http://informahealthcare.

com/ddi
ISSN: 0363-9045 (print), 1520-5762 (electronic)
Drug Dev Ind Pharm, Early Online: 17
! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/03639045.2014.991399

RESEARCH ARTICLE

Prediction of in vivo drug performance using in vitro dissolution coupled

with STELLA: a study with selected drug products
Sumon Chakraborty*, Lokesh Yadav*, and Deepika Aggarwal

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

Centre of Excellence-Biostudies, Integrated Product Development Organization, Dr. Reddys Laboratories, Hyderabad, Andhra Pradesh, India

Abstract

Keywords

Prediction of the in vivo performance of the drug product from the in vitro studies is the major
challenging job for the pharmaceutical industries. From the current regulatory perspective,
biorelevant dissolution media should now be considered as quality control media in order to
avoid the risk associated. Physiological based pharmacokinetic models (PBPK) coupled with
biorelevant dissolution medium is widely used in simulation and prediction of the plasma drug
concentration and in vivo drug performance. The present investigation deals with the
evaluation of biorelevant dissolution media as well as in vivo drug performance by PBPK
modelling using STELLA simulation software. The PBPK model was developed using STELLA
using dissolution kinetics, solubility, standard gastrointestinal parameters and post-absorptive
disposition parameters. The drug product selected for the present study includes Linezolid filmcoated immediate-release tablets (Zyvox), Tacrolimus prolonged-release capsules (Advagraf),
Valganciclovir tablets (Valcyte) and Mesalamine controlled-release capsules (Pentasa) each
belonging to different biopharmaceutics classification system (BCS). The simulated plasma drug
concentration was analyzed and pharmacokinetic parameters were calculated and compared
with the reported values. The result from the present investigation indicates that STELLA
when coupled with biorelevant dissolution media can predict the in vivo performance of the
drug product with prediction error less than 20% irrespective of the dosage form (immediate
release versus modified release) and BCS Classification. Thus, STELLA can be used for in vivo
drug prediction which will be helpful in generic drug development.

Biorelevant dissolution media, BCS

classification, PBPK modelling, STELLA

Introduction
Prediction of in vivo pharmacokinetics of drug products is one of
the major challenges for both innovator and generic pharmaceutical industries. Accurate to approximate prediction of in vivo
behavior of drug product will reduce failure rate in drug product
development as well as huge resources involved in the same.
Prediction of pharmacokinetics from in vitro and in silico tools
has been used since few decades and is based on the certain
physicochemical and biopharmaceutical properties of drug molecules and in vitro release rates. In vitro dissolution tests using
suitable media, i.e. biorelevant media has proven to be successful
methodology for in vivo performance prediction of drug products1,2. Recent literature reflects the wide use of simulated gastric
fluids or so-called biorelevant media for better in vivo drug
performance prediction3,4. Physiologically based pharmacokinetic
(PBPK) models predict the plasma concentration time profile
from the in vitro dissolution and in silico data inputs5,6. PBPK
modeling takes into account the factors influencing the absorption, distribution and elimination7,8. The insights given by the

*These authors contributed equally to this work.

Address for correspondence: Mr Lokesh Yadav, MS Pharm, CPPK-Bio,
IPDO, Dr. Reddys Laboratories, Hyderabad 500 072, Andhra Pradesh,
India. E-mail: lokeshyadav@drreddys.com

History
Received 29 May 2014
Revised 4 September 2014
Accepted 12 November 2014
Published online 12 December 2014

PBPK models are integrative and far better than that of noncompartmental or compartmental analysis. A report by Parrott
et al.9,10 suggests that PBPK models are superior to the other
empirical methods for interspecies scaling and prediction of
human pharmacokinetics. Prediction of pharmacokinetic behaviour of the drug products from the physicochemical and dissolution data assists the generic pharmaceutical companies to select
the optimized formulation when the qualitative and/or information
is unavailable. Currently, regulatory governing bodies are also
utilizing the PBPK models for better understanding of the drug
products11,12. Therefore, the generic pharmaceutical industries
are also utilizing the same concept of PBPK modelling using
dissolution specifications (of both test and reference listed drug)
for the pharmacokinetics prediction in order to avoid the
resources and time spent on the pilot biostudies.
Among the various in silico software packages for building
PBPK models (GastroPlus , PKSim , SimCyp and MatLab),
STELLA (Structural Thinking, Experimental learning
Laboratory with Animation) has been proven as useful
approach to simulate the in vivo behavior of orally administered
drugs13,14. It is an icon-based model building and simulation
tool that was first introduced in the 1980s. In contrast to
commercially available PBPK models available, through the use
of a graphical interface, the user constructs and runs the
simulation by building a graphical representation of the model
using STELLA 13 (Figures 1 and 2). STELLA is user

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

S. Chakraborty et al.

Drug Dev Ind Pharm, Early Online: 17

Figure 1. General model structure used to simulate the pharmacokinetic profiles of the selected drugs using STELLA .
Figure 2. STELLA model map used to
simulate the plasma profiles of the selected
drugs. SI, small intestine; Vfluid, volume of
co-administered fluid; API, active pharmaceutical ingredient.

interactive software and model building can be done based on

technical and scientific knowledge of the users. Many scientists
find the applicability of the STELLA in determining the
pharmacokinetics of poorly soluble and permeable drugs15,
food effects influencing the drug pharmacokinetics16, which in
turn helps in designing the formulation and pharmacokinetic

strategies. STELLA uses the dissolution inputs (for better

prediction, physiological relevant dissolution inputs) to predict
the plasma/serum drug profiles. The software requires in vitro
dissolution parameter (z-value), solubility, standard gastrointestinal parameters in accordance with dosing conditions and postabsorptive disposition pharmacokinetic parameters. The

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

DOI: 10.3109/03639045.2014.991399

pharmacokinetic parameters can be obtained from the published

literature. The present study aims to investigate the role of the
STELLA in predicting the in vivo pharmacokinetics using
drugs (in fasting condition) of BCS Class IIV both immediate
release (IR) and modified release (MR) drug products, thus
focusing the use of STELLA in better plasma profile
predictability of all type of drug products irrespective of their
physicochemical properties. Also, the study focuses the use
of biorelevant media for individual drug products, which helps
in better prediction of the in vivo performance. The drugs used
in the present study include Linezolid (BCS Class I),
Tacrolimus (BCS Class II), Valganciclovir (BCS Class III)
and Mesalamine (BCS Class IV). The results of the present
investigation reflect the better possible media (specific for
individual drug product used in the study), which can
accurately simulate the in vivo plasma profile using
STELLA software. Thus, STELLA can be used as in
silico tool for prediction of in vivo performance of the drug
irrespective of its physicochemical properties (BCS Class IIV)
and dosage form (IR versus MR).

Materials and methods

Chemicals and reagents
Linezolid film-coated IR tablets (Zyvox, 600 mg strength) were
purchased from Pharmacia & Upjohn (UK). Tacrolimus prolonged-release (PR) capsule (Advagraf, 5 mg strength) was
purchased from Astellas Pharma Europe Ltd. (Chertsey, Surrey,
UK). Valganciclovir tablets (Valcyte, 450 mg strength) was
purchased from Hoffmann-La Roche Ltd. (USA). Mesalamine
controlled-release (CR) capsule (Pentasa, 500 mg strength) was
purchased from Shire Pharmaceutical Ltd, Hampshire, UK.
Sodium phosphate, orthophosphoric acid (OPA), acetic acid and
sodium taurocholoate was purchased from SRL Chemicals
(Mumbai, India), SpectroChem Pvt Ltd (Mumbai, India),
Rankem (Haryana, India) and New Zealand Pharmaceuticals Ltd
(Palmerston North, New Zealand), respectively. Hydroxypropyl
cellulose (HPC; Klucel) was procured from Ashland Inc.
(Covington, KY) and sodium lauryl sulphate (SLS; Kolliphor
SLS; Texapon) from Signet chemicals (Thane, India). Sodium
chloride, sodium acetate and sodium dihydrogen phosphate was
purchased from Merck (Mumbai, India). Pepsin and lecithin was
purchased from HiMedia Laboratories (Mumbai, India).
Media preparation
All the reagents and media were prepared as per in-house standard
operating procedure (SOPs) and as per the standards of USP
pharmacopoeia. The fasting gradient media was prepared as
described by Garbacz and Fotaki17,18. The fasting gradient media
includes fasted state simulated gastric fluid (FaSSGF), fasted state
simulated intestinal fluid (FaSSIF), FaSSIF pH 7 (composition
is reduced to half that of FaSSIF, adjusted to pH 7), phosphate
buffer pH 7.2, phosphate buffer pH 7.5 and simulated colon
fluid (SCoF).
Solubility testing and dissolution analysis
The solubility of the selected drug products was determined in the
selected media using the shake flask method as per in-house SOP.
The dissolution media were selected on the drug product basis
(relevant to fasting condition and solubility). Linezolid filmcoated IR tablets dissolution was performed in 0.01 N HCl and 6.8
phosphate buffers in USP type II dissolution apparatus (Paddle) at
50 rpm in 900 mL. The dissolution of Tacrolimus PR capsules was
performed in pH 4.5 OPA with 0.005% HPC media recommended
in OGD (USFDA) as well as some trial and error media with

Prediction of drug performance using STELLA

optimized quantity of SLS, like in pH 4.5 acetate buffer with

0.05% SLS, 0.1 N HCl with 0.05% SLS, pH 6.8 phosphate buffer
with 0.05% SLS and pH 4.5 OPA with 0.05% SLS USP type II
dissolution apparatus (Paddle) at 50 rpm in 900 mL. SLS was used
in the dissolution media in order to avoid plug formation.
Valganciclovir tablets dissolution was performed in 0.1 N HCl in
900 mL media at 50 rpm and in pH 6.8 phosphate buffer in
900 mL at 25 rpm in USP II apparatus. The dissolution media for
Mesalamine CR capsules was selected on the basis of its release
pattern and solubility. The dissolution in fasting gradient media
(composition as directed in media preparation section) was
performed in USP type III (reciprocating cylinder) at 12 dips per
min in 250 mL media (each media), whereas dissolution in 0.1 N
HCl and pH 7.2 phosphate buffer was performed in USP type II
apparatus at 100 rpm with 900 mL media. Dissolution analysis
was performed using HPLC at different time points depending on
the drug product (IR versus MR).
Available pharmacokinetic data and assessments
The plasma drug concentrations of all the selected drugs were
obtained from available literature and patents (reported values are
given in Table 4)1925. Linezolid is the first approved oxazolidinone antibiotic, belonging to BCS Class I. Tacrolimus, a
macrolide lactone with potent immunosuppressive activity
belongs to BCS Class II, with limited solubility. The bioavailability of Tacrolimus (20%) is low due to limited solubility,
extensive first-pass metabolism and efflux by P-gp protein.
Valganciclovir, a prodrug of Ganciclovir has low oral bioavailability (60%) due to extensive metabolism (especially by
esterases in the intestinal membrane). Mesalamine is an antiinflammatory, locally acting drug for the treatment of ulcerative
colitis and similar intestinal diseases. It belongs to BCS Class IV,
with limited permeability and solubility. The available literature
plasma drug concentration was extracted using Plot Digitalizer
(version 1.9; Department of Physics, University of South
Alabama). The STELLA simulated plasma drug concentration
was analyzed using Phoenix/WinNonlin software (version
6.3.0.395, St. Louis, MO) and pharmacokinetic parameters were
calculated. Maximum plasma concentration (Cmax) was determined by direct inspection of the mean plasma concentration time
profile. AUC from time 0t (AUCt) was determined by trapezoidal method. The area under the plasma concentration versus
time curve from time zero to infinity was calculated using
following formula, AUC1 AUC0t + Ct/z, Ct is the last measurable concentration and z is the terminal elimination rate
constant. The elimination rate constant was determined for each
of the molecule by linear regression of the linear portion of the
curve [ln (concentration versus time curve)]. The post-absorptive
parameters (absorption rate constant and fraction absorbed) are
summarized in Table 3 for the selected drug products.
Computer simulations of the selected drug candidates
The plasma profiles of the drugs used were simulated using
STELLA software, in accordance with the previous reported
literature2,26. The model assumptions include negligible absorption from the stomach, simultaneous solid and liquid emptying
from the stomach with gastric emptying rate being constant, i.e.
2.8 h1 in fasted state (first-order gastric emptying rate) and
volume of fasted content as 250 mL (considering 240 mL water to
be administered with dosage form; 10 mL being the residual
intragastric fluid volume inclusive of gastric secretions and
gastric output)15. The formulation factor such as monolithic unit
dosage form (Linezolid film-coated IR tablets, Valganciclovir
tablets and Tacrolimus PR capsules) versus multiple unit pellets

S. Chakraborty et al.

Drug Dev Ind Pharm, Early Online: 17

(Mesalamine CR capsules) has been considered during the model

building. The bioavailability (F) and volume of distribution (V)
was used in the pharmacokinetic calculation of V/F values using
WinNonlin.
The in vivo dissolution kinetics was estimated from in vitro
dissolution data by applying modified NoyesWhitney equation
into the STELLA model. This model is based on the assumption
of dissolution of isometric, similarly sized spherical particles,
occurring under continuously decreasing surface area conditions
with the ratio D/d (diffusion coefficient and diffusion layer
thickness), which is assumed to be constant during the dissolution
process. The dissolution rate is given by the following equation:

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

dWt DGN 1=3 2

2

W 3 Xs Wt zW 3 Cs C
dt
Vp2=3
where Wt is the amount dissolved at time t; W is the amount of
drug remaining to be dissolved; Xs is the amount of drug which
saturates the volume V of the dissolution medium, Cs is the
solubility of drug, C is the concentration of the dissolved drug
at time t;  is the particle density, is the shape factor, N is the
number of the particles to be dissolved and z is the dissolution
parameter which is constant equivalent to D N1/3/d2/3.
Dissolution inputs are introduced into the model by assuming
the dissolution kinetics as the first-order kinetics, according to the
defined equation. Sink conditions are assumed for the dissolution
in the intestine. When in vitro dissolution was considered to occur
according to the equation and the dose administered in vivo was
the same as the dose tested in vitro, the z-value for the in vivo
dissolution process, z vivo, was identical to z vitro27. The
corresponding z-value is calculated for each drug product in
the respective dissolution media and is tabulated in Table 1. The
simulation of plasma drug concentration for each drug from the
dissolution media is done for doses 600 mg, 5 mg, 2*450 mg and
5*500 mg for Linezolid film-coated IR tablets, Tacrolimus PR
capsules, Valganciclovir tablets and Mesalamine CR capsules,
respectively.

in pH 6.8, which can affect the dissolution) dissolution media.

In both the dissolution media, the drug dissolved 480% in 15 min
(Figure 3A). Generally for BCS Class I drugs, dissolution is not
the rate-limiting step, therefore, the dissolution profile of
Linezolid-film coated IR tablets in two different media have
similar profile (Linezolid API does not show the pH dependent
solubility), the F2 value being 60.55.
The quantitative simulation of plasma profile of Linezolid-film
coated IR tablets in humans was performed with STELLA using
dissolution data of 0.01 N HCl and pH 6.8 phosphate buffer.
The plasma Linezolid concentration predicted is shown in
Figure 4(A), comparing the simulated plasma profile with
the observed (literature) plasma profile. The pharmacokinetic
parameters were calculated using WinNonlin and tabulated in
Tables 3 and 4. The tmax (time to reach the maximum plasma
concentration) was same for both the reported and simulated
plasma concentration. The % prediction error in the pharmacokinetic parameters (Cmax and AUC0t) was less (510%) in 0.01 N
HCl. Thus, 0.01 N HCl reflects the better simulated plasma
Linezolid drug concentration.
Dissolution and plasma concentration simulation of
Tacrolimus PR capsules
Tacrolimus API is BCS Class II (low solubility and high
permeability) drug. Tacrolimus is practically insoluble in water
(summary of product characteristics; SPC Tacrolimus PR capsules). Therefore, the solubility testing and dissolution was
performed using 0.05% SLS in the media (also to mimic the
in vivo surfactant property). The solubility was observed highest
Table 2. The observed solubility of the drugs in selected dissolution
medium.

Linezolid

Solubility, dissolution and plasma concentration

simulation of Linezolid film-coated IR tablets

Tacrolimus

The solubility of Linezolid API in 0.01 N HCl was observed

higher in comparison with pH 6.8 phosphate buffer (Table 2). The
dissolution of Linezolid film-coated IR tablets was performed in
0.01 N HCl and 6.8 phosphate buffer. Since, being BCS Class I
drug with high solubility across the GI pH, 0.01 N HCl can be
considered as biorelevant (since the drug has no precipitation

Valganciclovir
Mesalamine

Table 1. The calculated z-value of the drugs in selected dissolution

medium.

Linezolid

Dissolution medium

0.01 N HCl
pH 6.8 phosphate buffer
Tacrolimus
pH 4.5 OPA with 0.005% HPC
pH 4.5 OPA with 0.05% SLS
0.1 N HCl 0.05% SLS
pH 4.5 acetate buffer with 0.05% SLS
pH 6.8 phosphate buffer with 0.05% SLS
Valganciclovir 0.1 N HCl
pH 6.8 phosphate buffer
Mesalamine
Fasting gradient media
0.1 N HCl
pH 7.2 phosphate buffer

0.01 N HCl
pH 6.8 phosphate buffer
4.5 pH OPA with 0.005% HPC
4.5 pH OPA with 0.05% SLS
0.1 N HCl 0.05% SLS
4.5 pH acetate buffer with 0.05% SLS
6.8 pH phosphate buffer with 0.05% SLS
0.1 N HCl
pH 6.8 phosphate buffer
FaSSGF
FaSSIF (6.5)
FaSSIF pH 7.0
pH 7.2 phosphate buffer
pH 7.5 phosphate buffer
SCoF

4.38
2.53
0.007
0.068
0.002
0.082
0.079
602.9
625.9
6.20
4.52
15.32
9.87
11.35
5.25

Drug

Results and discussion

Drug

Dissolution medium

Solubility
(mg/mL, at
37 0.5 C)

z-Value
(mL/mg2/3/h)
0.890
0.952
0.280
0.425
0.239
0.525
0.511
0.758
0.815
0.525
0.925
0.345

Table 3. The pharmacokinetic disposition parameters of the selected

drugs.

Drugs
Linezolid
Tacrolimus
Valganciclovir
Mesalamine

V/F
mg/(ng/mL)

k10
(h1)

k12
(h1)

k21
(h1)

Bioavailability
(fraction)

0.004
1.3125
0.0518
0.0075

0.165
0.0223
0.1823
3.1048

0.1392
2.3125
0.251
5.6037

0.1257
1.7529
0.1295
9.1986

1
0.22
0.594
0.29

k10, elimination rate constant; k12, distribution rate constant for drug
transfer from first to second compartment; k21, distribution rate constant
for drug transfer from second to first compartment.

Prediction of drug performance using STELLA

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

DOI: 10.3109/03639045.2014.991399

in pH 4.5 acetate buffer with 0.05% SLS, whereas lowest in the

case of 0.1 N HCl with 0.05% SLS (Table 2). The dissolution
media was selected on trial and error basis along with the Office
of Generic Drug (OGD) recommended media. The biorelevant
media for this drug product seems to be pH 6.8 phosphate buffer
along with surfactant, since according to the scientific discussion;
EMEA, the drug molecule gets preferentially absorbed from
duodenum and jejunum. Therefore, dissolution in this media will
predict the in vivo pharmacokinetics of the drug. The pH 6.8
phosphate buffer with 0.05% SLS dissolved the tacrolimus 480%
within 2 h, whereas pH 4.5 OPA with 0.05% SLS and pH 4.5
acetate buffer with 0.05% SLS media dissolved480% in 6 h. 0.1 N
HCl with 0.05% SLS media dissolved 75% of the drug in 3 h,
and after that tacrolimus was degraded (the curve starts declining
after 3 h), whereas in 4.5 pH OPA with 0.005% HPC, the drug
release was incomplete (Figure 3B).
In the present study, all the above-mentioned media for
Tacrolimus PR capsules was used in simulating the plasma profile
using STELLA . The plasma drug concentration was simulated

till 72 h as shown in Figure 4(B) and the corresponding

pharmacokinetic parameters were calculated and tabulated in
Table 4. The predicted plasma profile using pH 6.8 phosphate
buffer with 0.05% SLS was in line with to the reported
pharmacokinetic parameters, with prediction error 515%,
whereas the % predicted error in the pharmacokinetic parameters
were high (420 in Cmax) in other dissolution media (Table 5).
Dissolution and plasma concentration simulation of
Valganciclovir tablets
Valganciclovir is a pro-drug with active metabolite as
Ganciclovir. Being BCS Class III molecule (high solubility, low
permeability) dissolution is not rate limiting for the absorption of
the drug. The solubility of the Valganciclovir API was very high
in both the selected medium (Table 2). The dissolution media
selected for simulation was 0.1 N HCl and pH 6.8 phosphate
buffer. Since, being BCS Class III drug with high solubility across
the GI pH (approximately same value in both the media 0.1 N HCl
and 6.8 phosphate buffer). Both the media shows more or less the

Figure 3. Dissolution profile of the selected drug products in the respective dissolution media. (A) Linezolid film-coated IR tablets, (B) Tacrolimus PR
capsules, (C) Valganciclovir tablets and (D) Mesalamine CR capsules.

Table 4. The observed (reported) and simulated (predicted using STELLA ) pharmacokinetic parameters of the drugs in selected dissolution medium.
Drug (BCS Class)
Linezolid film-coated IR tablets
Tacrolimus PR capsules
Valganciclovir tablets

Mesalamine CR capsules

Reported and predicted values

Cmax (ng/mL)

AUC0t (ng*h/ml)

AUC01 (ng*h/ml)

Reported values
pH 6.8 phosphate buffer
Reported values
6.8 pH phosphate buffer with 0.05% SLS
Reported values
0.1 N HCl
pH 6.8 phosphate buffer
Reported values
Fasting gradient media

13.68
15.03
11.57
9.82
4630
4450.23
4519.01
537.93
575.13

112
142.21
290.81
284.83
24 400
25 265
25 576.50
4831.05
5285.82

112.60
149.05
318.15
363.90
24 900.00
25 711.87
26 208.73
5401.84
6533.92

S. Chakraborty et al.

Drug Dev Ind Pharm, Early Online: 17

same dissolution profile and the same plasma profile prediction.

Hence either of the media can be considered as biorelevant.
In both the selected dissolution media, 480% drug dissolved
within 20 min with complete release in 30 min. 0.1 N HCl was
comparatively faster than 6.8 phosphate buffer, but similarity
factor F2 being 62.3 (Figure 3C). On simulation of plasma profile
with selected dissolution media using STELLA , the % prediction
error was 510% in all the pharmacokinetic parameters in both
the dissolution media (table). The predicted plasma profile was
in line with the observed plasma profile (Table 4 and Figure 4C).
Dissolution and plasma concentration simulation of
Mesalamine CR capsules

Conclusions
The present investigation indicates that the dissolution testing in
appropriate media, coupled with simulation software STELLA ,
can predict the plasma drug concentration of the selected drug

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

Mesalamine CR capsule (Pentasa) is gastrointestinal locally

acting drug for colitis. It belongs to BCS Class IV (low solubility
and low permeability). The solubility data in the selected
medium is shown in Table 2. The highest observed solubility
was in FaSSIF media, whereas least solubility was observed in

FaSSIF (pH 6.5). The dissolution was performed in three media,

the fasting gradient media (composition described in media
preparation section), and 0.1 N HCl and pH 7.2 phosphate buffer.
The dissolution was faster in 0.1 N HCl with480% drug dissolved
within 4 h, whereas in fasting gradient and pH 7.2 phosphate
buffer the 480% drug dissolved in 16 and 24 h, respectively
(Figure 3D). When the plasma drug concentration was simulated
using the above dissolution media using STELLA , the % prediction error was 510% (Cmax and AUCt) in fasting gradient
media, whereas in 0.1 N HCl and pH 7.2 phosphate buffer, the %
predicted error was more than 420% (Table 5). Thus, fasting
gradient media predicted plasma drug concentration was in line
with the observed value (Table 4 and Figure 4D).

Figure 4. Observed and simulated (using STELLA ) plasma drug concentration of the drugs. (A) Linezolid (600 mg tablets), (B) Tacrolimus (5 mg
capsules), (C) Valganciclovir (2*450 mg tablets) and (D) Mesalamine (2*500 mg capsules).
Table 5. The percentage prediction error in the observed (reported) and simulated (predicted using STELLA ) pharmacokinetic parameters of the
drugs in selected dissolution medium.

Drug (BCS Class)

Linezolid film-coated IR tablets
Tacrolimus PR capsules
Valganciclovir tablets
Mesalamine CR capsules

Media used for simulation

% Predicted
error (Cmax)

% Predicted
error (AUC0t)

% Predicted
error (AUC01)

0.01 N HCl
6.8 pH phosphate buffer with 0.05% SLS
0.1 N HCl
pH 6.8 phosphate buffer
Fasting gradient media

9
15
4
2
7

6
2
4
3
9

16
14
3
5
21

All the numerical values are expressed in percentage (%).

Drug Dev Ind Pharm Downloaded from informahealthcare.com by Merck KGAA on 01/08/15
For personal use only.

DOI: 10.3109/03639045.2014.991399

products. Dissolution testing, and modeling the dissolution data

for in vivo prediction is widely used in the present scenario in
pharmaceutical industries. Dissolution testing apart from serving
as quality control tool helps in predicting the in vivo performance
of drug products, provided the dissolution is performed in
physiological relevant media. Dissolution data when coupled
with physiological based pharmacokinetic models serves as a
better surrogate for simulation of the plasma drug concentration.
The present investigation focuses the use of simulation tool
STELLA for prediction of plasma drug concentration of the
selected drug products (one from each BCS Class, IR and MR
products), based on the biorelevant dissolution data. The
predicted plasma drug concentration simulated using dissolution
results were in line (comparable) with the observed plasma drug
concentration for all the selected drug products with % prediction
error 520%. Thus, STELLA can be used as a screening tool for
dissolution media as well as predictive tool for plasma drug
concentration when coupled with biorelevant dissolution results
for drug products (both IR and MR) irrespective of solubility
and permeability limitations.

Acknowledgements
Authors would like to thanks Dr. Reddys Laboratory for providing
assistance and support to publish this work. Also the authors acknowledge
the scientific inputs provided by Dr Hemant P. Joshi.

Declaration of interest
All the authors are employees of Dr. Reddys laboratory and
report no conflicts interest. The authors alone are responsible for
the content and writing of this article.

References
1. Dressman JB, Amidon GL, Reppas C, et al. Dissolution testing as a
prognostic tool for oral drug absorption: immediate release dosage
forms. Pharm Res 1998;15:1122.
2. Grass GM. Simulation models to predict oral drug absorption from
in vitro data. Adv Drug Deliv Rev 1997;23:199219.
3. Patel N, Forbes B, Murray J. Use of simulated intetsinal fluids with
Caco-2 cells and rat ileum. Drug Dev Ind Pharm 2006;32:15161.
4. Tandt LAGL, Stubbs C, Kanfer I. Level a in vitro/in vivo correlations: a quality control tool or bioequivalence predictor for
extended-release solid oral dosage forms? Drug Dev Ind Pharm
1995;21:889904.
5. Shono Y, Jantratid E, Dressman JB. Precipitation in the small
intestine may play a more important role in the in vivo performance
of poorly soluble weak bases in the fasted state: case example
nelfinavir. Eur J Pharm Biopharm 2011;79:34956.
6. Gerlowski LE, Jain RK. Physiologically based pharmacokinetic
modeling: principles and applications. J Pharm Sci 1983;72:
110327.
7. Nestorov I. Whole-body physiologically based pharmacokinetic
models. Expert Opin Drug Metab Toxicol 2007;3:23549.
8. Poulin P, Jones RD, Jones HM, et al. PHRMA CPCDC initiative on
predictive models of human pharmacokinetics, part 5: prediction of
plasma concentration-time profiles in human by using the physiologically-based pharmacokinetic modeling approach. J Pharm Sci
2011;100:412757.
9. Parrott N, Paquereau N, Coassolo P, et al. An evaluation of the
utility of physiologically based models of pharmacokinetics in early
drug discovery. J Pharm Sci 2005;94:232743.

Prediction of drug performance using STELLA

10. Parrott N, Davies B, Hoffmann G, et al. Development of a

physiologically based model for oseltamivir and simulation of
pharmacokinetics in neonates and infants. Clin Pharmacokinet 2011;
50:61323.
11. Rowland M, Balant L, Peck C, et al. Physiologically based
pharmacokinetics in Drug Development and Regulatory Science: a
workshop report (Georgetown University, Washington, DC,
May 2930, 2002). AAPS J 2004;6:5667.
12. Zhao P, Zhang L, Grillo JA, et al. Applications of physiologically
based pharmacokinetic (PBPK) modeling and simulation during
regulatory review. Clin Pharmacol Ther 2011;89:25967.
13. Kostewicz ES, Aarons L, Bergstrand M, et al. PBPK models for the
prediction of in vivo performance of oral dosage forms. Eur J Pharm
Sci 2013;57:30021.
14. Otsuka K, Shono Y, Dressman J. Coupling biorelevant dissolution
methods with physiologically based pharmacokinetic modelling to
forecast in-vivo performance of solid oral dosage forms. J Pharm
Pharmacol 2013;65:93752.
15. Shono Y, Jantratid E, Kesisoglou F, et al. Forecasting in vivo
oral absorption and food effect of micronized and nanosized
aprepitant formulations in humans. Eur J Pharm Biopharm 2010;
76:95104.
16. Shono Y, Jantratid E, Janssen N, et al. Prediction of food effects on
the absorption of celecoxib based on biorelevant dissolution testing
coupled with physiologically based pharmacokinetic modeling. Eur
J Pharm Biopharm 2009;73:10714.
17. Garbacz G, Klein S. Dissolution testing of oral modified-release
dosage forms. J Pharm Pharmacol 2012;64:94468.
18. Fotaki N, Vertzoni M. Biorelevant dissolution methods and their
applications in in vitroin vivo correlations for oral formulations.
Open Drug Deliv J 2010;4:213.
19. Bioequivalence study comparing linezolid 600 mg oral suspension
with 600 mg tablet in Chinese healthy male subjects. Available
from: http://www.clinicaltrials.gov/ct2/show/NCT01055769?term
Linezolid&rank6 [last accessed 30 April 2014].
20. International application published under Patent Co-operation
Treaty (PCT), Patent no. WO 2008/145/143 A1, Once daily oral
dosage form comprising tacrolimus, Applicant: LifeCycle Pharma
A/S, Denmark.
21. Clinical pharmacology and biopharmaceutics review of
Valganciclovir tablets (450 mg); Clinical Pharmacology and biopharmaceutics review: Centre for Drug Evaluation and Research;
Application no. 21-304; Roche Global Development; 2004.
22. Clinical pharmacology and biopharmaceutics review of Mesalamine
controlled release capsules (500 mg); Clinical Pharmacology and
biopharmaceutics review: Centre for Drug Evaluation and Research;
Application no. 020049; Shire Pharmaceuticals, UK; 1993.
23. Stalker DJ, Jungbluth GL, Hopkins NK, et al. Pharmacokinetics and
tolerance of single- and multiple-dose oral or intravenous linezolid,
an oxazolidinone antibiotic, in healthy volunteers. J Antimicrobial
Chemother 2003;51:123946.
24. Moller A, Iwasaki K, Kawamura A, et al. The disposition of
14
C-Labeled tacrolimus after intravenous and oral administration
in healthy human subjects. Drug Metab Dispos 1998;27:6336.
25. Myers B, Evans DNW, Rhodes J, et al. Metabolism and urinary
excretion of 5-amino salicylic acid in healthy volunteers when given
intravenously or released for absorption at different sites in the
gastrointestinal tract. Gut 1987;28:196200.
26. Fei Y, Kostewicz ES, Sheu MT, et al. Analysis of the enhanced oral
bioavailability of fenofibrate lipid formulations in fasted humans
using an in vitro-in silico-in vivo approach. Eur J Pharm Biopharm
2013;85:127484.
27. Nicolaides E, Symillides M, Dressman JB, et al. Biorelevant
dissolution testing to predict the plasma profile of lipophilic drugs
after oral administration. Pharm Res 2001;18:3808.