0090-9556/97/2510-1215-1218$02.00/0
DRUG METABOLISM AND DISPOSITION
Copyright © 1997 by The American Society for Pharmacology and Experimental Therapeutics
Vol. 25, No. 10
SHORT COMMUNICATION
Application of a First-Pass Effect Model to Characterize the
Pharmacokinetic Disposition of Venlafaxine after Oral Administration to
Human Subjects
 |
Abstract |
Venlafaxine (VEN), a drug used in the treatment of depression,
undergoes significant first-pass metabolism after oral dosing to
O-desmethylvenlafaxine (ODV), a metabolite with comparable therapeutic activity to that of parent drug. The pharmacokinetic disposition of VEN was characterized using a "first-pass" model that incorporates a presystemic compartment (liver) to account for the
first-pass metabolism of VEN to ODV. A series of differential equations
were simultaneously fitted to plasma concentrations of parent and
metabolite. A good fit of the model to observed data was demonstrated,
generating estimates for the following parameters:
ka (1.31 ± 0.009 hr
1),
VVEN (252 ± 87.6 liters),
CLint (65.8 ± 39.7 liters/hr),
RL (liver:plasma partition coefficient,
29.6 ± 18.3), VODV (181 ± 84.1 liters), and CLODV (23.5 ± 12.5 liters/hr). Parameter estimates correlated closely with those obtained
through noncompartmental methods. These results indicate that the
time-course disposition of a compound undergoing first-pass hepatic
metabolism after oral dosing can be successfully modeled.
| |
Article |
The consequences of presystemic metabolism on
the bioavailability of orally administered compounds are
well-established. In this regard, there are several pharmacokinetic
models of presystemic metabolism reported in the literature. Gibaldi
and Feldman introduced a three-compartment model to describe the
first-pass effect (1). Colburn and Gibaldi (2) later proposed a
pharmacokinetic perfusion model to describe the disposition of drugs
that are susceptible to both first-pass hepatic and gut wall
metabolism. Combined with subsequent papers
including those by Rowland
(3, 4), Wilkinson and Shand (5), and Pang and Rowland (6)these
investigators have provided the theoretical basis on which oral
bioavailability is mathematically described. However, pharmacokinetic
models that incorporate the effect of presystemic hepatic metabolism
have seldom been tested experimentally in terms of describing the
time-course disposition of drugs after oral dosing.
VEN1 is a
phenylethylamine derivative used clinically in the management of
depression (7, 8). After oral administration, VEN undergoes extensive
first-pass metabolism by the liver to two minor, less active
metabolites (N-desmethylvenlafaxine and N,O-didesmethylvenlafaxine) and a major
metabolite (ODV). ODV is a compound with antidepressant activity
comparable with the parent drug (9).
In this communication, we demonstrate that first-pass metabolism after
oral dosing can be successfully modeled. Plasma concentration-time data
for both VEN and ODV were simultaneously fitted using a pharmacokinetic model that accounts for first-pass hepatic metabolism of VEN to ODV.
Materials and Methods.
VEN, ODV, and IS (WY-45,818; IS) were supplied by Wyeth-Ayerst Research
(Philadelphia, PA). Acetonitrile was obtained from Mallinckrodt
Chemicals (Orlando, FL). Diethyl ether and sodium borate were purchased
from Sigma Chemical Company (St. Louis, MO).
Study Design.
The study was conducted in accordance with the provisions of the
Declaration of Helsinki and its amendments. Approval from both the Long
Island University Research Approval Committee and Brookdale University
Hospital and Medical Center's Research and Clinical Projects Committee
was obtained. Subjects gave written informed consent to participate in
the study. Five healthy male volunteers (ages 23-38) provided a
medical history and were given a physical examination, including blood
chemistry and hematological tests before initiation of the study.
After a 7-hr overnight fast, subjects were given a 1.5 mg/kg
dose of VEN tablets, rounded to the nearest 18.75 mg, with 6 oz of
water. Plasma samples were collected at time 0 (before drug administration), and at 0.5, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 10, 12, 24,
and 36 hr post dose. Samples were stored at 
20°C before analysis.
Drug Analysis.
VEN and ODV were quantitated in plasma samples by HPLC through slight
modification of a previously reported assay (10). To 1 ml of plasma
sample, 50 µl of IS (0.15 mg/ml), 300 µl of saturated sodium borate
solution (pH 9), and 5 ml of diethyl ether were added. The mixture was
vortexed and centrifuged at 2,500 rpm for 10 min. Three hundred
microliters of 0.01 N HCl was added to the organic phase, and the
mixture was vortexed and centrifuged at 2,500 rpm for 10 min. The
organic phase was then discarded, and the resultant solution was
aerated under mild heat to remove any dissolved ether. One hundred
microliters of the extract was injected into the HPLC system. The HPLC
consisted of a Thermo Separation P1000 Solvent Delivery Pump, a UV1000
Ultraviolet Detector, and a personal computer with PC1000 Integration
Software (Thermo Separation Products, Riviera Beach, FL). Separation
was accomplished with a Supelcosil LC8-DB deactivated base column
(Supelco, Bellafonte, PA) using a mobile phase consisting of 0.1 M
ammonium phosphate buffer (pH 4.4) and acetonitrile (25.5%). Mobile
phase was introduced at a flow rate of 1 ml/min. The detection
wavelength was 229 nm. VEN and ODV concentrations were calculated using
a peak height ratio (drug:IS) based on individual standard curves.
Minimum detectable concentrations of VEN and ODV were 10 and 25 ng/ml,
respectively.
Data Analysis.
A pharmacokinetic model was simultaneously fitted to both VEN and ODV
plasma concentrations using the least squares regression program
PCNONLIN (Statistical Consultants, Apex, NC). A schematic illustration
of the proposed model is provided in fig.
1. There are several underlying
assumptions to the model:
| 1. |
VEN is 100% metabolized by the liver. Although previous
investigations found that ~5% of VEN is excreted unchanged by the kidney (11), introduction of a renal clearance parameter into the
proposed model did not significantly reduce the weighted sums of
squares.
|
| 2. |
VEN is 55% metabolized to ODV (11, 12).
|
| 3. |
Hepatic blood flow is 90 liters/hr, and liver volume is 1.5 liters (13).
|
| 4. |
The blood:plasma partition coefficient for VEN and ODV is 1. After administration of 14C-venlafaxine, the ratio of total
radioactivity (venlafaxine plus metabolites) ranged from 0.9 to
1.1.2 Consequently, it was assumed that
plasma and blood concentrations were similar for both VEN and ODV.
|
| 5. |
One hundred percent of the administered dose is absorbed across
the gastrointestinal tract. A previous study found that 92% of an oral
VEN dose is absorbed (11). Therefore, it seems that this assumption
should not significantly affect the modeling results.
|
| 6. |
All clearance processes are first order. Whereas earlier
studies found that, after administration of multiple doses, the
metabolic pathway for VEN is saturable (14, 15), VEN and ODV have
exhibited linear pharmacokinetics over a daily dosage range of 75-450
mg of parent drug (data on file, Wyeth-Ayerst Research). Because subjects in the investigation received a single 1.5 mg/kg dose of drug
(range: 93.75-150 mg), linear pharmacokinetics was assumed.
|
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 1.
Schematic representation of a first-pass
effect model for VEN after oral administration.
See text for abbreviations.
|
|
The model consisted of four differential equations that were
simultaneously fitted to plasma VEN and ODV concentrations. These differential equations represented the change with respect to time of
the following variables: plasma VEN concentration
(Cven), plasma ODV concentration
(Codv), liver VEN concentration
(Lven), and amount of VEN at the absorption site
in the gut (Xgut). Overall, the model contained
six parameters: intrinsic hepatic clearance of VEN
(CLint), liver:plasma partition coefficient of
VEN (RL), volume of distribution of VEN in
plasma compartment (Vven), total clearance of
ODV (CLodv), volume of distribution of ODV
(Vodv), and the first order absorption rate
constant of VEN across the gut (ka). In
addition, there were three constants included in the model:
Q (90 liters/hr), fm (0.55), and
Vl (1.5 liters). The following differential
equations were used:
|
(1)
|
|
(2)
|
|
(3)
|
|
(4)
|
Initial conditions on all variables were 0, with the exception of
Xgut, which was equal to the administered dose.
From the final model parameter estimates, extraction ratio
(E) was estimated as the ratio of
CLint and the sum of Q and
CLint (16).
Noncompartmental Analysis.
As a means of evaluating parameter estimates generated with the
first-pass model, data were analyzed by noncompartmental methods. AUC
(0 
) was calculated using trapezoidal rule with residual AUC
estimated from the ratio of the last measured plasma concentration and
the terminal rate constant (k). This rate constant was
determined by linear regression of the terminal phase of a log-linear
plot of plasma concentration over time. VEN
CLint and CLODV
(CL/fm) were calculated as the ratio
of dose and AUC. Volumes of distribution of VEN
(V/F) and ODV
(V/fm) were calculated as the ratio
of clearance and k.
Results and Discussion.
A model of presystemic metabolism was used to describe the time-course
disposition of VEN after oral administration to humans. VEN is an ideal
compound for this type of modeling analysis, because it is nearly 100%
metabolized by the liver, has a relatively rapid clearance, and is
converted to a metabolite (ODV) that can be readily detected and
quantitated in the plasma. Adapted from those previously reported in
the literature (1-3, 17), the model is compartmental in nature, but
includes physiologically based constants, such as liver volume and
hepatic plasma flow. Unlike traditional compartmental models,
clearances and volumes were used as parameters in place of rate
constants. It is widely accepted that blood concentrations are required
for modeling hepatic drug disposition (3, 18). In the present study,
plasma concentrations were used in place of blood concentrations,
because it was assumed that the blood to plasma concentration ratio for
VEN and ODV was 1. The organ flow rate used in the model, however, was
liver blood flow (90 liters/hr).
Modeling results are presented in table
1. Values represent mean parameter
estimates generated from fitting the first-pass effect model to plasma
concentrations of individual subjects. In all cases, a good fit of the
model to the data was obtained, based on randomness of scatter of
residuals, coefficient of variation of final parameter estimates, and
95% confidence intervals. CLint, defined as the
ability of the liver to metabolize drug in the absence of flow
restrictions, was 65.8 ± 39.7 liters/hr. Extraction ratio was
estimated to be 0.40 ± 0.12. Although this value is not
indicative of a high extraction ratio compound (E > 0.70), there is evidence of extensive hepatic uptake of VEN.
Specifically, the liver:plasma partition coefficient
(RL) was 29.6 ± 18.3, suggesting that this
compound is efficiently sequestered and ultimately cleared by the
liver.
View this table:
[in this window]
[in a new window]
|
TABLE 1
Mean (SD) model parameter and secondary parameter estimates for VEN and
ODV obtained using first-pass model after oral administration of a
single dose (1.5 mg/kg) of VEN to human subjects
|
|
Plasma concentration-time profiles of VEN and ODV are shown in figs.
2 and 3,
respectively. Presented in these profiles are the mean observed plasma
concentrations of the individual subjects, along with concentrations
predicted by the proposed first-pass effect model. These predicted
concentrations were obtained by model simulation, using the mean model
parameter estimates listed in table 1. The model yield a good fit to
experimental data. Although a good correlation between observed and
predicted concentrations was achieved with for both compounds, observed
ODV concentrations after 12 hr declined much slower than those
predicted by the first-pass model. Inclusion of a "tissue"
compartment for ODV was able to describe better the terminal phase of
the concentration-time curve, but this more complex model (eight
parameters) neither improved the fit nor resulted in a significant
reduction in weighted sums of squares over the present model
(unpublished data).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 2.
VEN plasma concentrations vs. time after
administration of a single oral dose (1.5 mg/kg) to human subjects.
O, mean (SD) observed data; solid line, concentrations
predicted by the first-pass model using nonlinear least squares
regression analysis.
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 3.
ODV plasma concentrations vs. time after
administration of a single oral dose of VEN (1.5 mg/kg) to human
subjects.
O, mean (SD) observed data; solid line, concentrations
predicted by the first-pass model using nonlinear least squares
regression analysis.
|
|
Table 2 contains mean values of
parameters determined by noncompartmental analysis.
CLint was 59.6 ± 26.5 liters/hr, which is
similar to the model estimate of 65.8 ± 39.7 liters/hr. Although a disparity exists between estimates of Vven,
correction of the noncompartmental estimate (459 ± 192 liters)
for a apparent bioavailability of 60% (F = 1 E) makes this estimate comparable with the model generated
value of 252 ± 87.6 liters. Likewise, noncompartmental estimates
of CLODV and VODV must be
corrected for fm before a useful comparison can
be made between the two methods (assuming that 100% of the
administered dose is absorbed, a correction for F is not
necessary). Assuming fm to be 0.55, adjustment
of the noncompartmental estimate of 309 ± 60 liters correlates
closely with the value obtained with the first-pass model (181 ± 84.1). The corrected estimate of CLodv, however,
is lower than the model estimate of 23.5 ± 12.5 liters/hr. This
divergence may possibly be attributed to the previously discussed
assignment of "one-compartment pharmacokinetics" to describe ODV
disposition. Despite this limitation, a good correlation was observed
between both methods of analysis.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Mean (SD) pharmacokinetic parameters for VEN and ODV using
noncompartmental analysis after oral administration of a single dose (1.5 mg/kg) of VEN to human subjects
|
|
In addition to the previously stated assumptions of the model was
that the liver was solely responsible for the presystemic metabolism of
VEN. In vitro studies have identified two CYP isozymes involved in VEN metabolism: CYP2D6 and CYP3A4 (12). The majority of VEN
degradation proceeds via CYP2D6, including the
O-demethylation of VEN to form ODV.
N-demethylation of VEN involves the CYP3A4 enzyme, a minor
pathway for this compound. Although intestinal metabolism has been
attributed to the CYP3A enzyme system (19), it was considered of little
significance in the present study. Thus, a presystemic intestinal
compartment was not incorporated into the model, thereby attributing
all first-pass loss of drug to hepatic degradation.
In summary, the disposition of VEN and its active metabolite ODV after
oral administration was successfully characterized using a
pharmacokinetic model that accounts for the presystemic hepatic
metabolism of drug. Parameter estimates correlated closely with those
obtained through noncompartmental methods. The results demonstrate that
first-pass metabolism can be successfully modeled after oral drug
administration. In consideration of the model assumptions, however, it
should be noted that the ability to apply this model depends on the
particular characteristics of a specific drug. Although the model may
not be applied universally to all compounds that undergo first-pass
metabolism, it can potentially be adapted for other compounds whose
metabolites can be accurately measured. This communication validates
the use of a first-pass effect model as a pharmacokinetic tool for
exploring changes in presystemic metabolism and drug disposition
secondary to disease or drug-drug or drug-food interactions in human
subjects.
David R. Taft
Ganesh R. Iyer
Leon Behar
Robert V. Digregorio
Divisions of
Pharmaceutics and Industrial Pharmacy (D.R.T.,
G.R.I.) and
Pharmacy Practice (L.B., R.V.D.), Long Island
University; and
Department of Pharmacy Services (R.V.D.),
Brookdale Hospital
| |
Acknowledgments |
We thank Wyeth-Ayerst Research for supplying the compounds used in this
study. In addition, we acknowledge Dr. Steven Troy and Dr. Soong Chiang
for providing helpful insight regarding VEN disposition, which was
useful in the preparation of this manuscript.
| |
Footnotes |
Received March 5, 1997; accepted June 11, 1997.
2
S. Troy, Wyeth-Ayerst Research, personal
communication.
Send reprint requests to: Dr. David R. Taft, Division of
Pharmaceutics and Industrial Pharmacy, Long Island University, 1 University Plaza, Brooklyn, NY 11201.
| |
Abbreviations |
Abbreviations used are:
VEN, venlafaxine;
ODV, O-desmethylvenlafaxine;
IS, internal standard;
AUC, area
under the plasma concentration vs. time curve;
CYP, cytochrome P450.
| |
References |
| 1.
|
M. Gibaldi and
S. Feldman:
Pharmacokinetic basis for the influence of route of administration on the area under the plasma concentration-time curve.
J. Pharm. Sci.
58,
1477-1480 (1969)[Medline].
|
| 2.
|
W. A. Colburn and
M. Gibaldi:
Pharmacokinetic model of presystemic metabolism.
Drug Metab. Dispos.
6,
193-196 (1978)[Abstract].
|
| 3.
|
M. Rowland:
Influence of route of administration on drug availability.
J. Pharm. Sci.
61,
70-74 (1972)[Medline].
|
| 4.
|
M. Rowland,
L. Z. Benet, and
G. G. Graham:
Clearance concepts in pharmacokinetics.
J. Pharmacokinet. Biopharm.
1,
123-136 (1973)[Medline].
|
| 5.
|
G. R. Wilkinson and
D. G. Shand:
A physiologic approach to hepatic drug clearance.
Clin. Pharmacol. Ther.
18,
377-390 (1975)[Medline].
|
| 6.
|
K. S. Pang and
M. Rowland:
Hepatic clearance of drugs. I. Theoretical considerations of a "well-stirred" model and a "parallel tube" model. Influence of hepatic blood flow, plasma and blood cell binding, and the hepatocellular enzymatic activity on hepatic drug clearance.
J. Pharmacokinet. Biopharm.
5,
625-653 (1977)[Medline].
|
| 7.
| J. P. Feighner: The role of venlafaxine in antidepressant
therapy. J. Clin. Psychiatry 55 (Suppl. A),
62-68 (1994).
|
| 8.
| Y. Lecrubier: Clinical utility of venlafaxine in comparison
with other antidepressants. Int. Clin. Psychopharmacol.
10 (Suppl. 2), 29-35 (1995).
|
| 9.
|
F. A. Muth,
J. A. Moyer,
J. T. Haskins,
T. H. Andree, and
G. E. M. Husbands:
Biochemical, neurophysiological and behavioral effects of WY 45,233, it's enantiomers, and other identified metabolites of the antidepressant venlafaxine.
Drug Dev. Res.
23,
191-199 (1991).
|
| 10.
|
D. R. Hicks,
D. Wolaniuk,
A. Russell,
N. Cavanaugh, and
M. Kraml:
A high-performance liquid chromatographic method for the simultaneous determination of venlafaxine and o-desmethylvenlafaxine in biological fluids.
Ther. Drug Monit.
16,
100-107 (1994)[Medline].
|
| 11.
|
S. R. Howell,
G. E. M. Husbands,
J. A. Scantina, and
S. F. Sissenwine:
Metabolic disposition of 14C-venlafaxine in mouse, rat, dog, rhesus monkey and man.
Xenobiotica
28,
349-359 (1993).
|
| 12.
|
S. V. Otton,
S. E. Ball,
S. W. Cheung,
T. Inaba,
R. L. Rudolph, and
E. M. Sellers:
Venlafaxine oxidation in vitro is catalyzed by CYP2D6.
Br. J. Clin. Pharmacol.
41,
149-156 (1996)[Medline].
|
| 13.
|
M. Gibaldi and
D. Perrier:
"Pharmacokinetics." Marcel Dekker, New York, 1982.
|
| 14.
|
V. Parker,
L. Pruitt, and
K. Maloney:
Effect of age and sex on the pharmacokinetics of venlafaxine.
J. Clin. Pharmacol.
30,
832 (1990).
|
| 15.
|
K. J. Klamerulus,
K. Maloney,
R. L. Rudolph,
S. F. Sisenwine,
W. J. Jusko, and
S. T. Chiang:
Introduction of a composite parameter to the pharmacokinetics of venlafaxine and its active O-desmethyl metabolite.
J. Clin. Pharmacol.
32,
716-724 (1992)[Abstract].
|
| 16.
|
K. S. Pang and
J. R. Gillette:
A theoretical examination of the effects of gut wall metabolism, hepatic elimination, and enterohepatic recycling on estimates of bioavailability and hepatic blood flow.
J. Pharmacokinet. Biopharm.
6,
355-366 (1978)[Medline].
|
| 17.
|
J. G. Wagner:
"Pharmacokinetics for the Pharmaceutical Scientist." Technomic, Lancaster, PA, 1995.
|
| 18.
|
S. M. Pond and
T. N. Tozer:
First-pass elimination: basic concepts and clinical consequences.
Clin. Pharmacokinet.
9,
1-25 (1984)[Medline].
|
| 19.
|
K. E. Thummel,
D. O'Shea,
M. F. Paine,
D. D. Shen,
K. L. Kunze,
J. D. Perkins, and
J. R. Wilkinson:
Oral first-pass elimination of midazolam involves both gastrointestinal and hepatic CYP3A-mediated metabolism.
Clin. Pharmacol. Ther.
59,
491-502 (1996)[Medline].
|
Copyright © 1997 by The American Society for Pharmacology and Experimental Therapeutics
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Abstract
Article
Abstract
