Method and device for predicting variation of blood concentration transition and calculating parameter required for predicting drug interaction

Abstract

To provide a method and device for predicting plasma concentration transition in combination of multiple drugs, by easily determining Ki, fm in vivo, and making them into a database, since, there has been dynamic prediction using a physiologic medicine speed theory (PBPK) model, as a prediction method of a drug interaction for medical products, however, prediction using an inhibition constant (Ki) of in vitro data and a contribution ratio (fm) of a metabolic enzyme, do not always match increase of clinical AUC.SOLUTION: There is provided a method for determining a parameter of a compartment model from a pharmacokinetic parameter of an interacted drug and interaction drug, then converting the compartment parameter into a PBPK model parameter, then determining by simulation using the PBPK model, a Ki value and fm being AUC increase ratios observed clinically. The acquired parameter is made into a database, and simulation of blood concentration transition of the interacted drug and interaction drug is performed.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a method for calculating the values ​​of parameters required to predict drug interactions when an interacting drug is used in combination with an interacting drug using a simplified physiological pharmacokinetic model, and a method and device for simulating the progression of blood or plasma concentrations. [Background technology]

[0002] When multiple drugs are used concomitantly, the drugs may inhibit drug-metabolizing enzymes, uptake transporters, and excretion transporters involved in the absorption, metabolism, distribution, and excretion of the other drugs, resulting in a significant increase in blood concentrations and the risk of serious side effects. When a drug interaction results in a high risk of side effects, the combination of the two drugs is designated as contraindicated. If a high risk of drug interactions is discovered during drug development, development may be abandoned. Sales may also be discontinued after the drug is released to the market. Information on drug interactions is important for the appropriate use of medicines, but clinical trials are not conducted for all combinations, and predictive evaluation using computer simulations is considered useful.

[0003] To avoid drug interaction risks, the Ministry of Health, Labour and Welfare has issued drug interaction guidelines, which state that during the development of new drugs, clinical risks should be predicted using static and dynamic prediction methods, and if the risks cannot be ruled out, they should be confirmed through clinical trials.

[0004] To predict an increase in AUC due to a drug interaction, information is required regarding the inhibition constant (Ki) of the interacting drug (hereafter referred to as the inhibitor) against the metabolic enzyme of the interacting drug (hereafter referred to as the substrate) and the fraction metabolized by the metabolic enzyme of the substrate (fm). The Ki value is evaluated using human liver microsomes or recombinant enzymes. Furthermore, fm is estimated from the results of in vitro tests and clinical pharmacokinetic studies, but it is difficult to estimate a reasonable fm except for metabolic enzymes with genetic polymorphisms that indicate enzyme activity deficiency.

[0005] The Ki value determined in vitro is used to predict an increase in the substrate's AUC, but it does not necessarily coincide with the Ki value inhibited by the actually administered inhibitor in vivo. Using a simplified physiologically based pharmacokinetic (PBPK) model, analysis of the results of clinical interaction studies using the PBPK model revealed differences between the Ki values ​​obtained in vivo and in vitro for highly lipophilic drugs. The in vivo Ki value of ketoconazole for CYP3A is significantly different, at 1 / 200 of the in vitro Ki value (Non-Patent Document 1).

[0006] In static drug interaction prediction, the unbound concentration of the inhibitor in the liver and the Ki value are used to calculate the increase in the substrate AUC. There are multiple concentrations used as the inhibitor concentration in the liver, including the maximum circulating blood concentration and the maximum concentration at the liver inlet, and the prediction is overly accurate because it makes the unrealistic assumption that these concentrations will continue. Even under these conditions, if there is no increase in the substrate AUC, the risk of an interaction is deemed low. However, because there is a high frequency of interactions even when there is no interaction, it is recommended to use dynamic prediction, which has higher prediction accuracy.

[0007] Dynamic predictions using PBPK models can be highly accurate when in vivo Ki values ​​are used, but the analysis required to determine in vivo Ki values ​​is complex. The procedure involves reading data on the blood concentration trends of inhibitors and substrates from papers, analyzing those trends with a PBPK model, and then applying the substrate concentration trends during coadministration to obtain the Ki values ​​through calculations. This requires specialized knowledge and time, so reports of in vivo analysis results are limited. Furthermore, the fm of the substrate's metabolic enzymes is calculated under various assumptions, making it unclear whether the value is appropriate. These issues make it impossible to comprehensively predict drug interactions.

[0008] The CR-IR method is a comprehensive method for predicting drug interactions without using Ki values ​​(Clin Pharmacokinet. 2007; 46(8): 681-96). Using the ratio of increase in AUC of a substrate when used in combination with a representative strong inhibitor, the CR is used to indicate susceptibility to interactions, and the IR of the inhibitor is calculated from the ratio of increase in AUC when used in combination with a substrate with a known CR. While this method can predict interactions comprehensively, it has some problems, such as being limited to a single dosage regimen, unclear changes over time, and requiring calculation of CR and IR through clinical trials.

[0009] Another problem with drug interaction prediction methods is the selection of a liver model. The liver model used in general PBPK models is the well-stirred model. The well-stirred model is also used in the prediction of drug interaction guidelines. The well-stirred model is known to have poor predictive ability for high-clearance drugs. For high-clearance drugs, the parallel tube model or dispersion model is used. Furthermore, the 5-liver model, which divides the liver into five parts, is also used (Non-Patent Document 2). The 5-liver model is approximated by the dispersion model. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] Patent application 2024-35378 [Patent Document 2] Patent application 2024-64997 [Non-patent literature]

[0011] [Non-Patent Document 1] Pharm Res.2008;25(8):1891-901. [Non-patent document 2] CPT Pharmacometrics Syst Pharmacol.2022;11:919_933. Summary of the Invention [Problem to be solved by the invention]

[0012] In view of the above issues, we provide a method for easily determining the Ki value of an inhibitor and the fm of the substrate's metabolic enzyme by analysis using a simplified PBPK model from the pharmacokinetic parameters and the AUC increase ratio due to interactions described in the package inserts and academic papers of pharmaceuticals, a program for executing this method, and a program for predicting the blood or plasma concentration transition and AUC increase ratio due to drug interactions. [Means for solving the problem]

[0013] The present inventors have devised a method in which compartment model parameters are determined from the pharmacokinetic parameters of the substrate and inhibitor by the method described in Patent Document 1, the determined compartment parameters are converted into PBPK model parameters, and the Ki value of the inhibitor, which indicates the fold increase in AUC of the substrate observed in clinical practice, and the fm of the substrate are determined by analysis using a simplified PBPK model, and the following method in which the determined parameters are registered in a database and a simulation is performed of the changes in blood or plasma concentration when the registered substrate and inhibitor are used in combination.

[0014] In a first exemplary embodiment of the present invention, there is provided the following method for predicting the time course of blood or plasma concentrations of a drug when an inhibitor and a substrate are used in combination using a computer. This prediction method comprises the steps of calculating PBPK model parameters from the pharmacokinetic parameters of the inhibitor and the substrate via compartmental analysis and storing the parameters in a database, calculating the Ki value of the inhibitor by simulation using the PBPK model parameters of the substrate and the inhibitor and storing the calculated Ki value in a database, and simulating the time course of blood concentrations of the inhibitor and substrate whose interaction is to be predicted using the PBPK model.

[0015] In a second exemplary embodiment of the present invention, a database is provided that stores compartment model parameters, PBPK model parameters, Ki values, doses, administration schedules, and AUC increase folds of the substrate for inhibitors and substrates for carrying out the prediction method of the first embodiment.

[0016] In a third exemplary aspect of the present invention, there is provided a simulation device comprising a memory storing the program of the first aspect and a control circuit which is a computer, and which executes the computer program to calculate parameters in a compartment model and a PBPK model of an inhibitor and a substrate, and to predict the time course of the blood concentrations of the inhibitor and the substrate when the inhibitor and the substrate are used in combination using the parameters. [Effects of the Invention]

[0017] The Ki value of an inhibitor for a substrate's metabolic enzyme and the fm of the substrate's metabolic enzyme can be calculated from the pharmacokinetic parameters and the fold increase in the substrate's AUC described in drug package inserts, academic papers, academic paper abstracts, and academic conference presentation abstracts. By using the calculated PBPK parameters of many inhibitors and substrates, the Ki value of the inhibitor, and the fm of the substrate, it is possible to comprehensively predict the fold increase in the substrate's AUC when an inhibitor and substrate are used in combination, and to predict the risk of drug interactions for combinations that have not been clinically tested. [Brief explanation of the drawings]

[0018] [Figure 1] FIG. 1 is a flowchart showing the procedure of a drug interaction prediction system according to an embodiment of the present invention. [Figure 2] FIG. 1 is a diagram showing an example of a system for predicting changes in blood concentration of a drug according to an embodiment of the present invention. [Figure 3] FIG. 1 shows a simplified PBPK model. [Figure 4]FIG. 1 shows the plasma concentration profiles predicted by first-order absorption 2-compartment model parameters and PBPK model parameters estimated from the pharmacokinetic parameters of amenamevir, midazolam, and ketoconazole. [Figure 5] FIG. 1 is a diagram showing the relationship between the Ki value of ketoconazole and the fm of CYP3A, the metabolic enzyme of amenamevir. [Figure 6] FIG. 1 shows the relationship between the Ki value of ketoconazole corresponding to the fm of CYP3A, a metabolizing enzyme of midazolam, and the fold increase in AUC of midazolam in four interaction studies. [Figure 7] FIG. 1 shows the relationship between fm and Fg of CYP3A, the metabolic enzyme of amenamevir, based on the AUC increase ratio in an interaction study between ketoconazole and amenamevir. DETAILED DESCRIPTION OF THE INVENTION

[0019] The prediction method and apparatus of this embodiment will be described below with reference to the accompanying drawings as appropriate.

[0020] <Parameter definition> The parameters used in the following disclosure are defined as follows: D: Dosage <Pharmacokinetic parameters> Cmax: maximum plasma concentration Tmax: Time to reach maximum plasma concentration AUC: Area under the plasma concentration time curve HL: Half-life <compartment parameter> Vd: Volume of distribution in a one-compartment model F: Bioavailability Vd / F: apparent volume of distribution Tlag: Absorption lag time Ke: Disappearance rate constant Ka: absorption rate constant V1: Volume of distribution of compartment 1 in the 2-compartment K12: Rate constant for transfer from compartment 1 to compartment 2 K21: Rate constant for transfer from compartment 2 to compartment 1 K10: Disappearance measure constant from compartment 1 Vss: steady-state volume of distribution A / D: Dose-normalized concentration constant in the α phase. B / D: Dose-normalized concentration constant in the beta phase. α: Disappearance rate constant in the α phase β: Disappearance rate constant in the β phase <Physiologically Based Pharmacokinetic Model Parameters> Rb: Blood plasma concentration ratio (blood concentration / plasma concentration) Ksp: rate constant of transfer from the circulating compartment to the peripheral compartment Kps: rate constant of transfer from the peripheral compartment to the circulating compartment Vsys: Volume of distribution of the circulating compartment Vh: liver volume Kph: Liver-blood concentration ratio (liver concentration / blood concentration) Feu: Urinary excretion rate of unchanged substance Fef: Fecal excretion rate of unchanged drug Fa: Absorption rate Fg: small intestinal availability Fh: hepatic availability fm: Contribution rate of metabolic enzymes fb:Blood unbound fraction fp: Fraction of unbound form in plasma Qh: Hepatic blood flow velocity Qen: Blood flow rate to gastrointestinal epithelial cells CLh: hepatic clearance CLr: Renal clearance The compartment parameters CL / F, Vd / F, and V1 / F are based on plasma concentrations, and the PBPK model parameters CLr, CLh, and Vsys are based on blood concentrations.

[0021] The flowchart and system for the prediction method of the present invention are shown in Figures 1 and 2. If information about the target drug is not stored in the database, the dosage and the pharmacokinetic parameters (AUC, Cmax, Tmax, HL) at that dosage, as well as Fe, Fa or F, and Rb, are entered. Information about the increase in the substrate AUC when the substrate and inhibitor are used in combination, the administration frequency, administration interval, and timing of combination use are entered (S-1).

[0022] The PBPK model parameter calculation section calculates the parameters of a first-order absorption 1- or 2-compartment model from the pharmacokinetic parameters (S-2). In the case of a first-order absorption 1-compartment model, the 1-compartment model parameters are calculated using the following formula: Ke=1n2 / HL (1) CL / F=D / AUC (2) Vd / F=CL / F / Ke (3) Ka is calculated using the following equation (4) using a nonlinear analysis method such as the Newton-Raphson method. Tmax = 1n(Ka / Ke) / (Ka-Ke) (4) Using the obtained Vd / F, Ka, and Ke, Cmax is calculated by the following formula (5). Cmax=D*F / Vd*(Ka / Ke)^(Ke / (Ke-Ka)) (5) The ratio of the observed Cmax to the predicted Cmax is calculated and compared with a set standard to determine whether the model is a one-compartment model or a two-compartment model.

[0023] If a first-order absorption two-compartment model is selected, A1, A2, α, β, Ka, and Tlag are calculated from the pharmacokinetic parameters using the method described in Patent Document 1. In the two-compartment model, analysis is performed using the following equations. C=A1*exp(-K1*(t-Tlag))+A2*exp(-K2*(t-Tlag))-(A1+A2)*exp(-K3*(t-Tlag)) Although the case of Ka > α > β is usually assumed, the case of α > Ka > β (where α and β are the initial elimination rate constants upon intravenous administration) may also occur. The conversion from A1, A2, α, β, and Ka to microparameters is shown for both cases of Ka > α and α > Ka. Case of Ka > α > β A = A1*(K3 - K1) / K3 (6) B = A2*(K3 - K2) / K3 (7) α = K1 (8) β = K2 (9) Ka = K3 (10) Case of α > Ka > β A = (A1 + A2)*(K1 - K2) / K1 (11) B = A2*(K1 - K2) / K1 (12) α = K3 (13) β = K2 (14) Ka = K1 (15) K21 = (A*β + B*α) / (A + B) (16) K10 = α*β / K21 (17) K12 = α + β - K10 - K21 (18) V1 / F = D / (A + B) (19)

[0024] Convert from compartment parameters to PBPK model parameters (S-3). For both the substrate and the inhibitor, use the simplified PBPK model shown in Figure 3. Adopt the 5-liver model for the liver model. Use the compartment model parameters for Ka. Renal clearance: CLr = Ke*Vd / F*(1 - Fe) / Rb (20) CLr = K10*V1 / F(1 - Fe) / Rb (21) Total body clearance: <When F is known> In the case of the 1-compartment model CL = Ke*Vd / F*F / Rb (22) In the case of the 2-compartment model CL = K10*V1 / F*F / Rb (23) Hepatic clearance CLh = CL - CLr (24) Fh = 1 - CLh / Qh (25) FaFg = F / Fh (26) fb = fp / Rb (27) CLint = 5 * Qh * (1 - Fh^0.2) / Fh^0.2 / fb (28) In the case of a one - compartment model Vsys = Vd / F * F / Rb - Fh * Vh * Kph(29) Ksp = 0 (30) Kps = 0 (31) In the case of a two - compartment model Vsys = V1 / F * F / Rb - Fh * Vh * Kph(32) Ksp = V1 / F * F / Rb * K12 / Vsys (33) Kps = K21 (34) When metabolism by intestinal bacteria and degradation in the digestive tract can be ignored, Fa can be calculated from the fecal unchanged excretion rate. Fa = 1 - Fef (35) <When F is unknown and Fa is known> The value of CL / F and CLint indicating the known Fa are obtained using a method for solving non - linear equations such as the Newton - Raphson method with equations (22) - (25). By obtaining F, other parameters are also calculated. When F is unknown, since Fg cannot be calculated, a reasonable value is assumed and used. The obtained PBPK model parameters are stored in the database.

[0025] To simulate drug interactions and obtain the Ki value, fm or Fg, data on the increase ratio of AUC due to the combination of the PBPK model parameters, doses, dosing schedules of the substrate and inhibitor, and the subject are output from the database to the simulation section (S - 4).

[0026] The differential equations used in the simulation section are shown below. Substrate (drug being interacted with) Digestive tract dXa,s / dt = - Ka,s * Xa,s Circulating blood Vsys,s*dCsys,s=Qh*Ch5,s+Kps,s*Xp,s-(Vsys,s*Ksp,s+Qh)*Csys,s-CLr,s*Csys,s tip dXp,s / dt=Vsys,s*Ksp,s*Csys,s-Ksp,s*Xp,s Liver 1 0.2*Vh*dCh1,s / dt=Qh*(Csys,s-Ch1,s / Kph,s)-0.2*CLint,s*fb,s*Ch1,s / Kph,s+Fg,s*Fa,s*Ka,s*Xa,s Liver 2 0.2*Vh*dCh2,s / dt=Qh*(Ch1,s / Kph,s-Ch2,s / Kph,s)-0.2*CLint,s*fb,s*Ch2,s / Kph,s Liver 3 0.2*Vh*dCh3,s / dt= Qh*(Ch2,s / Khp,s-Ch3,s / Kph,s)-0.2*CLint,s*fb,s*Ch3,s / Kph,s Liver 4 0.2*Vh*dCh4,s / dt= Qh*(Ch3,s / Kph,s-Ch4,s / Kph,s)-0.2*CLint,s*fb,s*Ch4,s / Kph,s Liver 5 0.2*Vh*dCh5,s / dt= Qh*(Ch4,s / Kph,s-Ch5,s / Kph,s)-0.2*CLint,s*fb,s*Ch5,s / Kph,s Inhibitors (interaction agents) digestive tract dXa,i / dt=-Ka,i*Xa,i Circulating blood Vsys*dCsys,i=Qh*Ch5,i+Kphs,i*Xp,i-(Vsys,i*Ksp,i+Qh)*Csys,i-CLr,i*Csys,i tip dXp,i / dt=Vsys,i*Ksp,i*Csys,i-Ksp,i*Xp,i Liver 1 0.2 * Vh * dCh1,i / dt = Qh * (Csys,i - Ch1,i / Kph,i) - 0.2 * CLint,s * fb,s * Ch1,s / Kph,s + Fg,s * Fa,s * Ka,s * Xa,s Liver 2 0.2 * Vh * dCh2,i / dt = Qh * (Ch1,i / Kph,i - Ch2,i / Kph,i) - 0.2 * CLint,i * fb,i * Ch2,i / Kph,i Liver 3 0.2 * Vh * dCh3,i / dt = Qh * (Ch2,i / Kph,i - Ch3,s / Kph,i) - 0.2 * CLint,i * fb,i * Ch3,i / Kph,i Liver 4 0.2 * Vh * dCh4,i / dt = Qh * (Ch3,i / Kph,i - Ch4,i / Kph,i) - 0.2 * CLint,i * fb,i * Ch4,i / Kph,i Liver 5 0.2 * Vh * dCh5,i / dt = Qh * (Ch4,i / Kph,i - Ch5,i / Kph,i) - 0.2 * CLint,i * fb,i * Ch5,i / Kph,i Substrate + Inhibitor Digestive Tract dXa,s / dt = -Ka,s * Xa,s Circulating Blood Vsys * dCsys,s = Qh * Ch5,s + Kps * Xp,s - (Vsys,s * Ksp,s + Qh) * Csys,s - CLr,s * Csys,s Periphery dXp,s / dt = Vsys,s * Ksp * Csys,s - Ksp * Xp,s Liver 1 0.2 * Vh * dCh1,s / dt = Qh * (Csys,s - Ch1,s / Kph,s) - fm,s * 0.2 * CLint,s / (1 + fb,i * Ch1,i / Kph,i / Ki) * fb,s * Ch1,s / Kph,s - (1 - fm,s) * 0.2 * CLint,s * fb,s * Ch1,s / Kph,s + (1 - (1 - Fg,s) / (1 + Ka,i * Xa,i / Qen / Ki)) * Fa,s * Ka,s * Xa,s Liver 2 0.2*Vh*dCh2,s / dt= Qh*(Ch1,s / Kph,s-Ch2,s / Kph,s)-fm,s*0.2*CLint,s / (1+fb,i*Ch2,i / Kph,i / Ki)*fb,s*Ch2,s / Kph,s-(1-fm,s)*0.2*CLint,s*fb,s*Ch2,s / Kph,s Liver 3 0.2*Vh*dCh3,s / dt= Qh*(Ch2,s / Kph,s-Ch3,s / Kph,s)-fm,s*0.2*CLint,s / (1+fb,i*Ch3,i / Kph,i / Ki)*fb,s*Ch3,s / Kph,s-(1-fm,s)*0.2*CLint,s*fb,s*Ch3,s / Kph,s Liver 4 0.2*Vh*dCh4,s / dt= Qh*(Ch3,s / Kph,s-Ch4,s / Kph,s)-fm,s*0.2*CLint,s / (1+fb,i*Ch4,i / Kph,i / Ki)*fb,s*Ch4,s / Kph,s-(1-fm,s)*0.2*CLint,s*fb,s*Ch4,s / Kph,s Liver 5 0.2*Vh*dCh5,s / dt= Qh*(Ch4,s / Kph,s-Ch5,s / Kph,s)-fm,s*0.2*CLint,s / (1+fb,i*Ch5,i / Kph,i / Ki)*fb,s*Ch5,s / Kph,s-(1-fm,s)*0.2*CLint,s*fb,s*Ch5,s / Kph,s Ch is the liver concentration, and the numbers indicate the order of proximity to the liver inlet. The subscripts s and i represent substrate and inhibitor, respectively.

[0027] The target parameters to be calculated are the Ki value of the inhibitor, and fm and Fg for the substrate's metabolic enzyme. Fg is set to 1 if the substrate is not metabolized in the small intestine and is not calculated. All parameters other than the desired parameter are fixed, and only the target parameter is varied. The initial value of the target parameter is entered, and the blood concentration transition when that value is used is simulated to determine the increase in the substrate's AUC (S-5). The target parameter is varied until it reaches the AUC increase factor of the actually measured value. Methods for varying parameters to find a solution include the bisection method and the Newton-Raphson method. The obtained Ki value, fm, and Fg are stored in a database.

[0028] The PBPK model parameters for the inhibitor and substrate to be predicted, the Ki value of the inhibitor, and the fm value of the substrate are transferred from the database to the simulation section. The dose and administration schedule are entered, and the blood concentration transition and AUC increase ratio are simulated. The obtained blood concentration transition is converted to the plasma concentration transition by dividing it by the Rb value (S-6). [Example]

[0029] The present invention will be specifically described below with reference to examples, but the present invention is not limited to these examples.

[0030] This section provides an example of newly analyzing PBPK model parameters for inhibitors and substrates and registering them in a database.

[0031] For example, the reported interactions between ketoconazole and amenamevir and between ketoconazole and midazolam are shown. Ketoconazole is the inhibitor, and amenamevir and midazolam are the substrates.

[0032] Pharmacokinetic parameters after single oral administration of amenamevir, ketoconazole, and midazolam are shown in Table 1. [Table 1] Amenamevir: Clin Pharmacol Drug Dev.2019;8(5):595-602 Ketoconazole: Antimicrob Agents Chemother.1986;30(2):206-10 Midazolam:Cancer Chemother Pharmacol.2021;87(4):475-486.

[0033] The first-order absorption two-compartment model parameters obtained from the pharmacokinetic parameters are shown in Table 2. The predicted plasma concentration profiles are shown in Figure 4. [Table 2]

[0034] [Table 3] Amenamevir: Clin Pharmacol Drug Dev. 2019;8(5):595-602, Drug Interview Form Amenalif Tablets 200mg (9th Edition) Ketoconazole: Non-patent document 1 Midazolam: Goodman & Gilman Pharmacology 11th Edition, Biopharm Drug Dispos. 2010;31(5-6):286-97

[0035] The PBPK model parameters were calculated from the compartment model parameters using equations (20)-(35). Figure 4 shows the plasma concentration trends predicted using the obtained PBPK model parameters. The predictions from the PBPK model were nearly consistent with those predicted by the compartment model. Table 4 shows the PBPK model parameters. Here, the ketoconazole parameters are shown for K1 = α and K1 = Ka, with fm assumed to be 1. Kph is usually assumed to be 1, and when using predicted values, it is set under the condition Qh / Vh / Kph > 3*Ka. Qh, Qen, and Vh were set to 97 L / h, 18 L / h, and 1.5 L, respectively. [Table 4] The CLint values ​​for midazolam in the well-stirred model, parallel tube model, and dispersion model were 706 L / h, 595 L / h, and 621 L / h, respectively, and the 5-liver model closely matched the dispersion model.

[0036] <Analysis of the interaction between ketoconazole and amenamevir> When ketoconazole 400 mg was administered orally once daily for 11 days and amenamevir 400 mg was administered orally on the eighth dose, the AUC of amenamevir increased 2.58-fold (AdvTher.2017;34(11):2466-2480.).

[0037] Using the above conditions and the PBPK parameters in Table 4, the Runge-Kutta-Gill method was used to simulate the concentration profiles of ketoconazole and amenamevir in the circulating blood, liver, and peripheral compartments. The Ki value of ketoconazole resulting in a 2.58-fold increase in AUC was calculated using a dichotomy method based on the fm of CYP3A, the metabolic enzyme of amenamevir, under two conditions: K1 = α and K1 = Ka. The relationship between the resulting Ki value and fm is shown in Figure 5. The Ki values ​​calculated under both conditions depended on the fm of the substrate's metabolic enzyme, with the Ki value being lower when K1 = Ka.

[0038] <Analysis of the interaction between ketoconazole and midazolam> The reported interactions between ketoconazole and midazolam are shown in Table 5. [Table 5] The administration time of midazolam is indicated by the administration time of ketoconazole.

[0039] Using the PBPK model parameters for ketoconazole (K1 = α) and midazolam in Table 4 under the administration conditions in Table 5, the Ki value for ketoconazole against CYP3A was calculated for each set fm, at which the AUC increase factor for midazolam was the observed value, through simulation using the PBPK model. The relationship between the obtained Ki value and fm is shown in Figure 6. The geometric mean Ki value for ketoconazole when the fm for CYP3A in the metabolism of midazolam was 1 was 1.71 ng / mL.

[0040] <Analysis of Amenamevir fm> The Ki value of ketoconazole for CYP3A was 1.71 ng / mL when fm = 1 from the analysis results of the interaction study with midazolam, and the fm of CYP3A in amenamevir metabolism was calculated for each set Fg. The relationship between Fg and fm is shown in Figure 7. When Fg = 1, the fm was calculated to be 0.905. [Industrial Applicability]

[0041] This invention makes it possible to easily calculate parameters for predicting drug interactions from the pharmacokinetic parameters described in drug package inserts and academic papers and the AUC increase factor when an interaction is observed. This makes it possible to comprehensively predict the risks of concomitant use of drugs that have not undergone clinical trials, which is useful for ensuring the safe use of drugs. By assessing the risk of drug interactions in advance when developing a new drug, it is possible to determine early on whether or not to conduct drug interaction studies in new drug development.

Claims

1. A method for determining, using a computer, the inhibition constant of an interacting drug or the contribution rate of a metabolic enzyme of an interacting drug from the degree of drug interaction of the interacting drug after oral administration of the interacting drug and the interacting drug, comprising: a step of estimating first-order absorption one- or two-compartment model parameters from the pharmacokinetic parameters of the interacting drug and the interacting drug; A step of converting compartment model parameters of the interacting drug and the interacting drug into parameters of a simplified physiologically based pharmacokinetic model; determining the inhibition constant and the contribution rate of the metabolic enzyme from the fluctuation of the AUC of the interacting drug; creating a database of the pharmacokinetic parameters, compartment model parameters, physiological pharmacokinetic model parameters, inhibition constants, and metabolic enzyme contribution rates of the interacting drug and the interacting drug; a step of selecting a drug to be interacted with and a drug that interacts with the drug registered in the database, setting the dosage, administration period, and administration timing, and simulating the changes in blood and plasma concentrations and AUC of the drug to be interacted with and without the interacting drug; Forecasting methods, including:

2. 2. The method according to claim 1, wherein the first-order absorption one- or two-compartment model parameters estimated from the pharmacokinetic parameters are converted into parameters of a simplified physiologically based pharmacokinetic model.

3. The method according to claim 1, wherein the inhibition constant of the interacting drug, which indicates the AUC of the interacted drug due to clinical drug interactions, is determined by simulation using a simplified physiologically based pharmacokinetic model having a 5-liver model.

4. The method of claim 1, wherein the contribution rate of the metabolic enzyme of the interacted drug, which indicates the AUC of the interacted drug due to clinical drug interactions, is determined by simulation using a simplified physiological pharmacokinetic model having a 5-liver model.

5. A database of physiological pharmacokinetic model parameters of compounds obtained by the methods of claims 2-4, inhibition constants of inhibitors, and contribution rates of metabolic enzymes of interacting drugs.

6. A simulation device for carrying out the method according to any one of claims 1 to 4.

7. A computer program for carrying out the method according to any one of claims 1 to 4.

Images