- Email: [email protected]
Cerebral kinetics of oxycodone in conscious sheep
Cerebral Kinetics of Oxycodone in Conscious Sheep HANNE H. VILLESEN,1 DAVID J.R. FOSTER,2,4 RICHARD N. UPTON,3 ANDREW A. SOMOGYI,4,5 ALLISON MARTINEZ,3 CLIFF GRANT3 1
Department of Pharmacology and Pharmacotherapy, The Danish University of Pharmaceutical Sciences, Copenhagen 2100, Denmark 2
School of Pharmacy and Medical Sciences, University of South Australia, Adelaide 5000, Australia
3
Department of Anaesthesia and Intensive Care, Royal Adelaide Hospital and The University of Adelaide, Adelaide 5005, Australia
4
Department of Clinical and Experimental Pharmacology, The University of Adelaide, Adelaide 5005, Australia
5
Department of Clinical Pharmacology, Royal Adelaide Hospital, Adelaide 5005, Australia
Received 10 October 2005; revised 27 February 2006; accepted 1 March 2006 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20632
ABSTRACT: Oxycodone is an opioid analgesic that is administered orally or parenterally. The time-course of opioid action is a function of the systemic kinetics of the opioid, and the rate and extent of its entry into the brain and central nervous system. The latter is incompletely understood for oxycodone. Therefore, the cerebral kinetics of oxycodone was quantified using a conscious chronically instrumented sheep preparation. Five sheep were administered oxycodone as intravenous infusions (30 mg over 4 min). Using hybrid physiologically based kinetic models, cerebral kinetics was estimated from arteriosagittal sinus concentration gradients and cerebral blood flow (CBF). A two-compartment membrane-limited model best described the data. The volume of the first brain compartment was 35.4 mL with a half-life of equilibrium of 0.6 min. The brain:blood equilibration of oxycodone was relatively slow (half-life of 7.2 min), with a large deep cerebral distribution volume (222.8 mL) for the second compartment and a moderate membrane permeability of 54.8 mL/min, which exceeded the nominal CBF (40 mL/min). Drug retention in the brain was 1.3% after 45 min. In conclusion, pharmacokinetic modelling of oxycodone showed a delayed equilibration between brain and blood of a nature that would be affected by changes in both CBF and blood brain barrier permeability. ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 95:1666–1676, 2006
Keywords: oxycodone; pharmacokinetics; distribution; blood brain barrier; blood flow; physiological model; sheep
INTRODUCTION Oxycodone is an opioid analgesic usually administered for the treatment of moderate to severe pain for cancer pain or following surgery. It is functionally similar to morphine, but has a higher bioavailability after oral administration.1 The Correspondence to: Hanne H. Villesen (Telephone: (þ45) 35 30 6438; Fax: (þ45) 35 30 6050; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 95, 1666–1676 (2006) ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association
1666
effects of oxycodone, like those of morphine, lag behind its blood concentration-time profile.2–5 As reported for other opioids and anaesthetics, this delay in effect has been proposed to be similar in magnitude to the time required for equilibration of blood and central nervous system (CNS) concentrations.6 For some opioids, particularly when used parenterally, the delay in cerebral equilibration means that the cerebral kinetics of the opioid become an important factor in dictating their optimal clinical use.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
CEREBRAL KINETICS
The factors affecting the rate of cerebral uptake of opioids are complex. The ability of a drug to cross the blood brain barrier (BBB) is determined mainly by the degree of ionisation at physiological pH, molecular size, lipophilicity, affinity for efflux (i.e. P-glycoprotein), and uptake carrier mechanisms. Furthermore protein binding, cerebral blood flow (CBF), uptake into other tissues, cerebral metabolism and clearance rate also influence the net uptake.7 Different opioids are affected by these factors to varying extents. Pethidine, for example, has a large cerebral distribution volume and flowlimited kinetics, indicating that its cerebral kinetics are highly influenced by changes in CBF.8 In contrast, morphine has a small cerebral distribution volume and moderate BBB permeability,9 which might be due to P-glycoprotein mediated efflux transport.10,11 This suggests that the cerebral kinetics of morphine is influenced less by changes in CBF and more by the activity of efflux transport. Where oxycodone fits in this spectrum of cerebral kinetic behaviour is currently unknown. The aim of this study was therefore to elucidate the cerebral kinetics of oxycodone after a short intravenous infusion in a conscious sheep preparation. This preparation has previously been used to determine the cerebral kinetics of other opioids (e.g. alfentanil, pethidine, fentanyl, morphine, (R)- and (S)-methadone and levo-alphaacetyl-methadol (LAAM)). As a secondary aim, the cerebral kinetics of oxycodone was compared with the reported cerebral kinetics of these opioids using the same experimental study design.
MATERIALS AND METHODS Animal Preparation Sheep were chosen as an experimental animal as the relative perfusion of their brain and the control of CBF are similar to that in humans,12,13 afford good access to appropriate blood vessels and tolerate chronic instrumentation well. All experimental procedures were approved by the Animal Ethics Committee of the University of Adelaide and were conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. The surgical procedures for preparing the sheep and the details about the experiment has been described previously.8,13 Briefly, female Merino sheep of similar ages and body mass (mean weight 44.3 5.0 kg) were used. Instrumentation was DOI 10.1002/jps
1667
performed under general anaesthesia. Catheters were chronically implanted with their tips in the abdominal aorta (for sampling of arterial blood), in the right atrium (for drug administration and measurement of cardiac output (CO)), and in the dorsal sagittal sinus (for cerebral venous blood sampling).12 A Doppler transducer was placed over the sagittal sinus using a previously validated method to quantify the relative changes in CBF.12,13 After recovery from anaesthesia, the sheep were housed in metabolic crates or floor pens for a least a day before the study and their catheters maintained with a saline/heparin lock (0.9%/50 i.u./mL). Study Design Five sheep received oxycodone hydrochloride (MW 405,9 g/mol) (Boots, Nottingham, UK) 0.6 mg/kg as a 4-min i.v infusion. On the study day the instrumented sheep was placed in nonweight-bearing sling inside a metabolic crate and prepared for physiological measurements and blood sampling. After 5 min of baseline measurements, the infusion was commenced at time zero. Oxycodone hydrochloride 1 mg/mL was prepared in 0.9% isotonic saline by the Department Pharmacy of the Royal Adelaide Hospital prior to each study day. Infusion rate was adjusted accordingly to the weight of each sheep. The following procedures were performed during each study. Blood Sampling Prior to drug infusion, blank blood was sampled for use in the calibration curves for the quantification of oxycodone. Simultaneous blood samples of 2 mL volume were collected from the arterial and sagittal sinus catheters at 0.5, 1, 1.5, 2, 3, 4, 4.5, 5, 5.5, 6, 8, 10, 15, 20, 30, 45, 60 and 75 min after start of drug infusion. The blood samples were collected in 10 mL tubes with heparin as an anticoagulant. After the study the blood was centrifuged at 3250 rpm for 10 min and plasma was separated. The blood samples were stored at room temperature for <90 min before separation of plasma. The plasma samples were stored at 208C until analysis. Blood Gas, Cardiac Output and Cerebral Blood Flow Measurements Haemodynamic data (mean arterial blood pressure (MAP) and the CBF index) were continuously monitored using a computerised data JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1668
VILLESEN ET AL.
acquisition system. Relative changes in CBF were quantified using Doppler flow probe and a flow meter (Bioengineering, University of Iowa, IA) and MAP was recorded continuously via a pressure transducer on one of the arterial catheters connected to an analogue-to-digital card (Metrabyte DAS 16-G2) and a personal computer (486 based IBM compatible). Recording of these parameters was commenced 5 min before drug infusion (baseline) and throughout the study period. CO was measured in triplicate immediately prior to the oxycodone infusion, and at 4, 10, 30 and 60 min after the start of the infusion, using a thermodilution method.14 The values were averaged to obtain the mean at each time point. Additional arterial blood samples were taken at 0, 4, 10 and 30 min for blood gas measurements (ABL System 625, Radiometer, Sweden). High Performance Liquid Chromatography for Determination of Oxycodone The plasma samples containing oxycodone were assayed using a validated high performance liquid chromatography (HPLC) assay. The HPLC system comprised a LC-10AT pump (Shimadzu, Kyoto, Japan), a SIL-10A auto injector (Shimadzu), a RCM 8 10 compression module containing a Nova-Pak C-18 4 mm column (8 100 mm Radial-PakTM Cartridge, Waters Associates, Inc. Milford, MD) with a Guard C-18 5u precolumn (7.5 4.6 mm, Waters Associates) and a SPD-M10A diode array detector (Shimadzu) set at 210 nm. The mobile phase consisted of 28% acetonitrile, 0.1 M sodium dihydrogen orthophosphate and 230-mg/L lauryl sulphate. The final pH was adjusted to 2.4 with orthophosphoric acid. The flow rate was 1.25 mL/min giving retention times for hydromorphone (internal standard) and oxycodone at 7 and 14 min, respectively. All chemicals were of analytical or HPLC grade. Oxycodone plasma samples were processed using liquid–liquid extraction as previously described by Menelaou et al.15 with modifications. Briefly, in 10 mL flat bottom plastic tubes 1 mL of plasma and 70 ml of 10 mg/mL hydromorphone (internal standard) were alkalinised with 300 ml of 20% w/v sodium carbonate. Samples were extracted with 4 mL of dichloromethane for 10 min on a rotary mixer. The samples were centrifuged at 3250 rpm for 10 min and the upper aqueous layer was aspirated to waste. The remaining organic phase was transferred to clean plastic JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
tube containing 200 ml of 0.1 M hydrochloric acid. The samples were rotary-mixed for 10 min and centrifuged for 10 min at 3250 rpm. The acid bubble was transferred to an auto injector vial and 50 ml was injected on the HPLC system. Samples were injected in random order along with samples containing standards (0.05–5 mg/mL, n ¼ 8 standards). Quality controls (QC) were included in each run to monitor assay performance and were prepared in duplicate at three different concentrations: low (LQC, 0.15 mg/mL), medium (MQC, 1 mg/ mL) and high (HQC, 3.5 mg/mL). Assay Validation All assays were calibrated using eight-point standard curves prepared in plasma separated from blood taken from the same animal before drug administration. As a validation procedure, inter-assay variability was monitored with QC prior to assaying the oxycodone plasma samples. The slope of the calibration curves (mean SD, %coefficient of variation (%CV)) was 2.2 0.2, %CV 0.7 (n ¼ 6). The mean r2-value was 0.999 0.001, %CV 0.009 (n ¼ 6 assay days), while interassay accuracy and precision was 97.3 4.4, %CV 4.5 (n ¼ 11) (LQC), 99.3 3.3, %CV 3.3 (n ¼ 12) (MQC) and 101.4 4.7, %CV 4.7 (n ¼ 12) (HQC). Similarly, intra-assay (replicate samples) accuracy and precision were 102.4 5.2, %CV 5.1 (n ¼ 6) (LQC), 103.8 2.7, %CV 2.6 (n ¼ 6) (MQC) and 102.8 2.1, %CV 2.1 (n ¼ 6) (HQC). The assay was both precise and accurate at the limit of quantification (0.05 mg/mL), with inter-assay accuracy and precision being 106.3 12.1%, %CV 11.4 (n ¼ 8). The extraction recovery was 72.2 11.7%, %CV 16.2 (n ¼ 4). Pharmacokinetic Analysis All kinetic analyses were based on the mean concentrations of all sheep at each time points. This reduces the influence of random fluctuation in concentration on model fit. A hybrid modelling of kinetics was used to define the cerebral kinetics of oxycodone. Empirical forcing functions were used to represent inputs into the brain (Cin), and curve-fitting of the output of the brain (Cout) to physiologically realistic models used to determine model parameters.16 The forcing functions are essentially a device to interpolate the available data points, as solving the models require continuous functions. Provided they fit the available data, their form has no bearing on the discriminaDOI 10.1002/jps
CEREBRAL KINETICS
tion between various models of organ kinetics. Arterial input concentrations to the brain were interpolated using exponential forcing functions. If CBF showed statistically significant changes from baseline over time, these were interpolated in the forcing functions.12 The sagittal sinus output concentrations from the brain were curve fitted to determine model parameters using a curve fitting modelling software (Scientist for Windows, version 2.01, Micromath Scientific Software, UT).
1669
Four different kinetic models of the brain were fitted to the measured sagittal sinus oxycodone concentrations. The basic forms of the equations describing these models have been published previously.9,17 They are repeated below for convenience and shown in Figure 1. Cin and Cout are the afferent arterial and efferent sagittal sinus drug concentrations of the brain, respectively, and Q is CBF: A null model that tested the hypothesis that there was no oxycodone concentration gradient across the brain: Cout ¼ Cin
ð1Þ
A single flow-limited compartment defined by a single distribution volume and CBF: V1 dCout =dt ¼ Q ðCin Cout Þ
ð2Þ
A single flow-limited compartment with an apparent first-order loss (PSloss) representing either deep distribution or metabolism: V1 dCout =dt ¼ Q ðCin Cout Þ PSloss Cout ð3Þ A two-compartment membrane-limited model with a permeability term (PS) representing distribution into a deep compartment: V1 dCout =dt ¼ Q ðCin Cout Þ þ PS ðC2 Cout Þ V2 dC2 =dt ¼ PS ðCout C2 Þ ð4Þ
Figure 1. Graphical representation of kinetic models: (A) flow limited model (Eq. 2), (B) single flow-limited compartment with an apparent first order loss of drug (Eq. 3), and (C) a two-compartment membrane limited model with a permeability term describing distribution into a deep compartment (Eq. 4). DOI 10.1002/jps
V1 is the volume of the first compartment of the brain (which includes blood). The nominal volume of this vascular compartment in a sheep brain is 4.5 mL (5% of 90 mL). V2 and C2 are the volume of, and concentration in, the second compartment of the brain (if appropriate). The permeabilitysurface area coefficient (PS) is a term describing loss or exchange of drug from the first compartment. In the case of Eq. (3), PS represents a unidirectional loss of drug, such as metabolism or deep distribution, which is essentially irreversible. In the case of Eq. (4), PS represents the effective inter-compartmental clearance of drug between the two compartments. Selection criteria for the final model were a high model selection criterion (MSC) value (>3),16 a statistically (p < 0.05) significant improvement of the fit as determined by the F-ratio test based upon the residual sum of squares, and an even distribution of residuals. The data were also analysed using a non-linear mixed effect approach to establish the intersubject variability in parameter estimates as a JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1670
VILLESEN ET AL.
supplemental analysis. The best model of cerebral kinetics inferred from analysis of pooled data was fitted to population data (i.e. all individual sheep) using NONMEM (version V, level 1.1), with linear interpolation of arterial and CBF inputs for individual sheep, and FOCE with interaction analysis. Parameters for the brain kinetic model (V1, PS and V2) were assigned an exponential distribution across the population, and a combined additive and proportional residual error model was used. Calculated Variables Model parameters were used to calculate secondary variables to facilitate comparison with other opioids and literature values. Brain equilibration times for oxycodone were calculated by using the final cerebral kinetic model and best parameter values to simulate the time course of the brain concentrations for a step increase in the afferent arterial blood concentration from 0 to 1. The times required for the brain concentration to reach 50 and 95% of the arterial blood concentration were recorded. The apparent permeability of the BBB (PS) was compared with the CBF. The apparent brain:blood partition coefficient (R) was calculated from V2 and a nominal real volume of 63 mL for the region of the brain drained by the sagittal-sinus catheter (70% of 90 mL, whole brain).12 Drug retention (R%) in the brain was calculated as follows: 1 AUCsag 100% ð5Þ R% ¼ AUCart
measured at 1 s intervals, a 10 s moving average from each individual animal was calculated at each of the pharmacokinetic blood sampling timepoints. CO, MAP, arterial carbon dioxide tension (PaCO2) and arterial oxygen tension (PaO2), and relative changes from baseline (100%) in CBF were calculated as mean and 95% confidence intervals (CI). Statistically significant changes from base line for pharmacodynamic data were tested using one-way repeated measures ANOVA with Bonferroni correction. Time was the repeated measure. A a-value of 0.05 was set for statistically significant differences. All data are reported as mean SD or mean (95% CI).
RESULTS Pharmacodynamic Data CBF, CO and MAP changed from baseline during oxycodone infusion. Initially after the start of the infusion CBF decreased for a short period of time, thereafter it increased to 6.9 10.6% above baseline (44 mL/min) during the infusion and continued to increase to a maximum of 22.6 14.1% (p < 0.0001) of baseline at 5.5 min, where it remained for the rest of the study (Fig. 2). The mean baseline CO was 5.4 0.8 L/min. Neither CO nor MAP changed significantly from baseline
where the total arterial (AUCart) and sagittal sinus (AUCsag) area under the blood concentration-time curves to 45 min were calculated using the trapezoidal rule. The drug retention indicates the amount of drug that had entered the brain via the arterial blood, but had not left the brain via the efferent venous blood by the end of the period of interest. Retention in the brain could be the result of metabolism or deep distribution from which efflux is relatively slow.
Statistical Analysis GraphPad Prism (GraphPad Software version 4.02 for Windows, GraphPad Software, San Diego, CA) was used for all statistical analyses. In the case of CBF and MAP, which were continuously JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
Figure 2. Mean (open circles) and 95% confidence intervals (dotted lines) of cerebral blood flow expressed as a percentage of baseline values in five sheep. The solid line is baseline value. DOI 10.1002/jps
CEREBRAL KINETICS
Table 1.
Mean SD
Mean Baseline Values for Cardiovascular and Blood Gas Measurements CO (L/min)
MAP (mm Hg)
PaO2 (mm Hg)
PaCO2 (mm Hg)
O2 sat (%)
PH
HCO 3 (mmol/L)
5.4 0.8
92.2 18.3
108.0 10.6
30.8 1.9
101.2 1.5
7.5 0.02
23.0 2.2
(p ¼ 0.75). Additionally no significant changes from baseline of PaO2 (p ¼ 0.40) or PaCO2 (p ¼ 0.54) could be demonstrated at any time point during the study. The dose of oxycodone was not associated with significant respiratory or cardiovascular side effects. In all sheep, oxycodone was noted to produce a mild degree of dysphoria (agitation, mouthing of crate, bleating). Table 1 shows the mean baseline values for CO, MAP and blood gas measurements. The recorded changes in CBF were interpolated and incorporated in the input forcing functions used in the subsequent modelling. Pharmacokinetic Data Figure 3 shows the mean observed plasma oxycodone HCl concentrations. During the infusion period (first 4 min) a substantial extraction
Figure 3. Measured arterial (open circles) and sagittal-sinus (open invert triangles) blood concentrations of oxycodone HCl in five sheep. Data are shown as mean at each time point. Dotted and solid lines are the best fits of the kinetic model chosen, which was the membranelimited model (Table 2). DOI 10.1002/jps
1671
across the brain was seen. Peak arterial concentration occurred at 4 min and was 0.72 mg/mL (95% CI 0.49–0.95 mg/mL), and sagittal sinus Cmax was 0.38 mg/mL (95% CI 0.23–0.53 mg/mL) and occurred at 4.5 min. The differences in the concentration across the brain are shown in Figure 4. The concentrations of oxycodone HCl were below the limit of quantification after 45 min in all sheep; consequently kinetic analysis was only conducted using concentration-time data for the 0–45 min period. The results of the kinetic modelling for the various models are summarised in Table 2. The null model, the flow-limited as well as the flowlimited with a loss, gave equally poor descriptions of the data. The permeability term was best described by transport of oxycodone into a deep compartment as in the membrane-limited model (higher MSC). The volume of the first compartment (nominally the vascular space between the sampling catheters) was 35.4 mL and the t½ of
Figure 4. Mean (open circles) and 95% CI (dotted lines) of the individual concentration differences across the brain (arterial-sagittal sinus) in five sheep. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1672
VILLESEN ET AL.
Table 2. Model Parameter Estimated for Brain Kinetics Model
MSC
V1 (mL)
PS (mL/min)
Pooled data Null Flow Flow þ loss Membrane
1.20 0.71 1.73 3.32
182.4 (26.2) 69.8 (11.7) 35.4 (8.0)
25.0 (3.6) 54.8 (7.8)
Population data Membrane Parameter Inter-animal variability (%CV)
V2 (mL)
222.8 (23.4)
— 28.4 (14%) 12.8% (168%)
54.4 (8.5%) —
236 (9.5%) —
For the pooled data analysis, data are mean (SD) given by the curve fitting modelling program, with SD in this case indicating the reliability of the parameter estimate. The MSC is the model selection criteria—the higher the number, the better the fit. For the population data analysis, data are the value of the population estimate, with data in parentheses in this case indicating the % standard error of the parameter estimate indicating the reliability of the parameter estimate.
equilibrium through this compartment was 0.6 min. The permeability (PS) of oxycodone across the membrane barrier (nominally the BBB) was 54.8 mL/min, which was comparable with the CBF (baseline 40 mL/min). Drugs for which membrane permeability is of similar magnitude to CBF generally show characteristics of both bloodflow and membrane limitation in their cerebral uptake.16 The final volume of the deep compartment (nominally brain parachyma) was 222.8 mL. The times to reach 50 and 95% of equilibration were 7.2 and 29.5 min, respectively. Drug retention (R%) in the brain was negligible (1.3% after 45 min). Assuming a nominal real volume of the region of the brain drained by the sagittal sinus catheter of 63 mL (70% of 90 mL), the volume of the deep compartment equates to a brain:blood partition coefficient of approximately 3.5. Initial analyses found that inter-animal variability PS and V2 was very small (<0.1%), and an additive error component was not supported. As a consequence inter-animal variability was removed from these parameters (without a decrease in model fit) and a proportional error model only was employed to obtain the most parsimonious model. In the final model, the standard error of the population parameters estimates is acceptable, with a small residual error (% CV 18). Standard diagnostic plots of both the population model and individual Bayesian fits, supported an excellent fit to the data. The results of the mixed effect analysis using NONMEM is shown in Table 2. The population parameters of the model did not differ greatly from the results of the pooled analysis. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
DISCUSSION Cerebral Pharmacokinetic Data As oxycodone is often used in a similar clinical role as morphine, it is of interest to compare the two opioids. Like morphine, a membrane-limited model was found to best describe the cerebral kinetics of oxycodone. However, in comparison with morphine,9 the permeability of oxycodone across the BBB was relatively high (54.8 vs. 7.44 mL/min, respectively), as was its cerebral distribution volume (222.8 vs. 92.4 mL, respectively). The relatively fast permeation of oxycodone through the blood-brain barrier was followed by a relatively short equilibration half-life of the deep compartment (7.2 min) in comparison with morphine (10.3 min).9 The population pharmacokinetic parameters of the model did not differ greatly from the results of the pooled analysis. Reasons to the very small inter-animal variability in kinetic parameters might include the highly controlled nature of the experiment and data collection, such that extraneous variability was essentially removed form the data. For example, the inputs (arterial blood concentrations and CBF) for each animal were quantitated and described, the sheep were all from the same flock, and the assay was reasonably precise and accurate. These data suggest that there was minimal inter-individual variability in the cerebral kinetic parameters of oxycodone in the sheep studied, and supports the process of using pooled data to distinguish between various models. DOI 10.1002/jps
CEREBRAL KINETICS
Brain compartment volumes have not been normalized for mass of tissue, because the sheep model employed in the present study is representative of human cerebral kinetics compared species, due to a remarkably similar CBF/g tissue (0.44 mL/g tissue) to humans (0.54 mL/g tissue). Analgesia has been described to be achieved faster with oxycodone than with morphine after iv administration in both animal (15 and 30 min, respectively)18 and human studies (28 4.8 and 46 6.9 min, respectively).19 In the present study it was shown that this faster onset might be explained by a faster equilibration between blood and brain for oxycodone. Morphine and oxycodone are believed acting on different classes of opioid receptors with different affinities and different kinetics. This might represent part of the explanation for the different antinociceptive effect seen between the two opioids, however, it still remains controversial whether or not morphine and oxycodone do possess different receptor binding profiles.
Role of Transporters Another difference to be considered when comparing the two opioids is the role active transporters may play in permeation of the BBB. Many opioids, such as morphine and methadone, are substrates for P-glycoprotein transporters,10 which influence the uptake into the brain and thereby the effect.20 However, based on results from studies in rats oxycodone does not seem to be a substrate for P-glycoprotein.21 This suggests that only physicochemical factors such as lipophilicity, molecular size, ionisation degree, (all of which influence passive diffusion) govern the rate of cerebral uptake of oxycodone. It is worthwhile to consider the interplay of these factors, and to place them in context by comparison of oxycodone with other opioids.
Polarity, Lipophilicity and Cerebral Uptake The physicochemical properties, such as pKa, lipophilicity and water solubility of oxycodone are similar to those of morphine.22,23 Oxycodone is a weakly basic drug with a pKa value of 8.9, thus it is 97% ionised at physiological pH. Like morphine, oxycodone is a hydrophilic molecule and thus possesses a low lipophilicity; the octanol:buffer distribution coefficient is 1.27, 1.64 and 706 for morphine, oxycodone and fentanyl, respectively at 378C and pH 7.4.24 DOI 10.1002/jps
1673
Only one study has so far included oxycodone studies of opioid dose-response relationships for antinociceptive effect with regard to polarity of the substance. Oxycodone and morphine, along with other opioids, were given intrathecally to male Porton rats. The onset of antinociception was faster for oxycodone than for morphine, but only slightly slower than seen after administration of the more lipophilic opioid, fentanyl in both the Hot Plate (6.2, 24.9 and 5.3 min, respectively) and the Tail–Flick tests (6.3, 16.1 and 5.3 min, respectively),24 which is consistent with the results from human studies described by Po¨yhia¨ and Kalso19 and Leow and Smith.25 The above indicates that lipophilicity and polarity are important factors for the distribution of drugs within the CNS. Bernards and coworkers have stated that permeability increases with increasing lipophilicity until a certain octanol:buffer distribution coefficient,26 and it has been demonstrated that despite the large difference in octanol:buffer distribution coefficient between morphine and fentanyl, the permeability of fentanyl is only slightly higher than the permeability of morphine.27 This parabolic relationship between lipophilicity and permeability is also seen for other physiological membranes such as skin and gastro-intestinal mucosa. Different factors such as molecular weight have been incorporated into the fitting of the relationship and found to improve the correlations seen.28,29 A summary of the brain kinetics of different opioids is shown in Table 3 along with relevant physico-chemical parameters. With exception of methadone and LAAM, the proposed relationship applies for the opioids shown in Table 3. For fentanyl and pethidine, the time to reach 50% of equilibrium of the deep tissue compartment (tequil50%) is closer to that of oxycodone even though there is a large difference in the log D-values (2.8, 1.7 and 0.9, respectively). This could be explained by the biphasic relationship seen between lipophilicity and permeability. From Figure 5 it can be seen that there is an inverse bell shaped relationship between equilibrium halflife (tequil50%) and lipophilicity. Oxycodone has a lipophilicity that predicts a higher permeability than that of morphine and fentanyl and therefore a shorter tequil50%. For both alfentanil and pethidine the lipophilicity facilitates a permeation of the BBB that leads to an even shorter tequil50%. Neither methadone nor LAAM fits smoothly into this inverse bell shaped relationship. Both drugs structurally belong to the phenylheptylamines, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
80 97 89 91 95 98 99 8.0, 9.9 8.9 6.5 8.4 8.7 8.3 8.3 1.4 3.5 0.72 7.5 5.6 8.2 11.1
Figure 5. The inverse bell shaped relationship between lipophilicity and equilibrium half-life.
which are very different in chemical structure from both the phenylpiperidines (alfentanil, fentanyl and pethidine) and phenanthrene (morphine and oxycodone). Therefore, it is likely that the inverse bell shaped curve represent a structureaffinity/activity relationship for transporters. Interestingly, the rank order of the opioids in the present study is very similar to that reported by Dagenais et al.30 using an in situ mouse brain perfusion assay.
CONCLUSION Rbrain:blood: the apparent brain:blood partitioning coeffecient.
92.4 (59.2) 222.8 (23.4) — 490 (57) — 530.8 (52.6) 724.0 (58.5) 4.5 (fixed) 35.4 (8.0) 47 (8) 4.5 (fixed) 364 (17) 23.3 (3.4) 22.8 (2.9) Morphine9 Oxycodone Alfentanil8 Fentanyl32 Pethidine8 (R)-Methadone33 LAAM34
Membrane Partial membrane Flow Partial membrane Flow Partial membrane Partial membrane
7.44 (1.60) 54.8 (7.8) >1000 240 (72) >1000 49.3 (2.1) 56.1 (1.7)
10.3 7.2 0.8 10.0 6.3 17.1 22
44.3 29.5 3.5 43.0 27.3 73.2 94
% Ionised at pH 7.4 pKa (208C)31 Rbrain:blood Tequil95% (min) Tequil50% (min) V2 (ml) PS (ml/min) V1 (ml) Model Opioid
Summary Over Cerebral Kinetic for Different Opioids With Regard to Their Physico-Chemical Properties Table 3.
1.4 0.9 1.4 2.8 1.7 2.4 3.2
VILLESEN ET AL.
Log D at pH 7.4
1674
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
The delay seen in cerebral equilibration in this study might to some degree explain the clinically observed delay in time from administration until onset of action in the effects of oxycodone. A model that has both membrane and CBF limitation describes the kinetics of this process. However, the physico-chemical properties of oxycodone (e.g. higher lipophilicity) facilitate a faster penetration of the BBB and therefore a faster t½ compared to morphine.
REFERENCES 1. Kalso E. 2005. Oxycodone. J Pain Symptom Manage 29:S47–S56. 2. Leow KP, Smith MT, Williams B, Cramond T. 1992. Single-dose and steady-state pharmacokinetics and pharmacodynamics of oxycodone in patients with cancer. Clin Pharmacol Ther 52:487–495. DOI 10.1002/jps
CEREBRAL KINETICS
3. Po¨yhia¨ R, Olkkola KT, Seppala T, Kalso E. 1991. The pharmacokinetics of oxycodone after intravenous injection in adults. Br J Clin Pharmacol 32: 516–518. 4. Po¨yhia¨ R, Seppala T, Olkkola KT, Kalso E. 1992. The pharmacokinetics and metabolism of oxycodone after intramuscular and oral administration to healthy subjects. Br J Clin Pharmacol 33:617– 621. 5. Leow KP, Cramond T, Smith MT. 1995. Pharmacokinetics and pharmacodynamics of oxycodone when given intravenously and rectally to adult patients with cancer pain. Anesth Analg 80:296– 302. 6. Upton RN, Semple TJ, Macintyre PE. 1997. Pharmacokinetic optimisation of opioid treatment in acute pain therapy. Clin Pharmacokinet 33:225– 244. 7. Minn A, Leclerc S, Heydel J-M, Minn A-L, Denizot C, Cattarelli M, Netter P, Gradinaru D. 2002. Drug transport into the mammalien brain: The nasal pathway and its specific metabolic barrier. J Drug Target 10:285–296. 8. Upton RN, Ludbrook GL, Gray EC, Grant C. 1997. The cerebral pharmacokinetics of meperidine and alfentanil in conscious sheep. Anesthesiology 86: 1317–1325. 9. Upton RN, Ludbrook GL, Martinez AM, Grant C, Milne RW. 2003. Cerebral and lung kinetics of morphine in conscious sheep after short intravenous infusions. Br J Anaesth 90:750–758. 10. Thompson SJ, Koszdin K, Bernards CM. 2000. Opiate-induced analgesia is increased and prolonged in mice lacking P-glycoprotein. Anesthesiology 92:1392–1399. 11. Letrent SP, Pollack GM, Brouwer KR, Brouwer KLR. 1999. Effects of a potent and specific Pglycoprotein inhibitor on the blood-brain barrier distribution and antinociceptive effect of morphine in the rat. Drug Metab Dispos 27:827–834. 12. Doolette DJ, Upton RN, Grant C. 1999. Agreement between ultrasonic doppler venous outflow and Kety and Schmidt estimates of cerebral blood flow. Clin Exp Pharmacol Physiol 26:736–740. 13. Upton RN, Grant C, Ludbrook GL. 1994. An ultrasonic Doppler venous outflow method for the continuous measurement of cerebral blood flow in conscious sheep. J Cereb Blood Flow Metab 14: 680–688. 14. Runciman WB, Ilsley AH, Mather LE, Carapetis R, Rao MM. 1984. A sheep preparation for studying interactions between blood flow and drug disposition. I: Physiological profile. Br J Anaesth 56:1015– 1028. 15. Menelaou A, Hutchinson M, Quinn I, Christensen A, Somogyi AA. 2003. Quantification of the O- and N-demethylated metabolites of hydrocodone and oxycodone in human liver microsomes using
DOI 10.1002/jps
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
1675
liquid chromatography with ultraviolet absorbance detection. J Chromatogr B Biomed Appl 785:81–88. Upton RN, Ludbrook GL, Grant C, Doolette DJ. 2000. The effect of altered cerebral blood flow on the cerebral kinetics of thiopental and propofol in sheep. Anesthesiology 93:1085–1094. Huang YF, Upton RN, Zheng D, Mclean C, Gray EC, Grant C. 1998. The enantiomer-specific kinetics and dynamics of verapamil after rapid intravenous administration to sheep: Physiological analysis and modeling. J Pharmacol Exp Ther 284:1048–1057. Po¨yhia¨ R, Kalso E. 1992. Antinociceptive effects and central nervous system depression caused by oxycodone and morphine in rats. Pharmacol Toxicol 70:125–130. Kalso E, Poyhia R, Onnela P, Linko K, Tigerstedt I, Tammisto T. 1991. Intravenous morphine and oxycodone for pain after abdominal surgery. Acta Anaesthesiol Scand 35:642–646. Wang JS, Ruan Y, Taylor R, Donovan J, Markowitz J, DeVane CL. 2004. Brain penetration of methadone (R)- and (S)-enantiomers is greatly increased by P-glycoprotein deficiency in the blood-brain barrier of Abcb1a gene knockout mice. Psychopharmacology (Berl) 173:132–138. Bostro¨m E, Simonsson USH, Hammarlund-Udenaes M. 2005. Oxycodone pharmacokinetics and pharmacodynamics in the rat in the presence of the P-glycoprotein inhibitor PSC833. J Pharm Sci 94: 1060–1066. Leow KP, Wright AW, Cramond T, Smith MT. 1993. Determination of the serum protein binding of oxycodone and morphine using ultrafiltration. Ther Drug Monit 15:440–447. Po¨yhia¨ R, Seppala T. 1994. Liposolubility and protein binding of oxycodone in vitro. Pharmacol Toxicol 74:23–27. Plummer JL, Cmielewski PL, Reynolds GD, Gourlay GK, Cherry DA. 1990. Influence of polarity on dose-response relationships of intrathecal opioids in rats. Pain 40:339–347. Leow KP, Smith MT. 1994. The antinociceptive potencies of oxycodone, noroxycodone and morphine after intracerebroventricular administration to rats. Life Sci 54:1229–1236. Bernards CM. 2002. Understanding the physiology and pharmacology of epidural and intrathecal opioids. Best Pract Res Clin Anaesthesiol 16:489– 505. Bernards CM, Hill H. 1992. Physical and chemical properties of drug molecules governing their diffusion through the spinal meninges. Anesthesiology 77:750–756. Pe´hourcq F, Matoga M, Bannwarth B. 2004. Diffusion of arylpropionate non-steroidal antiinflammatory drugs into the cerebrospinal fluid:
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1676
VILLESEN ET AL.
A quantitative structure-activity relationship approach. Fundam Clin Pharmacol 18:65–70. 29. Clement R, Malinovsky JM, Hildgen P, Dollo G, Estebe JP, Chevanne F, Le Verge R, Le Corre P. 2004. Spinal disposition and meningeal permeability of local anesthetics. Pharm Res 21:706– 716. 30. Dagenais C, Graff CL, Pollack GM. 2004. Variable modulation of opiid brain uptake by P-glycoprotein in mice. Biochem Pharmacol 67:269–276. 31. Clarke’s Analysis of Drugs and Poisons. 2005. Clarke’s analysis of drugs and poisons. Moffat AC, Osselton MD, Widdop B, editors.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
32. Upton RN, Grant C, Martinez AM, Ludbrook GL. 2004. Recirculatory model of fentanyl disposition with the brain as target organ. Br J Anaesth 93: 687–697. 33. Foster D, Upton R, Somogyi A, Grant C, Martinez A. 2005. The acute disposition of (R)- and (S)methadone in brain and lung of sheep. J Pharmacokinet Pharmacodynamics 32:547–570. 34. Foster DJR, Jensen ML, Upton RN, Somogyi AA, Grant C, Martinez A. 2006. Blood-brain equilibration kinetics of levo-[alpha]-acetyl-methadol using a chronically instrumented sheep preparation. Br J Pharmacol 147:209–217.
DOI 10.1002/jps
Department of Pharmacology and Pharmacotherapy, The Danish University of Pharmaceutical Sciences, Copenhagen 2100, Denmark 2
School of Pharmacy and Medical Sciences, University of South Australia, Adelaide 5000, Australia
3
Department of Anaesthesia and Intensive Care, Royal Adelaide Hospital and The University of Adelaide, Adelaide 5005, Australia
4
Department of Clinical and Experimental Pharmacology, The University of Adelaide, Adelaide 5005, Australia
5
Department of Clinical Pharmacology, Royal Adelaide Hospital, Adelaide 5005, Australia
Received 10 October 2005; revised 27 February 2006; accepted 1 March 2006 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20632
ABSTRACT: Oxycodone is an opioid analgesic that is administered orally or parenterally. The time-course of opioid action is a function of the systemic kinetics of the opioid, and the rate and extent of its entry into the brain and central nervous system. The latter is incompletely understood for oxycodone. Therefore, the cerebral kinetics of oxycodone was quantified using a conscious chronically instrumented sheep preparation. Five sheep were administered oxycodone as intravenous infusions (30 mg over 4 min). Using hybrid physiologically based kinetic models, cerebral kinetics was estimated from arteriosagittal sinus concentration gradients and cerebral blood flow (CBF). A two-compartment membrane-limited model best described the data. The volume of the first brain compartment was 35.4 mL with a half-life of equilibrium of 0.6 min. The brain:blood equilibration of oxycodone was relatively slow (half-life of 7.2 min), with a large deep cerebral distribution volume (222.8 mL) for the second compartment and a moderate membrane permeability of 54.8 mL/min, which exceeded the nominal CBF (40 mL/min). Drug retention in the brain was 1.3% after 45 min. In conclusion, pharmacokinetic modelling of oxycodone showed a delayed equilibration between brain and blood of a nature that would be affected by changes in both CBF and blood brain barrier permeability. ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 95:1666–1676, 2006
Keywords: oxycodone; pharmacokinetics; distribution; blood brain barrier; blood flow; physiological model; sheep
INTRODUCTION Oxycodone is an opioid analgesic usually administered for the treatment of moderate to severe pain for cancer pain or following surgery. It is functionally similar to morphine, but has a higher bioavailability after oral administration.1 The Correspondence to: Hanne H. Villesen (Telephone: (þ45) 35 30 6438; Fax: (þ45) 35 30 6050; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 95, 1666–1676 (2006) ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association
1666
effects of oxycodone, like those of morphine, lag behind its blood concentration-time profile.2–5 As reported for other opioids and anaesthetics, this delay in effect has been proposed to be similar in magnitude to the time required for equilibration of blood and central nervous system (CNS) concentrations.6 For some opioids, particularly when used parenterally, the delay in cerebral equilibration means that the cerebral kinetics of the opioid become an important factor in dictating their optimal clinical use.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
CEREBRAL KINETICS
The factors affecting the rate of cerebral uptake of opioids are complex. The ability of a drug to cross the blood brain barrier (BBB) is determined mainly by the degree of ionisation at physiological pH, molecular size, lipophilicity, affinity for efflux (i.e. P-glycoprotein), and uptake carrier mechanisms. Furthermore protein binding, cerebral blood flow (CBF), uptake into other tissues, cerebral metabolism and clearance rate also influence the net uptake.7 Different opioids are affected by these factors to varying extents. Pethidine, for example, has a large cerebral distribution volume and flowlimited kinetics, indicating that its cerebral kinetics are highly influenced by changes in CBF.8 In contrast, morphine has a small cerebral distribution volume and moderate BBB permeability,9 which might be due to P-glycoprotein mediated efflux transport.10,11 This suggests that the cerebral kinetics of morphine is influenced less by changes in CBF and more by the activity of efflux transport. Where oxycodone fits in this spectrum of cerebral kinetic behaviour is currently unknown. The aim of this study was therefore to elucidate the cerebral kinetics of oxycodone after a short intravenous infusion in a conscious sheep preparation. This preparation has previously been used to determine the cerebral kinetics of other opioids (e.g. alfentanil, pethidine, fentanyl, morphine, (R)- and (S)-methadone and levo-alphaacetyl-methadol (LAAM)). As a secondary aim, the cerebral kinetics of oxycodone was compared with the reported cerebral kinetics of these opioids using the same experimental study design.
MATERIALS AND METHODS Animal Preparation Sheep were chosen as an experimental animal as the relative perfusion of their brain and the control of CBF are similar to that in humans,12,13 afford good access to appropriate blood vessels and tolerate chronic instrumentation well. All experimental procedures were approved by the Animal Ethics Committee of the University of Adelaide and were conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. The surgical procedures for preparing the sheep and the details about the experiment has been described previously.8,13 Briefly, female Merino sheep of similar ages and body mass (mean weight 44.3 5.0 kg) were used. Instrumentation was DOI 10.1002/jps
1667
performed under general anaesthesia. Catheters were chronically implanted with their tips in the abdominal aorta (for sampling of arterial blood), in the right atrium (for drug administration and measurement of cardiac output (CO)), and in the dorsal sagittal sinus (for cerebral venous blood sampling).12 A Doppler transducer was placed over the sagittal sinus using a previously validated method to quantify the relative changes in CBF.12,13 After recovery from anaesthesia, the sheep were housed in metabolic crates or floor pens for a least a day before the study and their catheters maintained with a saline/heparin lock (0.9%/50 i.u./mL). Study Design Five sheep received oxycodone hydrochloride (MW 405,9 g/mol) (Boots, Nottingham, UK) 0.6 mg/kg as a 4-min i.v infusion. On the study day the instrumented sheep was placed in nonweight-bearing sling inside a metabolic crate and prepared for physiological measurements and blood sampling. After 5 min of baseline measurements, the infusion was commenced at time zero. Oxycodone hydrochloride 1 mg/mL was prepared in 0.9% isotonic saline by the Department Pharmacy of the Royal Adelaide Hospital prior to each study day. Infusion rate was adjusted accordingly to the weight of each sheep. The following procedures were performed during each study. Blood Sampling Prior to drug infusion, blank blood was sampled for use in the calibration curves for the quantification of oxycodone. Simultaneous blood samples of 2 mL volume were collected from the arterial and sagittal sinus catheters at 0.5, 1, 1.5, 2, 3, 4, 4.5, 5, 5.5, 6, 8, 10, 15, 20, 30, 45, 60 and 75 min after start of drug infusion. The blood samples were collected in 10 mL tubes with heparin as an anticoagulant. After the study the blood was centrifuged at 3250 rpm for 10 min and plasma was separated. The blood samples were stored at room temperature for <90 min before separation of plasma. The plasma samples were stored at 208C until analysis. Blood Gas, Cardiac Output and Cerebral Blood Flow Measurements Haemodynamic data (mean arterial blood pressure (MAP) and the CBF index) were continuously monitored using a computerised data JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1668
VILLESEN ET AL.
acquisition system. Relative changes in CBF were quantified using Doppler flow probe and a flow meter (Bioengineering, University of Iowa, IA) and MAP was recorded continuously via a pressure transducer on one of the arterial catheters connected to an analogue-to-digital card (Metrabyte DAS 16-G2) and a personal computer (486 based IBM compatible). Recording of these parameters was commenced 5 min before drug infusion (baseline) and throughout the study period. CO was measured in triplicate immediately prior to the oxycodone infusion, and at 4, 10, 30 and 60 min after the start of the infusion, using a thermodilution method.14 The values were averaged to obtain the mean at each time point. Additional arterial blood samples were taken at 0, 4, 10 and 30 min for blood gas measurements (ABL System 625, Radiometer, Sweden). High Performance Liquid Chromatography for Determination of Oxycodone The plasma samples containing oxycodone were assayed using a validated high performance liquid chromatography (HPLC) assay. The HPLC system comprised a LC-10AT pump (Shimadzu, Kyoto, Japan), a SIL-10A auto injector (Shimadzu), a RCM 8 10 compression module containing a Nova-Pak C-18 4 mm column (8 100 mm Radial-PakTM Cartridge, Waters Associates, Inc. Milford, MD) with a Guard C-18 5u precolumn (7.5 4.6 mm, Waters Associates) and a SPD-M10A diode array detector (Shimadzu) set at 210 nm. The mobile phase consisted of 28% acetonitrile, 0.1 M sodium dihydrogen orthophosphate and 230-mg/L lauryl sulphate. The final pH was adjusted to 2.4 with orthophosphoric acid. The flow rate was 1.25 mL/min giving retention times for hydromorphone (internal standard) and oxycodone at 7 and 14 min, respectively. All chemicals were of analytical or HPLC grade. Oxycodone plasma samples were processed using liquid–liquid extraction as previously described by Menelaou et al.15 with modifications. Briefly, in 10 mL flat bottom plastic tubes 1 mL of plasma and 70 ml of 10 mg/mL hydromorphone (internal standard) were alkalinised with 300 ml of 20% w/v sodium carbonate. Samples were extracted with 4 mL of dichloromethane for 10 min on a rotary mixer. The samples were centrifuged at 3250 rpm for 10 min and the upper aqueous layer was aspirated to waste. The remaining organic phase was transferred to clean plastic JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
tube containing 200 ml of 0.1 M hydrochloric acid. The samples were rotary-mixed for 10 min and centrifuged for 10 min at 3250 rpm. The acid bubble was transferred to an auto injector vial and 50 ml was injected on the HPLC system. Samples were injected in random order along with samples containing standards (0.05–5 mg/mL, n ¼ 8 standards). Quality controls (QC) were included in each run to monitor assay performance and were prepared in duplicate at three different concentrations: low (LQC, 0.15 mg/mL), medium (MQC, 1 mg/ mL) and high (HQC, 3.5 mg/mL). Assay Validation All assays were calibrated using eight-point standard curves prepared in plasma separated from blood taken from the same animal before drug administration. As a validation procedure, inter-assay variability was monitored with QC prior to assaying the oxycodone plasma samples. The slope of the calibration curves (mean SD, %coefficient of variation (%CV)) was 2.2 0.2, %CV 0.7 (n ¼ 6). The mean r2-value was 0.999 0.001, %CV 0.009 (n ¼ 6 assay days), while interassay accuracy and precision was 97.3 4.4, %CV 4.5 (n ¼ 11) (LQC), 99.3 3.3, %CV 3.3 (n ¼ 12) (MQC) and 101.4 4.7, %CV 4.7 (n ¼ 12) (HQC). Similarly, intra-assay (replicate samples) accuracy and precision were 102.4 5.2, %CV 5.1 (n ¼ 6) (LQC), 103.8 2.7, %CV 2.6 (n ¼ 6) (MQC) and 102.8 2.1, %CV 2.1 (n ¼ 6) (HQC). The assay was both precise and accurate at the limit of quantification (0.05 mg/mL), with inter-assay accuracy and precision being 106.3 12.1%, %CV 11.4 (n ¼ 8). The extraction recovery was 72.2 11.7%, %CV 16.2 (n ¼ 4). Pharmacokinetic Analysis All kinetic analyses were based on the mean concentrations of all sheep at each time points. This reduces the influence of random fluctuation in concentration on model fit. A hybrid modelling of kinetics was used to define the cerebral kinetics of oxycodone. Empirical forcing functions were used to represent inputs into the brain (Cin), and curve-fitting of the output of the brain (Cout) to physiologically realistic models used to determine model parameters.16 The forcing functions are essentially a device to interpolate the available data points, as solving the models require continuous functions. Provided they fit the available data, their form has no bearing on the discriminaDOI 10.1002/jps
CEREBRAL KINETICS
tion between various models of organ kinetics. Arterial input concentrations to the brain were interpolated using exponential forcing functions. If CBF showed statistically significant changes from baseline over time, these were interpolated in the forcing functions.12 The sagittal sinus output concentrations from the brain were curve fitted to determine model parameters using a curve fitting modelling software (Scientist for Windows, version 2.01, Micromath Scientific Software, UT).
1669
Four different kinetic models of the brain were fitted to the measured sagittal sinus oxycodone concentrations. The basic forms of the equations describing these models have been published previously.9,17 They are repeated below for convenience and shown in Figure 1. Cin and Cout are the afferent arterial and efferent sagittal sinus drug concentrations of the brain, respectively, and Q is CBF: A null model that tested the hypothesis that there was no oxycodone concentration gradient across the brain: Cout ¼ Cin
ð1Þ
A single flow-limited compartment defined by a single distribution volume and CBF: V1 dCout =dt ¼ Q ðCin Cout Þ
ð2Þ
A single flow-limited compartment with an apparent first-order loss (PSloss) representing either deep distribution or metabolism: V1 dCout =dt ¼ Q ðCin Cout Þ PSloss Cout ð3Þ A two-compartment membrane-limited model with a permeability term (PS) representing distribution into a deep compartment: V1 dCout =dt ¼ Q ðCin Cout Þ þ PS ðC2 Cout Þ V2 dC2 =dt ¼ PS ðCout C2 Þ ð4Þ
Figure 1. Graphical representation of kinetic models: (A) flow limited model (Eq. 2), (B) single flow-limited compartment with an apparent first order loss of drug (Eq. 3), and (C) a two-compartment membrane limited model with a permeability term describing distribution into a deep compartment (Eq. 4). DOI 10.1002/jps
V1 is the volume of the first compartment of the brain (which includes blood). The nominal volume of this vascular compartment in a sheep brain is 4.5 mL (5% of 90 mL). V2 and C2 are the volume of, and concentration in, the second compartment of the brain (if appropriate). The permeabilitysurface area coefficient (PS) is a term describing loss or exchange of drug from the first compartment. In the case of Eq. (3), PS represents a unidirectional loss of drug, such as metabolism or deep distribution, which is essentially irreversible. In the case of Eq. (4), PS represents the effective inter-compartmental clearance of drug between the two compartments. Selection criteria for the final model were a high model selection criterion (MSC) value (>3),16 a statistically (p < 0.05) significant improvement of the fit as determined by the F-ratio test based upon the residual sum of squares, and an even distribution of residuals. The data were also analysed using a non-linear mixed effect approach to establish the intersubject variability in parameter estimates as a JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1670
VILLESEN ET AL.
supplemental analysis. The best model of cerebral kinetics inferred from analysis of pooled data was fitted to population data (i.e. all individual sheep) using NONMEM (version V, level 1.1), with linear interpolation of arterial and CBF inputs for individual sheep, and FOCE with interaction analysis. Parameters for the brain kinetic model (V1, PS and V2) were assigned an exponential distribution across the population, and a combined additive and proportional residual error model was used. Calculated Variables Model parameters were used to calculate secondary variables to facilitate comparison with other opioids and literature values. Brain equilibration times for oxycodone were calculated by using the final cerebral kinetic model and best parameter values to simulate the time course of the brain concentrations for a step increase in the afferent arterial blood concentration from 0 to 1. The times required for the brain concentration to reach 50 and 95% of the arterial blood concentration were recorded. The apparent permeability of the BBB (PS) was compared with the CBF. The apparent brain:blood partition coefficient (R) was calculated from V2 and a nominal real volume of 63 mL for the region of the brain drained by the sagittal-sinus catheter (70% of 90 mL, whole brain).12 Drug retention (R%) in the brain was calculated as follows: 1 AUCsag 100% ð5Þ R% ¼ AUCart
measured at 1 s intervals, a 10 s moving average from each individual animal was calculated at each of the pharmacokinetic blood sampling timepoints. CO, MAP, arterial carbon dioxide tension (PaCO2) and arterial oxygen tension (PaO2), and relative changes from baseline (100%) in CBF were calculated as mean and 95% confidence intervals (CI). Statistically significant changes from base line for pharmacodynamic data were tested using one-way repeated measures ANOVA with Bonferroni correction. Time was the repeated measure. A a-value of 0.05 was set for statistically significant differences. All data are reported as mean SD or mean (95% CI).
RESULTS Pharmacodynamic Data CBF, CO and MAP changed from baseline during oxycodone infusion. Initially after the start of the infusion CBF decreased for a short period of time, thereafter it increased to 6.9 10.6% above baseline (44 mL/min) during the infusion and continued to increase to a maximum of 22.6 14.1% (p < 0.0001) of baseline at 5.5 min, where it remained for the rest of the study (Fig. 2). The mean baseline CO was 5.4 0.8 L/min. Neither CO nor MAP changed significantly from baseline
where the total arterial (AUCart) and sagittal sinus (AUCsag) area under the blood concentration-time curves to 45 min were calculated using the trapezoidal rule. The drug retention indicates the amount of drug that had entered the brain via the arterial blood, but had not left the brain via the efferent venous blood by the end of the period of interest. Retention in the brain could be the result of metabolism or deep distribution from which efflux is relatively slow.
Statistical Analysis GraphPad Prism (GraphPad Software version 4.02 for Windows, GraphPad Software, San Diego, CA) was used for all statistical analyses. In the case of CBF and MAP, which were continuously JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
Figure 2. Mean (open circles) and 95% confidence intervals (dotted lines) of cerebral blood flow expressed as a percentage of baseline values in five sheep. The solid line is baseline value. DOI 10.1002/jps
CEREBRAL KINETICS
Table 1.
Mean SD
Mean Baseline Values for Cardiovascular and Blood Gas Measurements CO (L/min)
MAP (mm Hg)
PaO2 (mm Hg)
PaCO2 (mm Hg)
O2 sat (%)
PH
HCO 3 (mmol/L)
5.4 0.8
92.2 18.3
108.0 10.6
30.8 1.9
101.2 1.5
7.5 0.02
23.0 2.2
(p ¼ 0.75). Additionally no significant changes from baseline of PaO2 (p ¼ 0.40) or PaCO2 (p ¼ 0.54) could be demonstrated at any time point during the study. The dose of oxycodone was not associated with significant respiratory or cardiovascular side effects. In all sheep, oxycodone was noted to produce a mild degree of dysphoria (agitation, mouthing of crate, bleating). Table 1 shows the mean baseline values for CO, MAP and blood gas measurements. The recorded changes in CBF were interpolated and incorporated in the input forcing functions used in the subsequent modelling. Pharmacokinetic Data Figure 3 shows the mean observed plasma oxycodone HCl concentrations. During the infusion period (first 4 min) a substantial extraction
Figure 3. Measured arterial (open circles) and sagittal-sinus (open invert triangles) blood concentrations of oxycodone HCl in five sheep. Data are shown as mean at each time point. Dotted and solid lines are the best fits of the kinetic model chosen, which was the membranelimited model (Table 2). DOI 10.1002/jps
1671
across the brain was seen. Peak arterial concentration occurred at 4 min and was 0.72 mg/mL (95% CI 0.49–0.95 mg/mL), and sagittal sinus Cmax was 0.38 mg/mL (95% CI 0.23–0.53 mg/mL) and occurred at 4.5 min. The differences in the concentration across the brain are shown in Figure 4. The concentrations of oxycodone HCl were below the limit of quantification after 45 min in all sheep; consequently kinetic analysis was only conducted using concentration-time data for the 0–45 min period. The results of the kinetic modelling for the various models are summarised in Table 2. The null model, the flow-limited as well as the flowlimited with a loss, gave equally poor descriptions of the data. The permeability term was best described by transport of oxycodone into a deep compartment as in the membrane-limited model (higher MSC). The volume of the first compartment (nominally the vascular space between the sampling catheters) was 35.4 mL and the t½ of
Figure 4. Mean (open circles) and 95% CI (dotted lines) of the individual concentration differences across the brain (arterial-sagittal sinus) in five sheep. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1672
VILLESEN ET AL.
Table 2. Model Parameter Estimated for Brain Kinetics Model
MSC
V1 (mL)
PS (mL/min)
Pooled data Null Flow Flow þ loss Membrane
1.20 0.71 1.73 3.32
182.4 (26.2) 69.8 (11.7) 35.4 (8.0)
25.0 (3.6) 54.8 (7.8)
Population data Membrane Parameter Inter-animal variability (%CV)
V2 (mL)
222.8 (23.4)
— 28.4 (14%) 12.8% (168%)
54.4 (8.5%) —
236 (9.5%) —
For the pooled data analysis, data are mean (SD) given by the curve fitting modelling program, with SD in this case indicating the reliability of the parameter estimate. The MSC is the model selection criteria—the higher the number, the better the fit. For the population data analysis, data are the value of the population estimate, with data in parentheses in this case indicating the % standard error of the parameter estimate indicating the reliability of the parameter estimate.
equilibrium through this compartment was 0.6 min. The permeability (PS) of oxycodone across the membrane barrier (nominally the BBB) was 54.8 mL/min, which was comparable with the CBF (baseline 40 mL/min). Drugs for which membrane permeability is of similar magnitude to CBF generally show characteristics of both bloodflow and membrane limitation in their cerebral uptake.16 The final volume of the deep compartment (nominally brain parachyma) was 222.8 mL. The times to reach 50 and 95% of equilibration were 7.2 and 29.5 min, respectively. Drug retention (R%) in the brain was negligible (1.3% after 45 min). Assuming a nominal real volume of the region of the brain drained by the sagittal sinus catheter of 63 mL (70% of 90 mL), the volume of the deep compartment equates to a brain:blood partition coefficient of approximately 3.5. Initial analyses found that inter-animal variability PS and V2 was very small (<0.1%), and an additive error component was not supported. As a consequence inter-animal variability was removed from these parameters (without a decrease in model fit) and a proportional error model only was employed to obtain the most parsimonious model. In the final model, the standard error of the population parameters estimates is acceptable, with a small residual error (% CV 18). Standard diagnostic plots of both the population model and individual Bayesian fits, supported an excellent fit to the data. The results of the mixed effect analysis using NONMEM is shown in Table 2. The population parameters of the model did not differ greatly from the results of the pooled analysis. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
DISCUSSION Cerebral Pharmacokinetic Data As oxycodone is often used in a similar clinical role as morphine, it is of interest to compare the two opioids. Like morphine, a membrane-limited model was found to best describe the cerebral kinetics of oxycodone. However, in comparison with morphine,9 the permeability of oxycodone across the BBB was relatively high (54.8 vs. 7.44 mL/min, respectively), as was its cerebral distribution volume (222.8 vs. 92.4 mL, respectively). The relatively fast permeation of oxycodone through the blood-brain barrier was followed by a relatively short equilibration half-life of the deep compartment (7.2 min) in comparison with morphine (10.3 min).9 The population pharmacokinetic parameters of the model did not differ greatly from the results of the pooled analysis. Reasons to the very small inter-animal variability in kinetic parameters might include the highly controlled nature of the experiment and data collection, such that extraneous variability was essentially removed form the data. For example, the inputs (arterial blood concentrations and CBF) for each animal were quantitated and described, the sheep were all from the same flock, and the assay was reasonably precise and accurate. These data suggest that there was minimal inter-individual variability in the cerebral kinetic parameters of oxycodone in the sheep studied, and supports the process of using pooled data to distinguish between various models. DOI 10.1002/jps
CEREBRAL KINETICS
Brain compartment volumes have not been normalized for mass of tissue, because the sheep model employed in the present study is representative of human cerebral kinetics compared species, due to a remarkably similar CBF/g tissue (0.44 mL/g tissue) to humans (0.54 mL/g tissue). Analgesia has been described to be achieved faster with oxycodone than with morphine after iv administration in both animal (15 and 30 min, respectively)18 and human studies (28 4.8 and 46 6.9 min, respectively).19 In the present study it was shown that this faster onset might be explained by a faster equilibration between blood and brain for oxycodone. Morphine and oxycodone are believed acting on different classes of opioid receptors with different affinities and different kinetics. This might represent part of the explanation for the different antinociceptive effect seen between the two opioids, however, it still remains controversial whether or not morphine and oxycodone do possess different receptor binding profiles.
Role of Transporters Another difference to be considered when comparing the two opioids is the role active transporters may play in permeation of the BBB. Many opioids, such as morphine and methadone, are substrates for P-glycoprotein transporters,10 which influence the uptake into the brain and thereby the effect.20 However, based on results from studies in rats oxycodone does not seem to be a substrate for P-glycoprotein.21 This suggests that only physicochemical factors such as lipophilicity, molecular size, ionisation degree, (all of which influence passive diffusion) govern the rate of cerebral uptake of oxycodone. It is worthwhile to consider the interplay of these factors, and to place them in context by comparison of oxycodone with other opioids.
Polarity, Lipophilicity and Cerebral Uptake The physicochemical properties, such as pKa, lipophilicity and water solubility of oxycodone are similar to those of morphine.22,23 Oxycodone is a weakly basic drug with a pKa value of 8.9, thus it is 97% ionised at physiological pH. Like morphine, oxycodone is a hydrophilic molecule and thus possesses a low lipophilicity; the octanol:buffer distribution coefficient is 1.27, 1.64 and 706 for morphine, oxycodone and fentanyl, respectively at 378C and pH 7.4.24 DOI 10.1002/jps
1673
Only one study has so far included oxycodone studies of opioid dose-response relationships for antinociceptive effect with regard to polarity of the substance. Oxycodone and morphine, along with other opioids, were given intrathecally to male Porton rats. The onset of antinociception was faster for oxycodone than for morphine, but only slightly slower than seen after administration of the more lipophilic opioid, fentanyl in both the Hot Plate (6.2, 24.9 and 5.3 min, respectively) and the Tail–Flick tests (6.3, 16.1 and 5.3 min, respectively),24 which is consistent with the results from human studies described by Po¨yhia¨ and Kalso19 and Leow and Smith.25 The above indicates that lipophilicity and polarity are important factors for the distribution of drugs within the CNS. Bernards and coworkers have stated that permeability increases with increasing lipophilicity until a certain octanol:buffer distribution coefficient,26 and it has been demonstrated that despite the large difference in octanol:buffer distribution coefficient between morphine and fentanyl, the permeability of fentanyl is only slightly higher than the permeability of morphine.27 This parabolic relationship between lipophilicity and permeability is also seen for other physiological membranes such as skin and gastro-intestinal mucosa. Different factors such as molecular weight have been incorporated into the fitting of the relationship and found to improve the correlations seen.28,29 A summary of the brain kinetics of different opioids is shown in Table 3 along with relevant physico-chemical parameters. With exception of methadone and LAAM, the proposed relationship applies for the opioids shown in Table 3. For fentanyl and pethidine, the time to reach 50% of equilibrium of the deep tissue compartment (tequil50%) is closer to that of oxycodone even though there is a large difference in the log D-values (2.8, 1.7 and 0.9, respectively). This could be explained by the biphasic relationship seen between lipophilicity and permeability. From Figure 5 it can be seen that there is an inverse bell shaped relationship between equilibrium halflife (tequil50%) and lipophilicity. Oxycodone has a lipophilicity that predicts a higher permeability than that of morphine and fentanyl and therefore a shorter tequil50%. For both alfentanil and pethidine the lipophilicity facilitates a permeation of the BBB that leads to an even shorter tequil50%. Neither methadone nor LAAM fits smoothly into this inverse bell shaped relationship. Both drugs structurally belong to the phenylheptylamines, JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
80 97 89 91 95 98 99 8.0, 9.9 8.9 6.5 8.4 8.7 8.3 8.3 1.4 3.5 0.72 7.5 5.6 8.2 11.1
Figure 5. The inverse bell shaped relationship between lipophilicity and equilibrium half-life.
which are very different in chemical structure from both the phenylpiperidines (alfentanil, fentanyl and pethidine) and phenanthrene (morphine and oxycodone). Therefore, it is likely that the inverse bell shaped curve represent a structureaffinity/activity relationship for transporters. Interestingly, the rank order of the opioids in the present study is very similar to that reported by Dagenais et al.30 using an in situ mouse brain perfusion assay.
CONCLUSION Rbrain:blood: the apparent brain:blood partitioning coeffecient.
92.4 (59.2) 222.8 (23.4) — 490 (57) — 530.8 (52.6) 724.0 (58.5) 4.5 (fixed) 35.4 (8.0) 47 (8) 4.5 (fixed) 364 (17) 23.3 (3.4) 22.8 (2.9) Morphine9 Oxycodone Alfentanil8 Fentanyl32 Pethidine8 (R)-Methadone33 LAAM34
Membrane Partial membrane Flow Partial membrane Flow Partial membrane Partial membrane
7.44 (1.60) 54.8 (7.8) >1000 240 (72) >1000 49.3 (2.1) 56.1 (1.7)
10.3 7.2 0.8 10.0 6.3 17.1 22
44.3 29.5 3.5 43.0 27.3 73.2 94
% Ionised at pH 7.4 pKa (208C)31 Rbrain:blood Tequil95% (min) Tequil50% (min) V2 (ml) PS (ml/min) V1 (ml) Model Opioid
Summary Over Cerebral Kinetic for Different Opioids With Regard to Their Physico-Chemical Properties Table 3.
1.4 0.9 1.4 2.8 1.7 2.4 3.2
VILLESEN ET AL.
Log D at pH 7.4
1674
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
The delay seen in cerebral equilibration in this study might to some degree explain the clinically observed delay in time from administration until onset of action in the effects of oxycodone. A model that has both membrane and CBF limitation describes the kinetics of this process. However, the physico-chemical properties of oxycodone (e.g. higher lipophilicity) facilitate a faster penetration of the BBB and therefore a faster t½ compared to morphine.
REFERENCES 1. Kalso E. 2005. Oxycodone. J Pain Symptom Manage 29:S47–S56. 2. Leow KP, Smith MT, Williams B, Cramond T. 1992. Single-dose and steady-state pharmacokinetics and pharmacodynamics of oxycodone in patients with cancer. Clin Pharmacol Ther 52:487–495. DOI 10.1002/jps
CEREBRAL KINETICS
3. Po¨yhia¨ R, Olkkola KT, Seppala T, Kalso E. 1991. The pharmacokinetics of oxycodone after intravenous injection in adults. Br J Clin Pharmacol 32: 516–518. 4. Po¨yhia¨ R, Seppala T, Olkkola KT, Kalso E. 1992. The pharmacokinetics and metabolism of oxycodone after intramuscular and oral administration to healthy subjects. Br J Clin Pharmacol 33:617– 621. 5. Leow KP, Cramond T, Smith MT. 1995. Pharmacokinetics and pharmacodynamics of oxycodone when given intravenously and rectally to adult patients with cancer pain. Anesth Analg 80:296– 302. 6. Upton RN, Semple TJ, Macintyre PE. 1997. Pharmacokinetic optimisation of opioid treatment in acute pain therapy. Clin Pharmacokinet 33:225– 244. 7. Minn A, Leclerc S, Heydel J-M, Minn A-L, Denizot C, Cattarelli M, Netter P, Gradinaru D. 2002. Drug transport into the mammalien brain: The nasal pathway and its specific metabolic barrier. J Drug Target 10:285–296. 8. Upton RN, Ludbrook GL, Gray EC, Grant C. 1997. The cerebral pharmacokinetics of meperidine and alfentanil in conscious sheep. Anesthesiology 86: 1317–1325. 9. Upton RN, Ludbrook GL, Martinez AM, Grant C, Milne RW. 2003. Cerebral and lung kinetics of morphine in conscious sheep after short intravenous infusions. Br J Anaesth 90:750–758. 10. Thompson SJ, Koszdin K, Bernards CM. 2000. Opiate-induced analgesia is increased and prolonged in mice lacking P-glycoprotein. Anesthesiology 92:1392–1399. 11. Letrent SP, Pollack GM, Brouwer KR, Brouwer KLR. 1999. Effects of a potent and specific Pglycoprotein inhibitor on the blood-brain barrier distribution and antinociceptive effect of morphine in the rat. Drug Metab Dispos 27:827–834. 12. Doolette DJ, Upton RN, Grant C. 1999. Agreement between ultrasonic doppler venous outflow and Kety and Schmidt estimates of cerebral blood flow. Clin Exp Pharmacol Physiol 26:736–740. 13. Upton RN, Grant C, Ludbrook GL. 1994. An ultrasonic Doppler venous outflow method for the continuous measurement of cerebral blood flow in conscious sheep. J Cereb Blood Flow Metab 14: 680–688. 14. Runciman WB, Ilsley AH, Mather LE, Carapetis R, Rao MM. 1984. A sheep preparation for studying interactions between blood flow and drug disposition. I: Physiological profile. Br J Anaesth 56:1015– 1028. 15. Menelaou A, Hutchinson M, Quinn I, Christensen A, Somogyi AA. 2003. Quantification of the O- and N-demethylated metabolites of hydrocodone and oxycodone in human liver microsomes using
DOI 10.1002/jps
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
1675
liquid chromatography with ultraviolet absorbance detection. J Chromatogr B Biomed Appl 785:81–88. Upton RN, Ludbrook GL, Grant C, Doolette DJ. 2000. The effect of altered cerebral blood flow on the cerebral kinetics of thiopental and propofol in sheep. Anesthesiology 93:1085–1094. Huang YF, Upton RN, Zheng D, Mclean C, Gray EC, Grant C. 1998. The enantiomer-specific kinetics and dynamics of verapamil after rapid intravenous administration to sheep: Physiological analysis and modeling. J Pharmacol Exp Ther 284:1048–1057. Po¨yhia¨ R, Kalso E. 1992. Antinociceptive effects and central nervous system depression caused by oxycodone and morphine in rats. Pharmacol Toxicol 70:125–130. Kalso E, Poyhia R, Onnela P, Linko K, Tigerstedt I, Tammisto T. 1991. Intravenous morphine and oxycodone for pain after abdominal surgery. Acta Anaesthesiol Scand 35:642–646. Wang JS, Ruan Y, Taylor R, Donovan J, Markowitz J, DeVane CL. 2004. Brain penetration of methadone (R)- and (S)-enantiomers is greatly increased by P-glycoprotein deficiency in the blood-brain barrier of Abcb1a gene knockout mice. Psychopharmacology (Berl) 173:132–138. Bostro¨m E, Simonsson USH, Hammarlund-Udenaes M. 2005. Oxycodone pharmacokinetics and pharmacodynamics in the rat in the presence of the P-glycoprotein inhibitor PSC833. J Pharm Sci 94: 1060–1066. Leow KP, Wright AW, Cramond T, Smith MT. 1993. Determination of the serum protein binding of oxycodone and morphine using ultrafiltration. Ther Drug Monit 15:440–447. Po¨yhia¨ R, Seppala T. 1994. Liposolubility and protein binding of oxycodone in vitro. Pharmacol Toxicol 74:23–27. Plummer JL, Cmielewski PL, Reynolds GD, Gourlay GK, Cherry DA. 1990. Influence of polarity on dose-response relationships of intrathecal opioids in rats. Pain 40:339–347. Leow KP, Smith MT. 1994. The antinociceptive potencies of oxycodone, noroxycodone and morphine after intracerebroventricular administration to rats. Life Sci 54:1229–1236. Bernards CM. 2002. Understanding the physiology and pharmacology of epidural and intrathecal opioids. Best Pract Res Clin Anaesthesiol 16:489– 505. Bernards CM, Hill H. 1992. Physical and chemical properties of drug molecules governing their diffusion through the spinal meninges. Anesthesiology 77:750–756. Pe´hourcq F, Matoga M, Bannwarth B. 2004. Diffusion of arylpropionate non-steroidal antiinflammatory drugs into the cerebrospinal fluid:
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1676
VILLESEN ET AL.
A quantitative structure-activity relationship approach. Fundam Clin Pharmacol 18:65–70. 29. Clement R, Malinovsky JM, Hildgen P, Dollo G, Estebe JP, Chevanne F, Le Verge R, Le Corre P. 2004. Spinal disposition and meningeal permeability of local anesthetics. Pharm Res 21:706– 716. 30. Dagenais C, Graff CL, Pollack GM. 2004. Variable modulation of opiid brain uptake by P-glycoprotein in mice. Biochem Pharmacol 67:269–276. 31. Clarke’s Analysis of Drugs and Poisons. 2005. Clarke’s analysis of drugs and poisons. Moffat AC, Osselton MD, Widdop B, editors.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
32. Upton RN, Grant C, Martinez AM, Ludbrook GL. 2004. Recirculatory model of fentanyl disposition with the brain as target organ. Br J Anaesth 93: 687–697. 33. Foster D, Upton R, Somogyi A, Grant C, Martinez A. 2005. The acute disposition of (R)- and (S)methadone in brain and lung of sheep. J Pharmacokinet Pharmacodynamics 32:547–570. 34. Foster DJR, Jensen ML, Upton RN, Somogyi AA, Grant C, Martinez A. 2006. Blood-brain equilibration kinetics of levo-[alpha]-acetyl-methadol using a chronically instrumented sheep preparation. Br J Pharmacol 147:209–217.
DOI 10.1002/jps