Physicochemical Properties, Pharmacokinetics, and Pharmacodynamics of a Reformulated Microemulsion Propofol in Rats

The assessment of plasma bradykinin generation in healthy volunteers was approved by the Institutional Review Board (Asan Medical Center, Seoul, Korea), and written informed consent was obtained from all participants. The pharmacokinetic and pharmacodynamic analyses in rats were approved by the Institutional Animal Care and Use Committee of the Asan Institute for Life Sciences (Seoul, Korea).

The formula of microemulsion propofol was as follows: a mixture of 1% propofol, 10% PP188 (Daebong LS Co., LTD, Seoul, Korea), 0.7% polyethylene glycol 660 hydroxystearate (Solutol HS 15; BASF Company Ltd., Seoul, Korea), 1% glycerin, 0.0008% disodium ethylenediaminetetraacetic acid (EDTA), and 0.01% sodium ascorbate. Propofol was obtained from Clariant LSM Inc. (Viale Europa, Italy). PP188 is comprised of a single chain (block) of hydrophobic polyoxypropylene flanked by two chains (blocks) of hydrophilic polyoxyethylene.6Propofol in microemulsion propofol is surrounded by a corona of PP188 and polyethylene glycol 660 hydroxystearate and serves as the lipid oil core of the microemulsion. Oxidative process in microemulsion propofol resulted in propofol dimerization and microemulsion yellowing. Microemulsion yellowing is caused by the formation of propofol dimer quinone.20Sodium ascorbate and disodium EDTA can function as a water-soluble antioxidant and a chelating agent, respectively, which completely inhibit the oxidative process in microemulsion propofol, and hence microemulsion remained visibly white at all times.

Microemulsion propofol was prepared by dispersing propofol at 70°C in a solution of PP188 using a constant speed stirrer at 1,500 rpm. When propofol is dissolved, polyethylene glycol 660 hydroxystearate, glycerin, sodium EDTA, and sodium ascorbate were added. Microemulsion propofol contained in an ampoule was autoclaved for 20 min at 121°C.

The pH, particle size, and osmolarity were measured using a pH meter (model 720A; Orion Research Inc., Boston, MA), particle size analyzer (ELS-Z2; Otsuka Electronics Co., Ltd., Osaka, Japan), and cryoscopic osmometer (Osmomat 030; Gonotec GmbH, Berlin, Germany), respectively.

In this study, the LCT propofol used as a comparator was 1% Diprivan® (AstraZeneca, London, United Kingdom). As described in a previous study,4propofol preparations were dialyzed using a dialysis membrane (Dialysis tubing benzoylated®; Sigma-Aldrich Co., St. Louis, MO) with a cutoff molecular weight of approximately 3,500–4,000 Da. A solution of 2.25% (wt/vol) glycerin (LG household & Health Care, Seoul, Korea) in water was used as the release medium. Microemulsion and lipid emulsion propofol (5 ml) were prepared to measure the free propofol concentration in the aqueous phase of each formulation.

To evaluate whether agents that reduce propofol-induced pain decrease the free propofol concentration, six mixtures were prepared: 5.5 ml of each propofol formulation (5 ml) and 2% lidocaine (0.5 ml),215.2 ml of each propofol formulation (5 ml) and ketamine (10 mg in 0.2 ml),226 ml of each propofol formulation (5 ml) and metoclopramide (5 mg in 1 ml),237 ml of each propofol formulation (5 ml) and ondansetron (4 mg in 2 ml),247 ml of each propofol formulation and thiopental (50 mg in 2 ml),25and 5.4 ml of each propofol formulation (5 ml) and ephedrine (2 mg in 0.4 ml)26(n = 10). As controls, 5.2, 5.4, 5.5, 6.0, and 7.0 ml of each propofol formulation (5 ml) and saline (containing similar volumes of saline as the corresponding pain-reducing agents) were prepared (n = 10).

All samples were transferred to a dialysis membrane bag, and release media was added to produce a total volume of 10 ml. After sealing with a closer, the bag was immersed in 40 ml release medium and shaken for 100 strokes per min at 20°C in a water bath for 24 h.

Free propofol concentrations were measured using high-performance liquid chromatography (Agilent 1100 series; Agilent Technologies, Inc., Santa Clara, CA) with a C18 column (Xterra RP18, 5 μm, 4.6 × 150 mm; Waters Corporation, Milford, MA) and a tetrahydrofuran–water mixture as the mobile phase. The flow rate was 0.7 ml/min, and components of the column effluent were monitored using an ultraviolet detector with the wavelength set at 275 nm. Intraassay precision values were less than 2.44%. Interassay precision values were less than 7.27% (within day) and 4.57% (between day), respectively. Intraassay accuracy values were 89.88–101.64%, whereas interassay accuracy values were 90.27–101.46% of the nominal value.

Six adult healthy volunteers (M:F = 5:1) aged 29–39 yr with no medical history or medication were enrolled in this experiment.

Seven plastic syringes were prepared to contain the following samples at room temperature: 1.5 ml saline, LCT propofol, 10% lipid solvent (Intralipid; Kabi Pharmacia AB, Stockholm, Sweden), microemulsion propofol, 10% PP188, 0.7% polyethylene glycol 660 hydroxystearate, and a mixture of other ingredients in microemulsion propofol minus propofol, PP188, and polyethylene glycol 660 hydroxystearate. Saline (1.5 ml) was used as the control.

A 20-gauge angiocatheter was placed in a vein of the antecubital area, and 3.5 ml venous blood aspirated over 10 s from the catheter using prepared syringes. All samples were transferred to plain tubes containing 0.5 ml inhibition solution composed of aprotinin (10,000 KIU/ml), soybean trypsin inhibitor (800 μg/ml), and polybrene (4 mg/ml) for inactivation of plasma and glandular kallikrein and other kinin-producing enzymes, as well as 1,10-phenanthroline (10 mg/ml) and EDTA (20 mg/ml), for kinin-destroying enzymes.27Using this inhibition solution, bradykinin generation by propofol formulations and their components can be measured.

After shaking gently for 20 s, samples were centrifuged at 1,600g  for 15 min at 4°C. The plasma was collected and stored at −70°C until assay.20The procedure for the extraction of peptides from plasma is described as follows: plasma (2 ml) was acidified with an equal amount of buffer A (1% trifluoroacetic acid; Sigma-Aldrich, Inc., St. Louis, MO), and centrifuged at 15,000g  for 20 min at 4°C. The supernatant was loaded on to the pretreated separation column (Strata C18-E; Phenomenex, Inc., Torrance, CA), which was slowly washed with buffer A (3 ml, twice). The peptide was slowly eluted with 3 ml buffer B (buffer A containing 60% acetonitrile; Fisher Scientific, Pittsburgh, PA), and the eluent was collected in a polypropylene tube. The organic layer in the eluent was removed using a centrifugal concentrator (Savant Speedvac SPD2010; Thermo Fisher Scientific, Inc., Waltham, MA) for 15 min, and the remaining sample was subjected to freeze-drying overnight using a lyophilizer (Freeze Dry System; Labconco Corporation, Kansas City, MO).

The bradykinin concentration in samples was measured with a radioimmunoassay kit (Phoenix Pharmaceuticals, Inc., Burlingame, CA). Standard peptide was reconstituted with radioimmunoassay buffer provided by the manufacturer. Concentrated (lyophilized) bradykinin powder was dissolved in 250 μl radioimmunoassay buffer, and divided into 100-μl portions in duplicate. Appropriate dilutions of standard peptides (1,280, 640, 320, 160, 80, 40, 20, and 10 pg/ml) were used to generate a standard curve and were assayed in quadruplicate. Aliquot samples were assayed in duplicate. Assays were performed in 12 × 75-mm polystyrene tubes. After the addition of primary antibody, each tube was vortexed and incubated for 16–24 h at 4°C. Next, ¹²⁵I-peptide was added to each tube, and the procedure was repeated. Goat anti-rabbit immunoglobulin G and normal rabbit serum were added to each tube, vortexed, and incubated at room temperature for 90 min. Radioimmunoassay buffer was added and centrifuged at 1,700g  for 20 min after gentle vortexing. The supernatant was aspirated, and assay tubes evaluated as counts per minute using a gamma counter (Packard Cobra Gamma Counters, Downers Grove, IL). A standard curve (r  ²≥ 0.993) was obtained by plotting standard peptide concentrations and was used to determine the bradykinin concentration. The measured bradykinin level was converted into plasma bradykinin (pg/ml) using a factor of 8. Specifically, lyophilized powder was made from 2 ml plasma and dissolved into 250 μl radioimmunoassay buffer. In cases where the measured bradykinin concentration was above or below the range of the standard curve, samples were diluted or concentrated accordingly. The sensitivity level was 0.81 pg/ml, and coefficients of variation for the intraassay and interassay were 1.8–5.5% and 1.4–7.1%, respectively.

Thirty adult male Sprague-Dawley rats (weight: mean ± SD, 450.9 ± 57.9 g; range, 347–560 g) were included in the study. The rats were housed with a controlled light–dark cycle (light on between 6:00 am and 6:00 pm), an ambient temperature of 21°–22°C, and unlimited access to standard laboratory rat diet and water.

The animals were anesthetized with sevoflurane in 100% O²(2–3 vol%). A cannula for drug infusion was inserted into the right jugular vein and placed into the superior vena cava 0.5 cm above the right atrium. The cannula was tunneled under the pelt and externalized on the dorsal surface of the neck. The right femoral artery was cannulated to measure systemic arterial pressure and to collect blood samples. Tracheostomy was performed for artificial ventilation. For proper monitoring of the anesthetic status of rat, previous method was modified.28,29Briefly, electroencephalography was monitored over the lateral left parietal cortex (Ch 1) and dorsal hippocampus cortex (Ch 2). For this purpose, rats were placed in a stereotaxic frame (model 900; David Kopf Instruments, Tujunga, CA), four burr holes (1-mm diameter) were drilled into the skull, and electrodes (Bone Anchor Screw; Stoelting, Wood Dale, IL) were threaded as follows. One was placed midline in the frontal bone and used as a reference; one was placed midline in the occipital bone and used as a ground. One electrode per parietal bone was placed 3–5 mm lateral and 4–5 mm posterior to the bregma. For better electrical conduction between the dura and electrodes, the burr holes were filled with paste for electroencephalography (Elefix; Nihon Kohden, Tokyo, Japan). Electrodes were connected to commercial eight-pin serial computer connector, and the entire electrode and connector assembly was insulated and bonded to the skull with dental cement. Vecuronium (2 mg/kg) was administered intravenously to paralyze the animal. Then, the animal was artificially ventilated using a ventilator for small animals (Inspira ASV; Harvard Apparatus, Holliston, MA). The animal was placed in a supine position, and the head was extended and firmly immobilized in a stereotaxic frame. The electroencephalography connector was connected to the electroencephalography receiver and computer. Mean arterial pressures (MAPs) were continuously recorded from the arterial cannula using a patient monitor (Datex-Ohmeda S/5; Planar Systems, Inc., Beaverton, OR). Body temperature was monitored rectally and was maintained by a heating lamp. After the completion of initial instrumentation, an additional 1 h served as a stabilization period. Anesthesia was maintained with sevoflurane adjusted to keep MAP within ±15% of baseline values during the stabilization period. For 20 min before the administration of microemulsion propofol, sevoflurane was washed out. Depending on the raw electroencephalographic signal, sevoflurane was readministered after discontinuation of microemulsion propofol.

Blood gas analysis from an arterial sample was performed shortly before and approximately 20 min after administration of the study drug as well as at the end of the experiment.

The pharmacokinetics of microemulsion propofol were determined upon intravenous administration of 0.5, 1.0, and 1.5 mg · kg⁻¹· min⁻¹(n = 10 at each dose) during 20 min. The microemulsion propofols were administered using an infusion pump (Pump 11 Pico Plus; Harvard Apparatus) and a Hamilton syringe (Hamilton Company, Reno, NV).

For propofol analysis, arterial blood samples of 300 μl each were collected immediately before and at 2, 5, 7, 10, 15, 20, 30, 45, 60, and 90 min after administration. Samples were collected in heparinized tubes and were stored at −70°C until assay. Drawn blood was substituted by twice the volume of Hartman solution for maintenance of blood pressure.

Propofol was isolated from whole blood by extraction using pretreatment with deproteinization and was determined by high-performance liquid chromatography with fluorescence detection. Plasma proteins were precipitated with acetonitrile. The supernatants were analyzed by high-performance liquid chromatography using a Capcell Pak C18 UG120 column (Shiseido Fine Chemicals, Tokyo, Japan) and a mixture of acetonitrile and 0.1% trifluoroacetic acid in water (56:44, vol/vol) as a mobile phase. The components of the column effluent were monitored by a fluorometric detector with excitation and emission wavelengths set at 276 and 310 nm, respectively.

The lower limit of quantification of propofol was 50 ng/ml. The calibration curve was linear over the range of 50–5,000 ng/ml, with the coefficients of determination (R  ²) greater than 0.999 for all cases. Intraassay precision values were less than 5.74%. Interassay within-day and between-day precision values were less than 3.65% and 9.33%, respectively. Intraassay accuracy values were 86.66–103.94% of the nominal value. Interassay accuracy values were 99.06–103.72% of the nominal value.

Pharmacokinetic parameters were calculated by noncompartmental methods (WinNonlin Professional 5.2; Pharsight Corporation, Mountain View, CA). The area under the curve from the time of administration to the last measured concentration (AUC^last) was estimated by linear trapezoidal integration (linear interpolation). The area under the curve from administration to infinity (AUC^inf) was calculated as the sum of AUC^last+ C^last/λ^Z, in which C^lastis the last measured concentration and λ^Zis the apparent terminal rate constant estimated by unweighted linear regression for the linear portion of the terminal log concentration–time curve. The maximal concentration (C^max) and the time to reach C^max(t^max) after an intravenous infusion of microemulsion propofol were determined from the observed data. Summary statistics were determined for each parameter. To evaluate dose linearity, dose-normalized AUC^infvalues among each dose group were compared using one-way analysis of variance (ANOVA).30If the differences among three groups are statistically insignificant, we conclude that pharmacokinetics show dose linearity. In addition, we used a power model and confidence interval criteria approach as follows.30 

where dose linearity implies that β¹= 1 in equation 1and PK denotes a pharmacokinetic variable.

One-, two-, and three-compartment models with linear pharmacokinetics were fitted using an ADVAN 6 subroutine and the first-order conditional estimation procedure of NONMEM® VI (GloboMax LLC, Ellicott City, MD). The interindividual random variability on each of the model parameters was modeled using a log-normal model. A diagonal matrix was estimated for the different distributions of ηs, where η is interindividual random variability with mean zero and variance ω². A constant coefficient of variation model was used for the residual random variability. In addition, the time course of the propofol concentration was modeled using nonlinear models with Michaelis-Menten elimination and the well-stirred model.31The latter is one of hepatic clearance models to describe hepatic elimination of drugs, in which drug concentration is assumed to be constant throughout the hepatic compartment and equal to the outflow concentration.

For Michaelis-Menten elimination, clearance is defined as follows:

where V^maxis the maximum metabolic rate and K^mis the Michaelis-Menten constant in equation 2.

For the well-stirred model,

where Cl^His hepatic clearance, Q^His hepatic perfusion rate, f^uis the free fraction of a drug, and Cl^intis intrinsic clearance in equation 3. For simplicity, we can denote f^u· Cl^intby Cl^int.

The electroencephalographic activity of two-channel was continuously recorded by QEEG-8 (LXE3208; Laxtha Inc., Daejeon, Korea). The sampling frequency of the analyzed channel was 256 Hz. Baseline electroencephalographic activity was recorded for 5 min before continuous infusion of microemulsion propofol.

The raw electroencephalographic signal was filtered between 0.5 and 50 Hz for on-line calculation of the electroencephalographic approximate entropy (ApEn; Telescan version 2.85 and Complexity version 2.7; Laxtha Inc.) which quantifies the regularity of the data time series and is known to show better baseline stability as a measure of the arousal state of the central nervous system than other univariate descriptors.17,32To calculate ApEn, the length of the epoch (N) was 2,056, the number of previous values (m) used to predict the subsequent values was 2, and a filtering level (r) was 15% of the SD of the amplitude values. Smoothing by means of a simple moving average was applied to the calculation of ApEn. The number of neighboring points was seven. Serious artifacts were excluded by checking the maximum amplitude for each epoch; if the amplitude was greater than 200 μV, the epoch was excluded. The appropriateness of artifact rejection was manually confirmed. Artifact rejection and analysis of each electroencephalographic parameter were performed by a single, blinded, experienced analyst.

To account for the time lag (hysteresis) between the onset of the effect and the course of the plasma concentration, the pharmacodynamics were described using an effect compartment model in which k^e0, a first-order elimination rate constant characterizing effect site equilibration, was used to estimate the apparent effect site concentrations. For each animal, the relation of ApEn with propofol effect site concentration was analyzed using a sigmoid E^maxmodel:

where Effect is ApEn, E⁰is the baseline ApEn when no drug is present, E^maxis the ApEn value for maximum possible drug effect, Ce is the calculated effect site concentration of propofol, Ce⁵⁰is the effect site concentration associated with 50% maximal drug effect, and γ is the steepness of the concentration-versus -response relation in equation 4.

Interindividual random variability of Ce⁵⁰and k^e0was modeled using a log-normal model, and no interindividual variability of E⁰, E^max, or γ was assumed. A diagonal matrix was estimated for the different distributions of ηs. Residual random variability was modeled using an additive error model.

The models were evaluated using statistical and graphical methods. The minimal value of the objective function (equal to minus twice the log likelihood) provided by NONMEM® was used as the goodness-of-fit characteristic to discriminate between hierarchical models using the log likelihood ratio test.33A P  value of 0.05, representing a decrease in objective function value of 3.84 points, was considered statistically significant (chi-square distribution, df = 1). R (version 2.6.1; R Foundation for Statistical Computing, Vienna, Austria) was used for graphical model diagnosis.

To compare the distribution of ηs to the normal distribution, we calculated the correlation of the points in the normal probability plot in which the null hypothesis of the test is that the data come from a normal distribution.34,35If we obtain a P  value greater than 0.05, we fail to reject the null hypothesis and conclude that normality is a reasonable assumption.

The weighted residual was calculated as (measured − predicted)/predicted. The median weighted residual and median absolute weighted residual were calculated to examine the quality of the prediction of the pharmacokinetic models for the population. We used the median absolute residual and its percentage of pharmacodynamic range (E⁰− E^max) of the final model to describe the quality of prediction for the population.

For both the final pharmacokinetic and pharmacodynamic models, a nonparametric bootstrap analysis was performed as an internal model validation, using the software package Wings for NONMEM® VI (Holford N, version 600; Auckland, New Zealand).36This process was repeated 2,000 times. The final model parameter estimates were compared with the median parameter values and with the 2.5–97.5 percentile of the nonparametric bootstrap replicates of the final model.

The statistical analyses were performed with SPSS (version 12.0; SPSS Inc., Chicago, IL). The concentrations of free propofol in the aqueous phase in microemulsion and LCT propofol were analyzed using a t  test or Mann–Whitney rank sum test, as appropriate. Differences in measured concentrations of bradykinin in plasma between the groups were compared using one-way ANOVA. Changes in blood pressure were tested for statistical significance using repeated-measures ANOVA and the Tukey post hoc  test. Baseline (E⁰) and maximally decreased ApEn (E^max) and the difference between E⁰and E^maxvalues of the lateral left parietal cortex and dorsal hippocampus cortex were compared using a t  test or Mann–Whitney rank sum test, as appropriate. E⁰and E^maxand the difference between E⁰and E^maxvalues among each infusion rate were compared using one-way ANOVA or Kruskal–Wallis one-way ANOVA on ranks, as appropriate. In all comparisons, a P  value less than 0.05 was considered statistically significant. Data are the mean ± SD unless stated otherwise.

The pH, osmolarity, and particle size of microemulsion propofol were 7.5, 280 mOsm/l, and 67.0 ± 28.5 nm, respectively.

The free propofol concentrations in the aqueous phase of LCT propofol and microemulsion propofol with or without agents reducing propofol-induced pain are shown in table 1. The aqueous free propofol concentration of microemulsion propofol was five times higher than that of LCT propofol, suggesting the possibility of more severe and frequent pain on injection. Notably, no agents showed any influence on the level of the aqueous free propofol in either formulation.

Plasma bradykinin concentrations when LCT propofol and microemulsion propofol and their components were mixed with human blood are shown in figure 1. No significant differences among all agents used in this experiment were observed, which is contradictory to other studies.37,38These findings suggest that bradykinin generation induced by contact of both formulations with plasma may not be associated with injection pain.

Changes of MAP over time for all animals are shown in figure 2. MAP decreased significantly at the infusion rates of 1.0 and 1.5 mg · kg⁻¹· min⁻¹, compared with baseline MAPs and those at the infusion rate of 0.5 mg · kg⁻¹· min⁻¹. MAP at the infusion rates of 1.0 and 1.5 mg · kg⁻¹· min⁻¹was recovered to baseline values approximately 90 min after start of infusion. However, MAP between 1.0 and 1.5 mg · kg⁻¹· min⁻¹at each time point did not show statistically significant difference. The SD of MAP at each time point ranged from 9.2 to 16.4 mmHg for 0.5 mg · kg⁻¹· min⁻¹, from 13.2 to 30.3 mmHg for 1.0 mg · kg⁻¹· min⁻¹, and from 8.3 to 27.8 mmHg for 1.5 mg · kg⁻¹· min⁻¹, suggesting large interindividual variability of MAP at the infusion rates of 1.0 and 1.5 mg · kg⁻¹· min⁻¹. Normocapnia (arterial carbon dioxide tension = 35.5 ± 6.6 mmHg) and normothermia (36.5°± 0.4°C) were maintained throughout the study period.

The infusion of microemulsion propofol was started 244.1 ± 47.3 min after the beginning of the surgical preparation. The time course of the propofol concentration is illustrated in figure 3.

In all animals, at least 80% of the total area under the curve was covered by measured concentrations. Table 2shows the pharmacokinetic parameters that are calculated by noncompartmental methods. The relation between the total dose and the total area under the curve after log transformation is presented in figure 4and revealed dose nonlinearity. These nonlinear pharmacokinetics were also evidenced by the finding that the areas under the plasma concentration–time curve from time 0 to infinity normalized by dose (AUC^inf/dose) for the infusion rates of 0.5, 1.0, and 1.5 mg · kg⁻¹· min⁻¹were 29.9 ± 8.4, 48.6 ± 15.8, and 52.1 ± 10.2 min ·μg · ml⁻¹· mg⁻¹, respectively (P < 0.001).

For microemulsion propofol, the decline of the plasma concentration was biphasic, with an initial fast decline and a slower terminal elimination. A two-compartment model best described the time course of the propofol concentration, whereas nonlinear models with Michaelis-Menten elimination and the well-stirred model were not successful to be fitted to the pharmacokinetic data. We also tried to include MAP as a time varying covariate for CL1 in table 3but did not show statistical significance despite successful minimization. This might be attributed to large SDs of MAP, particularly at the infusion rates of 1.0 and 1.5 mg · kg⁻¹· min⁻¹. Table 3summarizes the results of the final population pharmacokinetic model. All of the ηs except η for k¹⁰and k¹²were normally distributed. Measured and predicted concentrations of propofol over time are depicted in figure 5.

The stability of baseline (E⁰) and maximally decreased ApEn (E^max) and the difference between E⁰and E^maxvalues of the lateral left parietal cortex and dorsal hippocampus cortex (both referenced by frontal cortex) in rats are shown in table 4. The overall baseline stability and the overall difference between E⁰and E^maxof the lateral left parietal cortex were better and larger than those of dorsal hippocampus cortex, respectively. Especially the difference between E⁰and E^maxvalues of the lateral left parietal cortex tended to increase more consistently in a dose-dependent manner. Therefore, ApEn derived from the lateral left parietal cortex (Ch 1) was chosen for pharmacodynamic modeling. The time course of the electroencephalographic entropy (ApEn) derived from the lateral left parietal cortex during and after the infusion of microemulsion propofol in rats is shown in figure 6.

Estimates of the population parameters of the final pharmacodynamic model for microemulsion propofol are summarized in table 5. The equilibration half-life, t^1/2ke0, was approximately 0.68 min. The ηs of Ce⁵⁰and k^e0in the final pharmacodynamic model were normally distributed. Observed and predicted ApEn values over time are shown in figure 7.

Free propofol concentration in the aqueous phase of the reformulated microemulsion propofol was five times higher than those of LCT propofol. Both formulations and their components did not generate bradykinin in plasma on mixing with human blood. Our in vitro  studies provide the evidence that both free propofol and lipid solvent do not activate the plasma kallikrein–kinin system and lipid solvent in LCT propofol decreases free propofol concentration in the aqueous phase, and hence plays an important role in decreasing propofol-induced pain. Injection pain of a number of sedative and hypnotic drugs is possibly caused by formulations of extremely unphysiologic osmolality.39Because the pH and osmolarity/osmolality of microemulsion and lipid emulsion are almost within physiologic ranges (pH 6–8.5 and 0.303 Osm/kg for lipid emulsion propofol40), these factors are unlikely to cause injection pain. Therefore, propofol-induced pain may be attributed to the direct irritation of free nerve endings by free propofol in the aqueous phase.

The exact mechanism of action of lidocaine, ketamine, metoclopramide, ondansetron, thiopental, and ephedrine reducing propofol-induced pain is still unclear. In this study, the mixtures of LCT propofol and microemulsion propofol with these agents neither decreased nor increased the level of free propofol in the aqueous phase. Considering this result as well as plasma bradykinin study, these agents may act via  local effects to reduce the irritation effect of free propofol.

In a previous study, the central volume of distribution (V^c) and the volume of distribution at steady state (V^ss) of LCT propofol were 0.13 and 0.8 l/kg, respectively.14Compared with these results, microemulsion propofol showed similar V^c(0.143 l/kg) but slightly smaller V^ss(0.565 l/kg), which may be due to the influence of formulation on the pharmacokinetics of propofol or a different infusion scheme as well as a different sampling schedule. Despite differences in contents, the formulation effect on the pharmacokinetics of propofol was also demonstrated in our previous study in which microemulsion propofol may be less extensively distributed to peripheral tissues because of a relatively lower tissue/blood partition coefficient of microemulsion system as a propofol vehicle.5 

Linear kinetic models permit the application of the superposition principle, in that doubling the dose will consequently lead to a doubling of the concentrations.31In this study, dose-normalized area under the curve was increased with increasing dose, suggesting nonlinear disposition, i.e. , a decrease in clearance. Propofol has a hepatic extraction ratio of nearly 0.8, and hepatic clearance accounts for a majority of the total propofol clearance.41Propofol reduces its own clearance with an increasing dose by reducing cardiac output and hepatic blood flow. The larger decrease in MAP at 1.0 and 1.5 mg · kg⁻¹· min⁻¹(−23% and −35% vs.  baseline MAP) compared with 0.5 mg · kg⁻¹· min⁻¹(−4% vs.  baseline MAP) in our study could potentially reduce hepatic perfusion and alter the pharmacokinetic profile of microemulsion propofol.

To describe nonlinear pharmacokinetics of a drug with high hepatic extraction ratio such as propofol, Michaelis-Menten elimination may be inappropriate, because it assumes that metabolism and/or renal secretion usually approach the capacity of the elimination system and the elimination rate is no longer strictly a constant.31The well-stirred model, also called the venous equilibrium model, may be more appropriate to describe nonlinear kinetics caused by changes in the perfusion of the liver.31The failure to fit the well-stirred model to pharmacokinetic data in this study may be attributed to either the small size of sample or insufficient and/or inhomogeneous decrease of MAP and hence hepatic blood flow, which may be supported by the large SDs of MAP at the infusion rates of 1.0 and 1.5 mg · kg⁻¹· min⁻¹. For highly cleared drugs, also called high-extraction compounds (hepatic extraction ratio ≥ 0.7), f^u· Cl^intin equation 3is much greater than Q^H, and then f^u· Cl^intcancels out, which means that hepatic clearance is nearly equal to hepatic perfusion rate.42When assuming this for simplicity at the circumstance that hepatic blood flow is unknown or cannot be expressed as a function of MAP, the well-stirred model in this study is not distinguished mathematically from the two-compartment model with linear pharmacokinetics.

When discussing nonlinear kinetics, the relevant dose range should be considered because nearly every drug can have nonlinearity at an extreme dose. The infusion rates used in this study are comparable to the doses used in the studies for LCT propofol, in which the infusion rate of 0.5 to 1.3 mg · kg⁻¹· min⁻¹in rats can produce light anesthesia and deep anesthesia.14,43,44Because of the potential for nonlinear kinetics, caution should be taken when extrapolating the pharmacokinetic data in this study before making predictions regarding the pharmacokinetic behavior of microemulsion propofol at other doses in rats. However, the median weighted residual and median absolute weighted residual of the pharmacokinetic model in this study were comparable to those in a previous study using a two-compartment model with Michaelis-Menten elimination in rats, in which the median weighted residual and median absolute weighted residual were −1.6% and 30.5%, respectively.14 

The estimates of γ and Ce⁵⁰of microemulsion propofol were lower than those of a previous study, in which modified median frequency was used as a surrogate measure of the effect of LCT propofol on the central nervous system.14Apparently, microemulsion propofol may be more potent than LCT propofol, and ApEn seemed to change less steeply than modified median frequency did. However, different surrogate measurement used for pharmacodynamic modeling might lead to different estimation of the pharmacodynamic parameters such as γ and Ce⁵⁰. Therefore, caution should be taken when comparing the results of both studies.

Because the placement of electroencephalographic electrodes in rats was not standardized, we selected and modified the methods that generally have been used in rat sleep research29,45,46and an epileptic model.28Although stainless steel electroencephalographic needle electrodes were placed occipito-occipitally in other studies,16,47the time course of ApEn in this study showed dose-dependent changes, suggesting the global effects of microemulsion propofol on the central nervous system in rats.

In conclusion, microemulsion propofol may have the potential to produce frequent and severe pain on injection because it contains higher concentrations of free propofol in the aqueous phase compared with those of the LCT propofol. Therefore, when microemulsion propofol is used in clinical settings, the techniques known to reduce pain on injection must be implemented. At the dose range used in this study, microemulsion propofol showed nonlinear pharmacokinetics, which may be caused by the prominent hemodynamic depression.

The authors thank Ae-Kyung Hwang, B.S. (Technician), and Hyun-Jeong Park, B.S. (Technician), in the Clinical Research Center of Asan Medical Center (Seoul, Korea) for measurement of propofol concentrations.