Sex‐Related Differences in Physiologically‐Based Biopharmaceutics Modeling Authors: Marianela Chavarría‐Rojas, Mubtasim Murshed, Marianela Lorier, Nikoletta Fotaki, Manuel Ibarra Published: 2026 ABSTRACT ABSTRACTPhysiologically based pharmacokinetic (PBPK) and physiologically‐based biopharmaceutics (PBBM) modeling are valuable tools in drug development, allowing mechanistic predictions of drug absorption and disposition. However, sex‐related differences in gastrointestinal physiology are often underrepresented in virtual populations, potentially limiting prediction accuracy. This study aimed to evaluate how sex‐related physiological differences are incorporated into commonly used PBPK platforms and to illustrate their impact on pharmacokinetic predictions using ketoprofen as a case study. Three PBPK platforms were systematically reviewed to assess predefined sex‐specific gastrointestinal parameters. All three platforms incorporated sex‐related differences in general anatomy and physiology but overlooked sex‐specific variability in gastrointestinal tract parameters. In addition, three PBBM models of ketoprofen were developed and verified in males and subsequently extrapolated to females using default and refined sex‐specific parameters. Under default female settings, the models overpredicted Cmax and underestimated Tmax, resulting in concentration–time profiles that were nearly indistinguishable from those of males. Refining gastrointestinal tract parameters for females population improved prediction performance and better reflected observed sex differences. These findings indicate that current PBPK platforms may require user‐defined adjustments to adequately represent sex‐specific gastrointestinal physiology and that incorporating such parameters could lead to more representative PBBM applications.Keywords: gastrointestinal tract physiology, ketoprofen, physiologically based biopharmaceutics modeling, physiologically based pharmacokinetic modeling, women physiology Introduction Physiologically based pharmacokinetic (PBPK) modeling has become a valuable tool in drug development, firmly established within the model‐informed drug development (MIDD) paradigm [1]. By integrating physiological, anatomical, and biochemical parameters specific to the individual; drug dependent physicochemical and pharmacokinetic (PK) parameters; and formulation properties, PBPK models constitute a quantitative, mechanistic platform suitable for hypothesis testing and extrapolation to inform drug‐development and pharmacotherapeutic decisions [2]. Over the past decade, the utility of PBPK modeling has extended into the realm of drug product development, particularly through its application in physiologically based biopharmaceutic modeling (PBBM). PBBM is an extension of PBPK modeling that enables the integration of in vitro dissolution data and drug product quality attributes with physiological parameters to assess how specific drug and formulation characteristics affect in vivo drug performance. By leveraging PBPK principles, PBBM supports the development of drug formulations through in silico prediction of oral drug bioavailability, dissolution, and absorption profiles. This approach frequently incorporates biorelevant in vitro dissolution profiles—critical formulation‐related quality attributes—to mechanistically represent drug performance at the site of administration and predict subsequent drug absorption [3, 4]. PBBM enables prediction of oral bioavailability across populations, evaluation of PK responses dependent on physiological, demographic, or sex‐related factors, and assessment of food effects and potential drug–drug interactions on absorption. Furthermore, PBBM supports pharmaceutical development and regulatory decision‐making, allowing for the simulation of complex scenarios and offering a powerful approach for designing optimized drug products and assessing pharmaceutical quality [5]. The strength of PBPK and PBBM approaches lies in their ability to integrate knowledge of the drug, individual physiology, and the formulation to perform pharmacokinetic predictions across various unobserved scenarios, such as different populations, doses, formulations, or routes of administration, often with limited experimental data. Accurate definition of variables such as organ sizes, blood flow rates, and enzymatic activity is crucial for predicting drug absorption and disposition in the body. Although this methodology has evolved rapidly over the past decades, there is still room for improvement, as greater understanding of PK processes and their variability across subpopulations can be attained. In particular, due to the complexity of gastrointestinal physiology, the accurate prediction of the rate and extent of oral drug absorption remains challenging, particularly for drugs of class II, III, and IV in the biopharmaceutical classification system, and for modified‐release formulations, where both drug release and absorption are highly dependent on formulation characteristics and can vary substantially across different gastrointestinal physiological conditions. Among the physiological factors requiring further exploration, sex‐related differences represent a particularly underexplored source of variability. Women have historically been under‐represented in biomedical research, partly due to the assumption that results obtained from men could be extrapolated across sexes [6]. This underrepresentation has led to a gap in our understanding of how sex‐related differences affect PK, pharmacodynamics, and disease progression, perpetuated by the tendency to treat men and women as a homogeneous population. Accumulating evidence demonstrates that this assumption often does not hold, particularly in the case of oral bioavailability and bioequivalence evaluations. Factors such as gastric emptying time (GET), intestinal transit, enzymatic activity, and drug transporter expression differ between men and women, affecting drug absorption and bioavailability and may explain the observed variations in clinical responses to medicines, including differences in safety and efficacy [7, 8, 9]. These physiological differences underscore the importance of accounting for sex‐related variability during drug development and regulatory assessments. PBPK models supporting drug development should therefore accurately incorporate sex‐specific physiology. This work had three objectives: (1) to review how sex‐related physiological differences are represented in the default virtual populations of widely used PBPK platforms; (2) to assess platform flexibility for incorporating user‐defined sex‐specific parameters; and (3) to analyze the impact of these differences on PK extrapolations between the sexes. To illustrate these aspects, we present a case study examining ketoprofen PK after single oral dose administration in men and women. Ketoprofen is a weakly acidic drug classified as a BCS class II compound, characterized by low solubility and high permeability [10]. As a weak acid, its solubility and absorption are strongly influenced by the pH of the gastrointestinal tract (GIT). Furthermore, in this case study, ketoprofen is administered as a gastro‐resistant formulation, whose performance is particularly dependent on both pH and gastrointestinal transit times, which may vary between the sexes. Thus, ketoprofen represents a suitable model compound for investigating sex‐related differences in oral absorption within a PBPK framework. This comprehensive analysis aims to improve bioavailability predictions across the sexes and enhance the integration of sex‐specific differences into PBPK and PBBM models. Methods PBPK Platform Assessment The evaluation methodology was adapted from Demeester et al. [11], who assessed the capacity of PBPK platforms to incorporate well‐reported physiological changes in elderly populations that may impact oral drug absorption. The analysis included a review of sex differences in physiological factors relevant to drug absorption, such as pH across different segments of the GIT, transit times, fluid volumes, bile acid synthesis, active transport, and metabolism, followed by a qualitative analysis of the default physiological parameters included in different PBPK platforms. The objective was to examine how each PBPK software accounts for sex‐related physiological differences in the absorption process. To isolate intrinsic sex‐related differences, PBPK parameters for virtual subjects of the same body weight and aged 18–50 years were compared. Finally, the analysis assessed whether users could manually adjust these parameters to create virtual subjects that account for reported sex‐related differences. Case Study: Ketoprofen PBPK Model An in‐house clinical study provided plasma concentration–time profiles of ketoprofen in eight men and seven women following administration of a single dose of an enteric‐coated (EC) formulation (100 mg) [12]. This dataset was obtained from a larger four‐period balanced crossover study evaluating an EC tablet and a bilayer modified‐release tablet under fasting and fed conditions in eight men and eight women. Fasting condition data from the EC formulation were selected to minimize the confounding effects of food and sex‐related physiological differences on gastrointestinal transit, and because the simpler drug‐release mechanism of the EC tablet is more suitable for PBPK modeling within the scope of this work. Complete fasting data were available for eight men and seven women, as one female subject did not attend the fasting administration period. This dataset offered an opportunity to evaluate the ability of PBPK models to predict sex‐related differences in ketoprofen PK. The approach involved first developing a male‐specific PBPK model and then assessing its predictive performance when extrapolated to the female population using (i) default PBPK platform parameters and (ii) user‐defined parameters that account for reported sex‐specific differences in gastrointestinal physiology. This analysis was performed independently in each evaluated PBPK platform. Software The most commonly utilized PBPK platforms were evaluated: PK‐Sim and MoBi (Open Systems Pharmacology Suite, version 11.2, https://www.open‐systems‐pharmacology.org/), GastroPlus (Simulation Plus Inc., version 10.1), and Simcyp (Certara LLC, version 22.0.288.0). Published plasma concentration–time profiles were digitized using PlotDigitizer (https://plotdigitizer.com/). Plots were generated using the R environment (version 4.5.0, R Foundation for Statistical Computing; https://www.r‐project.org/) and RStudio (RStudio Team (2024): Integrated Development Environment for R. Posit Software, PBC, Boston, MA; http://www.posit.co/). PBPK Model Development Three independent PBPK models were developed using PK‐Sim‐MoBi, GastroPlus, and Simcyp platforms following the general workflow described in (Figure 1). Initial model development for the male population incorporated drug‐specific physicochemical and PK parameters (Table 1), along with reported plasma concentration–time profiles from previously published studies. Model development incorporated data from four clinical studies (Table S1): one intravenous study (n = 20 men and n = 9 women) [17] and three oral administration studies (tablet and solution; n = 12, 12, and 7, respectively) conducted exclusively in male subjects [14, 18, 19]. General workflow of the ketoprofen physiologically‐based pharmacokinetic model utilized in all PBPK platforms. TABLE 1Summary of input data for the ketoprofen physiologically based pharmacokinetic model.ParameterValueReferencesMolecular Weight254.28 g/mol[13]Log P3.30[10, 13]pKa4.45[10, 13]Compound typeMonoprotic acidFraction unbound – plasma1%[10]Major binding proteinAlbumin[10]Specific intestinal permeability8.4 × E‐4 cm/s[10]Solubility in water0.010 mg/mL[10]Solubility pH 1.20.13 mg/mL[10]Solubility pH 5.00.38 mg/mL[10]Solubility pH 7.41.4 mg/mL[10]Liver plasma clearance1.01 mL/min/kg[14]Model for estimating organ‐to‐plasma partition coeficients (PK‐Sim)Rodgers and Rowland[15, 16]Kp method (GastroPlus)Rodgers, Leahy, and Rowland[15, 16]Partition coefficient intracellular:plasma (PK‐Sim)0.02 factorOptimizedPartition coefficient Kp (GastroPlus)0.5 factorOptimized Summary of input data for the ketoprofen physiologically based pharmacokinetic model. The model development and evaluation process consisted of four stages as represented in (Figure 1). In Stage 1, a male‐specific ketoprofen model was developed in each platform using literature‐sourced parameters and verified against published clinical data obtained primarily in male subjects. The formulation input for the capsule administered simulations was based on the dissolution profile at pH 6.8 reported by Roda [19]. In Stage 2, the models were refined using the male population data from an in‐house study, with particular attention to the effect of GET on drug release from the EC formulation. From this stage forward, the formulation input across all platforms was based on the EC tablet dissolution profile. In Stage 3, the verified male models were extrapolated to predict female PK using each platform's default settings for female physiology. Finally, Stage 4 involved refinement of the female ketoprofen PBPK models through adjustment of gastrointestinal parameters. In each software, a virtual population of 100 individuals was generated. The male population was based on a 30‐year‐old reference individual (73 kg, 176 cm, and BMI of 23.57 kg/m2), with an age range of 18–50 years, a body weight range of 51.70–105.76 kg, and a BMI range of 19–32 kg/m2. The female population was based on a 30‐year‐old reference individual (60 kg, 163 cm, BMI 22.58 kg/m2), with an age range of 18–50 years, a body weight range of 43.91–106.81 kg, and a BMI range of 17–35 kg/m2. Model evaluation was performed using both graphical and quantitative approaches. Simulated plasma concentration–time profiles were compared with clinical data, and goodness‐of‐fit plots were generated to compare predicted and observed PK metrics. Quantitative evaluation was based on predicted‐to‐observed ratios and the absolute average fold error (AAFE) for the area under the curve (AUC) and the maximum plasma concentration (Cmax). A 1.25‐fold error was used as the stricter acceptance criterion, whereas a 2‐fold error was applied as the more flexible criterion. The AAFE was calculated as follows: Results PBPK Platform Assessment Reported sex‐related differences in GIT physiology and their incorporation into PBPK platforms are summarized in (Table 2). All three platforms incorporate sex‐related differences in anatomical parameters, such as organ volumes and GIT dimensions. In contrast, functional GIT parameters show limited differentiation between sexes in the default configurations. Parameters such as pH, fluid volumes, and bile acid concentrations do not reflect sex‐specific differences in any of the evaluated platforms. Transit times show variable implementation: although gastric and small intestinal transit times lack sex differentiation in all platforms, Simcyp incorporates a sex‐specific colonic transit time in its default settings. Similarly, for metabolic enzymes and transporters, only Simcyp allows optional activation of CYP3A4 expression differences between the sexes. Importantly, all platforms provide the flexibility for manual adjustment of these parameters to reflect reported sex‐related physiological differences. A detailed comparison of literature‐reported values is provided in the Supporting Information. TABLE 2Incorporation of sex‐related differences in GIT characteristics in the PBPK modeling platforms.ParametersLiterature reported valuesInclusion of sex‐related differences in default configurationsCommentsWomenMenPK‐SimGastroPlusSimcypStomach pH (fasted state) [6, 20, 21, 22]2.792.16NoNoNoAll platforms include segment‐specific pH values, but default settings do not differentiate between sexes. User‐defined adjustments are possible.Duodenum pH [20]7.166.75NoNoNoJejunum pH [20, 23]7.107.2NoNoNoDescending colon pH [20]6.517.18NoNoNoFluids volume [20, 24]Stomach: 90 mL Duodenum: 138 mLStomach:135 mL Duodenum: 259 mLNoNoNoAll platforms model GIT fluid volumes (PK‐Sim and GastroPlus use segment‐specific volumes with transfer rates; Simcyp employs a dynamic fluid volume model), but default values do not reflect sex‐specific differences. User‐defined adjustments are possible.GIT transit times [20, 25, 26, 27, 28, 29, 30]LiquidsNoNoPartialAll platforms model GIT transit times with platform‐specific approaches (PK‐Sim: gastric emptying, small and large intestine transit times with segment transfer rates; GastroPlus: segment‐specific transit times, some weight‐dependent; Simcyp: mean gastric, small intestine, and colon residence times). Only Simcyp includes sex‐specific differences in default colonic transit time. User‐defined adjustments are possible in all platforms.Stomach: 53.8 minStomach: 30.3 minSolidsStomach: 92.4 min Small intestine: 3.30 h Large intestine: 26 hStomach: 59.8 min Small intestine: 3.02 h Large intestine: 42 hBile salts concentration [20, 23] Total: 0.3 mM Jejunum: 3.2 mM Total: 0.2 mM Jejunum: 2.6 mM NoNoNoBile salt modeling varies across platforms: PK‐Sim does not include bile salts concentration; GastroPlus and Simcyp model segment‐specific bile salt concentrations, but default values do not differentiate between sexes. User‐defined adjustments are possible where applicable.GIT organs size and volume [20] Duodenum: 25.2 cm Jejunum: 573.8 cm Ascending colon: 23 cm Transverse colon: 48 cm Descending colon: 23 cm Duodenum: 27.8 cm Jejunum: 643.9 cm Ascending colon: 23 cm Transverse colon: 40 cm Descending colon: 25 cmYesYesYesAll platforms include sex‐specific organ dimensions in default settings. User‐defined adjustments are possible.Active transport, metabolism, and enzyme expression [6, 20, 31]CYPs 2D6, 1A2 and 2E1 activities are increased in men. CYP 3A4 activity is increased in women. P‐gp expression is higher in men.NoNoPartialAll platforms model GIT enzymes and transporters. Simcyp allows activation of sex‐specific CYP3A4 content differences in default settings; GastroPlus includes weight‐based differences for specific enzymes (e.g., CYP2C8); PK‐Sim does not include sex‐specific enzyme expression in default settings. User‐defined adjustments are possible in all platforms. Incorporation of sex‐related differences in GIT characteristics in the PBPK modeling platforms. Total: 0.3 mM Jejunum: 3.2 mM Total: 0.2 mM Jejunum: 2.6 mM Duodenum: 25.2 cm Jejunum: 573.8 cm Ascending colon: 23 cm Transverse colon: 48 cm Descending colon: 23 cm Case study: Ketoprofen PBPK model Model Development and Evaluation The ketoprofen PBPK model was developed in PK‐Sim, GastroPlus, and Simcyp using seven clinical data sets, including both intravenous and oral doses (solution and capsule). All models incorporated the parameters listed in (Table 1). For initial model development, the capsule dissolution profile at pH 6.8 reported by Roda was included as a Weibull function with the following parameters: time to50%dissolutiont₅₀=0.21h, and Weibull shape parameterβ=0.88 [19]. The predicted and observed plasma concentration‐time profiles are shown in (Figure 2). Good model performance was observed across all platforms, as illustrated in (Figure 3), which presents the model evaluation results for the three software tools (Tables S2–S4). Specifically, 64% of predictions fell within the stricter 1.25‐fold error, with an additional 22% within 1.5‐fold. Overall, 98% met the 2‐fold criterion. Predicted and observed plasma concentration–time profiles of ketoprofen from GastroPlus (left column), PK‐Sim (middle column) and Simcyp (right column). Black solid lines represent mean predicted profiles in healthy subjects (n = 100); points show mean observed data with the corresponding standard deviation (if reported in the respective publication); shaded area represents 90% prediction intervals. Ketoprofen model evaluation. (A) Forest plots: Solid line represents a predicted/observed ratio equal to 1, and each clinical study ratio is shown as a single data point. (B) Goodness‐of‐fit plots: Solid lines mark the lines of identity, dotted lines indicate 1.25‐ fold, dashed lines indicate 1.5‐ fold and long‐dashed lines 2‐ fold deviation. The predicted absolute bioavailability was 92% in PK‐Sim and GastroPlus and 95% in Simcyp. Additionally, in all three platforms, the predicted clearance and volume of distribution values were consistent with those reported in the literature, with clearance ranging from 1.16 to 1.5 mL/min/kg [14, 17, 18, 19] and volume of distribution from 0.1 to 0.2 L/kg [10]. These results are presented in Tables S5 and S6. Simulations in Male and Female Populations The validated model from each PBPK platform was utilized to simulate EC tablet administration in both male and female populations. Drug release from this formulation was incorporated into the model by means of a Weibull function with parameters estimated from the in vitro dissolution profile at pH 6.8: time to50%dissolutiont₅₀=0.52h, and Weibull shape parameterβ=5.07. However, the approach to describe the onset of drug release from the EC formulation (i.e., the lag time due to the enteric coating) varied across platforms due to the different model structures available in each software. In PK‐Sim, drug release is triggered when the tablet reaches the duodenum, with GET controlling the timing of release rather than an explicit lag time parameter. To do this, a dummy formulation was used to track the tablet's location along the GIT. This approach enabled the reflection of the link between subject variability in gastric emptying with the variability in the onset of drug dissolution. In GastroPlus and Simcyp, the formulation input includes the dissolution profile along with the formulation's lag time, which should reflect the GET for different populations. Simulations were carried out following the workflow described in (Figure 1). (Figure 4A–C) show the plasma concentration–time profiles for PK‐Sim, GastroPlus, and Simcyp, respectively, while (Table 3) summarizes the predicted‐to‐observed ratios and the AAFE values for AUC, Cmax, and Tmax across the three platforms. Figure S1 shows the relative prediction error plots for the PK metrics in each software. Plasma concentration–time profiles of ketoprofen following a single oral dose of 100 mg in (A) PK‐Sim, (B) GastroPlus, and (C) Simcyp (n = 100). The solid‐colored line represents the simulated median plasma concentration; the shaded area represents the simulated 5th and 95th percentiles, the solid black lines are the observed median profile, and the dots are the observed data. The blue profiles correspond to a male population, the orange profiles represent a female population using default platform settings, and the green profiles represent a female population after refinement of GIT parameters in each software. TABLE 3Prediction error of PBPK model simulations for men and women populations in PK‐Sim, GastroPlus and Simcyp.Cmax (mg/L)AUCt (mg·h/L)tmax (h)PredObsPred/ObsAAFEPredObsPred/ObsAAFEPredObsPred/ObsAAFEPK‐SimMale pop8.618.471.021.0216.0318.050.891.133.153.001.051.05Female pop10.007.511.331.3318.1919.450.941.072.914.000.731.37Female pop (adjusted)7.097.510.941.0615.9119.450.821.224.304.001.081.08GastroPlusMale pop8.428.470.991.0120.7318.051.151.152.633.000.881.14Female pop9.837.511.311.3120.0719.451.031.032.604.000.651.54Female pop (adjusted)6.327.510.841.1920.7519.451.071.074.474.001.121.12SimcypMale pop7.198.470.851.1821.2718.051.181.182.893.000.961.04Female pop7.267.510.971.0321.0519.451.081.082.904.000.731.38Female pop (adjusted)6.077.510.811.2420.3719.451.051.053.964.000.991.01 Note: AUCt: area under the concentration–time curve from time zero to the last measurable time point. Prediction error of PBPK model simulations for men and women populations in PK‐Sim, GastroPlus and Simcyp. Note: AUCt: area under the concentration–time curve from time zero to the last measurable time point. In the male population, GET was optimized to describe the observed data, with the parameter defined as 90 ± 25 min. Results demonstrated good predictive performance across the three platforms, with prediction errors below 20% (blue concentration–time profiles in Figure 4). The optimized male models were then extrapolated to the female population using default platform settings. However, this extrapolation resulted in overprediction of Cmax and underestimation of Tmax. Moreover, the predicted PK profiles for males and females nearly overlapped (blue and orange profiles in Figure 4), indicating that default configurations provide minimal differentiation between sexes. The models were therefore refined for females to better describe the observed data, focusing on the adjustment of GIT parameters. In PK‐Sim, GET for the female population was set to 120 ± 95 min, and small intestine transit time (SITT) was increased by a factor of 1.5 while maintaining the distribution established by the software, resulting in a mean value of 3.10 h. In GastroPlus and Simcyp, the lag‐time and GET for women were set to 180 ± 95 min. These values were defined based on each model's ability to describe the observed data, in particular the median Tmax of each sex. Gastrointestinal pH was also adjusted to reflect sex‐related differences across all three platforms: gastric pH to 2.79, duodenal pH to 7.16, and jejunal pH to 7.10 [20]. In GastroPlus, the intestinal permeability was adjusted to 8.0 × E‐4 cm/s to prevent Cmax overprediction. After these refinements, the predicted plasma profiles (green profiles in Figure 4) showed improved alignment with the observed in vivo data. Although each PBPK platform employed slightly different approaches to simulate the EC formulation, the refinements focused on adjusting GET and its variability. These parameters are especially relevant for EC formulations, as drug release is highly dependent on gastrointestinal transit dynamics. Women exhibit slower gastric emptying and greater interindividual variability in gastrointestinal transit times compared to men [20], factors that can influence both the onset and extent of absorption for such formulations. Incorporating these sex‐specific physiological differences enabled the models to better differentiate the PK profiles of males and females while improving biological plausibility, with prediction errors reduced to below 20% for key PK metrics in both populations (Table 3). Discussion Differences in GIT physiology between men and women have been extensively documented in aspects that can be regarded as critical for oral bioavailability, such as transit times, fluid composition and pH, as well as in the expression of gastrointestinal enzymes and transporters. For some drugs and formulations, these differences can have a major impact on drug absorption and disposition processes, leading to sex‐related pharmacokinetic differences even when exposure metrics are normalized by body weight. In addition to the interindividual variability associated with sex‐related physiological differences, intraindividual variability also occurs, influenced by physiological and hormonal changes such as menstrual cycle phases or prandial state, adding another level of complexity to PK studies. Given both inter‐ and intraindividual sources of variability, tools that mechanistically represent male and female physiology are particularly relevant. PBPK models offer a suitable approach for this purpose; in addition to integrating anatomical, physiological, and biopharmaceutical information, they also allow inter‐ and intraindividual variability to be reflected, thereby providing a more comprehensive framework to analyze how sex‐dependent factors influence the PK of different drugs and formulations. Our evaluation of three widely used PBPK platforms revealed that, while sex‐related anatomical differences (organ volumes, organ sizes, blood flow rates) are incorporated in default settings, most functional GIT parameters do not differentiate between sexes. However, all platforms allow manual parameter adjustment, enabling the adequate representation of sex‐specific physiology when needed. An important application of PBBM is virtual bioequivalence (vBE) studies, which predict through simulations whether two formulations will achieve similar bioavailability. This approach constitutes a mechanistic framework to support decision‐making during drug and product development and enables evaluation of specific scenarios involving subpopulation characteristics. The relevance of accounting for sex‐related differences in BE studies has been demonstrated in both clinical and modeling investigations. We previously found that although average equivalence was achieved in the overall population, stratification by sex revealed significant differences in relative bioavailability for Cmax and AUC in a significant proportion of BE studies [32], highlighting the importance of including sex as a variable in study design and the need for PBPK models to adequately capture sex‐related physiological differences. While conventional BE studies are designed to minimize variability to clearly detect differences between formulations, accounting for sex‐related variability remains important. Differences in gastrointestinal characteristics between males and females can result in subpopulation‐specific differences in formulation performance, given that formulation behavior is influenced by the conditions of the gastrointestinal environment, potentially affecting PK and BE outcomes. Accounting for sex‐related variability helps ensure that BE conclusions are representative of the entire population, supports prediction of real‐world performance, and provides mechanistic insight into observed interindividual PK variability. Recent PBPK applications in vBE illustrate different approaches to incorporating sex differences. Gundeti et al. [33] assessed the effect of sex on dextromethorphan tablet bioequivalence across different male/female ratio scenarios, demonstrating that the formulation did not exhibit significant sex‐dependent effects. Similarly, Boddu et al. [34] modeled the impact of food, sex, and drug–drug interactions on bioequivalence under fed and fasted conditions. In both studies, sex‐specific GIT physiological adjustments were not explicitly described, leaving unclear whether sex‐dependent differences were adequately represented in the simulations. In contrast, Titus et al. [35] emphasized the importance of intraindividual GIT variability, demonstrating that parameters such as GET and colonic transit time—particularly their variability in women—substantially impact bioequivalence predictions. Incorporating this variability improved concordance with clinical data and supported more representative study designs. The ketoprofen case study illustrates these principles. As a BCS Class II drug administered as an EC formulation, ketoprofen release is highly GET‐dependent. Validated PBPK models using default platform settings showed good performance in males (prediction errors < 20%) but substantial errors for Cmax and Tmax in females. Specifically, simulations performed using the default settings resulted in substantial prediction errors for both Cmax and Tmax in females. Cmax prediction errors exceeded +30% in PK‐Sim and GastroPlus (AAFE > 1.3), while Tmax prediction errors exceeded 27% in magnitude across all platforms (AAFE > 1.37). Adjustment of gastric pH and GET in female models substantially improved these predictions, reducing Cmax and Tmax errors to below 20% in all cases, thereby better capturing observed sex differences in exposure profiles. The limited incorporation of sex‐specific GIT parameters in default platform settings reflects the complexity of platform qualification rather than oversight. PBPK platforms qualified by regulatory agencies require substantial scientific evidence to modify default parameters, creating practical barriers to rapid updates as new physiological data emerge. This highlights an important consideration for users: default sex‐specific settings should not be assumed to capture all reported physiological differences. When modeling sex‐stratified populations, particularly for drugs and formulations sensitive to GIT conditions, users must critically evaluate and adjust relevant physiological parameters to ensure model validity and biological plausibility for their specific application. These findings underscore the practical importance of incorporating sex‐specific GIT physiology into PBPK models. For applications such as virtual bioequivalence, where predictions inform regulatory and development decisions, adequate representation of sex‐related differences in transit times, pH, and their associated variability is essential to ensure both predictive accuracy and biological plausibility. Some limitations of this work should be acknowledged. The virtual populations were restricted to adults aged 18–50 years, consistent with typical bioequivalence study designs, excluding menopausal and geriatric populations whose specific physiological changes may affect drug absorption. The ketoprofen case study involved a small in‐house dataset (n = 7–8 per sex), and the sampling schedule of the original study, while adequate for a formulation with delayed absorption onset, may have limited the precision of Cmax, Tmax, and AUC estimates. Future work with larger datasets in female populations and additional scenarios would further strengthen the conclusions drawn from this proof‐of‐concept analysis. Conclusions Current PBPK platforms incorporate sex‐related differences in anatomical parameters (organ volumes and dimensions), but their default configurations do not account for sex‐specific variations in functional GIT parameters such as pH, transit times, and fluid volumes. The ketoprofen case study demonstrated that incorporating sex‐specific GET and pH values substantially improved prediction accuracy, reducing prediction errors from over 30% to below 20% for key PK metrics and better capturing observed sex‐related differences. All platforms provide the flexibility to incorporate sex‐specific GIT physiology through user‐defined adjustments. For drugs and formulations for which absorption is influenced by GIT conditions—particularly modified‐release products—explicit consideration of sex‐related physiological differences and associated variability is essential for accurate predictions and biologically plausible outcomes, with particular relevance for applications such as virtual bioequivalence studies. Author Contributions All authors wrote the manuscript. Marianela Chavarría‐Rojas, Marianela Lorier, Nikoletta Fotaki, and Manuel Ibarra designed the research. All authors performed the research. Marianela Chavarría‐Rojas and Mubtasim Murshed analyzed the data. Funding The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Pharmacokinetics studies used for the ketoprofen physiologically based pharmacokinetic model development and validation. Table S2: Ketoprofen PBPK model performance in PK‐Sim. Table S3: Ketoprofen PBPK model performance in GastroPlus. Table S4: Ketoprofen PBPK model performance in Simcyp. Table S5: Summary of predicted PK parameters across different administration protocols. Table S6: Summary of PK reported parameters in literature. Figure S1: Relative prediction error plots for AUC, Cmax and Tmax in the males, females, and refined females populations across the three software: GastroPlus (red), PK‐Sim (green) and Simcyp (blue). The dashed black line indicates zero prediction error. Circles represent prediction errors calculated from individual observations, while triangles represent prediction errors based on median observed versus median predicted values. The shaded gray area denotes a prediction error within the range of −25% to +25%.