| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
BRIEF REPORTS/PHARMACOKINETICS |
From the Abbott Laboratories, Abbott Park, Illinois.
Address for reprints: Sandeep Dutta, PhD, Abbott Laboratories, Dept. R4PK, Bldg. AP13A, 100 Abbott Park Road, Abbott Park IL 60064-6104; e-mail: Sandeep.Dutta{at}abbott.com.
Key Words: Valproate • controlled-release • absorption • model • sustained-release
Divalproex sodium extended-release (divalproex-ER) tablet is a novel once-daily formulation that offers the potential advantages of improved medication adherence, increased patient convenience, and easier titration to maximum effective dose while possibly reducing the risk of side effects (by reducing peak-trough fluctuations in plasma concentrations). Divalproex-ER is based on hydrophilic matrix technology; drug release is controlled primarily by erosion of a water-soluble polymer (hydroxypropylmethyl cellulose) from the matrix.
We have previously quantified the mean input absorption rate parameter value for valproic acid (VPA) from orally administered divalproex-ER as 0.0431 mg/h/mg dose, representing a pseudo zero-order absorption rate from this formulation.1 Closer inspection of this VPA absorption profile from divalproex-ER revealed that our characterization could be optimized. In particular, we needed to distinguish this formulation's true absorption characteristics that cannot be characterized by simple absorption models (eg, first- and zero-order) and several atypical absorption models (mixed zero- and first-order, Weibull function, time- and/or gastrointestinal [GI]-location-dependent, exponential, and several other absorption models) that have been proposed for other drugs or formulations.2
The purpose of this investigation was to expand our initial observations and more fully characterize, using a unique model, the absorption characteristics and single-dose pharmacokinetics of divalproex-ER tablets.
METHODS
Study Design
This single-dose, nonfasting, open-label, single-center study was conducted according to a 4-period, crossover design in 16 healthy adult subjects (10 men and 6 women); mean (SD) age was 34 (11.3) years, and mean (SD) body weight was 73 (9.3) kg.1 Before the performance of any screening and study-specific procedures, applicable written informed consent was obtained from each subject. The Institutional Review Board of Victory Memorial Hospital approved the study protocol.
Subjects were randomly assigned in equal numbers to 4 sequences of study regimens, and subjects received each of the 4 regimens upon completion of the study. Doses in the consecutive periods were separated by 1 week. The 4 treatments were 500-mg doses of 3 test divalproex-ER formulations and a reference 12-hour intravenous (IV) VPA (Depacon, Abbott Laboratories, IL) infusion. Only data from 2 treatment arms, the reference IV-VPA and 1 of the 3 test divalproex-ER formulations that was subsequently developed and is currently marketed, are of relevance and are presented here.
Blood samples were collected into evacuated heparinized collection tubes before dosing (0 hour) and at 1, 2, 3, 4, 5, 6, 7.5, 9, 10.5, 12, 13, 15, 18, 24, 36, 48, and 72 hours after dosing in each regimen. Plasma samples were assayed for VPA concentrations using a validated gas chromatography method with flame ionization detection (GC/FID).1 The lower limit of quantification was 0.5 mg/L, and the assay was unbiased with fewer than 4% coefficient of variation.
In Vivo Absorption Rate Profile
The in vivo absorption rate profile of divalproex-ER was inferred by deconvolution (WinNonlin, Pharsight Corp, Mountain View, CA) of the divalproex-ER plasma concentration-time profile against the IV-VPA concentration-time profile. A 1-compartment model was fit to the observed IV plasma VPA concentration-time data for each subject to obtain the parameters of the disposition function (elimination rate constant, 
, and the IV bolus intercept, C0), as shown in equation 1.
 | (1) |
= elimination rate constant for a 1-compartment open model, and 
= infusion duration.
The observed divalproex-ER plasma VPA concentration-time data for each subject were deconvolved using that subject's disposition parameters ( and C0) using WinNonlin to obtain the input or absorption rate versus time profile (see equation 2).
 | (2) |
The rate of VPA absorption from the divalproex-ER tablets was determined by deconvolution analysis using the software WinNonlin. The methodology is based on linear system analysis with linearity defined in the general sense of the linear superposition principle.
The concentration, C(t), of a drug at the sampling space (eg, site of measurement such as blood or plasma) after oral administration may be represented as equation 3.
 | (3) |
) represents the time since the small amount of drug described by f() was introduced in the system over a small interval of time 
. f(t) is the function that, when integrated between the limits of t = 0 and t, yields the cumulative amount of drug delivered to the impulse input point; * denotes the convolution operation. Equation 3 is the convolution equation that forms the basis for the evaluation of the drug input rate f(t). The function g(t) is the unit impulse response (ie, characteristic response or disposition function). WinNonlin uses the basic principle of deconvolution through convolution to determine the input function.
Multiphasic Absorption Rate Model
The drug input rate profile, f(t), for a unit dose determined by deconvolution was modeled as
 | (4) |
, 
, and 
= sigmoidicity parameters, for model reduction purpose = 2 and = 4.
Plasma Valproic Acid Concentration-Time Profile of Divalproex-ER
The multiphasic absorption model was coupled with a monoexponential disposition model to describe the divalproex-ER plasma concentration-time profile using ADAPT.
 | (5) |
RESULTS AND DISCUSSION
The multiphasic absorption rate versus time and the plasma VPA concentration-time profiles of divalproex-ER are presented in Figure 1, panels a and b, respectively. The VPA pharmacokinetic parameters are presented in Table I. Valproic acid absorption from divalproex-ER starts upon tablet wetting and occurs slowly and continuously over an average 22-hour period without dose dumping. This extended absorption (rate) is characterized by multiple phases. This is the first report of a novel multiphasic absorption model developed to optimally describe and quantify VPA absorption from divalproex-ER.
|
|
The multiphasic absorption model may be considered as a forcing or interpolating function that describes a distinctive absorption profile of (1) an initial increase in absorption rate as the drug moves from the gastric space to the upper segments of the small intestine where the absorption rate is at its peak (Amax + A0, followed by (2) attenuation of this peak absorption rate as the drug travels down the small intestine and enters the large intestine, followed by (3) an extended zero-order absorption in the large intestine at a slower rate (A0), followed by (4) attenuation of the absorption rate as the drug leaves the GI tract.
In general, the model-estimated values of the transit times calculated herein were consistent with physiologic expectations and previously published work.3-8 The optimal model was chosen based on visual inspection of the fit of several different absorption models (eg, first- and zero-order, biphasic, Weibull, double-Weibull),2 coupled with a monoexponential disposition model, to the data, the 95% confidence interval of the estimated parameters, Akaike Information and Schwarz Bayesian Criteria, and physiologic relevance of the selected model. Only the absorption initiation time for the drug (AIT50) was not robustly estimated (not significantly different from zero). The short AIT50 is likely attributable to VPA dissolution in the gastric space, followed by immediate absorption in the stomach and upper segments of the small intestine. Although VPA dissolution in the gastric space may be limited by rate of drug release from the divalproex-ER polymer matrix and available absorption area, VPA dissolution and absorption starts immediately after oral ingestion. It should be noted that VPA, which is highly water soluble, is absorbed to some extent from the stomach.9,10 Hence, the true absorption initiation time is likely very short and difficult to estimate robustly. However, this portion of the model that characterizes drug entry into the small intestine, followed by increase in absorption rate, is consistent with physiologic expectation and is essential to enforce the plasma concentrations to rise from zero at time zero. The estimated oral clearance and volume of distribution values were consistent with results from other VPA studies.11
Several absorption models, including the simple (eg, first- and zero-order absorption) and more complex models (eg, biphasic, Weibull, and double-Weibull),2 were explored for optimally characterizing the deconvolved absorption-rate time profile. These models failed to adequately describe the absorption rate versus time profile. It should be noted that the absorption rate versus time profile has greater sensitivity in discriminating different absorption models than a cumulative amount absorbed versus time profile, for example, the double-Weibull absorption model was found to characterize the cumulative absorption profile quite well but was not able to describe absorption rate versus time profile adequately (data for the poor fit of these models are not shown here).
As evident from Figure 1a, divalproex-ER exhibited a complex, extended absorption rate versus time profile that had distinct phases. These multiple phases of absorption were consistent with GI physiology (transit times, absorption surface area, and water content). Therefore a multiphasic absorption model was developed to optimally characterize and quantify the absorption rate versus time profile. In this model, the sigmoidicity parameter was constrained to be larger than , and was constrained to be larger than . The characterizes the rapid rise of absorption rate as the dissolved drug enters the small intestine and is immediately absorbed, whereas the attenuates the absorption rate as the drug travels down the small intestine and gradually enters the large intestine. Because the initial rise to maximum absorption rate is not an impulse function (eg, an infusion rate that becomes zero when the infusion is stopped) in this model, the was constrained to be larger than to (1) compensate for the increase in absorption rate resulting from submodel characterizing initiation of absorption and (2) attenuate the maximum absorption rate (observed in the initial segments of the small intestine) as the drug travels down the GI tract. Similarly, was constrained to be larger than to further attenuate the absorption rate to zero as the drug leaves the GI tract. An alternative model in which the sigmoidicity parameters were constrained in an additive manner (ie, , + and + + ) was also explored. However, the current model ( = 2 and = 4) was chosen as the optimal model because (1) it was more stable (although slightly less flexible) than the additive sigmoid parameters model and (2) a single sigmoid parameter () was robustly estimated, whereas robust estimates of 3 independent sigmoid parameters could not be obtained.
Even though the model-estimated parameter values reported here are generally consistent with published GI transit times, the parameters should be interpreted with caution because the model describing the absorption phase of VPA from divalproex-ER has several highly nonlinear components, with many parameters being highly correlated. It should be recognized that the parameters of this model cannot be estimated robustly in the absence of appropriate IV data, which were used to tease apart the absorption rate profile in this study. In such situations, limited physiologic meaning should be assigned to the estimated parameters. Because the model's utility as a forcing function in simulation experiments is largely based on the model's predictive value, the lack of physiologic meaning of some of the robustly estimated model parameters may not be important in the context of simulations.
Our multiphasic absorption model is more suitable than any of the previously described atypical absorption models used to characterize complex, multiphasic absorption profiles.2 Although some extended-release formulations, including divalproex-ER, are specifically engineered to release the drug very slowly, a pure zero-order absorption rate is unlikely to be achieved because the physiology, distribution of drugmetabolizing enzymes and transporters, water content, and pH changes over the entire GI tract. In the case of VPA absorption from divalproex-ER, inferred from the deconvolved absorption-rate profile, the multiphasic absorption model optimally describes (1) an initial higher absorption rate in the small intestine where higher extent and rate of absorption occurs because of larger surface area and higher water content, (2) an extended zero-order absorption in the large intestine where lower extent and a slower rate of absorption is observed because of the relatively water-deficient environment and lower surface area (per unit length), and (3) a subsequent attenuation in the absorption rate as the drug is eliminated from the body.
This unique model has clinical and practical utility as an interpolating function and can be used to appropriately describe the absorption rate profile for the purpose of simulation of various dosing regimens and clinical situations. Furthermore, our model can be generalized to sustained-release products of other drugs. It is consistent with the physiologic expectations of rapid absorption in the small intestine over 2 to 3 hours and extended slower absorption in the large intestine over a longer duration. This model may be more optimal than the simplest absorption models (eg, first- and zero-order) or the other complex models2 for describing absorption rate versus time profiles from many sustained-release products.
Abbott Laboratories funded this study.
REFERENCES
1. Dutta S, Reed RC, Cavanaugh JH. Absolute bioavailability and absorption characteristics of divalproex sodium extended-release tablets in healthy volunteers. J Clin Pharmacol. 2004;44: 737-742.
2. Zhou H. Pharmacokinetic strategies in deciphering atypical drug absorption profiles. J Clin Pharmacol. 2003;43: 211-227.
3. Davis SS. Assessment of gastrointestinal transit and drug absorption. In: Prescott LF, Nimmo WS, eds. Novel Drug Delivery and Its Therapeutic Application. New York, NY: John Wiley & Sons Ltd; 1989: 89-101.
4. Davis SS, Hardy JG, Fara JW. Transit of pharmaceutical dosage forms through the small intestine. Gut. 1986;27: 886-892.
5. Madsen JL. Effects of gender, age, and body mass index on gastrointestinal transit times. Dig Dis Sci. 1992;37: 1548-1553.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
6. Argenyi EE, Soffer EE, Madsen MT, Berbaum KS, Walkner WO. Scintigraphic evaluation of small bowel transit in healthy subjects: inter- and intrasubject variability. Am J Gastroenterol. 1995;90: 938-942.[Web of Science][Medline] [Order article via Infotrieve]
7. Degen LP, Phillips SF. Variability of gastrointestinal transit in healthy women and men. Gut. 1996;39: 299-305.
8. Fallingborg J, Christensen LA, Ingeman-Nielsen M, et al. Measurement of gastrointestinal pH and regional transit times in normal children. J Pediatr Gastroentero Nutr. 1990;11: 211-214.
9. Bourin M, Guenzet J, Thomare P, Kergueris MF, Ortega A, Larousse C. Effects of administration route on valproate pharmacokinetics in the rabbit. Fundam Clin Pharmacol. 1991;5: 331-339.[Web of Science][Medline] [Order article via Infotrieve]
10. Yeomans ND, Vajda FJ, Baldas J. Bioavailability of valproate after gastric and direct intestinal administration in rats. Clin Exp Pharmacol Physiol. 1982;9: 173-177.[Web of Science][Medline] [Order article via Infotrieve]
11. Dutta S, Reed RC. Divalproex to divalproex-ER conversion. Clin Drug Invest. 2004;24: 495-508.[CrossRef]
CiteULike 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
