Clin Pharmacokinet DOI 10.1007/s40262-015-0256-4
SHORT COMMUNICATION
A Compartmental Analysis for Morphine and Its Metabolites in Young Children After a Single Oral Dose Nieves Velez de Mendizabal • Ricardo Jimenez-Mendez • Erin Cooke • Carolyne J. Montgomery Joy Dawes • Michael J. Rieder • Katarina Aleksa • Gideon Koren • Carlos O. Jacobo-Cabral • Rodrigo Gonzalez-Ramirez • Gilberto Castan˜eda-Hernandez • Bruce C. Carleton
•
Ó Springer International Publishing Switzerland 2015
Abstract Background and Objectives Currently, the majority of the surgical procedures performed in paediatric hospitals are done on a day care basis, with post-operative pain being managed by caregivers at home. Pain after discharge of these post-operative children has historically been managed with oral codeine in combination with paracetamol (acetaminophen). Codeine is an opioid, which elicits its analgesic effects via metabolism to morphine and codeine6-glucuronide. Oral morphine is a feasible alternative for outpatient analgesia; however, the pharmacokinetics of morphine after oral administration have been previously
Electronic supplementary material The online version of this article (doi:10.1007/s40262-015-0256-4) contains supplementary material, which is available to authorized users. N. Velez de Mendizabal Division of Clinical Pharmacology, Department of Medicine, Indiana University School of Medicine, Indianapolis, USA N. Velez de Mendizabal Indiana Clinical and Translational Sciences Institute (CTSI), West Walnut Street, Indianapolis, IN 46202, USA R. Jimenez-Mendez B. C. Carleton Division of Translational Therapeutics, Department of Paediatrics, Faculty of Medicine, University of British Columbia, Vancouver, Canada R. Jimenez-Mendez B. C. Carleton Pharmaceutical Outcomes Programme, BC Children’s Hospital, Vancouver, Canada R. Jimenez-Mendez B. C. Carleton (&) Child & Family Research Institute, 950 West 28th Avenue, Rm. A3-207, Vancouver, BC V5Z 4H4, Canada e-mail: [email protected]
described only sparsely, and there is little information in healthy children. Methods The clinical trial included 40 children from 2 to 6 years of age, with an American Society of Anaesthesiologists physical status classification of 1 or 2, who were undergoing surgical procedures requiring opioid analgesia. Morphine was orally administered prior to surgery in one of three doses: 0.1 mg/kg, 0.2 mg/kg and 0.3 mg/kg. Blood samples were collected for plasma morphine, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) concentrations at 30, 60, 90, 120, 180 and 240 min after administration. All analyses were performed with the non-linear mixed-effect modelling software NONMEM version 7.2, using the firstorder conditional estimation (FOCE) method. Results A pharmacokinetic model was developed to simultaneously describe the plasma profiles of morphine and E. Cooke C. J. Montgomery J. Dawes Pediatric Anesthesia Research Team, University of British Columbia, Vancouver, Canada E. Cooke C. J. Montgomery J. Dawes Department of Pediatric Anesthesia, BC Children’s Hospital, Vancouver, Canada M. J. Rieder Department of Physiology and Pharmacology, Western University, London, Canada K. Aleksa G. Koren The Motherisk Program, Division of Clinical Pharmacology/ Toxicology, The Hospital for Sick Children, Toronto, Canada C. O. Jacobo-Cabral R. Gonzalez-Ramirez G. Castan˜eda-Hernandez Departamento de Farmacologı´a, Centro de Investigacio´n y de Estudios Avanzados del Instituto Polite´cnico Nacional, Mexico, DF, Mexico
N. Velez de Mendizabal et al.
its metabolites M3G and M6G after a single dose of oral morphine in young children (2–6 years of age). The disposition of morphine, M3G and M6G in plasma was best described by a one-compartment model. M3G and M6G metabolite formation was best described by a delay transit compartment, indicating a delay in the appearance of these two major metabolites. Conclusion This model provides a foundation on which to further evaluate the use of oral morphine and its safety in young children. Longer follow-up time for morphine oral doses and incorporation of other important covariates, such as phenotype, will add value and will help overcome the limitations of the presented population pharmacokinetic analysis.
Key Points Most surgeries in young children (2–6 years of age) are done on a day care basis, with post-operative pain being managed on the basis of the parents’ and/ or caregivers’ judgment. Pain after discharge has historically been managed with oral codeine in combination with paracetamol (acetaminophen). Because of genetic polymorphisms of cytochrome P450 2D6, codeine pharmacokinetics are highly variable, with poor, intermediate, extensive and ultra-rapid metabolizers, generating important safety and efficacy concerns. Oral morphine might be an alternative, safer treatment for outpatient analgesia. The population pharmacokinetic model presented here simultaneously describes the observed pharmacokinetics of morphine, morphine-3glucuronide and morphine-6-glucuronide in young children (2–6 years of age) after a single dose of oral morphine. This analysis is the first in the public domain to describe a combined parent/metabolite pharmacokinetic model after oral morphine administration in children. This model provides the background needed for further efficacy and safety evaluations of the use of oral morphine as a treatment for outpatient analgesia in children.
1 Introduction At BC Children’s Hospital, Vancouver, Canada, 75 % of all surgical procedures are performed as day procedures, without an overnight stay in hospital. Most patients leave
the hospital within 2–6 h after the time of their procedure. This shortened period of post-operative monitoring means that the analgesic regimen is managed by the parents or caregivers of the child. Pain after discharge is common and is often inadequately treated by parents because of variables such as insufficient information, uncertainty about pain assessment and fear of adverse drug reactions [1, 2]. Post-operative pain has typically been managed with oral paracetamol (acetaminophen), ibuprofen and codeine. Codeine is an opioid, which elicits its main analgesic effects via its metabolism to morphine and codeine-6-glucuronide [3–5]. Codeine is ineffective in up to 30 % of patients because of genetic polymorphisms of cytochrome P450 (CYP) 2D6 [4, 6–10]. Such genetic polymorphisms lead to several clinically relevant phenotypes: poor, intermediate, extensive and ultra-rapid metabolizers [7, 9, 10]. This variability in metabolic activity, and hence conversion of codeine to morphine, accounts for much of the clinical variability seen in analgesia. Poor-metabolizers, who approximately represent up to 9 % of the Caucasian population, experience inappropriately low analgesic efficacy due to inadequate conversion of codeine to morphine. Conversely, lethal adverse reactions, such as respiratory depression, have been reported in ultra-rapid metabolizers, who have on average 50 % higher plasma concentrations of morphine and constitute approximately 1 % of the Caucasian population. However, in some ethnic groups— notably, patients with African ancestry—up to 40 % of patients may be ultra-rapid metabolizers and therefore susceptible to potentially fatal respiratory depression [11– 16]. Because of these safety concerns, regulatory agencies have recently warned practitioners about the use of codeine in children, urging them to consider alternative analgesic medications. These warnings have been greater for certain procedures; codeine is contraindicated in some jurisdictions for post-operative pain management after tonsillectomy or adenoidectomy [17]. Since the majority of the analgesic effect derived from codeine is the result of its bioconversion to morphine, a potential substitute for codeine could be morphine. We propose that the use of oral morphine could be a feasible and safe alternative for outpatient analgesia. Although morphine has been widely used for many years in the management of post-operative pain in both adults and children [8, 18, 19], the pharmacokinetics of morphine after oral administration are poorly characterized in healthy children undergoing day surgery. The recommended therapeutic doses of oral morphine are 0.1–0.3 mg/kg administered every 3–4 h, although post-operative use in children is often empirical rather than evidence based; plasma morphine concentrations after oral administration in children have not been studied completely [8]. Mazoit et al. [20] presented a complete pharmacokinetic and
PopPK Model for Morphine and Its Metabolites in Children
pharmacodynamic (PKPD) model for morphine and its metabolites, morphine-3-glucoronide (M3G) and morphine-6-glucoronide (M6G), in post-operative adult patients. Lotsch et al. [21] published another analytical model describing morphine and M6G plasma concentrations in healthy volunteers. For children younger than 3 years, two different population pharmacokinetic models for morphine, M3G and M6G have been published [22, 23]. Both models were developed on the basis of large data sets of intravenous morphine administration, but there were important discrepancies between the two models, as discussed at the end of this article. In the paediatric population, bodyweight is reported to be the most significant covariate for morphine clearance [22–25]. A complete analysis of the impact of weight on morphine and M3G pharmacokinetics was recently performed by Wang et al. [25] on the basis of recollected data from preterm and term neonates, infants, children, adolescents and adults, covering the entire paediatric age. We are not aware of any population pharmacokinetic model of oral morphine and its metabolites in children. Here, we present a population pharmacokinetic model able to simultaneously describe morphine, M3G and M6G plasma concentrations in children from 2 to 6 years of age after a single oral dose of morphine.
each, with five subjects from each group in each block. We used random number generation to create the randomization schedule. After the first batch was completed, it was seen that all of the serum levels were sub-therapeutic for group 1. As a result, we discontinued recruitment into this group. The oral morphine hydrochloride syrup (DoloralÒ 1 mg/mL; Atlas Laboratories, Winnipeg, MB, Canada) was administered approximately 30 min prior to surgery. Induction and maintenance of general anaesthesia was left to the discretion of the attending anaesthetist. Additional opioid analgesia was provided as required using other synthetic opioids, such as remifentanil and fentanyl. In addition, non-opioid analgesia, such as ketamine, non-steroidal anti-inflammatories (NSAIDs), paracetamol, local anaesthetic infiltration and regional blocks, were permitted. After induction of anaesthesia, a second intravenous cannula was inserted. Blood samples were drawn at 30, 60, 90, 120, 180 and 240 min. Blood samples were collected in heparinized tubes and refrigerated. Samples were centrifuged at 3000 bpm and stored at -80 °C until plasma concentrations of morphine, M6G and M3G were measured using validated high-performance liquid chromatography. More details regarding the bioanalytical assay are presented in Online Resource 1 in the Electronic Supplementary Material.
2 Methods
2.2 Data Analysis: Population Pharmacokinetic Modelling
2.1 Study Design Research approval was obtained from the University of British Columbia—Children’s & Women’s Health Centre of BC Research Ethics Board (H09-03286). The ClinicalTrials.gov identifier for the study is NCT01071499. Informed written consent was obtained from parents or guardians. This study was conducted in a tertiary care freestanding academic paediatric hospital. A multidisciplinary Data and Safety Monitoring Board (DSMB) reviewed the data and study progress at regular intervals. Children aged from 2 to 6 years with an American Society of Anaesthesiologists physical status classification of 1 or 2, undergoing procedures that typically required postoperative opioid analgesia, were invited to participate. The exclusion criteria included patients in whom morphine or oral premedication was contraindicated or patients with a documented allergy or adverse reaction to morphine. Also excluded were patients with a history of abnormal hepatic or renal function. Forty patients were enrolled and assigned to three groups receiving three different oral morphine doses: group 1, 0.1 mg/kg (n = 5); group 2, 0.2 mg/kg (n = 18); and group 3, 0.3 mg/kg (n = 17). Patients were initially block randomized to three blocks of 15 subjects
All analyses were performed with the non-linear mixedeffect modelling software NONMEM version 7.2 (Icon Development Solutions, Hanover, MD, USA) using the first-order conditional estimation (FOCE) method. The population analysis was performed simultaneously for the parent compound, morphine, and the metabolites, M3G and M6G. Inter-subject variability (ISV) was modelled using exponential functions. Additive models were used to describe residual variability for log-transformed data (morphine, M3G and M6G), which equals a proportional error on the linear scale. The disposition of morphine and its metabolites in plasma were described with compartmental models parameterized in apparent volumes of distribution, apparent inter-compartmental clearances and apparent metabolite formation clearances. Selection was made between one-, two-, and three-compartment models for morphine. Different absorption models were explored: first- and/or zeroorder rate absorptions, in combination with the presence of possible delays in the drug absorption process: discrete lag times and transit compartment models [26] were explored. Time- and/or dose-dependent parameters were also evaluated. Metabolite transformation from the plasma compartment was described using linear and non-linear
N. Velez de Mendizabal et al.
kinetics. The volume of the central metabolite compartment was parameterized as being equal to the volume of the parent central compartment. Distribution of metabolites was explored using one-, two- and three-compartment models. Once the structural and statistical model was selected, the effects of intrinsic (weight, age, height, sex) and extrinsic patient factors (batch) were evaluated. Potentially relevant patient covariates were evaluated on all of the pharmacokinetic parameters. The covariate selection criteria were (1) a statistically significant difference in the objective function value (OFV) reported by NONMEM (criterion: C6.63-point drop, p \ 0.01); or (2) a decrease in the ISV of at least 10 %.
2.3 Model Selection Criteria and Evaluation The minimum OFV provided by NONMEM, which corresponds approximately to a -2 9 log (likelihood) [-2LL], served as a criterion for model comparison during the model development process. A decrease in -2LL of 6.63 points for one additional parameter was regarded as a significant model improvement, corresponding to a p value of 0.01 for nested models. The Akaike information criterion (AIC), calculated as AIC = -2LL ? 2 9 NP, where NP is the number of parameters in the model, was used for selection among non-nested models [27]. The choice of the final model was based also on the OFV value, the precision of parameter estimates, goodness-of-fit plots and visual predictive checks (VPCs). The precision of parameter estimates, expressed as 2.5th to 97.5th percentiles, was computed from the analysis of 500 bootstrap data sets. The bootstrap analysis was performed using Perl-speaks-NONMEM [28]. Model parameter estimates were presented together with the corresponding relative standard error (RSE) as a measure of parameter imprecision, which were computed from the results obtained from bootstrap analysis. The degree of ISV was expressed as the coefficient of variation. Inspection of goodness-of-fit plots included conditional weighted residuals (CWRES) and normalized prediction distribution errors (NPDEs) [29, 30]. Model performance was evaluated with VPCs: 300 studies were simulated with the same design characteristics as those in the original study. At each time point, the 5th, 50th and 95th percentiles were calculated in every simulated study for morphine, M3G and M6G concentrations. Then, the 90 % confidence intervals from the resultant percentiles (5th, 50th and 95th) were computed and represented over time, together with the raw data. Results were stratified by dose. Predictive checks were performed using MATLAB environment (The Mathworks, Natick, MA, USA).
3 Results We developed a population pharmacokinetic model able to simultaneously describe plasma concentrations for morphine and its metabolites, M3G and M6G, in young children aged 2–6 years. The structure of the selected model is shown in Fig. 1. The corresponding differential equations (see Online Resource 2) and the NONMEM script for the selected model (see Online Resource 3) can be found in the Electronic Supplementary Material. Demographic data are shown in Table 1. The disposition of morphine, M3G and M6G in plasma was best described by a one-compartment model (Fig. 1). The inclusion of direct morphine clearance (unaccounted clearance) did not improve the model prediction. M3G and M6G metabolite formation was best described by a delay transit compartment (see Fig. 1), indicating that the metabolite formation is not an instant process and there is a delay in the appearance of these two major metabolites. ISV was found to be significant for the absorption rate (ka), the volume of distribution in the central compartment (VC/F), bioavailability (F), the formation clearance to M3G (CL2M3G/F), and the first-order rate constant for the
Fig. 1 Structural pharmacokinetic model. The boxes represent the different compartments: the morphine concentration (shaded in grey), transit compartment (CMT) for morphine metabolism, morphine-3glucoronide (M-3-G [M3G]) concentration (shaded in blue) and morphine-6-glucoronide (M-6-G [M6G]) concentration (shaded in green). The volumes of distribution of M3G and M6G (V3MG and V6MG) were parameterized as being equal to the apparent volume of distribution in the central compartment (VC/F). CL2M3G/F formation rate for M3G, CL2M6G/F formation rate for M6G, CLM3G/F elimination rate for M3G, CLM6G/F elimination rate for M6G, ka first-order absorption rate constant, kmf first-order rate constant for the metabolism transit compartment, PO oral
PopPK Model for Morphine and Its Metabolites in Children Table 1 Patient demographic characteristics Characteristic
Dose 0.1 mg/kg
0.2 mg/kg
0.3 mg/kg
n
5
18
17
Sex [n; male/female]
1/4
10/8
6/11
Age [years]a
3.57 (1.44)
4.44 (1.32)
4.21 (1.37)
16.10 (3.50)
17.97 (3.11)
17.41 (4.53)
12.82 (7.20)
13.56 (6.26)
12.85 (6.20)
Weight [kg]
a
Body mass indexa a
Average (standard deviation)
metabolism transit compartment (kmf). The g-shrinkage (%) values were 26.09 (ka), 15.36 (VC/F), 5.08 (F), 28.23 (CL2M3G/F) and 17.78 (kmf). None of the covariates evaluated in this study (weight, age, height, sex and batch) was significant in terms of the criteria specified in the methods (changes in the OFV and/or ISV). Model parameter estimates with the corresponding RSE values are presented in Table 2. The low RSE values in conjunction with the results obtained from the non-parametric bootstrap indicate that there was good precision of the estimates with no bias observed, except for the metabolite formation clearances CL2M3G/F (114.53 %) and CL2M6G/F (126.34 %). The poor precision is probably due to the short period of time (4 h) used in this study to follow the respective pharmacokinetics. Figure 2, which shows the goodness-of-fit plots for morphine, M3G and M6G, indicates that the selected
model properly describes the observations. Observations were compared with the individual model predictions (Fig. 2, left). No dose bias and mis-specifications were observed. Evaluation of the CWRES did not show any tendency or model mis-specification (Fig. 2, centre). The scatterplots of NPDEs versus the population predictions indicate that the model performed properly for the three kinds of observations: morphine, M3G and M6G (Fig. 2, right). Population and individual predictions, as well as observations for morphine, M3G and M6G, are also shown in Fig. S1 in Online Resource 4. The results from the VPCs for the morphine and its metabolites, M3G and M6G, stratified by dose, are shown in Fig. 3. The selected model is able to describe both typical profiles and data dispersion for every dependent variable (morphine, M3G and M6G) and for every dose (0.1, 0.2 and 0.3 mg/kg). The strict model selection criteria and the detailed model evaluation performed in this study demonstrate the reliability of this population pharmacokinetic model for morphine and its metabolites, M3G and M6G, in young children.
4 Discussion and Conclusion The pharmacokinetics of intravenously administered morphine and its two major active metabolites, M3G and M6G, have been widely studied in paediatrics, using different
Table 2 Population pharmacokinetic parameter estimates Parameters
Estimate (RSE %)
ISV (RSE %)
Bootstrap analysis: median (2.5–97.5th percentiles) Estimate
ISV
0.006 (0.004–0.007) 4.75 (2.409–6.964)
46.71 (29.06–62.46) 64.76 (38.74–86.83)
0.779 (0.144–3.128)
28.44 (14.75–36.57)
ka [min-1] VC/F [L]
0.006 (17.32) 4.96 (27.96)
F
1 fixed
76.09 (36.86)
CL2M3G/F [L/min]
0.776 (114.53)
29.44 (40.94)
CLM3G/F [L/min]
0.068 (15.97)
0.066 (0.051–0.086) 0.062 (0.01–0.254)
CL2M6G/F [L/min]
0.064 (126.34)
CLM6G/F [L/min]
0.042 (22.13)
kmf [min-1]
2.020 (14.20)
V3M [L]
VC/F
V6M [L] Res error for morphine [log (ng/mL)]
46.58 (44.65) 66.25 (44.21)
73.03 (50.92–94.92)
0.040 (0.027–0.057) 48.57 (47.32)
1.951 (1.528–2.450)
46.44 (28.76–64.14)
VC/F a
0.321 (38)
0.319 (0.150–0.534)
Res error for M3G [log (ng/mL)]a
0.180 (31.88)
0.177 (0.098–0.281)
Res error for M6G [log (ng/mL)]a
0.300 (38.23)
0.286 (0.139–0.518)
Parameters are listed as estimates, with 95 % confidence intervals from 500 bootstrap datasets in parenthesis. The metabolite volumes of distribution (V3M and V6M) were parameterized as being equal to the apparent volume of distribution in the central compartment (VC/F). Estimates of inter-subject variability (ISV) are shown as coefficients of variation CL2M3G/F formation rate for M3G, CL2M6G/F formation rate for M6G, CLM3G/F elimination rate for M3G, CLM6G/F elimination rate for M6G, F bioavailability, ka first-order absorption rate constant, kmf first-order rate constant for the metabolism transit compartment, M3G morphine-3glucoronide, M6G morphine-6-glucoronide, Res residual, RSE relative standard error a
Additive error model in log scale
N. Velez de Mendizabal et al.
Fig. 2 Goodness-of-fit plots of the selected population pharmacokinetic model for morphine (shown in grey) and its metabolites, morphine-3-glucoronide (M-3-G [M3G]; shown in blue) and morphine-6-glucoronide (M-6-G [M6G]; shown in green). The solid lines show identity lines in the first two columns and zero lines in the last
two columns. CWRES conditional weighted residuals, DV observed concentration, IPRED individual model-predicted concentration, NPDE normalized prediction distribution error, PRED population model-predicted concentration
approaches and modelling techniques (e.g. non-compartmental and compartmental models). For children younger than 3 years, two different population pharmacokinetic models for morphine and its metabolites have been published [22, 23]. Krekels et al. [24] published a paper comparing these models, as there are discrepancies between them. Both models were developed on the basis of large data sets of intravenous morphine administration. For children younger than 3 years, multiple covariates were identified as descriptors of variability in the pharmacokinetic profiles between patients. We tried to use both models to describe the data presented here. However, neither of these models was able to appropriately model the data we generated with oral morphine use. This might also be explained by the age differences between the studies: 0–3 years in the aforementioned studies [22, 23] and 2–6 years in this study. Bouwmeester et al. [23] identified a direct morphine clearance (unaccounted clearance); however, in our case, the inclusion of this process did not significantly improve the fit. This could be explained by the short period of time (only 4 h) for sample collection after administration. However, Knibbe et al. [22] concluded that none of the morphine elimination pathways other than the transformation to M3G and M6G play a significant role in morphine clearance. Morphine elimination though these
other routes was not found to be significantly different from zero. Knibbe et al. [22] described morphine pharmacokinetics with a two-compartment model, although in the study by Bouwmeester et al. [23], the inclusion of this second compartment (distribution to peripheral tissue) was not significant. In the model presented here, morphine distribution to other tissues was not identified. More recently, Krekels et al. [31] also published a review exploring the possible advantages and disadvantages of the different data analysis techniques that have been applied, with a specific focus on the accuracy of morphine clearance predictions by paediatric pharmacokinetic models. Other population pharmacokinetic models have been published in order to describe the pharmacokinetic profiles of morphine and its metabolites, but these have been in adults. Mazoit et al. [20] presented a very complete PKPD model of morphine, M3G and M6G in post-operative adult patients who received morphine as an intravenous titration, followed by intramuscular administration postoperatively. Lotsch et al. [21] published another model analysis to describe morphine and M6G plasma concentrations in healthy adult volunteers (aged 23–30 years) after intravenous bolus injection. Interestingly, they also proposed a transit compartment identifying the delay of the appearance of M6G in plasma.
PopPK Model for Morphine and Its Metabolites in Children Fig. 3 Visual predictive checks: results from 1000 simulated studies. The shaded areas represent the 90 % prediction intervals of the 5th, 50th and 95th percentiles. The dashed lines correspond to the observed medians, and the solid lines correspond to the 5th and 95th percentiles of the observed data. The circles are observed data. Morphine is shown in grey, morphine-3-glucoronide (M-3-G [M3G]) is shown in blue and morphine-6glucoronide (M-6-G [M6G]) is shown in green. The results are stratified by morphine dose: a 0.1 mg/kg, b 0.2 mg/kg and c 0.3 mg/kg
In the paediatric population, body weight has been reported to be the most significant covariate for morphine clearance [22–25]. However, in this study, the inclusion of this covariate did not significantly improve the fit. This is probably due to the short observation period (4 h) after oral administration. The formation clearances to M3G and M6G were calculated with poor precision, and other clearance pathways were estimated as zero. Here, we present the first population pharmacokinetic model for morphine and its metabolites, M3G and M6G, in children aged 2–6 years after a single oral dose. The low RSE values in conjunction with the results obtained from the non-parametric bootstrap indicate that there was very good precision of the estimates and no bias observed, except for the metabolite formation clearances CL2M3G/ F and CL2M6G/F. The poor precision of these formation clearance values was likely due to the short period of time used to follow the respective pharmacokinetics. Longer studies have to be performed in order to improve confidence in these estimated values. Although a good option might have been to focus just on the morphine
pharmacokinetic analysis, with the aim being to estimate the morphine clearance with better precision, it was decided to incorporate the three compounds into the study for several reasons. First, the VPCs capture the kinetics and data dispersion very well for the three compounds. This good result, in combination with the good precision of the rest of the parameters, makes the presented model a reliable one for describing morphine and metabolite pharmacokinetics for the first 4 h after a single oral dose. However, this may not accurately describe the kinetic profiles after 4 h. Highlighting the limitations of this study here helps to justify the design of longer studies in the future. Moreover, this study highlights the necessity of reviewing the therapeutic dose recommendation for oral doses of morphine (0.1–0.3 mg/kg administered every 3–4 h). This study provides a background to further evaluate and model morphine following an oral dose in this age group. Longer follow-up time and the incorporation of other important covariates, such as phenotype, will add value and will help us overcome the present limitations.
N. Velez de Mendizabal et al.
References 1. Finley GA, McGrath PJ, Forward SP, McNeill G, Fitzgerald P. Parents’ management of children’s pain following ‘minor’ surgery. Pain. 1996;64:83–7. 2. Jonas DA. Parent’s management of their child’s pain in the home following day surgery. J Child Health Care. 2003;7:150–62. 3. Williams DG, Patel A, Howard RF. Pharmacogenetics of codeine metabolism in an urban population of children and its implications for analgesic reliability. Br J Anaesth. 2002;89:839–45. 4. Williams DG, Hatch DJ, Howard RF. Codeine phosphate in paediatric medicine. Br J Anaesth. 2001;86:413–21. 5. Sindrup SH, Brosen K. The pharmacogenetics of codeine hypoalgesia. Pharmacogenetics. 1995;5:335–46. 6. Sistonen J, Sajantila A, Lao O, Corander J, Barbujani G, Fuselli S. CYP2D6 worldwide genetic variation shows high frequency of altered activity variants and no continental structure. Pharmacogenet Genomics. 2007;17:93–101. 7. Daly AK, Brockmoller J, Broly F, Eichelbaum M, Evans WE, Gonzalez FJ, Huang JD, Idle JR, Ingelman-Sundberg M, Ishizaki T, et al. Nomenclature for human CYP2D6 alleles. Pharmacogenetics. 1996;6:193–201. 8. Berde CB, Sethna NF. Analgesics for the treatment of pain in children. N Engl J Med. 2002;347:1094–103. 9. Marez D, Legrand M, Sabbagh N, LoGuidice JM, Spire C, Lafitte JJ, Meyer UA, Broly F. Polymorphism of the cytochrome P450 CYP2D6 gene in a European population: characterization of 48 mutations and 53 alleles, their frequencies and evolution. Pharmacogenetics. 1997;7:193–202. 10. Cascorbi I. Pharmacogenetics of cytochrome p4502D6: genetic background and clinical implication. Eur J Clin Invest. 2003;33(Suppl 2):17–22. 11. Ciszkowski C, Madadi P, Phillips MS, Lauwers AE, Koren G. Codeine, ultrarapid-metabolism genotype, and postoperative death. N Engl J Med. 2009;361:827–8. 12. de Wildt SN, Koren G. Re: apnea in a child after oral codeine: a genetic variant—an ultra-rapid metabolizer [corrected]. Paediatr Anaesth. 2008;18:273–4 (author’s reply 275–276). 13. Koren G, Cairns J, Chitayat D, Gaedigk A, Leeder SJ. Pharmacogenetics of morphine poisoning in a breastfed neonate of a codeine-prescribed mother. Lancet. 2006;368:704. 14. Madadi P, Koren G, Cairns J, Chitayat D, Gaedigk A, Leeder JS, Teitelbaum R, Karaskov T, Aleksa K. Safety of codeine during breastfeeding: fatal morphine poisoning in the breastfed neonate of a mother prescribed codeine. Can Fam Physician. 2007;53:33–5. 15. Madadi P, Ross CJ, Hayden MR, Carleton BC, Gaedigk A, Leeder JS, Koren G. Pharmacogenetics of neonatal opioid toxicity following maternal use of codeine during breastfeeding: a case–control study. Clin Pharmacol Ther. 2009;85:31–5. 16. Gasche Y, Daali Y, Fathi M, Chiappe A, Cottini S, Dayer P, Desmeules J. Codeine intoxication associated with ultrarapid CYP2D6 metabolism. N Engl J Med. 2004;351:2827–31.
17. Kuehn BM. FDA: no codeine after tonsillectomy for children. JAMA. 2013;309:1100. 18. Lugo RA, Kern SE. Clinical pharmacokinetics of morphine. J Pain Palliat Care Pharmacother. 2002;16:5–18. 19. Choonara IA, McKay P, Hain R, Rane A. Morphine metabolism in children. Br J Clin Pharmacol. 1989;28:599–604. 20. Mazoit JX, Butscher K, Samii K. Morphine in postoperative patients: pharmacokinetics and pharmacodynamics of metabolites. Anesth Analg. 2007;105:70–8. 21. Lotsch J, Skarke C, Schmidt H, Liefhold J, Geisslinger G. Pharmacokinetic modeling to predict morphine and morphine-6glucuronide plasma concentrations in healthy young volunteers. Clin Pharmacol Ther. 2002;72:151–62. 22. Knibbe CA, Krekels EH, van den Anker JN, DeJongh J, Santen GW, van Dijk M, Simons SH, van Lingen RA, Jacqz-Aigrain EM, Danhof M, Tibboel D. Morphine glucuronidation in preterm neonates, infants and children younger than 3 years. Clin Pharmacokinet. 2009;48:371–85. 23. Bouwmeester NJ, Anderson BJ, Tibboel D, Holford NH. Developmental pharmacokinetics of morphine and its metabolites in neonates, infants and young children. Br J Anaesth. 2004;92:208–17. 24. Krekels EH, van Hasselt JG, Tibboel D, Danhof M, Knibbe CA. Systematic evaluation of the descriptive and predictive performance of paediatric morphine population models. Pharm Res. 2011;28:797–811. 25. Wang C, Sadhavisvam S, Krekels EH, Dahan A, Tibboel D, Danhof M, Vinks AA, Knibbe CA. Developmental changes in morphine clearance across the entire paediatric age range are best described by a bodyweight-dependent exponent model. Clin Drug Investig. 2013;33:523–34. 26. Savic RM, Jonker DM, Kerbusch T, Karlsson MO. Implementation of a transit compartment model for describing drug absorption in pharmacokinetic studies. J Pharmacokinet Pharmacodyn. 2007;34:711–26. 27. Ludden TM, Beal SL, Sheiner LB. Comparison of the Akaike information criterion, the Schwarz criterion and the F test as guides to model selection. J Pharmacokinet Biopharm. 1994;22:431–45. 28. Lindbom L, Ribbing J, Jonsson EN. Perl-speaks-NONMEM (PsN)—a Perl module for NONMEM related programming. Comput Methods Progr Biomed. 2004;75:85–94. 29. Hooker AC, Staatz CE, Karlsson MO. Conditional weighted residuals (CWRES): a model diagnostic for the FOCE method. Pharm Res. 2007;24:2187–97. 30. Comets E, Brendel K, Mentre F. Computing normalised prediction distribution errors to evaluate nonlinear mixed-effect models: the NPDE add-on package for R. Comput Methods Progr Biomed. 2008;90:154–66. 31. Krekels EH, Tibboel D, Danhof M, Knibbe CA. Prediction of morphine clearance in the paediatric population: how accurate are the available pharmacokinetic models? Clin Pharmacokinet. 2012;51:695–709.
SHORT COMMUNICATION
A Compartmental Analysis for Morphine and Its Metabolites in Young Children After a Single Oral Dose Nieves Velez de Mendizabal • Ricardo Jimenez-Mendez • Erin Cooke • Carolyne J. Montgomery Joy Dawes • Michael J. Rieder • Katarina Aleksa • Gideon Koren • Carlos O. Jacobo-Cabral • Rodrigo Gonzalez-Ramirez • Gilberto Castan˜eda-Hernandez • Bruce C. Carleton
•
Ó Springer International Publishing Switzerland 2015
Abstract Background and Objectives Currently, the majority of the surgical procedures performed in paediatric hospitals are done on a day care basis, with post-operative pain being managed by caregivers at home. Pain after discharge of these post-operative children has historically been managed with oral codeine in combination with paracetamol (acetaminophen). Codeine is an opioid, which elicits its analgesic effects via metabolism to morphine and codeine6-glucuronide. Oral morphine is a feasible alternative for outpatient analgesia; however, the pharmacokinetics of morphine after oral administration have been previously
Electronic supplementary material The online version of this article (doi:10.1007/s40262-015-0256-4) contains supplementary material, which is available to authorized users. N. Velez de Mendizabal Division of Clinical Pharmacology, Department of Medicine, Indiana University School of Medicine, Indianapolis, USA N. Velez de Mendizabal Indiana Clinical and Translational Sciences Institute (CTSI), West Walnut Street, Indianapolis, IN 46202, USA R. Jimenez-Mendez B. C. Carleton Division of Translational Therapeutics, Department of Paediatrics, Faculty of Medicine, University of British Columbia, Vancouver, Canada R. Jimenez-Mendez B. C. Carleton Pharmaceutical Outcomes Programme, BC Children’s Hospital, Vancouver, Canada R. Jimenez-Mendez B. C. Carleton (&) Child & Family Research Institute, 950 West 28th Avenue, Rm. A3-207, Vancouver, BC V5Z 4H4, Canada e-mail: [email protected]
described only sparsely, and there is little information in healthy children. Methods The clinical trial included 40 children from 2 to 6 years of age, with an American Society of Anaesthesiologists physical status classification of 1 or 2, who were undergoing surgical procedures requiring opioid analgesia. Morphine was orally administered prior to surgery in one of three doses: 0.1 mg/kg, 0.2 mg/kg and 0.3 mg/kg. Blood samples were collected for plasma morphine, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G) concentrations at 30, 60, 90, 120, 180 and 240 min after administration. All analyses were performed with the non-linear mixed-effect modelling software NONMEM version 7.2, using the firstorder conditional estimation (FOCE) method. Results A pharmacokinetic model was developed to simultaneously describe the plasma profiles of morphine and E. Cooke C. J. Montgomery J. Dawes Pediatric Anesthesia Research Team, University of British Columbia, Vancouver, Canada E. Cooke C. J. Montgomery J. Dawes Department of Pediatric Anesthesia, BC Children’s Hospital, Vancouver, Canada M. J. Rieder Department of Physiology and Pharmacology, Western University, London, Canada K. Aleksa G. Koren The Motherisk Program, Division of Clinical Pharmacology/ Toxicology, The Hospital for Sick Children, Toronto, Canada C. O. Jacobo-Cabral R. Gonzalez-Ramirez G. Castan˜eda-Hernandez Departamento de Farmacologı´a, Centro de Investigacio´n y de Estudios Avanzados del Instituto Polite´cnico Nacional, Mexico, DF, Mexico
N. Velez de Mendizabal et al.
its metabolites M3G and M6G after a single dose of oral morphine in young children (2–6 years of age). The disposition of morphine, M3G and M6G in plasma was best described by a one-compartment model. M3G and M6G metabolite formation was best described by a delay transit compartment, indicating a delay in the appearance of these two major metabolites. Conclusion This model provides a foundation on which to further evaluate the use of oral morphine and its safety in young children. Longer follow-up time for morphine oral doses and incorporation of other important covariates, such as phenotype, will add value and will help overcome the limitations of the presented population pharmacokinetic analysis.
Key Points Most surgeries in young children (2–6 years of age) are done on a day care basis, with post-operative pain being managed on the basis of the parents’ and/ or caregivers’ judgment. Pain after discharge has historically been managed with oral codeine in combination with paracetamol (acetaminophen). Because of genetic polymorphisms of cytochrome P450 2D6, codeine pharmacokinetics are highly variable, with poor, intermediate, extensive and ultra-rapid metabolizers, generating important safety and efficacy concerns. Oral morphine might be an alternative, safer treatment for outpatient analgesia. The population pharmacokinetic model presented here simultaneously describes the observed pharmacokinetics of morphine, morphine-3glucuronide and morphine-6-glucuronide in young children (2–6 years of age) after a single dose of oral morphine. This analysis is the first in the public domain to describe a combined parent/metabolite pharmacokinetic model after oral morphine administration in children. This model provides the background needed for further efficacy and safety evaluations of the use of oral morphine as a treatment for outpatient analgesia in children.
1 Introduction At BC Children’s Hospital, Vancouver, Canada, 75 % of all surgical procedures are performed as day procedures, without an overnight stay in hospital. Most patients leave
the hospital within 2–6 h after the time of their procedure. This shortened period of post-operative monitoring means that the analgesic regimen is managed by the parents or caregivers of the child. Pain after discharge is common and is often inadequately treated by parents because of variables such as insufficient information, uncertainty about pain assessment and fear of adverse drug reactions [1, 2]. Post-operative pain has typically been managed with oral paracetamol (acetaminophen), ibuprofen and codeine. Codeine is an opioid, which elicits its main analgesic effects via its metabolism to morphine and codeine-6-glucuronide [3–5]. Codeine is ineffective in up to 30 % of patients because of genetic polymorphisms of cytochrome P450 (CYP) 2D6 [4, 6–10]. Such genetic polymorphisms lead to several clinically relevant phenotypes: poor, intermediate, extensive and ultra-rapid metabolizers [7, 9, 10]. This variability in metabolic activity, and hence conversion of codeine to morphine, accounts for much of the clinical variability seen in analgesia. Poor-metabolizers, who approximately represent up to 9 % of the Caucasian population, experience inappropriately low analgesic efficacy due to inadequate conversion of codeine to morphine. Conversely, lethal adverse reactions, such as respiratory depression, have been reported in ultra-rapid metabolizers, who have on average 50 % higher plasma concentrations of morphine and constitute approximately 1 % of the Caucasian population. However, in some ethnic groups— notably, patients with African ancestry—up to 40 % of patients may be ultra-rapid metabolizers and therefore susceptible to potentially fatal respiratory depression [11– 16]. Because of these safety concerns, regulatory agencies have recently warned practitioners about the use of codeine in children, urging them to consider alternative analgesic medications. These warnings have been greater for certain procedures; codeine is contraindicated in some jurisdictions for post-operative pain management after tonsillectomy or adenoidectomy [17]. Since the majority of the analgesic effect derived from codeine is the result of its bioconversion to morphine, a potential substitute for codeine could be morphine. We propose that the use of oral morphine could be a feasible and safe alternative for outpatient analgesia. Although morphine has been widely used for many years in the management of post-operative pain in both adults and children [8, 18, 19], the pharmacokinetics of morphine after oral administration are poorly characterized in healthy children undergoing day surgery. The recommended therapeutic doses of oral morphine are 0.1–0.3 mg/kg administered every 3–4 h, although post-operative use in children is often empirical rather than evidence based; plasma morphine concentrations after oral administration in children have not been studied completely [8]. Mazoit et al. [20] presented a complete pharmacokinetic and
PopPK Model for Morphine and Its Metabolites in Children
pharmacodynamic (PKPD) model for morphine and its metabolites, morphine-3-glucoronide (M3G) and morphine-6-glucoronide (M6G), in post-operative adult patients. Lotsch et al. [21] published another analytical model describing morphine and M6G plasma concentrations in healthy volunteers. For children younger than 3 years, two different population pharmacokinetic models for morphine, M3G and M6G have been published [22, 23]. Both models were developed on the basis of large data sets of intravenous morphine administration, but there were important discrepancies between the two models, as discussed at the end of this article. In the paediatric population, bodyweight is reported to be the most significant covariate for morphine clearance [22–25]. A complete analysis of the impact of weight on morphine and M3G pharmacokinetics was recently performed by Wang et al. [25] on the basis of recollected data from preterm and term neonates, infants, children, adolescents and adults, covering the entire paediatric age. We are not aware of any population pharmacokinetic model of oral morphine and its metabolites in children. Here, we present a population pharmacokinetic model able to simultaneously describe morphine, M3G and M6G plasma concentrations in children from 2 to 6 years of age after a single oral dose of morphine.
each, with five subjects from each group in each block. We used random number generation to create the randomization schedule. After the first batch was completed, it was seen that all of the serum levels were sub-therapeutic for group 1. As a result, we discontinued recruitment into this group. The oral morphine hydrochloride syrup (DoloralÒ 1 mg/mL; Atlas Laboratories, Winnipeg, MB, Canada) was administered approximately 30 min prior to surgery. Induction and maintenance of general anaesthesia was left to the discretion of the attending anaesthetist. Additional opioid analgesia was provided as required using other synthetic opioids, such as remifentanil and fentanyl. In addition, non-opioid analgesia, such as ketamine, non-steroidal anti-inflammatories (NSAIDs), paracetamol, local anaesthetic infiltration and regional blocks, were permitted. After induction of anaesthesia, a second intravenous cannula was inserted. Blood samples were drawn at 30, 60, 90, 120, 180 and 240 min. Blood samples were collected in heparinized tubes and refrigerated. Samples were centrifuged at 3000 bpm and stored at -80 °C until plasma concentrations of morphine, M6G and M3G were measured using validated high-performance liquid chromatography. More details regarding the bioanalytical assay are presented in Online Resource 1 in the Electronic Supplementary Material.
2 Methods
2.2 Data Analysis: Population Pharmacokinetic Modelling
2.1 Study Design Research approval was obtained from the University of British Columbia—Children’s & Women’s Health Centre of BC Research Ethics Board (H09-03286). The ClinicalTrials.gov identifier for the study is NCT01071499. Informed written consent was obtained from parents or guardians. This study was conducted in a tertiary care freestanding academic paediatric hospital. A multidisciplinary Data and Safety Monitoring Board (DSMB) reviewed the data and study progress at regular intervals. Children aged from 2 to 6 years with an American Society of Anaesthesiologists physical status classification of 1 or 2, undergoing procedures that typically required postoperative opioid analgesia, were invited to participate. The exclusion criteria included patients in whom morphine or oral premedication was contraindicated or patients with a documented allergy or adverse reaction to morphine. Also excluded were patients with a history of abnormal hepatic or renal function. Forty patients were enrolled and assigned to three groups receiving three different oral morphine doses: group 1, 0.1 mg/kg (n = 5); group 2, 0.2 mg/kg (n = 18); and group 3, 0.3 mg/kg (n = 17). Patients were initially block randomized to three blocks of 15 subjects
All analyses were performed with the non-linear mixedeffect modelling software NONMEM version 7.2 (Icon Development Solutions, Hanover, MD, USA) using the first-order conditional estimation (FOCE) method. The population analysis was performed simultaneously for the parent compound, morphine, and the metabolites, M3G and M6G. Inter-subject variability (ISV) was modelled using exponential functions. Additive models were used to describe residual variability for log-transformed data (morphine, M3G and M6G), which equals a proportional error on the linear scale. The disposition of morphine and its metabolites in plasma were described with compartmental models parameterized in apparent volumes of distribution, apparent inter-compartmental clearances and apparent metabolite formation clearances. Selection was made between one-, two-, and three-compartment models for morphine. Different absorption models were explored: first- and/or zeroorder rate absorptions, in combination with the presence of possible delays in the drug absorption process: discrete lag times and transit compartment models [26] were explored. Time- and/or dose-dependent parameters were also evaluated. Metabolite transformation from the plasma compartment was described using linear and non-linear
N. Velez de Mendizabal et al.
kinetics. The volume of the central metabolite compartment was parameterized as being equal to the volume of the parent central compartment. Distribution of metabolites was explored using one-, two- and three-compartment models. Once the structural and statistical model was selected, the effects of intrinsic (weight, age, height, sex) and extrinsic patient factors (batch) were evaluated. Potentially relevant patient covariates were evaluated on all of the pharmacokinetic parameters. The covariate selection criteria were (1) a statistically significant difference in the objective function value (OFV) reported by NONMEM (criterion: C6.63-point drop, p \ 0.01); or (2) a decrease in the ISV of at least 10 %.
2.3 Model Selection Criteria and Evaluation The minimum OFV provided by NONMEM, which corresponds approximately to a -2 9 log (likelihood) [-2LL], served as a criterion for model comparison during the model development process. A decrease in -2LL of 6.63 points for one additional parameter was regarded as a significant model improvement, corresponding to a p value of 0.01 for nested models. The Akaike information criterion (AIC), calculated as AIC = -2LL ? 2 9 NP, where NP is the number of parameters in the model, was used for selection among non-nested models [27]. The choice of the final model was based also on the OFV value, the precision of parameter estimates, goodness-of-fit plots and visual predictive checks (VPCs). The precision of parameter estimates, expressed as 2.5th to 97.5th percentiles, was computed from the analysis of 500 bootstrap data sets. The bootstrap analysis was performed using Perl-speaks-NONMEM [28]. Model parameter estimates were presented together with the corresponding relative standard error (RSE) as a measure of parameter imprecision, which were computed from the results obtained from bootstrap analysis. The degree of ISV was expressed as the coefficient of variation. Inspection of goodness-of-fit plots included conditional weighted residuals (CWRES) and normalized prediction distribution errors (NPDEs) [29, 30]. Model performance was evaluated with VPCs: 300 studies were simulated with the same design characteristics as those in the original study. At each time point, the 5th, 50th and 95th percentiles were calculated in every simulated study for morphine, M3G and M6G concentrations. Then, the 90 % confidence intervals from the resultant percentiles (5th, 50th and 95th) were computed and represented over time, together with the raw data. Results were stratified by dose. Predictive checks were performed using MATLAB environment (The Mathworks, Natick, MA, USA).
3 Results We developed a population pharmacokinetic model able to simultaneously describe plasma concentrations for morphine and its metabolites, M3G and M6G, in young children aged 2–6 years. The structure of the selected model is shown in Fig. 1. The corresponding differential equations (see Online Resource 2) and the NONMEM script for the selected model (see Online Resource 3) can be found in the Electronic Supplementary Material. Demographic data are shown in Table 1. The disposition of morphine, M3G and M6G in plasma was best described by a one-compartment model (Fig. 1). The inclusion of direct morphine clearance (unaccounted clearance) did not improve the model prediction. M3G and M6G metabolite formation was best described by a delay transit compartment (see Fig. 1), indicating that the metabolite formation is not an instant process and there is a delay in the appearance of these two major metabolites. ISV was found to be significant for the absorption rate (ka), the volume of distribution in the central compartment (VC/F), bioavailability (F), the formation clearance to M3G (CL2M3G/F), and the first-order rate constant for the
Fig. 1 Structural pharmacokinetic model. The boxes represent the different compartments: the morphine concentration (shaded in grey), transit compartment (CMT) for morphine metabolism, morphine-3glucoronide (M-3-G [M3G]) concentration (shaded in blue) and morphine-6-glucoronide (M-6-G [M6G]) concentration (shaded in green). The volumes of distribution of M3G and M6G (V3MG and V6MG) were parameterized as being equal to the apparent volume of distribution in the central compartment (VC/F). CL2M3G/F formation rate for M3G, CL2M6G/F formation rate for M6G, CLM3G/F elimination rate for M3G, CLM6G/F elimination rate for M6G, ka first-order absorption rate constant, kmf first-order rate constant for the metabolism transit compartment, PO oral
PopPK Model for Morphine and Its Metabolites in Children Table 1 Patient demographic characteristics Characteristic
Dose 0.1 mg/kg
0.2 mg/kg
0.3 mg/kg
n
5
18
17
Sex [n; male/female]
1/4
10/8
6/11
Age [years]a
3.57 (1.44)
4.44 (1.32)
4.21 (1.37)
16.10 (3.50)
17.97 (3.11)
17.41 (4.53)
12.82 (7.20)
13.56 (6.26)
12.85 (6.20)
Weight [kg]
a
Body mass indexa a
Average (standard deviation)
metabolism transit compartment (kmf). The g-shrinkage (%) values were 26.09 (ka), 15.36 (VC/F), 5.08 (F), 28.23 (CL2M3G/F) and 17.78 (kmf). None of the covariates evaluated in this study (weight, age, height, sex and batch) was significant in terms of the criteria specified in the methods (changes in the OFV and/or ISV). Model parameter estimates with the corresponding RSE values are presented in Table 2. The low RSE values in conjunction with the results obtained from the non-parametric bootstrap indicate that there was good precision of the estimates with no bias observed, except for the metabolite formation clearances CL2M3G/F (114.53 %) and CL2M6G/F (126.34 %). The poor precision is probably due to the short period of time (4 h) used in this study to follow the respective pharmacokinetics. Figure 2, which shows the goodness-of-fit plots for morphine, M3G and M6G, indicates that the selected
model properly describes the observations. Observations were compared with the individual model predictions (Fig. 2, left). No dose bias and mis-specifications were observed. Evaluation of the CWRES did not show any tendency or model mis-specification (Fig. 2, centre). The scatterplots of NPDEs versus the population predictions indicate that the model performed properly for the three kinds of observations: morphine, M3G and M6G (Fig. 2, right). Population and individual predictions, as well as observations for morphine, M3G and M6G, are also shown in Fig. S1 in Online Resource 4. The results from the VPCs for the morphine and its metabolites, M3G and M6G, stratified by dose, are shown in Fig. 3. The selected model is able to describe both typical profiles and data dispersion for every dependent variable (morphine, M3G and M6G) and for every dose (0.1, 0.2 and 0.3 mg/kg). The strict model selection criteria and the detailed model evaluation performed in this study demonstrate the reliability of this population pharmacokinetic model for morphine and its metabolites, M3G and M6G, in young children.
4 Discussion and Conclusion The pharmacokinetics of intravenously administered morphine and its two major active metabolites, M3G and M6G, have been widely studied in paediatrics, using different
Table 2 Population pharmacokinetic parameter estimates Parameters
Estimate (RSE %)
ISV (RSE %)
Bootstrap analysis: median (2.5–97.5th percentiles) Estimate
ISV
0.006 (0.004–0.007) 4.75 (2.409–6.964)
46.71 (29.06–62.46) 64.76 (38.74–86.83)
0.779 (0.144–3.128)
28.44 (14.75–36.57)
ka [min-1] VC/F [L]
0.006 (17.32) 4.96 (27.96)
F
1 fixed
76.09 (36.86)
CL2M3G/F [L/min]
0.776 (114.53)
29.44 (40.94)
CLM3G/F [L/min]
0.068 (15.97)
0.066 (0.051–0.086) 0.062 (0.01–0.254)
CL2M6G/F [L/min]
0.064 (126.34)
CLM6G/F [L/min]
0.042 (22.13)
kmf [min-1]
2.020 (14.20)
V3M [L]
VC/F
V6M [L] Res error for morphine [log (ng/mL)]
46.58 (44.65) 66.25 (44.21)
73.03 (50.92–94.92)
0.040 (0.027–0.057) 48.57 (47.32)
1.951 (1.528–2.450)
46.44 (28.76–64.14)
VC/F a
0.321 (38)
0.319 (0.150–0.534)
Res error for M3G [log (ng/mL)]a
0.180 (31.88)
0.177 (0.098–0.281)
Res error for M6G [log (ng/mL)]a
0.300 (38.23)
0.286 (0.139–0.518)
Parameters are listed as estimates, with 95 % confidence intervals from 500 bootstrap datasets in parenthesis. The metabolite volumes of distribution (V3M and V6M) were parameterized as being equal to the apparent volume of distribution in the central compartment (VC/F). Estimates of inter-subject variability (ISV) are shown as coefficients of variation CL2M3G/F formation rate for M3G, CL2M6G/F formation rate for M6G, CLM3G/F elimination rate for M3G, CLM6G/F elimination rate for M6G, F bioavailability, ka first-order absorption rate constant, kmf first-order rate constant for the metabolism transit compartment, M3G morphine-3glucoronide, M6G morphine-6-glucoronide, Res residual, RSE relative standard error a
Additive error model in log scale
N. Velez de Mendizabal et al.
Fig. 2 Goodness-of-fit plots of the selected population pharmacokinetic model for morphine (shown in grey) and its metabolites, morphine-3-glucoronide (M-3-G [M3G]; shown in blue) and morphine-6-glucoronide (M-6-G [M6G]; shown in green). The solid lines show identity lines in the first two columns and zero lines in the last
two columns. CWRES conditional weighted residuals, DV observed concentration, IPRED individual model-predicted concentration, NPDE normalized prediction distribution error, PRED population model-predicted concentration
approaches and modelling techniques (e.g. non-compartmental and compartmental models). For children younger than 3 years, two different population pharmacokinetic models for morphine and its metabolites have been published [22, 23]. Krekels et al. [24] published a paper comparing these models, as there are discrepancies between them. Both models were developed on the basis of large data sets of intravenous morphine administration. For children younger than 3 years, multiple covariates were identified as descriptors of variability in the pharmacokinetic profiles between patients. We tried to use both models to describe the data presented here. However, neither of these models was able to appropriately model the data we generated with oral morphine use. This might also be explained by the age differences between the studies: 0–3 years in the aforementioned studies [22, 23] and 2–6 years in this study. Bouwmeester et al. [23] identified a direct morphine clearance (unaccounted clearance); however, in our case, the inclusion of this process did not significantly improve the fit. This could be explained by the short period of time (only 4 h) for sample collection after administration. However, Knibbe et al. [22] concluded that none of the morphine elimination pathways other than the transformation to M3G and M6G play a significant role in morphine clearance. Morphine elimination though these
other routes was not found to be significantly different from zero. Knibbe et al. [22] described morphine pharmacokinetics with a two-compartment model, although in the study by Bouwmeester et al. [23], the inclusion of this second compartment (distribution to peripheral tissue) was not significant. In the model presented here, morphine distribution to other tissues was not identified. More recently, Krekels et al. [31] also published a review exploring the possible advantages and disadvantages of the different data analysis techniques that have been applied, with a specific focus on the accuracy of morphine clearance predictions by paediatric pharmacokinetic models. Other population pharmacokinetic models have been published in order to describe the pharmacokinetic profiles of morphine and its metabolites, but these have been in adults. Mazoit et al. [20] presented a very complete PKPD model of morphine, M3G and M6G in post-operative adult patients who received morphine as an intravenous titration, followed by intramuscular administration postoperatively. Lotsch et al. [21] published another model analysis to describe morphine and M6G plasma concentrations in healthy adult volunteers (aged 23–30 years) after intravenous bolus injection. Interestingly, they also proposed a transit compartment identifying the delay of the appearance of M6G in plasma.
PopPK Model for Morphine and Its Metabolites in Children Fig. 3 Visual predictive checks: results from 1000 simulated studies. The shaded areas represent the 90 % prediction intervals of the 5th, 50th and 95th percentiles. The dashed lines correspond to the observed medians, and the solid lines correspond to the 5th and 95th percentiles of the observed data. The circles are observed data. Morphine is shown in grey, morphine-3-glucoronide (M-3-G [M3G]) is shown in blue and morphine-6glucoronide (M-6-G [M6G]) is shown in green. The results are stratified by morphine dose: a 0.1 mg/kg, b 0.2 mg/kg and c 0.3 mg/kg
In the paediatric population, body weight has been reported to be the most significant covariate for morphine clearance [22–25]. However, in this study, the inclusion of this covariate did not significantly improve the fit. This is probably due to the short observation period (4 h) after oral administration. The formation clearances to M3G and M6G were calculated with poor precision, and other clearance pathways were estimated as zero. Here, we present the first population pharmacokinetic model for morphine and its metabolites, M3G and M6G, in children aged 2–6 years after a single oral dose. The low RSE values in conjunction with the results obtained from the non-parametric bootstrap indicate that there was very good precision of the estimates and no bias observed, except for the metabolite formation clearances CL2M3G/ F and CL2M6G/F. The poor precision of these formation clearance values was likely due to the short period of time used to follow the respective pharmacokinetics. Longer studies have to be performed in order to improve confidence in these estimated values. Although a good option might have been to focus just on the morphine
pharmacokinetic analysis, with the aim being to estimate the morphine clearance with better precision, it was decided to incorporate the three compounds into the study for several reasons. First, the VPCs capture the kinetics and data dispersion very well for the three compounds. This good result, in combination with the good precision of the rest of the parameters, makes the presented model a reliable one for describing morphine and metabolite pharmacokinetics for the first 4 h after a single oral dose. However, this may not accurately describe the kinetic profiles after 4 h. Highlighting the limitations of this study here helps to justify the design of longer studies in the future. Moreover, this study highlights the necessity of reviewing the therapeutic dose recommendation for oral doses of morphine (0.1–0.3 mg/kg administered every 3–4 h). This study provides a background to further evaluate and model morphine following an oral dose in this age group. Longer follow-up time and the incorporation of other important covariates, such as phenotype, will add value and will help us overcome the present limitations.
N. Velez de Mendizabal et al.
References 1. Finley GA, McGrath PJ, Forward SP, McNeill G, Fitzgerald P. Parents’ management of children’s pain following ‘minor’ surgery. Pain. 1996;64:83–7. 2. Jonas DA. Parent’s management of their child’s pain in the home following day surgery. J Child Health Care. 2003;7:150–62. 3. Williams DG, Patel A, Howard RF. Pharmacogenetics of codeine metabolism in an urban population of children and its implications for analgesic reliability. Br J Anaesth. 2002;89:839–45. 4. Williams DG, Hatch DJ, Howard RF. Codeine phosphate in paediatric medicine. Br J Anaesth. 2001;86:413–21. 5. Sindrup SH, Brosen K. The pharmacogenetics of codeine hypoalgesia. Pharmacogenetics. 1995;5:335–46. 6. Sistonen J, Sajantila A, Lao O, Corander J, Barbujani G, Fuselli S. CYP2D6 worldwide genetic variation shows high frequency of altered activity variants and no continental structure. Pharmacogenet Genomics. 2007;17:93–101. 7. Daly AK, Brockmoller J, Broly F, Eichelbaum M, Evans WE, Gonzalez FJ, Huang JD, Idle JR, Ingelman-Sundberg M, Ishizaki T, et al. Nomenclature for human CYP2D6 alleles. Pharmacogenetics. 1996;6:193–201. 8. Berde CB, Sethna NF. Analgesics for the treatment of pain in children. N Engl J Med. 2002;347:1094–103. 9. Marez D, Legrand M, Sabbagh N, LoGuidice JM, Spire C, Lafitte JJ, Meyer UA, Broly F. Polymorphism of the cytochrome P450 CYP2D6 gene in a European population: characterization of 48 mutations and 53 alleles, their frequencies and evolution. Pharmacogenetics. 1997;7:193–202. 10. Cascorbi I. Pharmacogenetics of cytochrome p4502D6: genetic background and clinical implication. Eur J Clin Invest. 2003;33(Suppl 2):17–22. 11. Ciszkowski C, Madadi P, Phillips MS, Lauwers AE, Koren G. Codeine, ultrarapid-metabolism genotype, and postoperative death. N Engl J Med. 2009;361:827–8. 12. de Wildt SN, Koren G. Re: apnea in a child after oral codeine: a genetic variant—an ultra-rapid metabolizer [corrected]. Paediatr Anaesth. 2008;18:273–4 (author’s reply 275–276). 13. Koren G, Cairns J, Chitayat D, Gaedigk A, Leeder SJ. Pharmacogenetics of morphine poisoning in a breastfed neonate of a codeine-prescribed mother. Lancet. 2006;368:704. 14. Madadi P, Koren G, Cairns J, Chitayat D, Gaedigk A, Leeder JS, Teitelbaum R, Karaskov T, Aleksa K. Safety of codeine during breastfeeding: fatal morphine poisoning in the breastfed neonate of a mother prescribed codeine. Can Fam Physician. 2007;53:33–5. 15. Madadi P, Ross CJ, Hayden MR, Carleton BC, Gaedigk A, Leeder JS, Koren G. Pharmacogenetics of neonatal opioid toxicity following maternal use of codeine during breastfeeding: a case–control study. Clin Pharmacol Ther. 2009;85:31–5. 16. Gasche Y, Daali Y, Fathi M, Chiappe A, Cottini S, Dayer P, Desmeules J. Codeine intoxication associated with ultrarapid CYP2D6 metabolism. N Engl J Med. 2004;351:2827–31.
17. Kuehn BM. FDA: no codeine after tonsillectomy for children. JAMA. 2013;309:1100. 18. Lugo RA, Kern SE. Clinical pharmacokinetics of morphine. J Pain Palliat Care Pharmacother. 2002;16:5–18. 19. Choonara IA, McKay P, Hain R, Rane A. Morphine metabolism in children. Br J Clin Pharmacol. 1989;28:599–604. 20. Mazoit JX, Butscher K, Samii K. Morphine in postoperative patients: pharmacokinetics and pharmacodynamics of metabolites. Anesth Analg. 2007;105:70–8. 21. Lotsch J, Skarke C, Schmidt H, Liefhold J, Geisslinger G. Pharmacokinetic modeling to predict morphine and morphine-6glucuronide plasma concentrations in healthy young volunteers. Clin Pharmacol Ther. 2002;72:151–62. 22. Knibbe CA, Krekels EH, van den Anker JN, DeJongh J, Santen GW, van Dijk M, Simons SH, van Lingen RA, Jacqz-Aigrain EM, Danhof M, Tibboel D. Morphine glucuronidation in preterm neonates, infants and children younger than 3 years. Clin Pharmacokinet. 2009;48:371–85. 23. Bouwmeester NJ, Anderson BJ, Tibboel D, Holford NH. Developmental pharmacokinetics of morphine and its metabolites in neonates, infants and young children. Br J Anaesth. 2004;92:208–17. 24. Krekels EH, van Hasselt JG, Tibboel D, Danhof M, Knibbe CA. Systematic evaluation of the descriptive and predictive performance of paediatric morphine population models. Pharm Res. 2011;28:797–811. 25. Wang C, Sadhavisvam S, Krekels EH, Dahan A, Tibboel D, Danhof M, Vinks AA, Knibbe CA. Developmental changes in morphine clearance across the entire paediatric age range are best described by a bodyweight-dependent exponent model. Clin Drug Investig. 2013;33:523–34. 26. Savic RM, Jonker DM, Kerbusch T, Karlsson MO. Implementation of a transit compartment model for describing drug absorption in pharmacokinetic studies. J Pharmacokinet Pharmacodyn. 2007;34:711–26. 27. Ludden TM, Beal SL, Sheiner LB. Comparison of the Akaike information criterion, the Schwarz criterion and the F test as guides to model selection. J Pharmacokinet Biopharm. 1994;22:431–45. 28. Lindbom L, Ribbing J, Jonsson EN. Perl-speaks-NONMEM (PsN)—a Perl module for NONMEM related programming. Comput Methods Progr Biomed. 2004;75:85–94. 29. Hooker AC, Staatz CE, Karlsson MO. Conditional weighted residuals (CWRES): a model diagnostic for the FOCE method. Pharm Res. 2007;24:2187–97. 30. Comets E, Brendel K, Mentre F. Computing normalised prediction distribution errors to evaluate nonlinear mixed-effect models: the NPDE add-on package for R. Comput Methods Progr Biomed. 2008;90:154–66. 31. Krekels EH, Tibboel D, Danhof M, Knibbe CA. Prediction of morphine clearance in the paediatric population: how accurate are the available pharmacokinetic models? Clin Pharmacokinet. 2012;51:695–709.
