Journal of Drug Delivery Science and Technology 69 (2022) 103152

Contents lists available at ScienceDirect

Journal of Drug Delivery Science and Technology

journal homepage: www.elsevier.com/locate/jddst

Applications of PBPK/PBBM modeling in generic product development: An

industry perspective
Yuvaneshwari K, Sivacharan Kollipara, Tausif Ahmed *, Siddharth Chachad
Biopharmaceutics Group, Global Clinical Management, Dr. Reddy’s Laboratories Ltd., Integrated Product Development Organization (IPDO), Bachupally, Medchal
Malkajgiri District, Hyderabad, 500 090, Telangana, India

A R T I C L E I N F O A B S T R A C T

Keywords: Physiologically based pharmacokinetic (PBPK) modeling and physiologically based biopharmaceutics modeling
PBPK modeling (PBBM) are invaluable tools with wide applications during drug discovery and development stages. At the drug
Biopharmaceutics discovery setting, PBPK modeling is used for first in human pharmacokinetics prediction, food effect assessment,
Food effect
and preclinical modeling. During generic formulation development, PBPK and PBBM modeling can be utilized for
Dissolution safe space
Generic product development
suitable formulation design in order to achieve bioequivalence with a Reference Listed Drug (RLD). The purpose
of this review is to lay down the applications of PBPK from the generic product development perspective.
Evolution of PBPK models and regulatory perspectives in terms of PBPK modeling and simulation guidelines from
the USFDA and EMA are critically reviewed. Applications of PBPK modeling in generic industry are discussed in
detail. Utility of PBPK modeling for formulation development for modified release formulation development,
biowaiver and API particle size impact, food effect studies are emphasized in detail with case studies. Appli­
cations related to dissolution safe space, justification of dissolution specifications and development of clinically
relevant dissolution specifications are discussed with specific case examples. Further, insight is provided into
modeling applications with regard to non-oral routes of administrations such as ocular, pulmonary, transdermal
and parenteral. Finally, future perspectives for modeling & simulations for both oral and non-oral formulations
and requirements for harmonization of PBPK guidelines are discussed.

classes of compounds. Such models are often built with API physico­
1. Introduction chemical properties (e.g. solubility, permeability, particle size), formu­
lation aspects (e.g. dosage form, formulation type, dissolution) and
The generic industry plays a vital role in controlling costs of essential physiological aspects (e.g. fasting/fed state, animal or human
medicine there by increasing the affordability & reach of these medi­ physiology).
cines to patients without compromising the quality. Development PBPK modeling has a long history and evolved slowly over the years.
timelines and the cost of a product is bound to increase and delay the During 1930 to 1960, simple mathematical equation-based modeling
market launch in case of complex products (e.g. complex API and such as compartmental models were used by fitting data into one, two or
mixture of API’s, complex dosage form and formulations, complex route three compartmental models. In these models, ADME processes are
of delivery). Regulatory agencies across the globe encourage and described using a first-order equation considering elimination of con­
welcome alternate innovative clinical development approaches that stant portion of drug per unit time. During the late 1960’s and early
could shorten the timelines and fasten early launch of generic across the 1970’s, mechanism based models such as whole body PBPK models were
globe. The Biopharmaceutics Classification System (BCS) based bio­ evolved as existing compartmental models were unable to describe non-
waiver for solid oral dosage forms is one such approach aimed to fasten linear ADME processes. These models enabled simulation of clinical
product filings [1]. Since a biowaiver is applicable only for Class I & III outcomes, organ based predictions and detailed drug behavior under­
compounds, other approaches such as physiologically based pharma­ standing. Later, to understand sources of observed variability in plasma
cokinetic modeling (PBPK) or physiologically based biopharmaceutics drug concentrations, a new set of empirical models termed as population
modeling (PBBM) can be utilized for justifying pharmacokinetic study PK models were used to describe variability in the pharmacokinetic data
waivers through the model based pharmacokinetic predictions for all and identification of covariates and such models are called top-down

* Corresponding author.
E-mail address: [email protected] (T. Ahmed).

https://doi.org/10.1016/j.jddst.2022.103152
Received 1 October 2021; Received in revised form 5 January 2022; Accepted 31 January 2022
Available online 2 February 2022
1773-2247/© 2022 Elsevier B.V. All rights reserved.
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

Abbreviations IVIVE in vitro in vivo extrapolation

IVIVR in vitro in vivo relationship
ADME Absorption distribution metabolism excretion M&S modeling & simulation
AUC Area under curve MR modified release
BE bioequivalence OCAT ocular compartmental absorption model
CMA critical material attribute PBAM physiologically based Absorption model
CPP critical process parameter PBBM physiologically based biopharmaceutics modeling
CQA critical quality attribute PBPK physiologically based pharmacokinetic modeling
CRDS clinical relevant dissolution specification PD pharmacodynamics
DDI drug drug interaction PK pharmacokinetics
DDDPlus® Dose disintegration and dissolution plus PSD particle size distribution
DR delayed release P-PSD Product particle size distribution
ER extended release P-gp Permeability glycoprotein
FASSIF Fasted state simulated intestinal fluid PSA Parameter sensitivity analysis
FeSSIF Fed state simulated intestinal fluid QbD Quality by design
GIT Gastro intestinal tract QC Quality control
IR Immediate release SmPC Summary of product characteristics
IV intravenous SR sustained release
IVIVC in vitro in vivo correlation VBE Virtual Bioequivalence

models. Other models such as, bottom-up models were utilized to un­
derstand the system characteristics that might impact the plasma drug
concentrations and exposures. The prediction errors in such models
helped understand the underlying mechanism of ADME processes. The
PBPK/PBBM models explained in this publication are good examples of
a bottom-up approach that are considered valuable tools in drug dis­
covery, development and generic product development [2].
Drug discovery, development, drug-drug interaction, pediatric pre­
dictions are few areas wherein PBPK models have been used widely over
the last 10 years. These models not only minimize the preclinical and
clinical studies, but to a major extent impacts the cost & product
development timelines during the clinical development of a NCE
molecule [3]. PBPK modeling not only plays a vital role in NDA sub­
missions, it has multiple applications in generic industry as well. These
include but not limited to justifying change in dissolution specifications,
developing clinically relevant dissolution methods, setting up clinically
relevant dissolution specifications and biowaivers etc. Recently regula­ Fig. 1. Overview of applications of PBBK modeling in generic product
tory agencies such as US FDA and EMA have also released PBPK development.
modeling guidelines along with report format for regulatory sub­
missions, emphasizing further the importance of this emerging area in 2. Regulatory perspective of PBPK/PBBM modeling
generic drug development [4–6].
Literature review indicated couple of articles published in the area of In the period of 2008–2013, there were 112 PBPK applications
applications of PBPK modeling in new and generic drug development submitted to USFDA and in 2014 alone there were 45 PBPK submissions
[7–15]. However none of these articles provided details about PBPK to the USFDA [16]. These models were used for addressing clinical is­
applications for generic industry exclusively. In this context, the sues, evaluation of effects of intrinsic or extrinsic factors on drug phar­
following article is an attempt to focus PBPK applications in major areas macokinetics. Considering such remarkable rise in the inclusion of PBPK
of generic pharmaceutical industry namely, modified formulation models in dossiers, various workshops have been organized among
development, food effect assessment in oral dosage form, bio relevant regulatory agencies, academia and industries to share knowledge and
dissolution conditions and specifications, particle size impact towards best practices.
bioequivalence and biowaivers (Fig. 1) with emphasis on regulatory USFDA and EMA conducted specific workshops on PBPK modeling in
perspectives. The intent is also to bring harmonization across different 2014 and 2017 to share knowledge about dose selections and model
regulatory agencies on the utilization of PBPK modeling in generic in­ predictive performances [17,18]. EMA view on PBPK models and impact
dustry. Along with applications, best practices adopted in generic in­ on product development was reviewed by E Luzon et al. [19]. The same
dustry namely incorporating model parameters, dissolution inputs and publication also had identified the need to have EU regulatory guidance
key aspects to consider while validating the model are also discussed in exclusively for PBPK modeling and eventually was published in Jul 2019
detail. Along with applications, the authors also made an effort to [4]. Once a general understanding and consensus among PBPK models
mention regulatory decision outcome if it is appropriately mentioned in was developed, focus was shifted towards specific topics such as ab­
the publication. The authors also would like to clarify that even though sorption modeling. In 2019, a workshop was conducted on dissolution
the manuscript contains few examples from NCE’s (40 generic product and translational modeling strategies towards establishing an in vitro, in
examples and 6 new drug examples), the PBBM modeling concept used vivo link [20]. In this workshop, patient centric product development
in such examples can be applied to generic products as well. along with applications of PBPK modeling using case studies were dis­
cussed to support regulatory decision making, product labelling,
bio-predictive dissolution methods, model informed formulation

2
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

selection, predicting clinical performance, establishing BE safe space for reproducibility), physiological factors (GIT variability, motility, food
generics. The latest workshop in the area of PBPK modeling and simu­ effect, pH, pH, regional absorption) adds further complexities to predict
lations was conducted in 2020 wherein representatives from regulatory in vivo behavior. While the release rate can be optimized mathemati­
agencies discussed about various topics [21]. Definitions of terminol­ cally, the GIT variability can’t be described with simple mathematical
ogies such as PBAM, PBPK, PBBM, MAM were debated and PBBM was equations. Hence to overcome these limitations, modeling & simulation
considered as a suitable terminology to predict clinical impact of vari­ tools & IVIVC holds significance. By using such tools, a target dissolution
ations in formulation CMAs and CPPs (however PBBM is still emerging profile to achieve BE can be calculated using traditional deconvolution
and hence both PBPK and PBBM are used interchangeably in this approach. While such deconvolution can be performed by tools such as
article). Reasons for model rejections by agencies were attributed to Phoenix WinNonlin, physiological aspects can be integrated with Gas­
“model not fit for purpose”, “tested variants were not representative of troPlus & SimCyp to account for physiological aspects and to establish
the questions”, “not providing enough model verification against clin­ PK/PD relationship [25]. The case studies listed in Table 1 provides an
ical data”. Other concepts such as safe space, utilizing biopredictive overview of applications of such tools in MR formulation development.
dissolution data in PBBM models, dissolution specifications relaxation, The utility of GastroPlus mechanistic absorption model with PD
use of bio in-equivalence batch data to demonstrate model robustness extension comparing onset and duration of action with different release
were discussed. Additionally, challenges such as unavailability of profiles of Adinazolam and dissolution profiles impact on in vivo aspects
intravenous data to define disposition parameters, use of biomarkers in of Metoprolol was performed by Lukacova et al. [26]. For Adinazolam,
study design and model development were discussed in detail. the developed model was used to predict the PD profile of MR formu­
Eventually both USFDA and EMA came up with exclusive guidance’s lation and helped in designing new formulation with desired PD profile.
for PBPK guidelines [4–6] in order to encourage industry to include such For Metoprolol, release profile was optimized to get desired in vivo
models in submissions. Both guidance’s cover general overview about profiles. The utility of PBPK model to guide MR formulation develop­
modeling & simulation approaches and cases where it can be success­ ment of highly soluble BCS class I compound with emphasis on regional
fully applied. These guidance’s include factors to be considered while absorption was evaluated by Filippos et al. [32]. PBPK model was
developing models, development strategies, data requirements, valida­ developed in dogs and rabbits and was used to select minipigs as
tion and verification approaches were provided in detail. Highlight of appropriate species for further testing in order to simulate behavior in
these guidance is the application of PBPK model development that aid in humans. The utility of PBPK models to predict in vivo effect on change in
dissolution method development, clinically relevant dissolution speci­ dissolution specifications was evaluated by Jereb et al. [34]. PK pa­
fications setting, quality risk assessment, supporting post approval rameters were predicted for hypothetical formulation with 0% drug
changes, reducing number of clinical studies, establishing quality release in 2 h. Results indicated insignificant effect of hypothetical
specifications and biowaivers along with report format for submitting formulation and helped to justify change in dissolution specifications.
these models in dossier. Japan MHLW has also drafted Guideline for This validated PBPK model was proposed as an adequate alternative for
PBPK model analysis and analysis reports and subjected it to public traditional IVIVC or mechanistic IVIVC, when such IVIVC’s could not
consultation from September to October 2019 and will be finalized soon have been developed due to lack of sufficient data. Jones et al. published
[22]. There also has been other collaborative initiatives amongst Brazil, a case study wherein PBPK model was utilized to evaluate P-gp mediated
Japan, China where implementation of PBPK guidance’s in regulatory DDI model for SR formulation of Fesoterodine [35]. Validated model
submissions were extensively discussed [21]. It is anticipated that in indicated that predicted concentrations with SR formulation were below
near future, almost all regulatory agencies will have their own PBPK the IC50 for P-g and DDI study was waived off by the agency. Investi­
guidance’s, however better approach would be harmonization of indi­ gation of factors influencing absorption of class IV compound from MR
vidual guidance’s into a single one. The latest USFDA publication on formulation was evaluated by Otsuka et al. [28]. Stella based PBPK
PBPK modeling for NCE’s further emphasized the significance of PBPK model was validated and parametric sensitivity analysis indicated that
modeling in NDA submissions [23]. With this overview about regulatory absorption rate, gastric emptying, release rate can impact absorption.
understanding of PBPK modeling in generic product development, the Utility of Gastroplus absorption models to understand the possibility of
subsequent sections provide a detailed review on specific formulation mitigation of adverse clinical events with MR formulation design was
related applications of PBPK modeling. evaluated by Good et al. [30]. A validated Gastroplus absorption model
was used to identify key formulation variables to achieve lower Cmax
3. Applications of PBPK/PBBM modeling on generic product without impacting AUC. These case studies clearly explain the impor­
development tance and utility of PBPK models in MR formulation development and
additional case studies is provided in Table 1.
3.1. Modified release formulation development Although preclinical studies can provide rank ordering into human
behavior, full translation into humans is always challenging. Such lim­
Delayed release, controlled release and targeted release dosage forms itations can be addressed using PBPK modeling by incorporating human
are the various classifications under modified release system wherein physiologies and utilization of allometric scaling to scale preclinical data
the time and release of drug is altered compared to conventional for­ to humans. IVIVC level A correlation is the most accepted among
mulations [24]. Ideally zero order release is preferred to match with agencies however it is always challenging to develop such correlation as
elimination rate and to maintain constant systemic levels, however only it requires minimum three formulations with different release. Alter­
osmotic dosage forms provide such release. All other dosage forms such natively, PBPK models can be utilized in the absence of a level A cor­
as reservoir, matrix formulations offer first order release, bringing dif­ relation and to predict the behavior of an unknown formulation. IVIVC
ficulty to predict in vivo behavior. Apart from release, for­ correlation is not accepted with molecules exhibiting regional absorp­
mulation/manufacturing aspects (e.g. polymers, manufacturing process, tion or an absorption window. In such cases PBPK can be used as a tool to

3
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

Table 1
PBPK/PBBM applications in MR formulation development.
Molecule/Formulation BCS Purpose Modeling approach Modeling outcome
Class

Adinazolam, Metoprolol/ I Develop PK/PD for MR Gastroplus models developed: PD model Adinazolam – predicted PD of MR formulation Metoprolol –
MR [26] formulations for Adinazolam & mechanistic PK model predicted new MR formulations PK & used to select in vitro
for Metoroprolol conditions to match in vivo
Sildenafil/CR [27] I Development of CR Gastroplus model developed using Optimum formulation was selected to achieve desired PK
formulations dissolution & validated against reference parameters
Furosemide/MR [28] IV Investigation of key factors Stella PBPK model developed & validated PSA of MR formulations indicated absorption rate, gastric
influencing absorption with IR & MR formulations emptying, release rate can impact absorption
Guanfacine/ER [29] II Dose adjustments in Simcyp PBPK model developed & Dose recommendations provided in children and adults in
presence of CYP3A4 validated using clinical PK with CYP3A4 presence of CYP3A4 inhibitors and inducers
inhibitors & inducers inhibitors and inducers
BMS-914392/MR [30] II MR formulation Gastroplus model developed & validated Three identified prototypes demonstrated low Cmax with
development to mitigate to identify suitable formulation factors equivalent exposures.
adverse events
MK-0491/MR (matrix, III Compare traditional IVIVC IVIVC developed with three batches with Both IVIVC & PBPK models resulted in prediction errors less
multiparticulate pellets) & PBPK modeling different release rates with optimized than 10–15% with similar level A correlation
[31] scaling factors for PBPK model.
Gaboxadol/CR [32] I Regional absorption Dog PBPK model developed to predict Minipig model predicted the observed plasma conc data &
prediction in lower intestine regional absorption. Minipig model resulted in preclinical IVIVC development
developed to predict CR product PK
Zolpidem/MR [33] I Understand negative food Simcyp & Gastroplus models developed & Simcyp & Gastroplus indicated average fold error of <1.5.
effect absorption assessed using Qgut model Release controlled by gastric emptying & incomplete release
with liver & gut extraction in fed may be reason for negative food effect.
Drug substance/MR [34] I BE prediction of MR capsule PBPK model & IVIVR developed to predict IVIVR & PBPK model predicted insignificant effect of
PK for hypothetical formulation (0% hypothetical formulation & supported change in dissolution
release in 2 h) specifications
Fesoterodine/SR [35] I Regulatory support for P- Gastroplus model developed & validated Predicted plasma conc were <IC50 for P-glycoprotein & DDI
glycoprotein mediated DDI for drug, metabolite and SR formulations study waiver obtained
PK predicted.

guide MR formulation development. However obtaining such PE’s especially from PBPK models can be
Development of PBPK model for MR formulations requires multiple challenging because of presence of physiological variability along with
model inputs as well as validation. Initially, literature plasma concen­ formulation variability and hence appropriate justifications can be made
tration time profiles for reference product can be deconvoluted to obtain if the PE’s are not met. Additionally due to the absence of number of
in vivo input rate. Development of in vitro dissolution method devel­ validation steps for PBPK model, as much as validation can be performed
opment can be initiated to target the in vivo input rate. Once such with literature formulations (both IR and MR) or different dosage forms
dissolution method is developed, it can be validated against literature (solution, suspension) along with in-house test formulations (minimum
information followed by understanding regional absorption and impact 3 formulations) to get sufficient insight into model acceptability. Apart
of different dissolution profiles on in vivo absorption. While using liter­ from dissolution, other physical aspects such as swelling index, polymer
ature data, it should be kept in mind that peer reviewed articles may not erosion behavior etc. couldn’t be captured in the model. If future ver­
contain necessary details (e.g. formulation, number of subjects, study sions of these software tools can capture such aspects, prediction power
design, data availability) to make a judgment regarding the quality of of such tools can be enhanced tremendously. Despite these challenges, a
the data. Hence the authors suggest to use literature data to understand well-developed, validated and verified models are definitely an asset to
general predictive behavior of the model. Alternatively, when necessary MR formulation development.
details are present, the peer reviewed article may be used for complete
model validation. During initial model development, the generated in
vitro dissolution data can be provided as input to predict plasma con­ 3.2. Justifying dissolution specifications/building safe space
centration profiles. Gastroplus considers dissolution data input into such
dosage forms as in vivo dissolution input rather than in vitro dissolution Dissolution or drug product quality attribute safe space is defined as
input. Hence the IVIVR developed can be utilized as a tool to convert in a region in which formulation batches are anticipated to be bio­
vitro dissolution to in vivo dissolution once it is developed. Additionally, equivalent to each other. This terminology has gained impetus with
dissolution can also be given as input using Weibull fitting in softwares CRDS (clinically relevant dissolution specifications), which is used as
such as Gastroplus [36]. Like any other PBPK model, the development of specifications to demonstrate bioequivalence. These specifications are
MR formulation model also requires detailed understanding about input established to identify and accept only batches that are intended to meet
parameters, these aspects are discussed in later sections of this manu­ bioequivalence with the reference product and dissolution profiles
script. A detailed flow chart for PBPK model development for MR for­ similarity testing is not required within this region. It is advised to
mulations is provided in Fig. 2. establish such specifications early in the formulation development so as
When it comes to validation of PBPK models, criteria for prediction to select suitable batches for the pivotal clinical studies for achieving BE
errors (PE) are not clearly defined in guidelines. In absence of such [38] using various approaches. In one approach, it is advised to have
criteria, the PE requirements for IVIVC (10–20% based on internal/ formulations with different dissolution release rates achieved by delib­
external prediction) can be considered as acceptable limits [37]. erate variations in formulation to demonstrate in vivo difference. In this
case, it is advised to include any failed BE batch so as to validate the

4
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

Fig. 2. Modified release formulation development with aid of PBPK/PBBM modeling: model development, model validation and model applications.

5
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

Fig. 3. Concept of safe space (Dissolution, API PSD, CPP’s/CMA’s).

dissolution safe space. In another approach, it is advised to have hypo­ molecule to establish CRDS [40]. It was concluded that if >85% release
thetical profiles with similar shape as that of the pilot formulation and is achieved in 45 min, no impact on PK/PD is expected. For BCS II weak
study the impact of these formulations on the in vivo exposure in order to acid, Konstantinidis et al. used virtual BE to establish clinically relevant
establish safe space. Alternatively, CRDS can also be defined as a limit dissolution specifications [41]. Using several dissolution profiles, virtual
beyond which the formulations will be bio-inequivalent to the reference BE studies were performed and results indicated that if >85% release
batch. When the quality control (QC) method is also made as a achieved in 90 min, no impact on BE is expected there by indicating
bio-indicative method, it becomes a powerful tool to assess formulation possibility of BCS biowaiver extension beyond Class I and III Establish­
batches from both quality and BE perspective. In all of these cases PBPK ing dissolution based bioequivalence safe space for highly soluble drug
modeling comes as an effective aid to establish CRDS or to define compound Oseltamivir with the help of PBPK model was performed by
dissolution safe space and concept of safe space is represented in Fig. 3 Miao et al. [42] in both adults and pediatric patients by deriving
and examples from literature is listed in Table 2. dissolution specifications in these populations. Pepin et al. used PBPK
Establishment of CRDS for BCS IV compound using PBPK model was modeling to assess impact of in vitro dissolution on in vivo performance
performed by Kato et al. [39] to distinguish BE and non-BE batches. of Class II drug. QC media dissolution was given as input into the model
Parasio et al. developed a model consisting of PK and PD for BCS I and results defined safe space of Q = 80% in 30 min [43] to justify

Table 2
PBPK/PBBM applications in justifying dissolution specifications/establishing dissolution bioequivalence safe space.
Molecule/Formulation BCS Purpose Modeling approach Modeling outcome
Class

Model drug substance/ IV Establish CRDS Gastroplus PBPK model developed with three Predicted profiles matched with observed profiles for all
IR [39] BE & non-BE batches & validated formulations. CRDS was established.
Zolpidem/IR [40] I Establish clinically Simcyp model developed using dissolution, PK If >85% drug released in 45 min, no impact on PK/PD
relevant dissolution safe & PD data & CRDS established expected.
space
Naproxen/IR [41] II Set clinically relevant Simcyp model developed with solubility, If >85% drug released in 90 min, no impact on BE expected,
specifications dissolution data & validated. Several in vitro extension of biowaiver for BCS I, III is possible
dissolutions were used to predict PK.
Oseltamivir/IR [42] I or III Establish dissolution BE PBPK model for drug & metabolite developed Product with 10% slower dissolution than reference could be
safe space various & validated. Virtual BE performed with BE in adults. In children, stringent boundaries were predicted
populations theoretical dissolution profiles (6% & 4% slower in adolescents & neonates)
Lesinurad/IR [43] II Assess impact of Gastroplus model developed & validated using Batches with Q = 80% in 30 min will be BE. Slower profiles
dissolution on in vivo QC dissolution,against literature than specification also met BE, indicating adequacy of current
specifications
Dextromethorphan/IR II Irrelevance of dissolution Gastroplus PBPK model with lysosomal Lysosomal trapping is critical to release drug to systemic
[44] specifications trapping to study the influence of dissolution circulation & hence dissolution is not relevant for BE.
Piroxicam/IR [45] II Prediction of PK in dogs & Gastroplus PBPK model developed in dogs in Dissolution space is much wider than anticipated. Model
scaling to humans fasting and fed state. extrapolated to humans for biowaiver of IR formulation
DRL ABC,/ER [46] II Justification of dissolution Gastroplus PBBM model developed using Virtual BE performed with various dissolution specifications
specifications change literature data & validated, IVIVR developed and were found to be BE with Reference.
with clinical study.

6
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

adequacy of established specifications. Justifying irrelevance of disso­ such data is obtained from in vitro experiments, combining this data
lution specifications for BCS II molecule dextromethorphan was along with physiology forms an attractive choice to predict fed BE.
explained by Bolger et al. [44]. This molecule undergoes lysosomal Literature case studies wherein PBPK modeling was used to predict food
trapping in portal vein and in plasma thus indicating that dissolution is effect is listed in Table 3.
not a critical factor to achieve BE. Rebeka et al. used PBPK model for generic formulations of both
These case studies in Table 2 indicate the significance of CRDS and crystalline and amorphous API to predict fed BE outcome [48]. Virtual
dissolution BE safe space so as to reduce human BE studies and to imbed simulations predicted fed BE outcome and differences in crystalline and
quality into drug product. Although it would be good to have bio in- amorphous forms was captured by the model. Rebeka et al. published
equivalent batch for defining safe space, in absence of such batch safe another case study wherein PBPK models were developed for six mole­
space can be established using profiles with varying rates. Potential cules across BCS I, II and III [49]. This study showed that PBPK models
aspects such as PD can be embedded into the model to further under­ were able to capture pass (5) and failed (1) BE studies successfully and
stand the impact on efficacy. Defining CRDS will broaden the scope of possibility of fed BE studies waiver was discussed. Food effect prediction
BCS III very rapid dissolution criteria (>85% in 15 min) by pushing of different formulations of BCS II drug Danirixin was developed by
borderline BCS III molecules into BCS I, there by facilitating many bio- Lloyd et al. [50]. Compendial and TIM-1 dissolution models were
exemptions. While dissolution is generally used for defining safe incorporated into the PBPK model and data from clinical studies was
space, it can also be defined by other parameters such as CPP’s and used to validate this model. While TIM-1 data indicated reduced
CMA’s by indirectly inputting them into the software using dissolution. bio-accessibility in fed state, same was reflected in the model as well and
While defining dissolution BE safe space is challenging for highly vari­ negative food effect was predicted. Tistaert et al. developed fed PBPK
able molecules, approaches such as parametric sensitivity analysis (PSA) models for six API’s across BCS I and II [51]. This study demonstrated
helps to identify the factor governing in vivo variability and dissolution. the utility of PBPK models for simulating new doses, formulations, API
Despite all of these limitations, defining dissolution BE safe space and forms in the view of clinical food effect study. Parrott et al. developed
CRDS provides more flexibility in terms of regulatory and PBPK model to explore impact of food and gastric pH changes on PK for
manufacturing aspects, as absence of such space will lead to specifica­ BCS II molecule Erectinib [52]. The model predicted negative food effect
tions solely based on the pivotal BE test batch. and minor pH dependent DDI’s from clinical perspective and was sug­
gested to have importance in answering potential regulatory questions.
3.3. Food effect and fed BE studies These case studies indicate the importance and utility of PBPK
models in predicting bioequivalence under fed conditions [54]. Model
Food can have impact on both physiology as well as dosage form inputs such as aqueous solubility, particle size, in vitro dissolution,
performance. Because of altered PK in presence of food, it is recom­ biorelevant solubilities (e.g. FaSSIF, FeSSIF) coupled with physiology
mended to conduct bioequivalence studies of such API’s in fed condition occupies utmost importance for predicting food effect for both crystal­
along with fasting condition. However mimicking food effect in in vitro line and amorphous API’s. Biorelevant permeability values can be
condition is not easy because of involvement of many physiological al­ incorporated into the model for predicting negative food effect due to
terations in presence of food such as secretion of bile salts, altered pH complexation [47]. Additionally, negative food effect due to complex­
change in gastric motility, viscosity and possibility of complexation with ation with multivalent cations present in food (e.g. milk), can also be
food components. predicted by integration of relevant in vitro dissolution tests into PBPK
Literature review indicated that PBPK/PBBM modeling has been model. Along with impact on PK parameters (Cmax, AUC), delayed Tmax
successfully utilized for prediction of food effect of various dosage due to food ingestion can be predicted by making necessary alterations
forms/BCS classes by accounting physiological changes in presence of into splanchnic blood flow with high first pass extraction [51]. Addi­
food. The data from in vitro experiments can be integrated into physi­ tionally, for molecules exhibiting high PK variability (BCS II with pH
ological models for predicting food effect. For this purpose, biorelevant dependent solubility or BCS IV with poor solubility across GI pH) food
media’s with suitable dissolution method selection (e.g. hydrodynamics, effect predictions can be performed with virtual BE simulations. Mole­
dissolution apparatus, volume’s) to reflect in vivo conditions occupies cules with low permeability (BCS III and IV) can have profound effect of
significant importance. Additionally, negative food effect can be esti­ transporters and incorporating transporters kinetics may increase pre­
mated by using apparatus that combines dissolution and permeation in diction accuracies of the model. For these scenario’s, literature infor­
case of drugs that forms complexes with food components [47]. Once mation is limited and hence more understanding is required in this area.

Table 3
PBPK/PBBM applications for determining food effect.
Molecule/ BCS Purpose Modeling approach Modeling outcome
Formulation Class

Generic/amorphous II Predict food effect & fed Crystalline & amorphous forms models developed for Virtual simulations predicted BE in fed condition &
[48] BE outcome fasting & fed and <10% PE was obtained. differences in crystalline and amorphous forms was
captured.
Various IR (six I, II, Predict BE in fed Gastroplus models developed and population Models predicted passing (5) & failed (1) BE studies.
molecules) [49] III condition simulations performed in fasting & fed conditions Possibility of waiving off fed BE studies based BCS class
assessed
Danirixin/IR [50] II Explore food effect using Gastroplus model developed with compendial & TIM-1 indicated reduced accecssibility in fed compared to
biorelevant dissolutions TIM-1 dissolutions & validated with clinical study fasting. Integration of TIM-1 data into PBPK model
outcome of various formulations. predicted negative food effect.
Various IR products I, II Propose PBPK workflow Gastroplus models developed in fasting & fed Demonstrated the utility of PBPK models for simulating
(four molecules) for food effect prediction condition and validated with clinical data. new doses, formulations, API forms for clinical food effect
[51] study
Entrectinib/IR [52] II Explore impact of food & Gastroplus model developed using in vitro, in silico & Predicted negligible food effect and minor pH dependent
gastric pH changes on PK dissolution data & validated with clinical study data. DDI’s for marketed formulation. Suggested to justify
regulatory queries with this approach.
Venetoclax/IR [53] IV Develop PBPK model for Simcyp model developed & validated against fasted, Model captured food effect with low-fat diet, but didn’t
food effect fed data. capture with high-fat diet, indicating the possibility of
additional mechanism.

7
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

In addition to evaluating impact of standard FDA breakfast/meal, models. As for reference formulation, input API PSD may not be avail­
influence of different meal types (e.g. low fat, low calorie) can be pre­ able, alternative approaches such as determined API PSD in finished
dicted by making appropriate changes in physiology in terms of transit product can be evaluated. Even though all PBPK models assumes that
times as suggested by Sutton et al. [55]. PSA also can be performed as a PSD can effect dissolution rate, physical aspects such as aggregation,
part of model verification for identified parameters that can have pro­ agglomeration, wettability are not considered into simulation. If future
found effect on the BE outcome in presence of food. However utilization versions of software’s are accounted for such physical aspects, it will
of such alternative meal compositions is very rate in the generic product lead to accurate estimation of PSD impact. Another challenge foreseen in
development for evaluating bioequivalence against reference formula­ terms of particle size is for weakly basic molecules that gets solubilized
tion as high fat and high calorie meal represents worst case. in acidic environment in stomach and later gets precipitated at basic pH
in intestinal environment. Depending on the rate of precipitation,
nucleation, precipitated API can form different PSD’s and if this can be
3.4. Particle size impact and PSD safe space accounted in current PBPK modeling tools, that can further increase
prediction accuracy. Despite these current limitations, an appropriately
Understanding and selecting suitable particle size as a CMA is a designed, validated and verified PBPK models can act as aid to justify
critical factor for attaining in vivo BE for especially for low soluble regulatory questions for PSD specifications, and helps to build safe space
molecules such as BCS II and IV as it governs the in vivo dissolution and for API PSD thereby acting as a valuable tool for QbD based formulation
absorption behavior. While during the formulation development, API development.
PSD can be optimized by experiments followed by dissolution evalua­
tion, Gastroplus based simulations provide in vivo impact of API PSD by
mechanistically integrating it in the absorption behavior. Additionally, 3.5. Biowaivers
PSA can be performed to understand impact of varying PSD on the PK
parameters and helps to select suitable PSD for the formulation opti­ Biowaivers or bio-exemption provide excellent opportunity to waive
mization. Utility of such tools for predicting impact of PSD is presented off clinical studies on the basis of meeting certain specific characteris­
in Table 4. Utilization of computer simulations for particle size specifi­ tics. There are different types of biowaivers such as BCS based, lower or
cations selection has been described by Willmann et al. [56]. Gastroplus upper strength biowaivers. While BCS biowaivers are granted for spe­
model for Class II molecule Cilastazol was developed and suspension cific set of molecules, biowaivers within drug product, such as for lower
exposures with different PSD were simulated. It was concluded that such strengths is possible based on the similarity in dissolution characteris­
models can be useful later during the development. Pepin et al. used tics, composition proportionality and pharmacokinetics linearity &
Gastroplus modeling to justify particle size specification for BCS II based on BE of higher strength between RLD and Test formulations.
molecule Lesinurad [43]. The formulation developed with selected PSD Biowaiver can also be obtained using conventional IVIVR wherein,
was found to be BE and appropriateness of this PSD was justified. Cris­ middle strength biowaivers granted based on BE between lower &
tofoletti et al. applied simultaneous diffusion and chemical reaction higher strengths. IVIVC is another approach for requesting biowaiver
model to estimate P-PSD (product specific PSD) for BCS II drug and IVIVC guidelines from USFDA provides detailed cases wherein
Ibuprofen to develop IVIVE-PBPK model [57]. This approach was used IVIVC is useful for obtaining biowaiver [37]. Recently a safe space
to waive off BE studies and to support major approval changes. Simi­ approach for biowaivers using PBPK modeling is being evolved when
larly, further utility of P-PSD was demonstrated by Pepin et al. using above mentioned approaches are not feasible. Literature examples
Acalabrutinib as a model drug [58]. The developed model was validated wherein PBPK approach has been used for biowaiver applications is
with different PSD and was used to develop a safe space for API PSD. summarized in Table 5 and conceptually depicted in Fig. 4.
Mitra et al. used effective PSD (when dissolution data can’t be directly Zhang et al. developed PBPK modeling for Isosorbide mononitrate
incorporated into the model) as a strategy to incorporate particle size BCS I compound to evaluate if rapid release is crucial for achieving BE
into Gastroplus model to predict BE of immediate release formulation of [60]. It was concluded that different release profiles had same in vivo
BCS II compound. The developed model was able to predict the BE be­ performance and batch at specifications (75% at 45 min) has clinical
tween two formulations. relevance. Mitra et al. evaluated utility of PBPK modeling to evaluate
The above mentioned examples clearly state utility of PBPK performance of two different formulations of Etoricoxib and possibility
modeling to identify suitable particle size for achieving BE with refer­ of biowaiver application [61]. The multimedia dissolution in 0.01 N HCl
ence formulations. Tools such as product PSD (P-PSD) and effective PSD indicated dissolution profiles similarity whereas pH 4.5 and 6.8 indi­
are novel approaches to incorporate PSD into such PBPK mechanistic cated that profiles are not similar. Simulations were carried out with

Table 4
PBPK/PBBM applications for understanding particle size impact.
Molecule/ BCS Purpose Modeling approach Modeling outcome
Formulation Class

Cilastazol/IR [56] II Predict particle size & Gastroplus PBPK models developed & PK of different PSD Influence of PSD on PK in fasting & fed was
impact on absorption suspensions simulated. predicted, usability of model for formulation
development demonstrated
Lesinurad/IR [43] II Justify particle size Gastroplus model developed with QC dissolutions & validation Simulations with specification at limit of
specification against literature data. specification indicated BE, confirming adequacy of
specification.
Ibuprofen/IR [57] II Usability of novel Novel P-PSD (product specific particle size distribution) used Novel approach suggested to support development
simulation tools for developing an IVIVE-PBPK model of BCS Class II or similar drugs
Acalabrutinib/IR II Develop PBPK model & P-PSD values were integrated into Gastroplus model & clinical Model used to define safe space for particle size
[58] use of P-PSD PK under normal, PPI administration were predicted
Test compound/IR IIb Predict BE & use of Gastroplus model developed & sensitivity analysis done. Z-factor based model used to predict BE between
[59] effective PSD Effective PSD & population simulations were conducted to test & reference
assess BE
Various IR (six I, II, III Predict BE in fed state Gastroplus models developed with effective PSD (with help of Effective PSD based models successfully predicted
molecules) [49] DDD+, in absence of disso data), population simulations passing (5) & failed (1) BE studies.
performed in fasting & fed.

8
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

Table 5
PBPK/PBBM applications for biowaivers.
Molecule/Formulation BCS Purpose Modeling approach Modeling outcome
Class

Isosorbide mononitrate/IR I Evaluate if rapid release is crucial PBPK model developed with clinical & non- Different release profiles met BE, batch at spec (Q
[60] for BE clinical data, virtual BE simulations performed 75%, 45 min) had clinical relevance
with various dissolution profiles
Etoricoxib/IR [61] II Explore biowaiver possibility of Gastroplus model developed & validated with Validated model indicated BE between
batches manufactured at two sites clinical data with (different dose, PPI formulations & can be used to demonstrate BE of
administration). Population simulations two different sites with 0.01 HCl as biopredictive
performed with 0.01 N HCl. media.
Etoricoxib/IR [62] II Justify biowaiver for class II Gastroplus model developed using transfer IVIVC established with Gastroplus & confirmed
solution & tablet model dissolution data & simulations were similarity of oral solution & tablet. Simulations
performed. can support biowaiver
Metronidazole/IR [63] I Predict BE using dissolution data Gastroplus PBPK model developed by BE achieved with different profiles (85% in 15
incorporating in vitro dissolution into the min–120 min), justifying possibility of extension
model. of BCS I specifications.
Propranolol, Metoprolol, I, III Predict oral absorption of Class I, Gastroplus model developed & virtual trials BCS I drugs met BE with up to 85% in 120 min.
Atenolol, Amoxicillin, III drugs with possible waiver were performed to predict drug absorption. Class III drugs met BE with 85% up to 45 or 60 min
Cimetidine [64] extension for Class III drugs dissolution, indicating extension possibility.
Nadolol, Acebutolol, III Assess site-specific absorption, Gastroplus simulations performed for 3 P-gp Site specific absorption & low Fa magnified
Atenolol [65] gastric emptying rate on substrates drugs modeled with site specific impact of dissolution on PK. Dissolution criteria
biowaiver of class III drugs absorption. can be relaxed (85% in 30 min) for highly
bioavailable drugs.
Amlodipine/IR [66] I Support similarity & dissimilarity f1 & f2 were calculated for generic products, Matching f1 & f2 factors indicated BE in
factors for lower strength Gastroplus model used to predict plasma conc simulations, confirming interchangeability
biowaivers profiles between brands

Carbamazepine/IR [67] II Use of gastrointestinal simulation Gastroplus model developed & validated against Different dissolution profiles indicated same PK,
technology & IVIVC for biowaiver clinical data. Plasma conc profiles generated provided rationale for Class II biowaivers, wider
extension with various dissolutions dissolution specifications

Fig. 4. PBPK/PBBM modeling utility to support biowaivers.

0.01 N HCl, model verification was performed against clinical study and [63]. PBPK model was developed by incorporating different dissolution
concluded that simulations can support biowaiver despite dissolution profiles and was concluded that despite different dissolution profiles, BE
criteria not being met. Okumu et al. performed PBPK simulations for was achieved. Evaluating very rapid dissolution criteria (>85% in 15
Etoricoxib to justify biowaiver for this BCS II compound [62]. Gastroplus min) and rapid dissolution criteria (>85% in 30 min) of BCS III and I
based IVIVC model was developed and validated against clinical data. It drugs was performed by Tsume et al. [64]. Gastroplus PBPK models were
was concluded that similar behavior of oral solution and tablet was developed for BCS I and III drugs indicated that BCS I drugs met BE up to
observed and this methodology can be used for supporting biowaivers. 120 min and III drugs up to 30 or 45 min, indicating possibility of
Utility of PBPK modeling to support biowaiver for BCS I API Metroni­ extension of dissolution specification criteria for BCS III and I drugs.
dazole with different dissolution profiles was evaluated by Zhang et al. Similar case for BCS III drugs was evaluated by Sun et al. and simulations

9
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

indicated that dissolution criteria can be extended up to 85% in 30 min models that considers these unique features and offers simulation of
[65]. variety of ocular formulations. These software tools (e.g. OCAT in Gas­
PBPK simulations can be used to support biowaiver in cases of troplus) support prediction of drug concentration in various eye tissues
dissolution dissimilarity between lower or higher strengths, f2 mismatch such as cornea, conjunctiva, aqueous humor and systemic concentra­
and challenges of IVIVC development due to lack of data. This meth­ tions along with evaluation of critical formulation attributes such as
odology can also extend current dissolution specifications of rapid and viscosity, particle size on the in vivo disposition [68]. Recently, regula­
very rapid dissolution criteria for BCS I and III molecules to enable more tory agencies such as USFDA and EMA had specific grants released to
waivers. Additionally, PBPK modeling can be utilized to support bio­ these companies for developing and validating ocular absorption models
waiver of BCS II weakly basic molecules (e.g. tyrosine kinase inhibitors) so as to support new and generic formulation development. Examples of
with high solubility in acid and precipitation potential in basic intestinal utility of these tools in ocular modeling is provided in Table 6. Merdy
condition. The simulations can be performed to demonstrate absence of et al. developed PBPK model for dexamethasone suspension to under­
in vivo impact due to low dissolution in intestinal conditions. With these stand impact of formulation variables on ophthalmic bioavailability in
futuristic thoughts, authors believe that using PBPK approaches, more rabbits [69]. PSA indicated interplay between formulation variables and
and more biowaivers can be granted by minimizing human experiments. ocular clearance that determines ocular availability and impact of PSD
and viscosity were discussed in detail. Merdy et al. developed OCAT
4. PBPK in other routes of administration: ocular, pulmonary, model to characterize ocular pharmacokinetics of Fluorometholone
transdermal, parenteral ointment in rabbits. The model was utilized to predict aqueous humor
concentrations, impact of surface area, application time, release area on
In addition to oral administration, applying modeling tools for non- drug passage was determined [70].
oral routes (e.g. ocular, pulmonary, transdermal and parenteral) is also Conducting ocular PK in humans and determining aqueous humor
gaining impetus recently. Formulation development of these advanced concentrations possess challenges from cost perspective and bio­
routes is considered as complex owing to complex interplay between analytical perspective due to low volume available for quantification. In
formulation factors and physiological factors. The physiological aspects such cases, these tools helps to predict behavior of formulations by
in these route of administration are comparatively not well understood reducing the number of studies to facilitate regulatory submissions.
and hence possessing challenges in terms of in vivo prediction. Because While more research is required in this area, a little acquaintance can be
of these concerns, mostly the product specific guidance’s (PSG) of these obtained from existing literature to develop OCAT models.
products requires demonstration of composition similarity (Q1/Q2) and
CQA’s similarity (Q3) between test and reference formulations to ensure
4.2. Pulmonary route of administration
similar in vivo exposures. In absence of these, utilization of simulation
tools will definitely help to understand formulation and physiological
While Pulmonary administration offers advantages in terms of local
aspects that governs in vivo behavior. A brief details about evolution of
delivery, minimal side effects, it also possess challenges in terms of
PBPK tools in these areas is provided below.
localized delivery, loss of dosage form during the administration pro­
cess, complex physiology coupled with different formulation types.
4.1. Ocular route of administration Though the advantage for pulmonary delivery is an attracting factor
over the last few decades, understanding of PK after pulmonary
Ocular route of administration offers uniqueness in terms of patient administration is significantly low as compared to oral dosage forms
demographics, limited space for local delivery, two distinct elimination [76]. Modeling and simulation tools from Gastroplus offers PCAT
process (eye drainage, systemic elimination), anatomy of ocular tissue. module with a possibility of simulation of variety of nasal and pulmo­
Simulation tools such as Gastroplus has inbuilt ocular physiological nary dosage forms by predicting local tissue concentrations along with

Table 6
PBPK/PBBM applications in other routes of administration (Ocular, Pulmonary, Transdermal, Parenteral).
Molecule/Formulation Route of Purpose Modeling approach Modeling outcome
administration

Dexamethasone/ Ocular Assess impact of formulation OCAT model developed with Gastroplus, Interplay between formulation, ocular
Suspension [69] properties on ophthalmic non-linear PK in ocular tissues, impact of clearance & impact on BA were evaluated.
bioavailability PSD, viscosity was evaluated. Influence of PSD and viscosity were
discussed
Fluorometholone/ Ocular Characterize ocular Aqueous humor conc were predicted with Application time impacted aqueous humor
Ointment [70] pharmacokinetics using OCAT OCAT model. Surface area, application time, conc, surface area & release rate influenced
model release rate on passage of drug was conc and exposure
evaluated.
Two API’s/Inhaled drug Pulmonary Simulate PK of inhaled PCAT model developed by fitting total lung PCAT model described systemic exposure of
product [71] combination product disposition, dissolution of reference, IVIVC two API’s. Dissolution based IVIVC under
demonstrated. predicted exposure due to systemic
variability
Hydroxychloroquine/ Pulmonary Develop PBPK model for Gastroplus nasal pulmonary module used to 25 mg BID inhalation dosing resulted in
Nebulized [72] nebulized formulation determine dosing regimen alveolar levels of 7 μM, predicted conc up to
5 days.
Diclofenac/Topical gel Transdermal Transdermal model Certara Simulator MechDermA model was The model developed, validated and agency
[73] development & study waiver developed as alternative for Q1, Q2 & Q3 accepted, waived off clinical study
formulation for waiver of clinical end point
study.
Rifampicin/Vesicular Transdermal In silico model development to Gastroplus model developed & used to Model predicted effect of various
elastic liposomes [74] support formulation investigate effect of physchem, formulation parameters & aided in formulation
development properties. development
Olanzapine, Huperzine, Parenteral IVIVC for long acting Level A IVIVC developed with Gastroplus. PK Significance of triple Weibull function, role
Ornitide/LAI’s [75] injectables & CR models integrated to develop IVIVC of optimization function, & interpolated
data utility in IVIVC were established

10
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

systemic concentrations. Examples of utility of M&S tools in pulmonary (LAI) for liposomes and peptides are considered as complex formula­
administration is provided in Table 6. tions to predict in vivo behavior. Additionally LAI’s comes up with
Lukacova et al. simulated absorption, distribution, PK of two inhaled complexity in terms of polymer characteristics, drug delivery systems,
API’s in a combination product [71]. Pulmonary (PCAT) Gastroplus release mechanisms, that gets altered in vitro and in vivo environment
model was developed by fitting total lung disposition, dissolution of [79]. Modeling and simulation tools also comes handy to predict the PK
reference and IVIVC was created. The PCAT/PBPK model described behavior of such formulations.
systemic exposure of two API’s and IVIVC under predicted exposure due A case study of simulating behavior of LAI’s using Gastroplus is
to systemic variability. Idkaidek et al. developed PBPK model for presented in Table 6. Shahraz et al. developed Gastroplus IVIVC models
hydroxychloroquine nebulization for Covid-19 development. Gastroplus for Olanzapine, Huperzine, Ornitide LAI’s integrating controlled release
nasal pulmonary module developed using physiological, drug disposi­ model and PK module. Results indicated significance of triple Weibull
tion, PK parameters indicated that 25 mg bid resulted in alveolar levels function, role of optimization function, and interpolated data utility in
of 7 μM [72]. successful IVIVC [75]. Currently only few case studies are available and
Using such pulmonary PBPK models, formulation variables such as much more work is required in this area to predict in vivo performance of
particle size can be optimized in inhaled suspensions. Data such as such complex formulations.
aerodynamic particle size distribution (APSD) from lung deposition Overall, as mentioned above the non-oral route of administration
models can also be integrated into these models to enhance the pre­ comes up with their own complexities in terms of physiology and for­
diction accuracy. However other critical aspects such as spray pattern, mulations. However, in silico tools stated above are less expensive, faster
plume geometry are not currently included as model inputs hence and safer compared to conduct of clinical study and represent an
trigger further optimization of these modules. Like ocular model, much important advancement for both innovator and generic drug companies.
work is required in this area however available tools can already be used
to guide formulation development and to predict in vivo behavior. 5. Conclusions and future perspective

4.3. Transdermal route of administration PBPK/PBBM modeling and simulation has evolved significantly over
the past two decades for applications in new drug and generic product
Transdermal delivery route is considered as an attractive approach development. Currently, the utility of modeling tools is being recognized
that offers extended release, mitigated dose, dose associated undesirable in generic industry to successfully develop formulations and to answer
effects, hepatic avoidance and relief from other gastrointestinal distur­ regulatory queries. PBPK/PBBM models can have multiple applications
bances, non-invasiveness and high patient compliance [77]. Modeling in terms of MR formulation development, establishing PSD specifica­
softwares such as Gastroplus and Simcyp can be used to predict the tions, dissolution safe space & CRDS, biowaivers and possibility of
behavior of formulation administered via transdermal route by using in extension of biowaivers for BCS II (weak bases). Additionally, PBPK
vitro drug release and permeation data. Although such PBPK models are modeling in non-oral routes of administration such as ocular, pulmo­
at a very preliminary stage, Certara with the support and grants from nary, transdermal and parenteral administration are also getting
regulatory bodies such as USFDA, developed a transdermal model for impetus and in future we will be able to witness more and more appli­
complex dermal generic product Diclofenac gel that requires BE clinical cations in these areas. More emphasis can also be made in terms of
end point study to demonstrate BE. PBPK modeling based virtual BE regulatory applications of the developed models in terms of mechanistic
trials [73,78] was used as an alternative for Q1/Q2/Q3 formulation and explanation of the model, validation with external clinical data and
waived an in vivo comparative clinical endpoint study. This approval internal data with different formulation variants/doses, defining pre­
was based on totality of evidence including qualitative & quantitative diction errors for PBPK simulations, utility of two media’s (biopredictive
sameness, physical & structural similarity to the reference product and and QC) in model development etc. These aspects were considered as
an in vivo PK end point BE study. This could be considered as the first and certain constraints, the authors believe that as PBPK/PBBM models will
only application where PBPK was accepted and approved for BE by be utilized for more and more products, and will further evolve to
waiving in vivo studies. The MechDerm model used in these simulations overcome these limitations. In this regard, it is vital for generic in­
is similar to oral absorption model and accounts for inclusion of API, dustries, regulatory agencies and academia to collaborate with each
formulation, physiology and environmental parameters for building other to sharing knowledge through workshops/meetings to disseminate
model for various formulations namely gel, emulsions, patches, sus­ the knowledge and to address limitations. While USFDA and EMA are
pensions and pastes and used to simulate drug partitioning and ab­ main regulatory agencies, who conducted and organized the workshops,
sorption through hair follicular pathway. Along with Simcyp, Gastroplus recently other agencies from developing markets such as Brazil, Japan,
also offers dermal module (TCAT) representing various skin tissues such China have also participated. At present the PBPK/PBBM guidance’s are
as stratum corneum, viable epidermis, dermis, subcutaneous tissue, available only from USFDA and EMA but the authors feels that the best
sebum, hair lipid and hair core and allows prediction of PK behavior of approach would be to have harmonized guidelines such as ICH guide­
various formulations: liquids, semi-solids, subcutaneous injections. Few lines, for implementation of modeling approaches to support drug
case examples of utility of transdermal simulation tools is presented in product quality and minimize in vivo studies as much as possible. All this
Table 6. Hussain et al. developed a transdermal model for Rifampicin will ultimately bring high quality affordable generic medicines within
vesicular elastic liposomes to aid formulation development. Gastroplus the reach of patients and within a short time.
model was developed and used to investigate effect of physicochemical
properties, formulation and physiology of drug on dissolution, absorp­ Funding
tion, access to portal vein [74] to aid in formulation development. While
these case studies indicate significance of such tools in formulation This research did not receive any specific grant from funding
development, more development is required in this area. agencies in the public, commercial, or not-for-profit sectors.

4.4. Parenteral route of administration Author contributions

Parenteral route of administration is considered either as simple or Yuvaneshwari K: Writing – original draft, Sivacharan Kollipara:
complex based on the type of formulation. While formulations such as Writing – original draft, Conceptualization, Tausif Ahmed: Conceptu­
intravenous bolus, infusion, intramuscular solution are considered as alization, Writing – review & editing, Supervision, Siddharth Chachad:
simple and easy to predict, formulations such as long acting injectables Writing – review & editing, Supervision.

11
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

Declaration of competing interest Pharmacometrics Syst. Pharmacol. 6 (2) (2017) 71–72, https://doi.org/10.1002/
psp4.12166.
[19] E. Luzon, K. Blake, S. Cole, A. Nordmark, C. Versantvoort, E.G. Berglund,
All the authors are employees of Dr. Reddy’s laboratory and report Physiologically based pharmacokinetic modeling in regulatory decision-making at
no conflicts interest. The authors alone are responsible for the content the European Medicines Agency, Clin. Pharmacol. Ther. 102 (1) (2016) 98–105,
and writing of this article. https://doi.org/10.1002/cpt.539.
[20] T. Heimbach, S. Suarez-Sharp, M. Kakhi, N. Holmstock, A. Olivares-Morales,
X. Pepin, F. Kesisoglou, Dissolution and translational modeling strategies toward
Acknowledgements establishing an in vitro-in vivo link—a workshop summary report, AAPS J. 21 (2)
(2019) 29, https://doi.org/10.1208/s12248-019-0298-x.
[21] A. Mitra, S. Suarez-Sharp, X.J.H. Pepin, T. Flanagan, Y. Zhao, E. Kotzagiorgis,
Authors would like to thank Dr. Reddy’s Laboratory for providing an A. Babiskin, Applications of physiologically based biopharmaceutics modeling
opportunity to publish this work. (PBBM) to support drug product quality: a workshop summary report,
J. Pharmaceut. Sci. 110 (2) (2021) 594–609, https://doi.org/10.1016/j.
xphs.2020.10.059.
References [22] S. Kijima, Y. Ochiai, A. Ishiguro, Meeting report: PMDA public workshop on
pharmacometrics at Japan, CPT Pharmacometrics Syst. Pharmacol. 9 (10) (2020)
[1] G.L. Amidon, H. Lennernas, V.P. Shah, J.R. Crison, A theoretical basis for a 550–552, https://doi.org/10.1002/psp4.12554.
biopharmaceutic drug classification: the correlation of in vitro drug product [23] F. Wu, H. Shah, M. Li, P. Duan, P. Zhao, S. Suarez, P. Seo, Biopharmaceutics
dissolution and in vivo bioavailability, Pharm. Res. 12 (1995) 413–420, https:// applications of physiologically based pharmacokinetic absorption modeling and
doi.org/10.1023/a:1016212804288. simulation in regulatory submissions to the U.S. Food and drug administration for
[2] S.A. Peters, Physiologically based modeling, in: S.A. Peters (Ed.), Physiologically- new drugs, AAPS J. 23 (2) (2021) 31. http://doi.org/10.1208/s12248-02
Based Pharmacokinetic (PBPK) Modeling and Simulations: Principles, Methods, 1-00564-2.
and Applications in the Pharmaceutical Industry, John Wiley & Sons, Inc., 2012, [24] E.O. Akala, Oral controlled release solid dosage forms, in: T.K. Ghosh, B.R. Jasti
pp. 13–16. (Eds.), Theory and Practice of Contemporary Pharmaceutics, Taylor Francis. CRC
[3] Z. Lin, J. Fisher, A history and recent efforts of selected physiologically based Press, 2005, p. 34.
pharmacokinetic modeling topics, in: J.W. Fisher, J.M. Gearhart, Z. Lin (Eds.), [25] G. Kaur, J. Grewal, K. Jyoti, U.K. Jain, R. Chandra, J. Madan, Oral controlled and
Physiologically Based Pharmacokinetic (PBPK) Modeling Methods and sustained drug delivery systems: concepts, advances, preclinical and clinical status,
Applications in Toxicology and Risk Assessment, Academic Press, 2020, pp. 1–26. in: A.M. Grumezescu (Ed.), Drug Targeting and Stimuli Sensitive Drug Delivery
[4] EMA, Guideline on the reporting of physiologically based pharmacokinetic (PBPK) Systems, William Andrew, 2018, pp. 567–626.
modelling and simulation. https://www.ema.europa.eu/en/documents/scientific [26] V. Lukacova, W.S. Woltosz, M.B. Bolger, Prediction of modified release
-guideline/guideline-reporting-physiologically-based-pharmacokinetic-pbpk-mod pharmacokinetics and pharmacodynamics from in vitro, immediate release, and
elling-simulation_en.pdf, Jul, 2019. (Accessed 20 September 2021). intravenous data, AAPS J. 11 (2) (2009) 323–334, https://doi.org/10.1208/
[5] USFDA, CDER, Guidance for industry, physiologically based pharmacokinetic s12248-009-9107-2.
analyses — format and content. https://www.fda.gov/files/drugs/published/Phys [27] M. Arafat, M. Sarftaz, S. AbuRuz, Development and in vitro evaluation of
iologically-Based-Pharmacokinetic-Analyses-%E2%80%94-Format-and-Content controlled release Viagra® containing poloxamer-188 using Gastroplus™ PBPK
-Guidance-for-Industry.pdf, August, 2018. (Accessed 20 September 2021). modeling software for in vivo predictions and pharmacokinetic assessments,
[6] USFDA, CDER, Guidance for industry, the use of physiologically based Pharmaceuticals 14 (5) (2021) 479, https://doi.org/10.3390/ph14050479.
pharmacokinetic analyses – biopharmaceutics applications for oral drug product [28] K. Otsuka, C. Wagner, A. Selen, J. Dressman, Prediction of in-vivo pharmacokinetic
development, Manufacturing changes and controls. https://www.fda.gov/me profile for immediate and modified release oral dosage forms of furosemide using
dia/142500/download, October 2020. (Accessed 20 September 2021). an in-vitro–in-silico–in-vivo approach, J. Pharm. Pharmacol. 67 (5) (2015)
[7] V.A. Ramchandani, J. Bolane, T.K. Li, S. O’Connor, A physiologically based 651–665, https://doi.org/10.1111/jphp.12365.
pharmacokinetic model for Alcohol facilitates Rapid BrAC clamping, Alcohol Clin. [29] A. Li, K. Yeo, D. Welty, H. Rong, Development of guanfacine extended-release
Exp. Res. 23 (4) (1999) 617–623. dosing strategies in children and adolescents with ADHD using a physiologically
[8] A. Garg, J.P. Balthasar, Physiological based pharmacokinetic model to predict IgG based pharmacokinetic model to predict drug-drug interactions with moderate
tissue kinetics in wild-type and FcRn-knockout mice, J. Pharmacokinet. CYP3A4 inhibitors or inducers, Pedatr. Drugs 20 (2) (2018) 181–194, https://doi.
Pharmacodyn. 34 (5) (2007) 687–709, https://doi.org/10.1007/s10928-007- org/10.1007/s40272-017-0270-0.
9065-1. [30] D.J. Good, R. Hartley, N. Mathias, J. Crison, G. Tirucherai, P. Timmins, N.K.
[9] P. Zhao, L. Zhang, J.A. Grillo, Q. Liu, J.M. Bullock, Applications of physiologically E. Fung, Mitigation of adverse clinical events of a narrow target therapeutic index
based pharmacokinetic modeling & simulation during Regulatory Review, Clin. compound through modified release formulation design: an in vitro, in vivo, in
Pharmacol. Ther. 89 (2) (2011) 259–267, https://doi.org/10.1038/clpt.2010.298. silico, and clinical pharmacokinetic analysis, Mol. Pharm. 12 (2015) 4434–4444,
[10] R. Leong, M.L. Vieira, P. Zhao, Y. Mulugeta, C.S. Lee, S.M. Huang, G.J. Burckart, https://doi.org/10.1021/acs.molpharmaceut.5b00624.
Regulatory experience with physiologically based pharmacokinetic modeling for [31] K. Filippos, B. Xia, N.G.B. Agarwal, Comparison of deconvolution-based and
pediatric drug trials, Clin. Pharmacol. Ther. 91 (5) (2012) 925–931, https://doi. absorption modeling IVIVC for extended release formulations of a BCS III drug
org/10.1038/clpt.2012.19. development candidate, AAPS J. 17 (6) (2015) 1492–1500, https://doi.org/
[11] H.M. Jones, K. Mayawala, P. Poulin, Dose selection based on physiologically based 10.1208/s12248-015-9816-7.
pharmacokinetic (PBPK) approaches, AAPS J. 15 (2013) 377–387, https://doi.org/ [32] K. Filippos, A. Balakrishnan, K. Manser, Utility of PBPK absorption modeling to
10.1208/s12248-012-9446-2. guide modified release formulation development of gaboxadol, a highly soluble
[12] J.E. Sager, J. Yu, I. Ragueneau-Majlessi, N. Isoherranen, Physiologically based compound with region-dependent absorption, J. Pharmaceut. Sci. 105 (2016)
pharmacokinetic modeling and simulation approaches: a systemic review of 722–728, https://doi.org/10.1002/jps.24674.
published models, applications and model verification, Drug Metab. Dispos. 43 [33] C.J. Andreas, X. Pepin, C. Markopoulos, M. Vertzoni, C. Reppas, J.B. Dressman,
(11) (2015) 1823–1837, https://doi.org/10.1124/dmd.115.065920. Mechanistic investigation of the negative food effect of modified release zolpidem,
[13] X. Zhang, Y. Yang, M. Grimstein, J. Fan, J.A. Grillo, S.M. Huang, Y. Wang, Eur. J. Pharmaceut. Sci. 1020 (2017) 284–298, https://doi.org/10.1016/j.
Application of PBPK modeling & simulation for Regulatory decision making and its ejps.2017.03.011.
impact on US prescribing information: an update on the 2018-2019 submissions to [34] R. Jereb, J. Opara, I. Legen, B. Petek, D.G. Peklar, In vitro–in vivo relationship and
the US FDA’s office of Clinical pharmacology, J. Clin. Pharmacol. 60 (1) (2020) bioequivalence prediction for modified-release capsules based on a PBPK
S160–S178, https://doi.org/10.1002/jcph.1767. absorption model, AAPS PharmSciTech 21 (2020) 18, https://doi.org/10.1208/
[14] C. Wagner, Y. Pan, V. Hsu, V.A. Grillo, L. Zhang, K.S. Reynolds, P. Zhao, Predicting s12249-019-1566-x.
the effect of cytochrome P450 inhibitors on substrate drugs: analysis of [35] H.M. Jones, M. Dickins, K. Youdim, J.R. Gosset, N.J. Attkins, T.L. Hay, Application
physiologically based pharmacokinetic modeling submissions to the US food and of PBPK modeling in drug discovery and development at Pfizer, Xenobiotica 42 (1)
drug administration, Clin. Pharmacokinet. 54 (2015) 117–127, https://doi.org/ (2012) 94–106, https://doi.org/10.3109/00498254.2011.627477.
10.1007/s40262-014-0188-4. [36] H. Ali, P.R.P. Verma, S.K. Dubey, J. Venkatesan, In vitro–in vivo and
[15] C. Wagner, Y. Pan, V. Hsu, V. Sinha, P. Zhao, Predicting the effect of CYP3A pharmacokinetic evaluation of solid lipid nanoparticles of furosemide using
Inducers on the Pharmacokinetics of substrate drugs using physiologically based Gastroplus™, RSC Adv. 53 (2017) 33314–33326, https://doi.org/10.1039/
pharmacokinetic modeling: an analysis of PBPK submissions to the USFDA, Clin. C7RA04038E.
Pharmacokinet. 55 (2016) 475–483, https://doi.org/10.1007/s40262-015-0330-y. [37] USFDA, CDER, Guidance for industry, extended release oral dosage forms:
[16] X. Zhuang, C. Lu, PBPK modeling and simulation in drug research and development, evaluation, and application of in vitro/in vivo correlations.
development, Acta Pharm. Sin. B. 6 (5) (2016) 430–440, https://doi.org/10.1016/ https://www.fda.gov/media/70939/download, Sep 1997. (Accessed 20 September
j.apsb.2016.04.004. 2021).
[17] C. Wagner, P. Zhao, Y. Pan, V. Hsu, J. Grillo, S.M. Huang, V. Sinha, Application of [38] A. Abend, T. Heimbach, M. Cohen, F. Kesisoglou, Dissolution and Translational
physiologically based pharmacokinetic modeling to support dose selection: report modeling strategies enabling patient-centric drug product development: the M-
of an FDA Public Workshop on PBPK, CPT Pharmacometrics Syst. Pharmacol. 4 (4) CERSI Workshop Summary Report, AAPS J. 20 (3) (2018) 60, https://doi.org/
(2015) 226–230, https://doi.org/10.1002/psp4.33. 10.1208/s12248-018-0213-x.
[18] P. Zhao, Report from the EMA workshop on qualification and reporting of [39] T. Kato, H. Nakagawa, T. Mikkaichi, T. Miyano, Establishment of a clinically
physiologically based pharmacokinetic (PBPK) modeling and simulation, CPT relevant specification for dissolution testing using physiologically based

12
Y. K et al. Journal of Drug Delivery Science and Technology 69 (2022) 103152

pharmacokinetic (PBPK) modeling approaches, Eur. J. Pharm. Biopharm. 151 inhibitor drug interactions, and administration with acidic juices, Eur. J.
(2020) 45–52, https://doi.org/10.1016/j.ejpb.2020.03.012. Pharmaceut. Sci. 142 (2019) 435–448, https://doi.org/10.1016/j.
[40] R.L.M. Paraiso, R.H. Rose, N. Fotaki, M. McAllister, The use of PBPK/PD to ejpb.2019.07.011.
establish clinically relevant dissolution specifications for zolpidem immediate [59] A. Mitra, B. Petek, A. Bajc, R. Velagapudi, I. Legen, Physiologically based
release tablets, Eur. J. Pharmaceut. Sci. 155 (2020) 105534, https://doi.org/ absorption modeling to predict bioequivalence of controlled release and immediate
10.1016/j.ejps.2020.105534. release oral products, Eur. J. Pharm. Biopharm. 134 (2019) 117–125, https://doi.
[41] I.L. Konstantinidis, R. Cristofoletti, N. Fotaki, D.B. Turner, Establishing virtual org/10.1016/j.ejpb.2018.11.019.
bioequivalence and clinically relevant specifications using in vitro biorelevant [60] F. Zhang, Y. Zhou, N. Wu, R. Jia, A. Liu, B. Liu, H. Wang, In silico prediction of
dissolution testing and physiologically-based population pharmacokinetic bioequivalence of Isosorbide Mononitrate tablets with different dissolution profiles
modeling. case example: Naproxen, Eur. J. Pharmaceut. Sci. 143 (2020) 105170, using PBPK modeling and simulation, Eur. J. Pharm. Biopharm. 157 (1) (2021)
https://doi.org/10.1016/j.ejps.2019.105170. 105618, https://doi.org/10.1016/j.ejps.2020.105618.
[42] L. Miao, Y.M. Mousa, L. Zhao, K. Raines, P. Seo, F. Wu, Using a physiologically [61] A. Mitra, F. Kesisoglou, P. Dogterom, Application of absorption modeling to predict
based pharmacokinetic absorption model to establish dissolution bioequivalence bioequivalence outcome of two batches of etoricoxib tablets, AAPS PharmSciTech
safe space for Oseltamivir in adult and pediatric populations, AAPS J. 22 (2020) 16 (1) (2015) 76–84, https://doi.org/10.1208/s12249-014-0194-8.
107, https://doi.org/10.1208/s12248-020-00493-6. [62] A. Okumu, M. DiMaso, R. Lobenberg, Computer simulations using GastroPlus to
[43] X.J.H. Pepin, T.R. Flanagan, D.J. Holt, A. Eidelman, D. Treacy, C.E. Rowlings, justify a biowaiver for etoricoxib solid oral drug products, Eur. J. Pharm.
Justification of drug product dissolution rate and drug substance particle size Biopharm. 72 (1) (2009) 91–98, https://doi.org/10.1016/j.ejpb.2008.10.019.
specifications based on absorption PBPK modeling for Lesinurad immediate release [63] S. Zhang, M. Fang, Q. Zhang, X. Li, T. Zhang, Evaluating the bioequivalence of
tablets, Mol. Pharm. 13 (9) (2016) 3256–3269, https://doi.org/10.1021/acs. metronidazole tablets and analyzing the effect of in vitro dissolution on in vivo
molpharmaceut.6b00497. absorption based on PBPK modeling, Drug Dev. Ind. Pharm. 45 (10) (2019)
[44] M.B. Bolger, J.S. Macwan, M. Sarfraz, M. Almukainzi, R. Lobenberg, The 1646–1653, https://doi.org/10.1080/03639045.2019.1648502.
irrelevance of in vitro dissolution in setting product specifications for drugs like [64] Y. Tsume, G.L. Amidon, The biowaiver extension for BCS class III drugs: the effect
dextromethorphan that are subject to lysosomal trapping, J. Pharmaceut. Sci. 108 of dissolution rate on the bioequivalence of BCS class III immediate-release drugs
(1) (2019) 268–278, https://doi.org/10.1016/j.xphs.2018.09.036. predicted by computer simulation, Mol. Pharm. 7 (4) (2010) 1235–1243, https://
[45] X. Li, Y. Yang, Y. Zhang, C. Wu, Q. Jiang, W. Wang, H. Li, Justification of biowaiver doi.org/10.1021/mp100053q.
and dissolution rate specifications for piroxicam immediate release products based [65] L. Sun, J. Sun, Z. He, Exploring the feasibility of biowaiver extension of BCS class
on physiologically based pharmacokinetic modeling: an in-depth analysis, Mol. III drugs with site-specific absorption using gastrointestinal simulation technology,
Pharm. 16 (9) (2019) 3780–3790, https://doi.org/10.1021/acs. Eur. J. Drug Metab. Pharmacokinet. 42 (2017) 471–487, https://doi.org/10.1007/
molpharmaceut.9b00350. s13318-016-0361-2.
[46] S. Jaiswal, T. Ahmed, S. Kollipara, M. Bhargava, S. Chachad, Development, [66] A.N. Zaid, N. Shraim, A. Radwan, N. Jaradat, S. Hirzallah, I. Issa, A. Khraim, Does
validation and application of physiologically based biopharmaceutics model to GastroPlus support similarity and dissimilarity factors of in vitro-in vivo prediction
justify the change in dissolution specifications for DRL ABC extended release in biowaiver studies? A lower strength amlodipine as a model drug, Drug Res. 68
tablets, Drug Dev. Ind. Pharm. 47 (5) (2021) 778–789, https://doi.org/10.1080/ (11) (2018) 625–630, https://doi.org/10.1055/a-0611-4927.
03639045.2021.1934870. [67] I. Kovacevic, J. Parojcic, I. Homsek, M.T. Grozdanis, Justification of biowaiver for
[47] C. Markopoulos, F. Thoenen, D. Preissig, M. Symillides, M. Vertzoni, N. Parrott, carbamazepine, a low soluble high permeable compound, in solid dosage forms
G. Imanidis, Biorelevant media for transport experiments in the Caco-2 model to based on IVIVC and gastrointestinal simulation, Mol. Pharm. 6 (1) (2009) 40–47,
evaluate drug absorption in the fasted and the fed state and their usefulness, Eur. J. https://doi.org/10.1021/mp800128y.
Pharm. Biopharm. 86 (3) (2014) 438–448, https://doi.org/10.1016/j. [68] H.J. Gukasyan, S. Hailu, T.K. Karami, R. Graham, Ocular biopharmaceutics: impact
ejpb.2013.10.017. of modeling and simulation on topical ophthalmic formulation development, Drug
[48] J. Rebeka, O. Jerneja, L. Igor, P. Bostjan, B. Aleksander, Z. Simon, K. Albin, PBPK Discov. Today 24 (8) (2019) 1587–1597, https://doi.org/10.1016/j.
absorption modeling of food effect and bioequivalence in fed state for two drudis.2019.04.002.
formulations with crystalline and amorphous forms of BCS 2 class drug in generic [69] M.L. Merdy, J. Fan, M. Bolger, V. Lukacova, J. Spires, E. Tsakalozou, L. Zhao,
drug development, AAPS PharmSciTech 20 (2) (2019) 59, https://doi.org/ Application of mechanistic ocular absorption modeling and simulation to
10.1208/s12249-018-1285-8. understand the impact of formulation properties on ophthalmic bioavailability in
[49] J. Rebeka, A. Kristl, A. Mitra, Prediction of fasted and fed bioequivalence for rabbits: a case study using dexamethasone suspension, AAPS J. 21 (4) (2019) 65,
immediate release drug products using physiologically based biopharmaceutics https://doi.org/10.1208/s12248-019-0334-x.
modeling (PBBM), Eur. J. Pharmaceut. Sci. 155 (1) (2020) 105554, https://doi. [70] M.L. Merdy, J. Spires, V. Lukacova, M.L. Tan, A. Babiskin, X. Xu, M.B. Bolger,
org/10.1016/j.ejps.2020.105554. Ocular physiologically based pharmacokinetic modeling for ointment
[50] R.S. Lloyd, M.I. Hingle, J.C. Bloomer, S.J. Charles, J.M. Butler, A. Paul, C. Ambery, formulations, Pharm. Res. 37 (2020), https://doi.org/10.1007/s11095-020-02965-
Negative food effect of Danirixin: use of PBPK modelling to explore the effect of y. Article number: 245.
formulation and meal type on clinical PK, Pharm. Res. 37 (2020) 233, https://doi. [71] Lukacova-DDL2017-Pulmonary-IVIVC, in: Drug Delivery to the Lungs Conference
org/10.1007/s11095-020-02948-z. 2017, December 6-8, 2017. Edinburgh, Scotland.
[51] C. Tistaert, T. Heimbach, B. Xia, N. Parrott, T.S. Samant, F. Kesisoglou, Food effect [72] N. Idkaidek, F. Hawari, Y. Dodin, N. Obeidat, Development of a physiologically-
projections via physiologically based pharmacokinetic modeling: predictive case based pharmacokinetic (PBPK) model of nebulized hydroxychloroquine for
studies, J. Pharmaceut. Sci. 108 (2019) 592–602, https://doi.org/10.1016/j. pulmonary delivery to COVID-19 patients, Drug Res. 71 (5) (2021 May) 250–256,
xphs.2018.05.024. https://doi.org/10.1055/a-1325-0248.
[52] N. Parrott, C. Stillhart, M. Lindenberg, B. Wagner, K. Kowalski, E. Guerini, [73] E. Tsakalozou, A. Babiskin, L. Zhao, Physiologically-based Pharmacokinetic
G. Meneses-Lorente, Physiologically based absorption modelling to explore the Modeling to Support Bioequivalence and Approval of Generic Products: A Case for
impact of food and gastric pH changes on the pharmacokinetics of entrectinib, Diclofenac Sodium Topical Gel 1%, CPT Pharmacometrics Syst. Pharmacol, 2021
AAPS J. 22 (4) (2020) 78, https://doi.org/10.1208/s12248-020-00463-y. https://doi.org/10.1002/psp4.12600.
[53] A.E. Riedmaier, D.J. Lindley, J.A. Hall, S. Castleberry, R.T. Slade, P. Stuart, [74] A. Hussain, M.A. Altamimi, S. Alsheheri, S.S. Imam, S.K. Singh, Vesicular elastic
M. Nijsen, Mechanistic physiologically based pharmacokinetic modeling of the liposome for transdermal delivery of Rifampicin: in vitro, in vivo and in silico
dissolution and food effect of a biopharmaceutics classification system IV GastroPlus prediction studies, Eur. J. Pharm. Biopharm. 151 (2020) 105411,
compound-the venetoclax story, J. Pharmaceut. Sci. 107 (1) (2018) 495–502, https://doi.org/10.1016/j.ejps.2020.105411.
https://doi.org/10.1016/j.xphs.2017.09.027. [75] A. Shahraz, et al., Development of in vitro-in vivo correlation for long acting
[54] F. Kesisoglou, Can PBPK modeling streamline food effect assessments? J. Clin. injectable microsphere formulations, in: AAPS Annual Meeting and Exposition,
Pharmacol. 50 (S1) (2020) S98–S104, https://doi.org/10.1002/jcph.1678. November 12-15, 2017, San Diego, CA, 2017.
[55] S.C. Sutton, R. Nause, K. Gandelman, The impact of gastric pH, volume, and [76] T. Sou, I.K. Ibrulj, F. Soukarieh, S. Halliday, R.C. Levesque, P. Williams, C.A.
emptying on the food effect of ziprasidone oral absorption, AAPS J. 19 (2017) S. Bergstrom, Model based drug development in pulmonary delivery: PK analysis of
1084–1090, https://doi.org/10.1208/s12248-017-0065-9. noval drug candidates for treatment of Pseudomonas aeruginosa Lung Infection,
[56] S. Willmann, K. Thelen, C. Becker, J.B. Dressman, J. Lippert, Mechanism-based J. Pharmaceut. Sci. 108 (1) (2019) 630–640, https://doi.org/10.1016/j.
prediction of particle size-dependent dissolution and absorption: cilostazol xphs.2018.09.017.
pharmacokinetics in dogs, Eur. J. Pharm. Biopharm. 76 (2010) 83–94, https://doi. [77] T. Tanner, R. Marks, Delivering drugs by the transdermal route: review and
org/10.1016/j.ejpb.2010.06.003. comment, Skin Res. Technol. 14 (3) (2008) 249–260, https://doi.org/10.1111/
[57] R. Cristofoletti, B. Hens, N. Patel, V.V. Esteban, S. Schmidt, J.B. Dressman, j.1600-0846.2008.00316.x.
Integrating drug- and formulation-related properties with gastrointestinal tract [78] USFDA, Draft guidance on diclofenac sodium (https://www.accessdata.fda.go
variability using a product-specific particle size approach: case example ibuprofen, v/drugsatfda_docs/psg/Diclofenac%20Sodium_draft_Topical%20gel_RLD%
J. Pharmaceut. Sci. 108 (12) (2019) 3842–3847, https://doi.org/10.1016/j. 2022122_RC07-18.pdf, Accessed Sep 20, 2021).
xphs.2019.09.012. [79] K. Chaudhary, M.M. Patel, P.J. Mehta, Long-acting injectables: current perspectives
[58] X.J.H. Pepin, A.J. Moir, J.C. Mann, N.J. Sanderson, R. Barker, E. Meehan, J. and future promise, Crit. Rev. Ther. Drug. 36 (2) (2019) 137–181, https://doi.org/
G. Slatter, Bridging in vitro dissolution and in vivo exposure for acalabrutinib. Part 10.1615/CritRevTherDrugCarrierSyst.2018025649.
II. A mechanistic PBPK model for IR formulation comparison, proton pump

13