varvel JR, Donoho DL, Shafer SL. Measuring the predictive performance of computer-controlled infusion pumps. J Pharmacokinet Biopharm 1992; 20: 6394[ISI][Medline]
15 Westmoreland CL, Hoke JF, Sebel PS, Hug CC Jr, Muir KT. Pharmacokinetics of remifentanil (GI87084B) and its major metabolite (GI90291) in patients undergoing elective inpatient surgery. Anesthesiology 1993; 79: 893903[ISI][Medline]
16 Duthie DJ, Stevens JJ, Doyle AR, et al. Remifentanil and pulmonary extraction during and after cardiac anesthesia. Anesth Analg 1997; 84: 7404[Abstract]
17 Schnider TW, Minto CF, Gambus PL, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 117082[ISI][Medline]
18 Glass PS, Hardman D, Kamiyama Y, et al. Preliminary pharmacokinetics and pharmacodynamics of an ultra-short-acting opioid: remifentanil (GI87084B). Anesth Analg 1993; 77: 103140[Abstract]
19 Chism JP, Rickert DE. The pharmacokinetics and extra-hepatic clearance of remifentanil, a short acting opioid agonist, in male beagle dogs during constant rate infusions. Drug Metab Dispos 1996; 24: 3440[Abstract]
20 Grosse CM, Davis IM, Arrendale RF, Jersey J, Amin J. Determination of remifentanil in human blood by liquidliquid extraction and capillary GC-HRMS-SIM using a deuterated internal standard. J Pharm Biomed Anal 1994; 12: 195203[CrossRef][ISI][Medline]
21 Hermann DJ, Egan TD, Muir KT. Influence of arteriovenous sampling on remifentanil pharmacokinetics and pharmacodynamics. Clin Pharmacol Ther 1999; 65: 51118[ISI][Medline]
22 Johnson KB, Kern SE, Hamber EA, McJames SW, Kohnstamm KM, Egan TD. Influence of hemorrhagic shock on remifentanil: a pharmacokinetic and pharmacodynamic analysis. Anesthesiology 2001; 94: 32232[CrossRef][ISI][Medline]