Article

From In Vivo Predictive Dissolution to Virtual Bioequivalence:

A GastroPlus®-Driven Framework for Generic Candesartan
Cilexetil Tablets
Hao Ruan 1,2,† , Xiaoting Geng 2,† , Zijing Situ 3 , Qian Shen 2 , Tianjian Ye 4 , Xin Chen 4, * and Weike Su 1, *

1 College of Pharmaceutical Sciences, Zhejiang University of Technology, Hangzhou 310014, China;

[email protected]
2 Zhejiang Key Laboratory of Biopharmaceutical Contact Materials, NMPA Key Laboratory for Core Technology
of Generic Drug Evaluation, Zhejiang Institute for Food and Drug Control, Hangzhou 310052, China
3 School of Pharmacy, China Pharmaceutical University, Nanjing 210009, China
4 Zhejiang Yongning Pharmaceutical Co., Ltd., Taizhou 318020, China
* Correspondence: [email protected] (X.C.); [email protected] (W.S.); Tel.: +86-139-5763-9868 (X.C.);
+86-0571-8718-0366 (W.S.)
† These authors contributed equally to this work.

Abstract: Background: Candesartan cilexetil, a Biopharmaceutics Classification System

(BCS) II prodrug, demonstrates compromised bioavailability attributable to its limited aque-
ous solubility coupled with P-glycoprotein (P-gp)-mediated efflux and hepatic first-pass
metabolism, thereby introducing complexities in generic drug bioequivalence assessments.
With the rapid advancement of computational technologies, the integration of biorelevant
dissolution methodologies with physiologically based pharmacokinetic (PBPK) modeling is
emerging as a transformative paradigm in advancing bioequivalence evaluation strategies
for generic drug products. This study presents a GastroPlus® -driven framework integrating
in vivo predictive dissolution (IPD) and virtual bioequivalence (VBE) to evaluate the quality
consistency of generic candesartan cilexetil tablets. Methods: By developing an oral PBPK
model in GastroPlus® , we established an IPD method using a phosphate-buffer-based
flow-through cell dissolution apparatus. In vitro dissolution profiles of generic tablets
Academic Editor: Serge Mordon from four manufacturers were measured and incorporated into the model to perform VBE
Received: 19 March 2025
simulations. Results: The results demonstrated that only the product from Company
Revised: 6 April 2025 A achieved virtual bioequivalence with the reference product, aligning with real-world
Accepted: 9 April 2025 quality consistency assessments. Conclusions: The proposed framework exhibited robust
Published: 11 April 2025 predictive capability, bridging in vitro dissolution data to in vivo bioequivalence outcomes,
Citation: Ruan, H.; Geng, X.; Situ, Z.; thereby offering a cost-effective and efficient strategy for formulation optimization and
Shen, Q.; Ye, T.; Chen, X.; Su, W. From preclinical bioequivalence evaluation of generic drugs.
In Vivo Predictive Dissolution to
Virtual Bioequivalence: A Keywords: candesartan cilexetil; GastroPlus®; in vivo predictive dissolution; flow-through cell;
GastroPlus® -Driven Framework for
virtual bioequivalence
Generic Candesartan Cilexetil Tablets.
Pharmaceuticals 2025, 18, 562.
https://doi.org/10.3390/
ph18040562
1. Introduction
Copyright: © 2025 by the authors.
Licensee MDPI, Basel, Switzerland. Dissolution testing remains a cornerstone of oral solid dosage form development,
This article is an open access article enabling rational formulation screening, manufacturing process refinement, and quality
distributed under the terms and assurance across production batches [1]. Its role has expanded under modern regulatory
conditions of the Creative Commons paradigms, where dissolution profiles increasingly serve as surrogates for therapeutic
Attribution (CC BY) license
equivalence assessments—particularly when in vitro–in vivo correlations (IVIVC) exist [2].
(https://creativecommons.org/
licenses/by/4.0/).
Nevertheless, conventional bioequivalence (BE) studies continue to depend on costly

Pharmaceuticals 2025, 18, 562 https://doi.org/10.3390/ph18040562

Pharmaceuticals 2025, 18, 562 2 of 17

and ethically challenging clinical trials [3]. This limitation proves especially acute for
Biopharmaceutics Classification System (BCS) II drugs (low solubility/high permeability),
where dissolution-driven absorption variability complicates equivalence predictions [4].
In recent years, the rapid advancement of computer technology has provided many
new ideas for drug research and development [5]. One example is pharmacokinetic studies
integrating in vivo predictive dissolution (IPD) with physiologically based pharmacoki-
netic (PBPK) models. These approaches integrate in vitro dissolution data with in vivo
absorption kinetics to simulate drug absorption and bioequivalence profiles, effectively
circumventing ethical dilemmas while enhancing the scientific rigor of therapeutic equiv-
alence assessments [6–10]. Notably, the flow-through cell system coupled with dynamic
pH modulation has demonstrated exceptional accuracy in predicting dissolution for BCS II
drugs [11–14]. The advantages of the flow-through cell lie in its capability to dynamically
adjust the dissolution medium composition and flow rate during experiments. For dosage
forms prone to floating, this method secures the sample at the center of the flow cell to
ensure full contact with the dissolution medium. In the case of poorly soluble drugs, the
open-loop configuration of the flow-through cell theoretically allows unlimited volumes
of dissolution medium, making it superior to basket or paddle methods in maintaining
sink conditions. Furthermore, by integrating fluid dynamics with biorelevant media (e.g.,
simulating gastrointestinal pH gradients), the flow-through cell effectively bridges IVIVC
for challenging compounds [11]. When conducting dissolution testing, the selection of
appropriate dissolution media is critical. According to the Chinese Pharmacopoeia (ChP) [15],
the media should be specified in the individual monograph for each drug product, typically
including: water (deaerated), 0.01~0.1 mol·L−1 hydrochloric acid (HCl) solution, and pH-
adjusted buffer solutions (e.g., phosphate or acetate buffers). In contrast, the United States
Pharmacopeia (USP) [16] mandates the use of standardized media such as pH 1.2 HCl
solution, pH 4.5 acetate buffer, and pH 6.8 phosphate buffer. Additionally, surface-active
agents (e.g., sodium dodecyl sulfate, SDS) may be incorporated into the media to enhance
solubility for poorly water-soluble drugs if necessary.
Commercial software platforms, including GastroPlus® , PK-Sim® , Stella® , and GI-
®
Sim , now enable comprehensive simulations of drug absorption, distribution, metabolism,
and excretion (ADME). Among these, GastroPlus® integrates computational models of phar-
macokinetics and pharmacodynamics while enabling the construction of ADME models to
simulate drug behavior under diverse administration routes [17–19]. Through PBPK mod-
eling, GastroPlus® facilitates personalized simulations by accounting for interindividual
variability [20–22], thereby supporting clinical dose optimization, drug safety evaluation,
and regulatory compliance [23–28]. The multifaceted strengths of GastroPlus® under-
pin its broad applicability. Primarily, it incorporates a comprehensive database of drug
physicochemical and biopharmaceutical parameters, ensuring robust model foundation.
Secondly, its computational framework spans the entire drug disposition cascade, enabling
holistic simulation of in vivo drug behavior. Moreover, the software’s flexibility allows
customized parameter adjustments to meet specific research objectives. For example, when
investigating pharmacokinetics in special populations (e.g., pediatric, geriatric, or hepa-
torenal impairment cohorts), users can adjust physiological and disposition parameters to
predict drug responses, thereby informing precision medicine strategies. This personalized
modeling capability enables GastroPlus® to play a pivotal role throughout drug develop-
ment phases, ranging from early-stage drug design to late-phase clinical trial planning and
regulatory submissions, thereby substantially enhancing both the efficiency and success
rate of pharmaceutical research and development.
Candesartan, a selective angiotensin II receptor blocker predominantly administered
as its prodrug candesartan cilexetil, is a first-line antihypertensive agent also indicated for
Pharmaceuticals 2025, 18, 562 3 of 18

Pharmaceuticals 2025, 18, 562 Candesartan, a selective angiotensin II receptor blocker predominantly administered 3 of 17
as its prodrug candesartan cilexetil, is a first-line antihypertensive agent also indicated for
cardiovascular conditions such as congestive heart failure [29,30]. However, its therapeu-
cardiovascular
tic potential is conditions
hindered bysuch
poorasaqueous
congestive heart failure
solubility (pKa [29,30]. However, P-glycoprotein
6.0), pronounced its therapeutic
potential is hindered by poor aqueous solubility (pKa 6.0), pronounced P-glycoprotein
(P-gp)-mediated efflux, and extensive hepatic first-pass metabolism, collectively limiting
(P-gp)-mediated efflux, and extensive hepatic first-pass metabolism, collectively limiting
oral bioavailability to 14% [31,32]. Marketed as Blopress® by Takeda Pharmaceutical Co .
oral bioavailability to 14% [31,32]. Marketed as Blopress® by Takeda Pharmaceutical Co.
(Tokyo, Japan), candesartan cilexetil tablets—a BCS II drug characterized by low solubil-
(Tokyo, Japan), candesartan cilexetil tablets—a BCS II drug characterized by low solubility
ity and high permeability—exhibit dissolution-limited absorption, making them an ideal
and high permeability—exhibit dissolution-limited absorption, making them an ideal
candidate for establishing IVIVC [33,34].
candidate for establishing IVIVC [33,34].
In this study, candesartan cilexetil tablets were selected as a model drug to develop
In this study, candesartan cilexetil tablets were selected as a model drug to develop a
a biorelevant flow-through cell dissolution method, construct a PBPK oral absorption
biorelevant flow-through cell dissolution method, construct a PBPK oral absorption model
model via GastroPlus® for translating in vitro dissolution data into in vivo release kinetics,
via GastroPlus® for translating in vitro dissolution data into in vivo release kinetics, and
and perform VBE analysis. This integrated framework establishes a cost-efficient strategy
perform VBE analysis. This integrated framework establishes a cost-efficient strategy to
to advance formulation development and therapeutic equivalence evaluations for BCS II
advance formulation development and therapeutic equivalence evaluations for BCS II
drugs, addressing critical challenges in current regulatory practices (Figure 1).
drugs, addressing critical challenges in current regulatory practices (Figure 1).

Figure1.1. Framework
Figure Frameworkof
ofthis
thisresearch.
research.

2.
2. Results
Results and
and Discussion
Discussion
2.1.
2.1. Methodological
Methodological Experiments
Experiments
For the in vitro dissolution testing requiring analytical quantitation, method validation
For the in vitro dissolution testing requiring analytical quantitation, method valida-
studies were conducted on the established High-Performance Liquid Chromatography (HPLC)
tion studies were conducted on the established High-Performance Liquid Chromatog-
analytical procedure in accordance with pharmacopeial requirements. The calibration curve
raphy (HPLC) analytical procedure in accordance with pharmacopeial requirements. The
demonstrating excellent linear correlation (y = 0.5483x − 0.0009, R2 = 1.0000) was established
calibration curve demonstrating excellent linear correlation (y = 0.5483x − 0.0009, R2 =
for candesartan cilexetil over the concentration range of 0.010 to 12.077 µg·mL− 1 , with the
1.0000) was established for candesartan cilexetil over the concentration range of 0.010 to
peak area showing proportional response to analyte concentration. The concentration of
12.077 μg·mL−1, with the peak area showing proportional response to analyte concentra-
the limit of quantification (LOQ) was 0.010 µg·mL− 1 , and the blank excipient showed no
tion. The concentration of the limit of quantification (LOQ) was 0.010 μg·mL−1, and the
significant interference. The recovery rates in different dissolution media are all between
blank excipient showed no significant interference. The recovery rates in different disso-
98.1% and 99.2%, and the precision RSD (n = 6) is between 0.11% and 0.26%, which meets
lution media are all between 98.1% and 99.2%, and the precision RSD (n = 6) is between
the requirements [15,35] for accuracy and precision. These results confirm the analytical
0.11% and 0.26%, which meets the requirements [15,35] for accuracy and precision. These
procedure’s suitability for dissolution testing of candesartan cilexetil tablets.
 Pharmaceuticals 2025, 18, 562 4 of 18

Pharmaceuticals 2025, 18, 562 4 of 17

results confirm the analytical procedure’s suitability for dissolution testing of candesartan
cilexetil tablets.
2.2. USP Apparatus 2 Dissolution Test
InUSP
2.2. the Apparatus
USP apparatus 2 dissolution
2 Dissolution Test testing of candesartan cilexetil tablets, we con-
ducted comparative analysis2between
In the USP apparatus thetesting
dissolution reference listed drug cilexetil
of candesartan (RLD) and generic
tablets, products
we con-
(Figure
ducted 2).comparative
Dissolutionanalysis
profiles were statistically
between the referenceevaluated
listed drugusing
(RLD)theandf2 similarity
generic prod-factor
ucts (Figure
method (Table2).1).Dissolution
The analysisprofiles were marked
revealed statistically evaluatedbetween
disparities using thegeneric
f2 similarity fac-
formulations
tor method (Table 1). The analysis revealed marked disparities between
from a certain company (notably Company D and B) and the RLD. Products of Company generic formula-
tions
B, D, andfrom a certain company
C demonstrated (notably
lower Companydissolution.
cumulative D and B) and the RLD. Products
Company of Com-exhib-
D’s product
pany B, D, and C demonstrated lower cumulative dissolution. Company D’s product ex-
ited an initial rapid dissolution followed by a slowdown, while Company C’s product
hibited an initial rapid dissolution followed by a slowdown, while Company C’s product
consistently showed slower dissolution rates compared to other formulations. Although
consistently showed slower dissolution rates compared to other formulations. Although
Company
Company B’sB’s
product displayed
product displayeddissolution
dissolution kinetics
kineticssimilar
similartotothe reference
the referenceproduct
productwithin
the within
first 10the
min,
firsta 10
subsequent reduction
min, a subsequent in dissolution
reduction rate was
in dissolution rateobserved. The f2
was observed. similarity
The f2
factor analysis
similarity indicated
factor analysisthat only that
indicated Company A’s product
only Company exhibited
A’s product dissolution
exhibited behavior
dissolution
comparable to the reference
behavior comparable to the product
referenceacross
product allacross
dissolution media.media.
all dissolution

Figure 2. Dissolution profile comparison between the RLD and generic drug product of candesartan
Figure 2. Dissolution profile comparison between the RLD and generic drug product of candesartan
cilexetil tablets in different dissolution media. (A) pH 1.0 hydrochloric acid solution (1.0% Tween
cilexetil tablets in different dissolution media. (A) pH 1.0 hydrochloric acid solution (1.0% Tween 20),
20), (B) pH 4.5 acetate buffer (1.0% Tween 20), (C) pH 6.5 phosphate buffer (0.25% Tween 20), (D)
(B) pH 4.5 acetate buffer (1.0% Tween 20), (C) pH 6.5 phosphate buffer (0.25% Tween 20),
pH 6.5 phosphate buffer (0.35% Tween 20), and (E) water (1.0% Tween 20).
(D) pH 6.5 phosphate buffer (0.35% Tween 20), and (E) water (1.0% Tween 20).

Table 1. Summary of f2 factor results.

Table 1. Summary of f2 factor results.
pH 1.0 Hydro- pH 6.5 Phos- pH 6.5 Phos-
pH 4.5 Acetate
pH 1.0 Hydrochloric chloric
pHAcid
4.5 Acetate
So- Buffer pH 6.5phate
Phosphate Buffer
Buffer pH 6.5
So- phate Phosphate
Buffer So- Buffer
Water (1.0%
Water (1.0%
Manufacturer Manufacturer
Acid Solution Buffer Solution Solution
Solution Solution
(1.0% Tween 20)
lution (1.0%
(1.0% Tween 20)
lution
(0.25%
(0.25%
Tween 20)
lution (0.35%
(0.35% Tween
Tween 20)
20)
Tween 20)
(1.0% Tween 20)
Tween 20) Tween 20) Tween 20)
A 73 80 74 58 69
B A 29 73 28 80 48 74 58 49 69 28
C B 38 29 40 28 26 48 49 28 28 39
D 32 41 44
C 38 40 26 28 41 39 39

D 32 41 44 41 39
2.3. Construction of GastroPlus® Model
2.3.1. Intravenous Prediction Model Results
We incorporated literature-reported intravenous PK data of candesartan into the
GastroPlus® disposition model by setting corresponding clinical parameters for PK profile
simulation. Observed data points represent urinary excretion amounts of candesartan,
while simulated curves depict plasma concentrations (dark blue) and urinary excretion
levels (light blue) (Figure 3). As candesartan is a P-glycoprotein (P-gp) substrate, it under-
goes minimal metabolic clearance, with the majority of elimination occurring via biliary
 2.3.1. Intravenous Prediction Model Results
We incorporated literature-reported intravenous PK data of candesartan into the
GastroPlus® disposition model by setting corresponding clinical parameters for PK profile
Pharmaceuticals 2025, 18, 562 5 of 17
simulation. Observed data points represent urinary excretion amounts of candesartan,
while simulated curves depict plasma concentrations (dark blue) and urinary excretion
levels (light blue) (Figure 3). As candesartan is a P-glycoprotein (P-gp) substrate, it under-
excretion. So, in the intravenous PBPK model, hepatic clearance was set to 0.748 L·h−1
goes minimal metabolic clearance, with the majority of elimination occurring via biliary
(40% of total
excretion. So, inclearance) and renal
the intravenous PBPKclearance to 1.12
model, hepatic L·h−1was
clearance (60%setof
to total clearance),
0.748 L·h −1 (40% result-
ing in aclearance)
total systemic clearance − 1
L·h−1 (60% − 1
mL·minresulting − 1
·kg ),in consistent
of total and renal clearanceofto1.868
1.12 L·h (about 0.37
of total clearance), a
with literature values. The calculated volume of distribution (10.479
total systemic clearance of 1.868 L·h (about 0.37 mL·min ·kg ), consistent with literature
−1 −1 −1 L) aligned with
literature-reported
values. The calculated data
volume − 1
(0.13ofL·distribution
kg × 84.25 kg). L)
(10.479 The model
aligned accurately
with predicted urinary
literature-reported
data (0.13
drug L·kg−1 patterns,
excretion × 84.25 kg). Therenal
with modelelimination
accurately predicted
proportion urinary drug excretion
reflecting candesartan’spat- in vivo
terns, with renal elimination proportion reflecting candesartan’s in vivo
clearance characteristics. Note that plasma PK curve comparison was omitted due to clearance charac-
teristics. Note that plasma PK curve comparison was omitted due to unavailable concen-
unavailable concentration–time data in the literature, though clearance and volume param-
tration–time data in the literature, though clearance and volume parameters remained lit-
eters remained literature-consistent. These results validate the accuracy of the Lukacova
erature-consistent. These results validate the accuracy of the Lukacova (Rodgers–Singh)
(Rodgers–Singh) method for predicting candesartan’s distribution. Consequently, the oral
method for predicting candesartan’s distribution. Consequently, the oral PBPK model of
PBPK modelcilexetil
candesartan of candesartan cilexetilmethodology.
adopted identical adopted identical methodology.

Figure Prediction
3.Prediction
Figure 3. results
results of candesartan
of candesartan American
American intravenous
intravenous PBPK model.
PBPK model.

2.3.2. OralPrediction
2.3.2. Oral Prediction Model
Model Results
Results
Basedon
Based onPK
PK data
data from
from the the regulatory
regulatory submission
submission documents
documents of candesartan
of candesartan cilexetil cilexetil
tablets from Takeda and actual measurement data, we
measurement data, we developed
developedand
andvalidated
validatedPBPKPBPK models
models
for 8 mgfor 8 mgoral
fasted fasted oral administration
administration in bothinJapanese
both Japanese and Chinese
and Chinese populations.
populations. Both models
Both models utilized the Opt logD Model SA/V 6.1 ASF model, with comparable
utilized the Opt logD Model SA/V 6.1 ASF model, with comparable intestinal first-pass intestinal
first-pass
effects buteffects but total
distinct distinct total clearance
clearance values,values,
whilewhile maintaining
maintaining consistent
consistent hepatic
hepatic (40%) and
(40%) and renal (60%) contributions to total clearance. Predictive accuracy (PE%) was cal-
renal (60%) contributions to total clearance. Predictive accuracy (PE%) was calculated by
culated by comparing simulated and observed PK profiles to comprehensively evaluate
Pharmaceuticals 2025, 18, 562 comparing simulated and observed PK profiles to comprehensively evaluate the 6 ofmodel’s
18
the model’s ability to characterize absorption, distribution, and metabolism processes
ability to characterize absorption, distribution, and metabolism processes (Figure 4).
(Figure 4).
For all PK parameters in the 8 mg oral dose models, PE% values remained below 20%
(Table 2), confirming the reliability of the physicochemical parameters, absorption model,
disposition parameters, and computational algorithms in reflecting the in vivo behavior
of candesartan cilexetil across ethnic populations.

Figure 4. PBPK
Figure 4. PBPKmodel
modelpredictions
predictions of
of 8 mgcandesartan
8 mg candesartan cilexetil
cilexetil tablets
tablets administered
administered via preprandial
via preprandial
oraloral
route across
route ethnic
across populations.
ethnic populations. (A) Japaneseand
(A) Japanese and(B)
(B) Chinese.
Chinese.

Table 2. Summary data of PK parameters for predicted results versus measured results, and assess-
ment of prediction accuracy.

Cmax AUC0-inf AUC0-t

Category Tmax (h)
(ng·mL−1) (ng·h·mL−1) (ng·h·mL−1)
 Pharmaceuticals 2025, 18, 562 6 of 17

For all PK parameters in the 8 mg oral dose models, PE% values remained below 20%
(Table 2), confirming the reliability of the physicochemical parameters, absorption model,
disposition parameters, and computational algorithms in reflecting the in vivo behavior of
candesartan cilexetil across ethnic populations.

Table 2. Summary data of PK parameters for predicted results versus measured results, and assess-
ment of prediction accuracy.

Cmax AUC0-inf AUC0-t

Category Tmax (h)
(ng·mL−1 ) (ng·h·mL−1 ) (ng·h·mL−1 )
Literature Value 83.5 4.01 795.2 780.4
Candesartan Cilexetil
Predicted Value 88.88 3.26 795.2 783.7
po 8 mg in Japanese
PE% 1 6.44 18.70 0.00 0.42
Measured Value 90.9 3.5 1118.5 1071.5
Candesartan Cilexetil
Predicted Value 93.85 3.36 1123.4 1111.9
po 8 mg in Chinese
PE% 1 3.25 −4.00 0.44 3.77
1 PE% = |Predicted-Measured|/Measured × 100.

2.4. In Vivo Dissolution and Absorption Evaluation Based on GastroPlus®

Using the established PBPK model for fasted oral administration of 8 mg candesar-
tan cilexetil in Chinese subjects, we predicted the corresponding in vivo dissolution and
absorption profiles (Figure 5). The results showed gradual drug dissolution in vivo, achiev-
ing complete dissolution within approximately 3 h after dosing. The dissolved drug
exhibited rapid and complete transcellular absorption into enterocytes, confirming can-
desartan’s high permeability and indicating that permeability does not constitute the
rate-limiting step in systemic absorption. This mechanistic understanding supports the
feasibility of establishing IVIVC for dissolution behavior. By aligning in vitro dissolution
methods with the predicted in vivo profiles, we further validated the utility of such correla-
tions for predicting pharmacokinetics and conducting VBE assessments. These findings
Pharmaceuticals 2025, 18, 562 7
underscore the potential of leveraging IPD to streamline formulation development and
bioequivalence evaluation.

Figure
Figure PBPK-model-predicted pharmacokinetic
5. 5.PBPK-model-predicted pharmacokineticprofiles of 8 mg
profiles of candesartan cilexetil tablets
8 mg candesartan cilexetil table
following fasted oral administration in Chinese.
lowing fasted oral administration in Chinese.
Furthermore, our investigation of in vivo absorption characteristics quantified the
Furthermore,
regional absorption our investigation
percentages of in cilexetil
of candesartan vivo absorption
tablets acrosscharacteristics quantifie
intestinal segments
regional absorption percentages of candesartan cilexetil tablets across intestinal segm
(Figure 6). During absorption, candesartan cilexetil undergoes conversion to candesa
a known P-gp substrate. While the current model does not explicitly simulate P-gp-m
 Furthermore, our investigation of in vivo absorption characteristics quantified
regional absorption percentages of candesartan cilexetil tablets across intestinal segme
Pharmaceuticals 2025, 18, 562 (Figure 6). During absorption, candesartan cilexetil undergoes conversion to 7 ofcandesart
17

a known P-gp substrate. While the current model does not explicitly simulate P-gp-me
ated (Figure
efflux 6).
effects
Duringonabsorption,
candesartan, this transport
candesartan mechanism
cilexetil undergoes was indirectly
conversion accounted
to candesartan,
through intestinal
a known first-passWhile
P-gp substrate. effectthe
parameterization.
current model does Simulation results
not explicitly under
simulate P-gp- fasted co
ditions indicate
mediated effluxcomplete absorptionthis
effects on candesartan, of transport
candesartan cilexetil,
mechanism with primary
was indirectly accountedabsorpt
for through intestinal first-pass effect parameterization. Simulation
sites localized to the jejunum, ileum, and cecum. However, approximately results under fasted 90% of
conditions indicate complete absorption of candesartan cilexetil, with primary absorp-
sorbed candesartan undergoes P-gp-dependent efflux back into the intestinal lumen, u
tion sites localized to the jejunum, ileum, and cecum. However, approximately 90% of
mately beingcandesartan
absorbed excreted inundergoes
the feces. This extensive
P-gp-dependent enteric
efflux recycling
back into resultslumen,
the intestinal in an absol
bioavailability of approximately
ultimately being 11%.
excreted in the feces. This extensive enteric recycling results in an absolute
bioavailability of approximately 11%.

Figure 6. The absorption percentages of 8 mg candesartan cilexetil tablets in different intestinal

Figure 6. The absorption percentages of 8 mg candesartan cilexetil tablets in different intestinal s
segments after oral administration before meals in Chinese.
ments after oral administration before meals in Chinese.
2.5. Establishment of In Vivo Predictive Dissolution Method
We compared
2.5. Establishment of Inthe in vivo
Vivo dissolution
Predictive profile ofMethod
Dissolution the reference product predicted by
GastroPlus® software (version 9.5) simulation with the in vitro dissolution curve obtained
We compared
using the in vivo
the USP apparatus dissolution
2 (Figure profile
7). The results of the reference
demonstrated productcilex-
that candesartan predicted
GastroPlus ® software (version 9.5) simulation with the in vitro dissolution curve obtain
etil tablets exhibited a slow release process in vivo, with approximately 87% released at
using the USP apparatus
150 min, and 2 (Figure
the dissolution rate was7). The results
significantly demonstrated
lower thatincandesartan
than that observed the in vitro cilex
dissolution curve. This indicates that, although the paddle method can differentiate prod-
uct quality under different dissolution media to some extent, its insufficient biorelevance
prevents it from reflecting the actual in vivo dissolution behavior of candesartan cilexetil
tablets. Therefore, based on the in vivo release profile predicted by the GastroPlus® model,
we conducted dissolution testing using the more biorelevant flow-through cell method. By
adjusting factors such as flow rate and pH gradient variations to optimize the alignment
between the RLD’s in vitro dissolution profile and the predicted in vivo release profile,
the methodological exploration is detailed in Table 3, with the resulting curves shown in
Figure 8. The results demonstrated that under method-three conditions, the RLD exhib-
ited the closest correlation between in vitro dissolution profiles and in vivo release curves,
thus method three was selected as the discriminatory dissolution method for subsequent
comparative dissolution testing of generic drug products from four companies (Figure 9).
 hibited the closest correlation between in vitro dissolution profiles and in vivo release
curves, thus method three was selected as the discriminatory dissolution method for sub-
tablets exhibited
sequent a slow dissolution
comparative release process in vivo,
testing with approximately
of generic drug products 87%
fromreleased at 150
four companies
Pharmaceuticals 2025, 18, 562 min, and the dissolution rate was significantly lower than that observed in the in vitro 8 of 17
(Figure 9).
dissolution curve. This indicates that, although the paddle method can differentiate prod-
uct quality under different dissolution media to some extent, its insufficient biorelevance
prevents it from reflecting the actual in vivo dissolution behavior of candesartan cilexetil
tablets. Therefore, based on the in vivo release profile predicted by the GastroPlus® model,
we conducted dissolution testing using the more biorelevant flow-through cell method.
By adjusting factors such as flow rate and pH gradient variations to optimize the align-
ment between the RLD’s in vitro dissolution profile and the predicted in vivo release pro-
file, the methodological exploration is detailed in Table 3, with the resulting curves shown
in Figure 8. The results demonstrated that under method-three conditions, the RLD ex-
hibited the closest correlation between in vitro dissolution profiles and in vivo release
curves, thus method three was selected as the discriminatory dissolution method for sub-
sequent comparative dissolution testing of generic drug products from four companies
(Figure 9).

Figure 7. Comparative analysis between the in vivo release profile of the RLD and in vitro dissolution
Figure 7. Comparative analysis between the in vivo release profile of the RLD and in vitro dissolu-
profiles obtained via USP apparatus 2 under varied biorelevant media conditions.
tion profiles obtained via USP apparatus 2 under varied biorelevant media conditions.

Table 3. Flow-through cell method exploration parameters.

Selected Method Flow Rate (mL·min−1 ) Medium Sampling Time (min)

pH 1.2 Hydrochloric Acid Solution (0.2% Tween 20) 10, 20
pH 4.5 Acetate Buffer Solution (0.2% Tween 20) 30, 45
Method One (Open Loop) 4 pH 5.7 Phosphate Buffer Solution (0.2% Tween 20) 60, 75, 90, 105, 120
pH 6.8 Phosphate Buffer Solution (0.3% Tween 20) 135, 150, 180, 210, 240
pH 1.2 Hydrochloric Acid Solution (0.2% Tween 20) 10, 20
pH 4.5 Acetate Buffer Solution (0.2% Tween 20) 30, 45
Method Two (Open Loop) 6 pH 5.7 Phosphate Buffer Solution (0.2% Tween 20) 60, 75, 90, 105, 120
pH 6.8 Phosphate Buffer Solution (0.3% Tween 20) 135, 150, 180, 210, 240
pH 1.2 Hydrochloric Acid Solution (0.2% Tween 20) 10, 20
Method pH 4.5 Acetate Buffer Solution (0.2% Tween 20) 30, 45
Three (Open Loop) 6 pH 5.7between
Phosphate
Figure 7. Comparative analysis the Buffer
in vivoSolution (0.2% Tween
release profile of the 20)
RLD and 60, 75, 90,dissolu-
in vitro 105, 120, 135, 150
pH 6.8 Phosphate Buffer Solution (0.3% Tween 20) 180, 210, 240
tion profiles obtained via USP apparatus 2 under varied biorelevant media conditions.

Figure 8. Comparative analysis of dissolution profiles under varied flow-through cell conditions.

Pharmaceuticals 2025, 18, 562 9 of 18

Figure 8. Comparative analysis of dissolution profiles under varied flow-through cell conditions.
Figure 8. Comparative analysis of dissolution profiles under varied flow-through cell conditions.

Dissolution
Figure9.9.Dissolution
Figure profile
profile characterization
characterization of candesartan
of candesartan cilexetilcilexetil tablets
tablets (RLD (RLD vs.
vs. generic generic drugs)
drugs)
usingthe
using theflow-through
flow-through
cell.cell.

Table 3. Flow-through cell method exploration parameters.

Selected Flow Rate

Medium Sampling Time (min)
Method (mL·min−1)
pH 1.2 Hydrochloric Acid Solution (0.2% Tween 20) 10, 20
Method
pH 4.5 Acetate Buffer Solution (0.2% Tween 20) 30, 45
One (Open 4
pH 5.7 Phosphate Buffer Solution (0.2% Tween 20) 60, 75, 90, 105, 120
Pharmaceuticals 2025, 18, 562 9 of 17

2.6. Virtual Bioequivalence Simulation

The in vitro dissolution profiles of the RLD and each generic formulation obtained
from the experiments were incorporated into PBPK models to conduct virtual bioequiva-
lence simulations for candesartan cilexetil tablets (Figure 10). The population simulation
involved 30 virtual subjects, with bioequivalence determined by whether the 90% confi-
dence intervals (CI) of the geometric mean ratios for Cmax, AUC, and AUCt fell within
80–125% of the RLD. The virtual bioequivalence simulation results revealed that only the
product from Company A demonstrated virtual bioequivalence to the RLD. Furthermore,
market investigation confirmed that Company A was the sole company whose product had
passed the national quality consistency evaluation, aligning with the simulation outcomes.
Detailed results are summarized in Table 4. Previous studies [32] have demonstrated the
utility of flow-through cell apparatus in developing predictive dissolution methods for
candesartan cilexetil, where deconvolution-derived absorption profiles were employed to
establish IVIVC models. In contrast, our approach leverages GastroPlus® -based simulations
to optimize dissolution conditions by aligning in vitro release profiles with physiologi-
cally informed in vivo prediction curves. Following parameter refinement, the validated
dissolution data were integrated into GastroPlus® to perform VBE simulations. While
conventional IVIVC methodologies rely on statistical correlations between dissolution
and absorption, they often overlook physiological variabilities such as gastrointestinal
Pharmaceuticals 2025, 18, 562 motility, pH gradients, and transporter-mediated processes. The PBPK-IPD-VBE 10 of 18platform

introduced in this study addresses these limitations by mechanistically incorporating for-

mulation properties and patient-specific physiological factors, thereby enabling robust
corporating formulation properties and patient-specific physiological factors, thereby en-
predictions under diverse clinical scenarios. This advancement holds particular promise
abling robust predictions under diverse clinical scenarios. This advancement holds par-
for complex generics
ticular promise where traditional
for complex generics where IVIVC frameworks
traditional face challenges.
IVIVC frameworks Furthermore,
face challenges.
as PBPK databases
Furthermore, and validation
as PBPK databases and frameworks continue continue
validation frameworks to expand—particularly
to expand—partic- for low-
ularly for
solubility, low-solubility, variable-permeability
variable-permeability (BCS II/IV) (BCS II/IV) compounds
compounds like candesartan
like candesartan cilexetil—the
cilexetil—the integration of in silico tools into bioequivalence assessments
integration of in silico tools into bioequivalence assessments is anticipated is anticipated
to become a
to become a cornerstone of modern generic drug development.
cornerstone of modern generic drug development.
Table 4. Virtual bioequivalence prediction results.
Table 4. Virtual bioequivalence prediction results.
Company A Company B Company C Company D
90% CI (MeanT-MeanR) 82.37~113.71
Company A 84.52~116.36
Company B 89.16~122.04
Company C82.05~113.62Company D
Cmax (GeomMeanT/GeomMeanR)
90% CI (MeanT-MeanR) × 100 99.09
82.37~113.71 101.3
84.52~116.36 106.3
89.16~122.04 98.6782.05~113.62
Cmax 90% CI (GeomMeanT/GeomMeanR)
(GeomMeanT/GeomMeanR) × 100 80.10~122.60
99.09 81.80~125.35
101.3 85.88~131.59
106.3 79.75~122.0798.67
90% CI (GeomMeanT/GeomMeanR)
90% CI (MeanT-MeanR) 80.10~122.60
82.21~115.44 81.80~125.35 87.30~121.83
87.35~121.84 85.88~131.59 79.75~122.07
86.25~120.42
90% CI (MeanT-MeanR) 82.21~115.44 87.35~121.84 87.30~121.83 86.25~120.42
AUC (GeomMeanT/GeomMeanR) × 100 99.87 105.2 105.1 104.1
AUC (GeomMeanT/GeomMeanR) × 100 99.87 105.2 105.1 104.1
90%90% CI (GeomMeanT/GeomMeanR)
CI (GeomMeanT/GeomMeanR) 80.40~124.04
80.40~124.04 84.66~130.84
84.66~130.84 84.55~130.73
84.55~130.7383.82~129.41
83.82~129.41
90%90%
CI (MeanT-MeanR)
CI (MeanT-MeanR) 82.30~115.34
82.30~115.34 87.39~121.65 87.42~121.76
87.39~121.65 87.42~121.7686.30~120.25
86.30~120.25
AUCt AUCt (GeomMeanT/GeomMeanR)
(GeomMeanT/GeomMeanR) × 100× 100 99.85
99.85 105.2
105.2 105.2 105.2 104.1 104.1
90% CI (GeomMeanT/GeomMeanR) 80.43~123.97 84.64~130.70 84.60~130.70 83.82~129.29
90% CI (GeomMeanT/GeomMeanR) 80.43~123.97 84.64~130.70 84.60~130.70 83.82~129.29

Figure 10. Cont.

Pharmaceuticals 2025, 18, 562 10 of 17

Figure
Figure 10. 10. Predicted VBE
Predicted VBE results
resultsof generic drug products.
of generic (A) Company
drug products. (A) A, (B) Company
Company A, B,
(B)(C)Company B,
Company C,
(C) Company C, (D)
(D)Company
Company D. D.

3. Materials and Methods

3.1. Materials
Acetonitrile (HPLC grade) was purchased from Meck (Darmstadt, Germany). Sodium
chloride, sodium acetate anhydrous, sodium hydroxide, potassium dihydrogen phosphate,
and phosphoric acid (analytical grade) were purchased from Sinopharm Chemical Reagent Co.,
Ltd. (Shanghai, China). Acetic acid (analytical grade) was purchased from Lingfeng Chemical
Reagent Co., Ltd. (Shanghai, China). Hydrochloric acid (analytical grade) was purchased
from Hannuo Chemical Technology Co., Ltd. (Zhejiang, China). Tween 20 (analytical grade)
was purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan).
The reference standard of candesartan cilexetil (batch number: 100685-201903, 99.8%
purity) was obtained from the National Institutes for Food and Drug Control (Beijing, China).
The reference products of candesartan cilexetil tablets were obtained from Tianjin
Takeda Pharmaceuticals Co., Ltd. (Tianjin, China). The generic drugs were obtained from
Company A (batch number: 1473A), Companies B (batch number: 2005126), Company C
(batch number: 200503), and Company D (batch number: 200401).

3.2. HPLC Conditions

The analysis was performed on the LC-20ADXR liquid chromatography system from
Shimadzu (Kyoto, Japan).
The quantification of candesartan was analyzed by HPLC-UV using a Kromasil 100-5
C18 column (5 µm, 4.6 × 50 mm) set at the controlled temperature of 35 ◦ C. The mobile
phase consisted of acetonitrile–glacial acetic acid–water (77:1:23, v/v/v). The flow rate was
1.0 mL/min, with an injection volume of 20 µL, and the detection wavelength was set
at 254 nm.

3.3. Method Validation

Approximately 50 mg of candesartan cilexetil reference standard was accurately
weighed and transferred to a 250 mL volumetric flask. The standard was dissolved in
ethanol (HPLC grade) via ultrasonication, diluted to volume with ethanol, and homog-
enized to yield a stock solution (200 µg·mL−1 ). A working solution (8 µg·mL−1 ) was
prepared by transferring 2.0 mL of the stock solution to a 50 mL volumetric flask and
diluting with dissolution medium.
Calibration standards were prepared at different concentrations ranging from
0.010 µg·mL−1 to 12.077 µg·mL−1 through serial dilution. A calibration curve was con-
structed by plotting the analyte peak area against nominal concentrations, with linear
Pharmaceuticals 2025, 18, 562 11 of 17

regression analysis performed using the least squares method. The LOQ was defined as
the lowest validated concentration meeting acceptable accuracy and precision criteria.
According to the analytical method validation guidelines of the ChP [15], samples
at the same concentration (equivalent to 100% concentration level) were used to evaluate
the accuracy and precision (n = 6). The recovery (%) and RSD (%) for the known spiked
amount was calculated.
The calculation formula for recovery rate is as follows:

Measured concentration in spiked sample

Recovery = × 100% (1)
Nominal spike concentration

Using a blank excipient solution prepared from Company A’s blank excipients, the
interference of the excipients on the test results was examined.

3.4. In Vitro Dissolution Tests

A comparative evaluation was conducted to assess the predictive capacity of the
flow-through cell versus the USP apparatus 2 in characterizing the in vivo dissolution
behavior of candesartan cilexetil tablets. The in vitro dissolution data derived from the
flow-through cell were subsequently integrated into a PBPK model for mechanistic analysis
of drug absorption dynamics.

3.4.1. USP Apparatus 2

The dissolution test of USP apparatus 2 was performed on the 709-DS automatic
dissolution system from Agilent (Santa Clara, CA, USA).
Dissolution profiles were evaluated using USP Apparatus 2 under the following
conditions: 900 mL of dissolution media (compositions detailed in Table 5), temperature
maintained at 37.0 ± 0.5 ◦ C, and agitation speed of 50 rpm. Twelve replicates per batch
were analyzed and samples were continuously filtered through a 0.45 µm membrane filter
and quantified via HPLC analysis (n = 12 for each batch).

Table 5. Parameter settings of USP apparatus 2.

Medium Flow Rate (r·min−1 ) Sampling Time (min)

Water (1.0% Tween 20) 50 10, 15, 30, 45, 60, 90, 120
pH 1.0 Hydrochloric Acid Solution (1.0% Tween 20) 50 10, 15, 30, 45, 60, 90, 120
pH 4.5 Acetate Buffer Solution (1.0% Tween 20) 50 10, 15, 30, 45, 60, 90, 120, 150
pH 6.5 Phosphate Buffer Solution (0.25% Tween 20) 50 10, 15, 20, 30, 45, 60
pH 6.5 Phosphate Buffer Solution (0.35% Tween 20) 50 10, 15, 20, 30, 45, 60

3.4.2. Flow-Through Cell

The dissolution test of the flow-through cell was performed on the SOTAX CE 7smart
system from sotax (Basel, Switzerland).
The parameters were configured as detailed in Table 6. The experimental setup
consisted of a ruby bead fixed at the bottom of a 12 mm cell, followed by standardized
scoop-loading of 1 mm glass beads. Tablets were secured in sample holders, with the cell
assembly capped by a GF/D (2.7 µm) and GF/F (0.7 µm) glass microfiber filter system
containing defatted cotton purchased from Glasai Life Sciences (Shanghai, China) Co., Ltd.
(Shanghai, China). Six parallel experiments (n = 6) were simultaneously performed un-
der controlled temperature conditions (37 ± 0.5 ◦ C). Samples were continuously filtered
through a 0.45 µm membrane filter and analyzed by HPLC using a validated method
consistent with USP apparatus 2 specifications.
Pharmaceuticals 2025, 18, 562 12 of 17

Table 6. Parameter settings of flow-through cell.

Flow Rate
Medium Sampling Time (min)
(mL·min−1 )
pH 1.2 Hydrochloric Acid Solution (0.2% Tween 20) 6 10, 20
pH 4.5 Acetate Buffer Solution (0.2% Tween 20) 6 30, 45
pH 5.7 Phosphate Buffer Solution (0.2% Tween 20) 6 60, 75, 90, 105, 120, 135, 150
pH 6.8 Phosphate Buffer Solution (0.3% Tween 20) 6 180, 210, 240

3.5. Construction of PBPK Model

The PBPK model of candesartan cilexetil was developed using commercially available
PBPK modeling software GastroPlus® (version 9.5, Simulations Plus, Inc., Lancaster, CA, USA)
and Digit literature data extraction software (version 1.0.4, Simulation Plus, Inc., Lan-
caster, CA, USA).
Physicochemical and biopharmaceutical parameters for model construction were inte-
grated from multiple sources, including literature-reported data, experimentally measured
values, software-default parameters, and quantitative structure-based predictions derived
from candesartan cilexetil and its active metabolite candesartan (Table 7).

Table 7. Parameter settings of candesartan cilexetil PBPK model.

Parameter Value Source

440.46 g·mol−1 (candesartan)
Molecular Weight
610.67 g·mol−1 (candesartan cilexetil)
3.42 (candesartan) [36]
logP
7.1 (candesartan cilexetil) [37]
1.45, 4.23 (candesartan) [36]
pKa
Acid: 6.0 (candesartan cilexetil) [38]
pH 1.0: 0.23 µg·mL−1
pH 4.5: 0.51 µg·mL−1
Solubility (pH-solubility) [36]
pH 6.5: 0.8 µg·mL−1
Water (pH 6.8): 1.4 µg·mL−1
Precipitation Time 900 s Default Value
Calculated based on the
Diffusion Coefficient 0.63 × 10−5 cm2 ·s−1
formula built in GastroPlus®
Particle Density 1.2 g·mL−1 Default Value
Particle Size 4 µm [39]
Human jejunal permeability coefficient:
Human Permeability [40]
3.6 × 10−4 cm·s−1
Whole Blood/Plasma Drug Predicted Value from ADMET
0.68
Concentration Ratio Predictor 9.5
Instruction Manual:
Plasma Unbound Drug
1% Candesartan is more than 99%
Fraction
bound to plasma proteins

The in vivo dissolution process of candesartan cilexetil tablets was examined using
the Johnson dissolution model, with the formula as follows:

dMd D (1+2s)
dt = ρhγt × S × (Cs − Cl ) × Mu,t (2)

where D is the diffusion coefficient, ρ is the particle density, r is the particle radius, h is
the diffusion layer thickness of the dissolution process, s is the shape factor (the ratio
of the particle length to diameter), C is the drug solubility, and Mu,t is the amount of
undissolved drug.
The in vivo absorption of candesartan cilexetil tablets was simulated by integrating
GastroPlus® ’s mechanistic absorption models with physiological gastrointestinal parame-
ters. The governing equations are expressed as:

dMabs,i    
= ASF trans,i × Ptrans,i × Vlum,i × c(t)lum,i − c(t)entU,i + ASF para,i × Ppara,i × Vlum,i × c(t)lum,i − c(t) pvU (3)
dt
Pharmaceuticals 2025, 18, 562 13 of 17

where ASFtrans,i and ASFpara,i are absorption-scale factors for transcellular and paracellular
diffusion, theoretical ASF is calculated as the surface area-to-volume ratio (equals 2 divided
by Ri , where Ri represents the radius of intestinal segment i), Ptrans,i and PPara,i are effective
permeability coefficients for transcellular and paracellular pathways, Vlum,i is luminal
volume in gastrointestinal compartment, C(t)lum,i is time-dependent drug concentration in
intestinal lumen of compartment, C(t)entU,i is unbound drug concentration in enterocytes of
compartment, and C(t)pvU is unbound drug concentration in hepatic portal vein.
The gastrointestinal physiological model was implemented using the Advanced Com-
partmental Absorption and Transit (ACAT) framework. The Absorption Scale Factor
(ASF) model employed the Opt logD Model SA/V 6.1 (default configuration), with key
physiological parameters as detailed in Supplementary Materials.
Candesartan cilexetil undergoes rapid and complete hydrolysis to candesartan during
gastrointestinal absorption, with only the active metabolite entering systemic circulation.
Consequently, the physicochemical parameters of candesartan were utilized in conjunction
with tissue-specific physiological parameters to simulate its distribution kinetics. The
steady-state volume of distribution was estimated using the Poulin method, incorporating
tissue volumes and tissue-to-plasma partition coefficients:

Vss = Vp + Ve × E : P + ∑ Vt × Kpt × (1 − ERt ) (4)

where Vp is plasma volume, Ve is erythrocyte volume, E:P is erythrocyte-to-plasma con-

centration ratio (calculated from blood-to-plasma ratio B/P and hematocrit), Vt is tissue
volume, Kpt is tissue-to-plasma partition coefficient (predicted via the Lukacova method,
also known as Rodgers–Singh methodology), and ERt is tissue-specific uptake rate.
The candesartan cilexetil intravenous and oral model was developed by replicating
literature-reported [41,42] administration protocols (e.g., identical route, dosage, and sub-
ject demographics) as detailed in Table 8. Model validation involved comparative analysis
between simulated PK profiles and observed clinical data, with particular emphasis on op-
timizing clearance (CL) and volume of distribution (Vd) parameters. Subsequently, human
oral PK profiles from published studies were incorporated into the GastroPlus® disposition
module through clinical parameter alignment. This enabled simulation-based evaluation
of model predictive accuracy for in vivo dissolution-absorption processes, culminating in
definitive characterization of ADME parameters and associated computational algorithms.

Table 8. Basic setup of intravenous and oral PK model for candesartan.

Model Dose (mg) Population Clearance Rate Source

American male
Candesartan (25 years old,
Cilexetil iv 4 mg in 4 84.25 Kg body / Literature report
American weight) under
fasted conditions
Japanese male Candesartan
Candesartan (30 years old, cilexetil tablet
Liver: 0.4 L·h−1
Cilexetil po 8 mg 8 62.57 Kg body registration
Kidney: 0.6 L·h−1
in Japanese weight) under documents from
fasted conditions Takeda
Chinese male
Candesartan (30 years old, 63 Kg Liver: 0.312 L·h−1
Cilexetil po 8 mg 8 body weight) Kidney: Measured data
in Chinese under fasted 0.468 L·h−1
conditions

Leveraging pharmacokinetic data from the reference formulation, the Johnson dissolu-
tion model was implemented to simulate IPD profiles in Chinese populations. The derived
Pharmaceuticals 2025, 18, 562 14 of 17

in vivo dissolution–absorption correlation matrix informed development of clinically rele-

vant dissolution testing methodologies.

3.6. Virtual Bioequivalence Evaluation

Virtual bioequivalence assessment serves as a critical tool for demonstrating phar-
maceutical quality and therapeutic performance throughout the product lifecycle, while
supporting formulation optimization through mechanistically grounded predictions.
Virtual bioequivalence trials provide scientific evidence for the safety and efficacy of
pharmaceutical products, ensuring quality and performance throughout the entire product
lifecycle and offering support for the development of drug formulations. By integrating
in vitro dissolution profiles into the established PBPK models, we simulated in vivo PK
curves using the Weibull equation and compared them with experimentally measured PK
profiles. The simulation involved a virtual population of 30 subjects over a 48 h period.
The Weibull equation is as follows:
    
T )b1 −(t− T )b2
%DoseRelease = Max × 1 − f 1 exp −(t− A1 − f 2 exp A2

b3
 (5)
− f 3 exp −(t−
A3
T)

where Max is the total dissolution percentage, T is the lag time (h), f 1, f 2, and f 3 are the
dissolution fractions of the 1st, 2nd, and 3rd phases (f 1 + f 2 + f 3 = 1), b1, b2, and b3 are the
shape factors for the 1st, 2nd, and 3rd phases, and A1 , A2 , and A3 are the corresponding
phase time scales.

4. Conclusions
The above experiments demonstrated that the flow-through cell method could better
simulate in vivo conditions. Based on this, we developed an IPD method for generic drug
consistency evaluation. By integrating this method with a PBPK model established using
GastroPlus® , we successfully simulated the absorption, distribution, and metabolism of
candesartan cilexetil in different populations. Furthermore, virtual bioequivalence results
indicated that only products of Company A exhibited bioequivalence with the reference
product. In conclusion, the integrated PBPK-IPD-VBE platform established in this study
provides a robust framework for simulating dissolution, absorption, distribution, and
metabolism of candesartan cilexetil tablets in Chinese populations. This approach enhances
quality consistency evaluation, formulation development, and preclinical bioequivalence
prediction, ultimately accelerating generic drug development cycles and reducing research
and development costs. While European and American PK data were not available in
the current study, future work will focus on model refinement and validation for these
populations through prospective clinical data collection.

Supplementary Materials: The following are available online at https://www.mdpi.com/article/10

.3390/ph18040562/s1, Table S1: Accuracy and precision assessment results obtained by HPLC-UV;
Figure S1: Physiological parameter diagram of candesartan cilexetil PBPK model.

Author Contributions: Conceptualization, X.C. and W.S.; Data curation, H.R., X.G. and Z.S.; Formal
analysis, X.G. and Q.S.; Funding acquisition, H.R.; Investigation, H.R.; Methodology, H.R., X.G. and
X.C.; Project administration, X.C. and W.S.; Resources, T.Y.; Software, H.R. and X.G.; Supervision,
X.C.; Validation, X.G. and Q.S.; Visualization, Z.S.; Writing—original draft, H.R. and X.G.; Writing—
review and editing, Z.S. and W.S. All authors have read and agreed to the published version of
the manuscript.
Pharmaceuticals 2025, 18, 562 15 of 17

Funding: This work was supported by the Zhejiang Basic Public Welfare Research Program (LGC22H300003);
Science and Technology Project of Zhejiang Medical Products Administration (2023018).

Institutional Review Board Statement: The study was conducted in accordance with the Declaration
of Helsinki, and approved by the Institutional Review Board of Human Subject Research Ethics
Committee of the Second Affiliated Hospital School of Medicine, Zhejiang University (LSYD No. 230,
26 July 2017).

Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement: The data presented in this study are available in this article.

Conflicts of Interest: Authors Tianjian Ye and Xin Chen were employed by the company Zhejiang
Yongning Pharmaceutical Co., Ltd. The remaining authors declare that the research was conducted
in the absence of any commercial or financial relationships that could be construed as a potential
conflict of interest.

References
1. Shah, V.P. Progressive Applications of Dissolution, Its Impact, and Implications in the Pharmaceutical World. J. Pharm. Sci. 2013,
102, 2895–2897. [CrossRef]
2. Liu, W.; Tu, L.X.; Yang, S.L.; Jin, Y. Research progress of in vitro and in vivo correlation evaluation method for generic oral solid
preparations. Drug Eval. Res. 2020, 43, 2565–2570.
3. Alomari, N.; Alhussaini, W. Update on the Advances and Challenges in Bioequivalence Testing Methods for Complex Topical
Generic Products. Front. Pharmacol. 2024, 15, 1330712. [CrossRef] [PubMed]
4. Honório, T.D.S.; Pinto, E.C.; Rocha, H.V.A.; Esteves, V.S.D.; Dos Santos, T.C.; Castro, H.C.R.; Rodrigues, C.R.; De Sousa, V.P.;
Cabral, L.M. In Vitro–In Vivo Correlation of Efavirenz Tablets Using GastroPlus® . AAPS PharmSciTech 2013, 14, 1244–1254.
[CrossRef]
5. Mondal, P.; Roy, S.; Loganathan, G.; Mandal, B.; Dharumadurai, D.; Akbarsha, M.A.; Sengupta, P.S.; Chattopadhyay, S.;
Guin, P.S. 1-Amino-4-Hydroxy-9,10-Anthraquinone—An Analogue of Anthracycline Anticancer Drugs, Interacts with DNA and
Induces Apoptosis in Human MDA-MB-231 Breast Adinocarcinoma Cells: Evaluation of Structure–Activity Relationship Using
Computational, Spectroscopic and Biochemical Studies. Biochem. Biophys. Rep. 2015, 4, 312–323. [CrossRef] [PubMed]
6. Shuai, W.; Cao, J.; Qian, M.; Tang, Z. Physiologically Based Pharmacokinetic Modeling of Vancomycin in Critically Ill Neonates:
Assessing the Impact of Pathophysiological Changes. J. Clin. Pharma 2024, 64, 1552–1565. [CrossRef]
7. Kalsoom, S.; Rasool, M.F.; Imran, I.; Saeed, H.; Ahmad, T.; Alqahtani, F. A Comprehensive Physiologically Based Pharmacokinetic
Model of Nadolol in Adults with Renal Disease and Pediatrics with Supraventricular Tachycardia. Pharmaceuticals 2024, 17, 265.
[CrossRef]
8. Kollipara, S.; Bhattiprolu, A.K.; Boddu, R.; Chougule, M.; Saha, P.; Ahmed, T. Demonstrating Discriminatory Power of a
Dissolution Method Using DDDPlus: Case Study of an Extended-Release Formulation and Use in Regulatory Justifications.
Dissolut. Technol. 2024, 31, 174–181. [CrossRef]
9. Statelova, M.; Vertzoni, M.; Kourentas, A. Simulation of Intraluminal Performance of Lipophilic Weak Bases in Fasted Healthy
Adults Using DDDPlusTM. AAPS J. 2022, 24, 89. [CrossRef]
10. Njoku, J.O.; Amaral Silva, D.; Mukherjee, D.; Webster, G.K.; Löbenberg, R. In Silico Tools at Early Stage of Pharmaceutical
Development: Data Needs and Software Capabilities. AAPS PharmSciTech 2019, 20, 243. [CrossRef]
11. D’Arcy, D.M.; Liu, B.; Corrigan, O.I. Investigating the Effect of Solubility and Density Gradients on Local Hydrodynamics and
Drug Dissolution in the USP 4 Dissolution Apparatus. Int. J. Pharm. 2011, 419, 175–185. [CrossRef] [PubMed]
12. Kushwah, V.; Arora, S.; Tamás Katona, M.; Modhave, D.; Fröhlich, E.; Paudel, A. On Absorption Modeling and Food Effect
Prediction of Rivaroxaban, a BCS II Drug Orally Administered as an Immediate-Release Tablet. Pharmaceutics 2021, 13, 283.
[CrossRef]
13. Milanowski, B.; Hejduk, A.; Bawiec, M.A.; Jakubowska, E.; Urbańska, A.; Wiśniewska, A.; Garbacz, G.; Lulek, J. Biorelevant In
Vitro Release Testing and In Vivo Study of Extended-Release Niacin Hydrophilic Matrix Tablets. AAPS PharmSciTech 2020, 21, 83.
[CrossRef]
14. Zarmpi, P.; Flanagan, T.; Meehan, E.; Mann, J.; Østergaard, J.; Fotaki, N. Biopharmaceutical Implications of Excipient Variability
on Drug Dissolution from Immediate Release Products. Eur. J. Pharm. Biopharm. 2020, 154, 195–209. [CrossRef] [PubMed]
15. Chinese Pharmacopoeia; Chinese Pharmacopoeia Commission: Beijing, China, 2020; 4, pp. 132, 480. Available online: https:
//db.ouryao.com/yd2020/view.php?id=f91d2ccf6a (accessed on 3 April 2025).
Pharmaceuticals 2025, 18, 562 16 of 17

16. United States Pharmacopeia (USP 43-NF 38). <711> Dissolution. 2020. Available online: https://www.usp.org/search (accessed
on 3 April 2025).
17. Kane, Z.; Picetti, R.; Wilby, A.; Standing, J.F.; Grassin-Delyle, S.; Roberts, I.; Shakur-Still, H. Physiologically Based Modelling of
Tranexamic Acid Pharmacokinetics Following Intravenous, Intramuscular, Sub-Cutaneous and Oral Administration in Healthy
Volunteers. Eur. J. Pharm. Sci. 2021, 164, 105893. [CrossRef] [PubMed]
18. Jeong, Y.-S.; Jusko, W.J. Meta-Assessment of Metformin Absorption and Disposition Pharmacokinetics in Nine Species. Pharma-
ceuticals 2021, 14, 545. [CrossRef]
19. Deb, S.; Reeves, A.A. Simulation of Remdesivir Pharmacokinetics and Its Drug Interactions. J. Pharm. Pharm. Sci. 2021,
24, 277–291. [CrossRef]
20. Kim, C.; Lo Re, V.; Rodriguez, M.; Lukas, J.C.; Leal, N.; Campo, C.; García-Bea, A.; Suarez, E.; Schmidt, S.; Vozmediano, V.
Application of a Dual Mechanistic Approach to Support Bilastine Dose Selection for Older Adults. CPT Pharmacom Syst. Pharma
2021, 10, 1006–1017. [CrossRef]
21. Ye, L.; You, X.; Zhou, J.; Wu, C.; Ke, M.; Wu, W.; Huang, P.; Lin, C. Physiologically Based Pharmacokinetic Modeling of
Daptomycin Dose Optimization in Pediatric Patients with Renal Impairment. Front. Pharmacol. 2022, 13, 838599. [CrossRef]
22. Xu, Y.; Chen, J.; Ruan, Z.; Jiang, B.; Yang, D.; Hu, Y.; Lou, H. Simulation of Febuxostat Pharmacokinetics in Healthy Subjects and
Patients with Impaired Kidney Function Using Physiologically Based Pharmacokinetic Modeling. Biopharm. Drug Dispos. 2022,
43, 140–151. [CrossRef]
23. Silva, T.M.D.; Honorio, T.D.S.; Chaves, M.H.D.C.; Duque, M.D.; Cabral, L.M.; Patricio, B.F.D.C.; Rocha, H.V.A. In Silico
Bioavailability for BCS Class II Efavirenz Tablets Using Biorelevant Dissolution Media for IVIVR and Simulation of Formulation
Changes. Drug Dev. Ind. Pharm. 2021, 47, 1342–1352. [CrossRef] [PubMed]
24. Najjar, A.; Punt, A.; Wambaugh, J.; Paini, A.; Ellison, C.; Fragki, S.; Bianchi, E.; Zhang, F.; Westerhout, J.; Mueller, D.; et al. Towards
Best Use and Regulatory Acceptance of Generic Physiologically Based Kinetic (PBK) Models for in Vitro-to-in Vivo Extrapolation
(IVIVE) in Chemical Risk Assessment. Arch. Toxicol. 2022, 96, 3407–3419. [CrossRef] [PubMed]
25. Rahim, N.; Naqvi, S.B.S. In Vitro In Vivo Extrapolation and Bioequivalence Prediction for Immediate-Release Capsules of
Cefadroxil Based on a Physiologically-Based Pharmacokinetic ACAT Model. AAPS PharmSciTech 2024, 25, 100. [CrossRef]
[PubMed]
26. Zhang, X.; Wang, X.; Li, R.; Zhang, C.; Du, J.; Zhao, H.; Wen, Q. Development and Application of a Physiologically Based
Pharmacokinetic Model for Elagolix in the Adult and Adolescent Population. Clin. Pharmacokinet. 2024, 63, 1357–1370. [CrossRef]
27. Wang, L.; Zhao, P.; Luo, T.; Yang, D.; Jiang, Q.; Chen, J.; Lou, H.; Ruan, Z.; Jiang, B. Physiologically Based Absorption Modeling to
Predict the Bioequivalence of Two Cilostazol Formulations. Clin. Transl. Sci. 2023, 16, 2323–2330. [CrossRef]
28. Cai, L.; Ke, M.; Wang, H.; Wu, W.; Lin, R.; Huang, P.; Lin, C. Physiologically Based Pharmacokinetic Model Combined with
Reverse Dose Method to Study the Nephrotoxic Tolerance Dose of Tacrolimus. Arch. Toxicol. 2023, 97, 2659–2673. [CrossRef]
29. Aly, U.F.; Sarhan, H.A.; Ali, T.F.; Sharkawy, H.A.E.-B. Applying Different Techniques to Improve the Bioavailability of Candesartan
Cilexetil Antihypertensive Drug. Drug Des. Dev. Ther. 2020, 14, 1851–1865. [CrossRef]
30. Tjandrawinata, R.R.; Setiawati, E.; Yunaidi, D.A.; Simanjuntak, R.; Santoso, I.D.; Susanto, L.W. Bioequivalence Study of Two
Formulations of Candesartan Cilexetil Tablet in Healthy Subjects under Fasting Conditions. Drug Des. Dev. Ther. 2013, 7, 841–847.
[CrossRef]
31. Anwar, W.; Dawaba, H.M.; Afouna, M.I.; Samy, A.M.; Rashed, M.H.; Abdelaziz, A.E. Enhancing the Oral Bioavailability of
Candesartan Cilexetil Loaded Nanostructured Lipid Carriers: In Vitro Characterization and Absorption in Rats after Oral
Administration. Pharmaceutics 2020, 12, 1047. [CrossRef]
32. Figueroa-Campos, A.; Sánchez-Dengra, B.; Merino, V.; Dahan, A.; González-Álvarez, I.; García-Arieta, A.; González-Álvarez, M.;
Bermejo, M. Candesartan Cilexetil In Vitro–In Vivo Correlation: Predictive Dissolution as a Development Tool. Pharmaceutics
2020, 12, 633. [CrossRef]
33. Dudhipala, N.; Veerabrahma, K. Candesartan Cilexetil Loaded Nanodelivery Systems for Improved Oral Bioavailability. Ther.
Deliv. 2017, 8, 79–88. [CrossRef] [PubMed]
34. Jain, S.; Reddy, V.A.; Arora, S.; Patel, K. Development of Surface Stabilized Candesartan Cilexetil Nanocrystals with Enhanced
Dissolution Rate, Permeation Rate across CaCo-2, and Oral Bioavailability. Drug Deliv. Transl. Res. 2016, 6, 498–510. [CrossRef]
[PubMed]
35. International Council for Harmonisation. Validation of Analytical Procedures: Text and Methodology Q2 (R1). 2005. Avail-
able online: https://www.cde.org.cn/ichWeb/guideIch/downloadAtt/1/34be5e106c19f419d10dfd32fb760ec6 (accessed on
3 April 2025).
36. Ma, H.Z. Study on Consistency of Dissolution and Simulation of Bioequivalence of Candesartan Cilexetil Tablets. Master’s Thesis,
China Medical University, Shenyang, China, 2018.
Pharmaceuticals 2025, 18, 562 17 of 17

37. De Campos, D.P.; Silva-Barcellos, N.M.; Lima, R.R.; Savedra, R.M.L.; Siqueira, M.F.; Yoshida, M.I.; Da Nova Mussel, W.; De
Souza, J. Polymorphic and Quantum Chemistry Characterization of Candesartan Cilexetil: Importance for the Correct Drug
Classification According to Biopharmaceutics Classification System. AAPS PharmSciTech 2018, 19, 3019–3028. [CrossRef]
38. Husain, A.; Azim, M.S.; Mitra, M.; Bhasin, P.S. A Review on Candesartan: Pharmacological and Pharmaceutical Profile. J. Appl.
Pharm. Sci. 2011, 1, 12–17.
39. Scholz, A.; Kostewicz, E.; Abrahamsson, B.; Dressman, J.B. Can the USP Paddle Method Be Used to Represent In-Vivo Hydrody-
namics? J. Pharm. Pharmacol. 2003, 55, 443–451. [CrossRef]
40. Haslam, I.S.; O’Reilly, D.A.; Sherlock, D.J.; Kauser, A.; Womack, C.; Coleman, T. Pancreatoduodenectomy as a Source of Human
Small Intestine for Ussing Chamber Investigations and Comparative Studies with Rat Tissue. Biopharm. Drug Dispos. 2011, 32,
210–221. [CrossRef] [PubMed]
41. van Lier, J.J.; van Heiningen, P.N.; Sunzel, M. Absorption, Metabolism and Excretion of 14 C-Candesartan and 14 C-Candesartan
Cilexetil in Healthy Volunteers. J. Hum. Hypertens. 1997, 11 (Suppl. 2), S27–S28.
42. Yu, J.; Ma, K.; Qi, J.W.; Jin, G.; Wang, Y.; Fang, S.G.; Li, G.H.; Xie, D. Pharmacokinetics and bioequivalence of candesartan cilexetic
tablet and capsule in healthy volunteers. Chin. J. Clin. Pharmacol. Ther. 2009, 14, 794.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.