Method for predicting drug concentration in the brain, drug administration support program, drug administration support device, and drug administration support system

Abstract

To provide a new PK / PD model which can more accurately predict the intracerebral concentration of a medicine.SOLUTION: A prediction method of the intracerebral concentration of a medicine to be continuously administered to a patient with intravenous injection includes the step of using a multi-compartment model in which a brain compartment is added to a three-compartment model containing a central compartment, a rapid equilibration compartment, and a slow equilibrium compartment to derive a predictive value of the intracerebral concentration on the basis of the product of the distribution rate constant k1e from the central compartment to the brain compartment and the distribution coefficient Kp which corresponds to the ratio of the intracerebral concentration to the plasma concentration of the medicine in the steady state. Consequently, the method can more accurately predict the intracerebral concentration of the medicine.SELECTED DRAWING: None

Description

[Technical Field]

[0001] This invention relates to a method for predicting drug concentrations in the brain, a drug administration support program, a drug administration support device, and a drug administration support system. [Background technology]

[0002] Intravenous anesthetics are used to regulate perioperative anesthesia depth and recovery using target-controlled infusion (TCI) methods programmed with pharmacokinetic and pharmacodynamic (PK / PD) models. TCI rapidly increases the anesthetic concentration to a target site concentration and maintains that concentration. Specifically, propofol is administered using this TCI method, with the goal of controlling the site concentration.

[0003] The pharmacokinetics of many drugs are represented by a multi-compartment model. The multi-compartment model is an arithmetic model for kinetically analyzing the dynamics of a drug after it has been administered to a living organism. It consists of a central compartment where the drug is injected (usually in the blood), several peripheral compartments where the drug is redistributed, and the site of action where the drug exerts its effect (effect compartment), and in addition to these, the elimination of the drug due to metabolism is also taken into account.

[0004] The pharmacokinetics of propofol are often described using a three-compartment model, specifically the Marsh model (Non-Patent Literature 1) and the Schnider model (Non-Patent Literature 2, 3), which include a central compartment, a rapid equilibration compartment (a compartment with abundant blood flow, such as muscles and internal organs), and a slow equilibration compartment (a compartment with poor blood flow, such as adipose tissue and bone).

[0005] The Marsh model is an original PK / PD model constructed from data from children (Non-Patent Literature 1), but the effect site equilibrium rate constant (k e0 ) was added later based on another study (Non-Patent Literature 4). Parameter k e0This is estimated from the electroencephalogram response to propofol (Non-Patent Literature 5). Numerous validations have been conducted on such PK / PD models, but these validations have been based primarily on measuring the plasma concentration equilibrated to the effect site concentration (Non-Patent Literature 1-3, 5-10). [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Marsh B, White M, Morton N, Kenny GNC. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991; 67: 41-8. [Non-Patent Document 2] Schnider TW, Minto CF, Gambus PL, et al. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998; 88: 1170-82. [Non-Patent Document 3] Schnider TW, Minto CF, Shafer SL, et al. The influence of age on propofol pharmacodynamics. Anesthesiology 1999; 90: 1502-16. [Non-Patent Document 4] Struys MMRF, De Smet T, Glen JIB, Vereecke HEM, Absalom AR, Schnider TW. The history of target-controlled infusion. Anesth Analg 2016; 122: 56-69. [Non-Patent Document 5] Wakeling HG, Zimmerman JB, Howell S, Glass PS. Targeting effect compartment or central compartment concentration of propofol: what predicts loss of consciousness? Anesthesiology 1999; 90: 92-7. Non-Patent Document 6 Kazama T, Ikeda K, Morita K, et al. Comparison of the effect-site keos of propofol for blood pressure and EEG bispectral index in elderly and younger patients. Anesthesiology 1999; 90: 1517-27. Non-Patent Document 7 Van Hese L, Theys T, Absalom AR, et al. Comparison of predicted and real propofol and remifentanil concentrations in plasma and brain tissue during target-controlled infusion: a prospective observational study. Anaesthesia 2020; 75: 1626-34. Non-Patent Document 8 Park JH, Choi SM, Park JH, et al. Population pharmacokinetic analysis of propofol in underweight patients under general anaesthesia. Br. J. Anaesth 2018; 121: 559-66. Non-Patent Document 9 Sahinovic MM, Eleveld DJ, Miyabe-Nishiwaki T, Struys MMRF, Absalom AR. Pharmacokinetics and pharmacodynamics of propofol: changes in patients with frontal brain tumors. Br. J. Anaesth 2017; 118: 901-9. [Non-Patent Document 10] Lee YH, Choi GH, Jung KW, et al. Predictive performance of the modified Marsh and Schnider models for propofol in underweight patients undergoing general anaesthesia using target-controlled infusion. Br. J. Anaesth 2017; 118: 883-91. [Overview of the Initiative] [Problems that the invention aims to solve]

[0007] The inventors focused on the fact that current PK / PD models do not evaluate the accuracy of effect site prediction (i.e., brain concentration). The inventors then simultaneously measured the plasma and brain concentrations of propofol in patients who underwent awake craniotomy. They found a significant positive correlation between the measured plasma concentrations and the predicted values ​​by the Marsh and Schnider models. However, the measured brain concentrations of propofol were higher than the predicted values ​​by the Marsh and Schnider models, as well as the measured plasma concentrations. In other words, this result indicated that the concentration at the site of action predicted by the Marsh and Schnider models did not reflect the actual brain concentration. However, the brain distribution rate constant used in the Marsh and Schnider models is k 1eTherefore, the predicted site-of-action concentration cannot be higher than the plasma concentration. In other words, the inventors faced the problem that the Marsh model and the Schnider model inherently have a structural flaw that prevents them from accurately reflecting brain concentrations.

[0008] Therefore, the objective of the present invention is to construct and provide a novel PK / PD model that can more accurately predict drug concentrations in the brain. [Means for solving the problem]

[0009] The inventors, considering not only steady-state brain concentration data but also sequential brain concentration data, have developed a three-compartment model based on the Marsh and Schnider models, further developing a distribution rate constant k for brain compartments. 1e The partition coefficient K corresponds to the ratio of the brain concentration to the plasma concentration in the steady state. p We constructed a modified multi-compartment model by combining these elements, and found that this new dynamic model based on measured concentrations can more accurately predict drug concentrations in the brain. This invention was completed by further investigation based on this finding.

[0010] In other words, the present invention provides inventions in the following embodiments. Item 1. A method for predicting the brain concentration of a drug continuously administered to a patient by intravenous injection. Using a multi-compartment model that adds a brain compartment to a three-compartment model including a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, Distribution rate constant k from the central compartment to the brain compartment 1e The partition coefficient K, which corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state, is also defined as follows: p A method comprising the step of deriving a predicted value of brain concentration based on the product of [a certain factor]. Item 2. The method according to Item 1, wherein the brain concentration is derived by the following formula: [Equation] In the formula, k 1e is the distribution rate constant from the central compartment to the brain compartment, K p is the distribution coefficient corresponding to the ratio of the brain concentration to the plasma concentration of the drug at a steady state, C central is the plasma concentration of the drug in the central compartment, k e0 ' is the elimination rate constant from the brain compartment, and C brain represents the brain concentration of the drug in the brain compartment. Item 3. The k 1e is a value calculated by dividing the inter-compartment clearance Q4 between the central compartment and the brain compartment by the volume V1 of the central compartment, the k e0 ' is the same value as the k 1e , The method according to Item 2. Item 4. The K p is 1.1 to 1.5, The method according to any one of Items 1 to 3. Item 5. The k 1e is 0.008 to 0.03, The method according to any one of Items 1 to 4. Item 6. The k 1e and the K p value have a product of 0.01 to 0.04, The method according to any one of Items 1 to 5. Item 7. The drug is selected from the group consisting of central nervous system active agents, The method according to any one of Items 1 to 6. Item 8. An auxiliary device for assisting administration of a drug that is continuously administered to a patient by intravenous injection, comprising a prediction unit for calculating a predicted value of the brain concentration of the drug, The prediction unit uses a multi-compartment model in which a brain compartment is added to a three-compartment model that includes a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, and the distribution rate constant k from the central compartment to the brain compartment is used. 1e The partition coefficient K, which corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state, is also defined as follows: p An auxiliary device that derives a predicted value of brain concentration based on the product of [two factors]. Item 9. The auxiliary apparatus according to item 8, further comprising a display processing unit for displaying the predicted value and / or the target concentration of the drug. Item 10. The auxiliary device according to item 8 or 9, further comprising a calculation unit for calculating the dosage of the drug based on the predicted value and the target concentration of the drug. Item 11. The auxiliary device according to item 10, further comprising a dispensing control unit for controlling the dosage of a drug to the aforementioned dosage. Item 12. A drug administration assistance system comprising an auxiliary device described in any of items 8 to 11 and a dispensing device for administering the drug to the patient. Item 13. A drug administration support program to assist in the administration of drugs continuously administered to patients by intravenous injection. Computers, A multi-compartment model is used in which a brain compartment is added to a three-compartment model that includes a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, and the distribution rate constant k from the central compartment to the brain compartment is used. 1e The partition coefficient K, which corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state, is also defined as follows: p A program designed to function as a predictor, deriving predicted values ​​of brain concentration based on the product of two factors. [Effects of the Invention]

[0011] According to the present invention, a novel PK / PD model is provided that can more accurately predict drug concentrations in the brain. [Brief explanation of the drawing]

[0012] [Figure 1] This document shows a modified multi-compartment model used in the method for predicting the brain concentration of drugs according to the present invention. [Figure 2] A block diagram of one embodiment of the drug administration assist device of the present invention is shown. [Figure 3] This is a flowchart illustrating the operation of one embodiment of the drug administration support system of the present invention. [Figure 4] The plasma concentration (A), brain concentration (B), and Kp value (C) of propofol after intravenous infusion are shown. Propofol concentrations were measured in all 57 patients. For each patient, plasma and brain samples (5-30 mg) were collected at up to five time points (immediately before propofol administration, and 5, 15, 25, and 75 minutes after discontinuation of administration). The Kp value is the ratio of brain concentration to plasma concentration. Thin lines represent data for each patient, and thick lines represent the average value. [Figure 5] This shows the correlation between measured and predicted plasma and brain propofol concentrations. It presents 157 plasma concentrations and 112 brain concentrations from 57 patients. The correlation between measured plasma and brain concentrations at all points is shown (A). Predicted concentrations based on the Marsh model (B) and the Schnider model (C) are shown alongside measured plasma and brain concentrations. The solid line represents the regression line. [Figure 6] The goodness-of-fit plot for propofol in the model of the present invention is shown. Measured concentration versus population predictor (PRED) and measured concentration versus individual predictor (IPRED) are shown in (A) and (B), respectively. PRED versus conditionally weighted residual (CWRES) and time after discontinuation of propofol administration versus conditionally weighted residual (CWRES) are shown in (C) and (D), respectively. The blank circles represent data from patients in the model construction population. They show 106 plasma concentrations and 80 brain concentrations from 29 patients. The solid line represents the regression line. [Figure 7]This shows a comparison of measured and predicted propofol concentrations in the validation population. The measured values ​​are shown against the concentrations predicted by the model of the present invention. Blank circles represent data from patients in the validation population. It shows 51 plasma concentrations and 32 brain concentrations from 28 patients. The solid line represents the regression line. [Figure 8] This shows a prediction-corrected visual predictive check (pcVPC) plot using predicted plasma and brain concentrations of propofol. Circles represent measured concentrations. Solid lines represent the 5th, 50th, and 95th percentiles of the measured concentrations. Dashed lines show the 5th, 50th, and 95th percentiles of the predicted concentrations based on the model of the present invention. Shaded areas show the 95th, 50th, and 5th percentile 95% confidence intervals based on simulations. [Figure 9] This shows the predicted propofol concentrations in a given case using the Marsh model (A), the Schnider model (B), and the model of the present invention (C). It also shows the predicted plasma and brain concentrations for one patient in the validation population based on the Marsh model (A), the Schnider model (B), and the novel model (C). The dotted and dashed lines represent the predicted plasma and brain concentrations, respectively. The solid line represents the actual propofol administration rate (mg / h). [Figure 10] This shows a conventional three-compartment model known as the Marsh model or Schnider. [Modes for carrying out the invention]

[0013] 1. Methods for predicting drug concentrations in the brain The present invention provides a method for predicting the brain concentration of a drug administered to a patient continuously by intravenous injection, and is characterized by including a step of deriving a predicted brain concentration value using a predetermined modified multi-compartment model.

[0014] Modified multi-compartment model The modified multi-compartment model used in this invention is based on a three-compartment model, with an additional brain compartment. A schematic diagram of this modified multi-compartment model is shown in Figure 1.

[0015] The underlying three-compartment model consists of a central compartment, a rapid equilibration compartment, and a slow equilibration compartment, and refers to the three-compartment model used in the model known as the Marsh model or Schnider, as shown in the schematic diagram in Figure 10.

[0016] In Figures 1 and 10, I represents the central compartment. II represents the rapid equilibration compartment, III represents the low-speed balancing compartment. Figure IV (10) represents the effect compartment. V (Figure 1) represents brain compartments, C central This represents the plasma concentration of the drug in the central compartment. C RD This represents the concentration of the drug in the rapid equilibration compartment (in blood-rich tissues such as muscles and internal organs), C SD This represents the concentration of the drug in the slow equilibration compartment (in tissues with poor blood flow, such as adipose tissue and bone). C brain (Figure 1) shows the brain concentration of drugs in brain compartments. C e (Figure 10) shows the brain concentration of the drug in the effect compartment. V1 represents the volume of the central compartment. V2 represents the volume of the rapid equilibration compartment. V3 represents the volume of the low-speed equilibrium compartment. Dose refers to a drug that is continuously administered to a patient by intravenous injection. k 12This represents the vanishing constant that describes the distribution of volume V1 in the direction of volume V2, k 21 This represents the vanishing constant that describes the distribution of volume V2 in the direction of volume V1, k 13 This represents the vanishing constant that describes the distribution of volume V1 in the direction of volume V3, k 31 This represents the vanishing constant that describes the distribution of volume V3 in the direction of volume V1, k 10 This represents the elimination constant of drugs from the body, k 1e (Figure 1) shows the distribution rate constant from the central compartment to the brain compartment. K p (Figure 1) shows the distribution coefficient, which corresponds to the ratio of the concentration of a drug in the brain to the concentration in the plasma at steady state. k e0 (Figure 1) shows the rate constant of disappearance from brain compartments. k 1e (Figure 10) shows the distribution rate constant from the central compartment to the effect compartment. k e0 (Figure 10) shows the proportional change in the concentration gradient between plasma and the effect site over each unit of time.

[0017] In other words, the conventional 3-compartment model shown in Figure 10 has a velocity constant k 1e While only is considered and brain concentration is predicted in the effect compartment, the modified compartment model shown in Figure 1 uses the rate constant k 1e In addition, the partition coefficient K, which corresponds to the ratio of the brain concentration to the plasma concentration of the drug at steady state. p The model configuration has been modified to take this into consideration, in which brain concentration is predicted in the brain compartment. In the conventional 3-compartment model shown in Figure 10, plasma concentration and brain concentration are inevitably almost the same value at all time points, whereas in the modified compartment model shown in Figure 1, the distribution coefficient K pBy including this component, the modified compartment model more accurately reflects the actual pharmacokinetics, where brain concentration changes with a delay compared to changes in plasma concentration (drug transfer from the central compartment to the brain compartment is sequential). As a result, the predictive ability of the modified compartment model to predict brain concentration is improved compared to the conventional three-compartment model.

[0018] Based on the modified compartment model in Figure 1, the drug follows the following path: After intravenous infusion, the drug rapidly reaches central compartment I by circulation, and then undergoes tissue-specific redistribution (constant k) from central compartment I to rapid equilibration compartment II (compartments with abundant blood flow, such as muscles and internal organs) and slow equilibration compartment III (compartments with poor blood flow, such as adipose tissue and bone). 12 , k 13 ) occurs, and the opposite (constant k 21 , k 31 ) also occurs. Drugs that reach central compartment I reach brain compartment V with a time lag. The parameter that reflects this drug movement is the distribution rate constant k from central compartment I to brain compartment V. 1e The partition coefficient K corresponds to the ratio of the concentration of a drug in the brain to the concentration in the plasma at a steady state. p The product of k 1e *K p ". Note that a steady state refers to a state in which equilibrium has been reached between each compartment. Furthermore, the body has a specific vanishing rate (constant k 10 ) Remove the drug from central compartment I.

[0019] In the prediction method of the present invention, the above product "k" 1e *K p Based on this, a predicted value of brain concentration is derived. A specific method for deriving brain concentration is the derivation method using the following formula.

number

[0020] k 1e This can be calculated by dividing the compartment clearance Q4 between the central compartment I and the brain compartment V by the volume V1 of the central compartment I. e0 ' is k 1e It can be set to the same value as . Furthermore, the density of brain tissue can be, for example, 0.95 to 1 It can be set between 0.05 g / mL and 0.05 g / mL.

[0021] In the prediction method of the present invention, k 1e and K p For the specific values ​​of each parameter except for one, the values ​​used in the conventional 3-compartment model can be applied. central This can be derived using the same method as the conventional three-compartment model, with the volumes of the central compartment I, the rapid equilibration compartment II, and the slow equilibration compartment III, as well as various parameters related to each compartment.

[0022] On the other hand, in the prediction method of the present invention, K p As for specific values, for example, 1.1 to 1.5, preferably 1.1 to 1.4, more preferably 1.2 to 1.35, and even more preferably 1.28 to 1.32 can be cited. k in the prediction method of the present invention 1e Specific values ​​include, for example, 0.008 to 0.03, preferably 0.01 to 0.02, and more preferably 0.012 to 0.016. 1e *K pSpecific values ​​for " include, for example, 0.01 to 0.04, preferably 0.012 to 0.03, more preferably 0.014 to 0.025, and even more preferably 0.016 to 0.02.

[0023] drugs The drugs to which the prediction method of the present invention is applied are not particularly limited, as long as they are drugs administered continuously to patients by intravenous injection. Since the modified multi-compartment model used in the prediction method of the present invention is configured to include a brain compartment, drugs that target the brain as their site of action are used. Furthermore, since the prediction method of the present invention uses a modified multi-compartment model based on the conventional three-compartment model, lipid-soluble drugs whose tissue concentration at the site of action tends to gradually increase in response to an increase in plasma concentration are typically used. Specific examples of drugs include central nervous system activators, and more specifically, anesthetics, antiepileptics, hypnotics, analgesics, etc.

[0024] Specific examples of anesthetics include propofol (Diprivan), midazolam (Dormicum), remimazolam (Anelem), remifentanil (Urtiva), and fentanil. These anesthetics may be used individually or in combination. Among these anesthetics, propofol is preferred.

[0025] Specific examples of antiepileptic drugs include phenytoin, perampanel, levetiracetam, and lacosamide.

[0026] patient The patients to whom the prediction method of the present invention is applied are not particularly limited as long as they are animals with a brain, for example, vertebrates, preferably mammals, more preferably humans, horses, cows, monkeys, pigs, dogs, cats, rabbits, foxes, raccoons, fennec foxes, ferrets, hamsters, guinea pigs, rats, mice, squirrels, etc., and particularly preferably humans.

[0027] 2. Drug administration support devices The drug administration assist device of the present invention is an assist device for assisting the administration of a drug continuously administered to a patient by intravenous injection, and includes a prediction unit for calculating a predicted value of the brain concentration of the drug, the prediction unit using a multi-compartment model in which a brain compartment is further added to a three-compartment model including a central compartment, a rapid equilibration compartment, and a slow equilibration compartment, and the distribution rate constant k from the central compartment to the brain compartment 1e The partition coefficient K, which corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state, is also defined as follows: p Based on the product of these two factors, a predicted value for brain concentration is derived.

[0028] Figure 2 shows a block diagram representing the configuration of one embodiment (drug administration assist device 1) of the drug administration assist device of the present invention. Drug administration assist device 1 includes a brain concentration prediction unit 101 as a prediction unit for calculating predicted values ​​of drug concentration in the brain. Except for the brain concentration prediction unit 101 shown in Figure 2, the components are optional components of the drug administration assist device of the present invention, and in the drug administration assist device of the present invention, one or more of these optional components may be combined with the brain concentration prediction unit 101 as appropriate. Drug administration assist device 1 will be described in detail below.

[0029] The drug administration assist device 1 comprises a control unit 10, which is composed of a CPU and the like, controls each part and executes various programs; a storage unit 11, which is composed of a storage medium such as an HDD and flash memory and stores information; an operation unit 12, which outputs operation signals to the control unit 10 according to the operation content; a display unit 13, which displays information using images and / or characters; and a communication unit 14, which communicates with an external device.

[0030] By executing the drug administration program 110, the control unit 10 functions as at least a brain concentration prediction unit 101, and in Figure 2, it further functions as a plasma concentration prediction unit 100, a dosage calculation unit 102, a drug dispenser control unit 103, and a display processing unit 104.

[0031] The plasma concentration prediction unit 100 preferably derives a predicted value of the plasma concentration of a drug from a plasma concentration prediction setting value 112 by executing a drug administration support program 110, and records the derived plasma concentration prediction value along with the time in a plasma concentration prediction value dataset 114.

[0032] The brain concentration prediction unit 101, by executing the drug administration support program 110, derives a predicted brain concentration of the drug from the brain concentration prediction setting value 113 and the plasma concentration prediction value dataset 114, and records the derived brain concentration prediction value along with the time in the brain concentration prediction value dataset 115.

[0033] The dosage calculation unit 102 calculates the dosage based on the target concentration setting value 111 and the predicted brain concentration value from the brain concentration prediction value dataset 115, and records it in the dosage dataset 116 along with the time. The dosage is usually calculated as the flow rate of the drug.

[0034] The dispensing device control unit 103 instructs the communication unit 14 to control the amount of drug administered from the dispensing device outside the drug administration support device 1 to match the amount from the dosage data set 116.

[0035] The display processing unit 104 processes the target concentration setting value 111 and / or the predicted brain concentration value from the brain concentration prediction value dataset 115, and / or the dosage from the dosage dataset 116, and displays them on the display unit 13 in real time or including history. Display methods include numerical values, graphs, colors, and combinations thereof.

[0036] The memory unit 11 stores a drug administration support program 110 that causes the control unit 10 to operate as at least the brain concentration prediction unit 101 described above, and in Figure 2, as a dosage calculation unit 102, a drug dispenser control unit 103, and a display processing unit 104, as well as a target concentration setting value 111, a setting value for brain concentration prediction 113, a brain concentration prediction value dataset 115, a dosage dataset 116, etc.

[0037] The drug administration support program 110 is a program designed to at least implement the prediction method described in "1. Method for predicting drug concentration in the brain" above.

[0038] The target concentration setting of 111 is the drug concentration set by the doctor according to the patient's weight, age, etc., and specifically refers to the drug concentration in the brain.

[0039] The plasma concentration prediction setting value 112 is preferably a value of various parameters used to predict plasma concentration in the execution of the drug administration support program 110, and specifically, as described in the "modified multi-compartment model" in "1. Method for predicting drug concentration in the brain" above, k 1e , K p and k e0 Other parameters and values ​​can be listed. The prediction setting value 112 may have at least some commonality with the brain concentration prediction setting value 113.

[0040] The setting values ​​113 for predicting brain concentration are values ​​of various parameters used for predicting brain concentration in the execution of the drug administration support program 110. Specifically, these are values ​​of various parameters used to execute the following formula, as described in the "Modified Multi-Compartment Model" in "1. Method for Predicting Brain Concentration of Drugs" above. Of these values, C central This is the same as the predicted plasma concentration recorded in the predicted plasma concentration dataset 114.

number

[0041] The plasma concentration prediction dataset 114 consists of predicted values ​​derived by the plasma concentration prediction unit 100, linked to time.

[0042] The brain concentration prediction dataset 115 consists of predicted values ​​derived by the brain concentration prediction unit 101, linked to time.

[0043] The dosage dataset 116 contains dosages calculated by the dosage calculation unit 102, linked to specific times.

[0044] 3. Drug administration support system The drug administration support system of the present invention includes a drug administration support device and a drug dispenser for administering a drug to a patient. The drug dispenser may be equipped with a pump for intravenous drug injection, preferably a syringe pump. The drug administration support device is described in detail in "2. Drug Administration Support Device" above. The drug dispenser is operated via the communication unit 14 according to the control contents of the drug dispenser control unit 103 described in "2. Drug Administration Support Device" above. The drug is described in detail in "1. Method for Predicting the Brain Concentration of a Drug" above. Furthermore, the drug administration support system of the present invention may also include various devices that can be used during drug administration, such as a ventilator and a vital signs monitor.

[0045] Figure 3 shows a flowchart illustrating the operation of one embodiment of the drug administration support system of the present invention.

[0046] In one embodiment of the drug administration support system of the present invention, first, the drug administration support device is activated, and drug administration is started by the dispenser via the communication unit (14 in Figure 2) (S1). The plasma concentration prediction unit (100 in Figure 2) of the drug administration support device predicts the plasma concentration C of the drug. central The predicted value is derived (S2), and furthermore, the brain concentration prediction unit of the drug administration assist device (101 in Figure 2) determines the distribution rate constant k 1e Distribution coefficient K p , the rate constant of disappearance from brain compartments k e0 , the plasma concentration of the drug C central Based on the predicted value, the drug concentration in the brain C brain Predicted values Derived (S3). The dosage calculation unit of the drug administration support device (102 in Figure 2) calculates the target concentration setting value and the brain concentration C of the drug. brainBased on the predicted value, the dosage is derived (S4). If the dosage determined in S4 is consistent with continuing administration (the derived dosage is greater than zero), the drug administration support device's injector control unit (103 in Figure 2) controls the injector via the communication unit (14 in Figure 2) to continue administration at that dosage, and the process returns to S2. If the dosage determined in S4 is inconsistent with continuing administration (the derived dosage is zero), the drug administration support device's injector control unit (103 in Figure 2) controls the injector via the communication unit (14 in Figure 2) to stop administration (S6), and the drug administration is terminated.

[0047] 4. Drug administration support program The drug administration support program of the present invention is a program that causes a computer to execute the above-described "1. Method for predicting the concentration of a drug in the brain".

[0048] In other words, the drug administration support program of the present invention is a drug administration support program for assisting the administration of drugs that are continuously administered to patients by intravenous injection. Computers, A multi-compartment model is used in which a brain compartment is added to a three-compartment model that includes a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, and the distribution rate constant k from the central compartment to the brain compartment is used. 1e The partition coefficient K, which corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state, is also defined as follows: p This program functions as a prediction unit that derives predicted brain concentration values ​​based on the product of [two factors].

[0049] The drug administration assistance program of the present invention can be used to operate the drug administration assistance device described in "2. Drug Administration Assistance Device" above. In the embodiment shown in Figure 2, the operation of each part constituting the control unit 10 of the drug administration assistance device 1 is realized by the drug administration assistance program 110 installed in the storage unit 11, but this may also be realized by hardware such as an ASIC. The drug administration assistance program of the present invention may be provided in a state stored on a recording medium such as a CD-ROM, may be provided by distribution via the Internet, or may be a program that operates on the cloud. [Examples]

[0050] The present invention will be described in more detail below with reference to examples, but the present invention is not limited to these examples.

[0051] method Study design and target patients Between May 2019 and June 2022, we recruited 29 patients with gliomas and 28 patients with other diseases (epilepsy, cavernous hemangioma, tuberculoma) to undergo awake craniotomy at Kyoto University Hospital. For each patient, plasma and brain samples (5-30 mg) were collected at up to five time points: immediately before propofol administration and at 5, 15, 25, and 75 minutes after discontinuation of propofol administration to induce wakefulness. For the lobectomy area around the brain tumor and the uncapping margin, normal brain samples showing normal brain signals on preoperative magnetic resonance imaging (MRI) were used.

[0052] The following data were collected for each patient: diagnosis, sex, age, propofol dosage, administration and sampling times, weight, height, body mass index (BMI), aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin, albumin levels, and creatinine clearance (Ccr). Ccr was calculated using serum creatinine with the Cockcroft-Gault equation. Data on brain tumors and epileptic lesions were also collected.

[0053] Sample preparation Plasma and brain samples were stored at -80°C until measurement. Brain samples were weighed, homogenized with 20 times the amount of physiological saline, and then used as samples. Plasma and homogenized solutions were used.

[0054] Calibration curves were prepared by adding appropriate amounts of standard solutions to pooled plasma to obtain calibration curve points corresponding to propofol concentrations of 0.5, 1, 3, 5, 7, 10, 15, 20, 25, and 30 mg / mL. Next, 50 μL of the sample was vortexed with 150 μL of acetonitrile containing acetic acid (final concentration of acetic acid: 0.1 v / v%), followed by centrifugation. The supernatant obtained after centrifugation was subjected to high-performance liquid chromatography (HPLC). Quantitative analysis was performed using a Prominence HPLC system (Shimadzu Corporation, Kyoto, Japan). Chromatographic separation was performed using a Chemco Pak Liquid Chromatography Column (150 mm × 4.6 mm I.D.) (Chemco Plus Scientific Co., Ltd., Osaka, Japan) maintained at 40°C. The mobile phase consisted of water containing acetic acid (35:0.1, volume ratio) (mobile phase A) and acetonitrile (mobile phase B). The flow rate was maintained at 1.0 mL / min. The detection wavelengths were 276 nm for excitation and 310 nm for fluorescence. The analysis time was 15 minutes, and the injection volume was 2 μL. LC solution software (Shimadzu Corporation) was used for quantitative analysis of all data.

[0055] Population pharmacokinetic (PPK) modeling Population pharmacokinetic (PPK) analysis was performed using the nonlinear mixed-effects modeling program NONMEM version 7.5.0 (ICON, Ellicott City, MD) with the first-order conditional estimation method (FOCE) involving interactions. For structural PK models, a three-compartment model and an effects-compartment model were used. An exponential error model was used for inter-individual variability. For intra-individual variability, proportional error models, additive error models, and mixed (proportional-additive) error models were used to compare intra-individual residual variability in serum or brain concentrations.

[0056] In addition, the PK parameters of the brain compartment were also estimated. The system of differential equations representing the distribution of the drug in the brain compartment is as follows. [Equation] In the formula, k 1e is the distribution rate constant from the central compartment to the brain compartment, Kp is the distribution coefficient corresponding to the ratio of the brain concentration to the plasma concentration of propofol at steady state, C central is the plasma concentration of propofol in the central compartment, k e0 ' is the elimination rate constant from the brain compart ment, and C brain represents the brain concentration of propofol in the brain compartment. Here, the k 1e value was calculated by dividing the inter-compartmental clearance (Q4) between the central compartment and the brain compartment by the volume (V1) of the central compartment, the k e0 ' value was set to the same value as k 1e , and the density of brain tissue was assumed to be 1 g / mL.

[0057] In order to evaluate the effects of continuous covariates such as height, body weight, BMI, AST, and ALT, covariate analysis was performed using the following formula. [Equation] In the formula, θ1 is the mean parameter to be estimated, and θ2 is the factor contributed by the covariate. COV is the value of each patient, COV median is the median of the clinical data.

[0058] In order to evaluate the effect of sex, covariate analysis was performed using the following formula. [Equation] In the formula, here, sex is set to 1 for female and 0 for male, and θ3 is the factor contributed by sex.

[0059] The impact of each covariate on Q4 was evaluated using the variable augmentation method, specifically the difference in objective function (OBJ) between the base model and the model including the covariates. This was later confirmed using the variable reduction method. In the variable augmentation method, a p-value of 0.05 or less (ΔOBJ > 3.84 with 1 degree of freedom assuming a chi-squared distribution) was considered statistically significant, while in the variable reduction method, a p-value of 0.005 or less (ΔOBJ > 7.88 with 1 degree of freedom assuming a chi-squared distribution) was considered statistically significant. Individual predicted PK parameters and concentrations were obtained using empirical Bayesian estimation in the FOCE method.

[0060] Model evaluation To evaluate model fit, the following diagnostic plots were used: measured concentration vs. population predictive value (PRED) or individual predictive value (IPRED), conditionally weighted residuals (CWRES) vs. PRED, or time after propofol cessation to identify bias corresponding to model misidentification. A predictive corrected visual prediction test (pcVPC) was performed to assess the median trend and variability of observed data predicted by the final model. A dataset was created by repeating the original dataset 1000 times using non-memory random resampling. The median and 90% prediction interval (PI) of the simulated concentration were plotted against the observed concentration. The accuracy of the final model parameter estimates was evaluated using asymptotic standard errors and nonparametric bootstraps. Bootstrapping was performed using 500 randomly sampled (with substitution) replicas from the 29 subjects in the original dataset. Model parameters were estimated for each bootstrap replica using non-memory.

[0061] Furthermore, the validity of the constructed model was evaluated using validation data that was not used in model construction. Specifically, for each patient in the validation cohort, the plasma and brain concentrations of propofol were simulated and compared with the observed concentrations. In addition, the mean prediction error (MPE) and root mean square error (RMSE) were calculated using the initial plasma or brain concentration sample for each patient, and the bias and accuracy of the predictions were estimated as follows.

number

[0062] result Patient characteristics 106 plasma concentration samples and 80 brain concentration samples from 29 patients were used for model building analysis. Patient characteristics are summarized in Table 1. In the table, each value represents either a numerical value or a median (range). P-values ​​were calculated using the Mann-Whitney U test for continuous values ​​and Fisher's exact test for categorical values. Glial glandular neuropathy Patients other than those with tumors (others) were diagnosed with epilepsy, cavernous hemangioma, and tuberculoma. BMI (body mass index); AST (aspartate aminotransferase); ALT (alanine aminotransferase); ALB (albumin); T-Bil (total bilirubin); Ccr (creatinine clearance).

[0063] Fifty-one plasma concentration samples and 32 brain concentration samples from 28 patients were used to validate the model. Diagnoses included glioma in 53 cases and other diagnoses in 4 cases. In the model-building data, 19 patients (65.5%) and 10 patients (34.5%) were adults (13-59 years) and elderly (60 years and older), respectively. There were no significant differences in patient characteristics between the groups.

[0064] [Table 1]

[0065] Measured and predicted values ​​of propofol plasma and brain concentrations. Figure 4 shows the plasma and brain concentrations measured in 57 subjects after discontinuation of propofol administration. Plasma and brain concentrations gradually decreased over time. The ratio of brain concentration to plasma concentration, K, is also shown. pThe values tended to increase.

[0066] The relationship between the measured concentration and the predicted concentration of propofol in 57 patients is shown in Figure 5. The measured brain concentration values were higher than the measured plasma concentration values. There was a good correlation between the measured plasma concentration values and the predicted values based on the Marsh model and the Schnider model. However, regarding the brain concentration, the correlation between the predicted concentration and the measured concentration was low, and the measured values were higher than the predicted values.

[0067] Modeling of population pharmacokinetics (PPK) The constructed model structure is shown in Figure 1. As shown in Figure 5, since the plasma concentration of propofol was accurately predicted by the Marsh model, k 10 , k 12 , k 13 , k 21 , k 31 , and the mean value of V1 were fixed at the mean parameter values used in the Marsh model (Fixed), and the parameters related to the brain compartment as well as the intra-individual and inter-individual variabilities were estimated. Regarding the residual variability, the proportional error model for plasma concentration and the additive error model for brain concentration showed better model fitting. For the predicted brain concentration of propofol, the model fitting performance was improved by including the ratio of the k p value to the K 1e value.

[0068] The final estimated values and 95% CI of the propofol PPK parameters are shown in Table 2 (RSE: relative standard error, Shrinkage: an index related to model accuracy and individual information volume). Among all the covariates examined, no improvement in the predictive performance of the model by adding covariates was found. Note that k 1e in the Marsh model is 0.260, and k 1e in the Schnider model is 0.456.

[0069] [Table 2]

[0070] Model evaluation Figure 6 shows the goodness-of-fit plot of the model of the present invention. The plots of the model's predicted concentrations and the measured concentrations of PRED and IPRED lie on the same line, indicating good agreement between the two. Furthermore, CWRES was evenly distributed around zero with respect to the time after discontinuation of PRED and propofol. This indicates that the model of the present invention appropriately and unbiasedly represents the variability.

[0071] Figure 7 shows a comparison of measured concentrations and predicted concentrations using the model of the present invention in the validation population. Figure 7 shows that the regression line approximates a straight line where the predicted value equals the measured value. Table 3 shows a comparison of the MPE values ​​of plasma and brain concentrations of propofol in the validation data using the model of the present invention and the Marsh model. As shown in Table 3, the model of the present invention was shown to have superior predictive capabilities.

[0072] [Table 3]

[0073] The pcVPC plot (Figure 8) showed that the plasma and brain concentrations of propofol were in good agreement with the concentrations simulated by the model. Furthermore, the bootstrap replication values ​​and the median of the final parameter estimates were similar, indicating that the final parameters were appropriately estimated (Table 1). These bootstrap statistics and the pcVPC plot support the accurate performance of the model of the present invention.

[0074] Furthermore, using data from one patient, the predicted concentrations of propofol according to the Marsh model, the Schnider model, and the model of the present invention are shown in Figure 9. As shown in Figure 9, in the Marsh model and the Schnider model, plasma concentrations and brain concentrations are almost the same at all time points. However, according to the model of the present invention, it was confirmed that the brain concentration changes with a delay compared to the change in plasma concentration (drug transfer from the central compartment to the brain compartment is sequential), which more accurately reflects the actual pharmacokinetics.

Claims

1. A method for predicting the brain concentration of a drug, performed by an auxiliary device for assisting the administration of a drug continuously given to a patient by intravenous injection, Using a multi-compartment model that adds a brain compartment to a three-compartment model including a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, Distribution rate constant k from the central compartment to the brain compartment 1e The distribution coefficient K corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state. p A method comprising the step of deriving a predicted value of brain concentration based on the product of the following: [Math 1] During the ceremony, k 1e is the distribution rate constant from the central compartment to the brain compartment. Kp is the partition coefficient corresponding to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state. C central is the plasma concentration of the drug in the central compartment. k e0' is the rate constant of disappearance from the brain compartment, and C brain represents the brain concentration of the drug in the brain compartment.

2. The aforementioned k 1e However, the compartment clearance Q between the central compartment and the brain compartment 4 The volume V of the central compartment 1 It is a value calculated by dividing by, The aforementioned k e0 ' is the aforementioned k 1e The method according to claim 1, which has the same value.

3. The aforementioned K p The method according to claim 1, wherein is 1.1 to 1.

5.

4. said k 1e is 0.008 to 0.03, the method according to claim 1.

5. The aforementioned k 1e and the aforementioned K p The method according to claim 1, wherein the product of is 0.01 to 0.

04.

6. The method according to claim 1, wherein the drug is a central nervous system activator.

7. This is an auxiliary device for assisting the administration of drugs that are continuously given to patients by intravenous injection. It includes a prediction unit for calculating a predicted value of the brain concentration of the drug, The prediction unit uses a multi-compartment model in which a brain compartment is added to a three-compartment model that includes a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, and the distribution rate constant k from the central compartment to the brain compartment is k. 1e The distribution coefficient K corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state. p Based on the product of the following, the following formula is used to derive a predicted value of brain concentration: (Auxiliary device) [Math 2] During the ceremony, k 1e is the distribution rate constant from the central compartment to the brain compartment. Kp is the partition coefficient corresponding to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state. C central is the plasma concentration of the drug in the central compartment. k e0' is the rate constant of disappearance from the brain compartment, and C brain represents the brain concentration of the drug in the brain compartment.

8. The auxiliary device according to claim 7, further comprising a display processing unit for displaying the predicted value and / or the target concentration of the drug.

9. The auxiliary device according to claim 7, further comprising a calculation unit that calculates the dosage of the drug based on the predicted value and the target concentration of the drug.

10. The auxiliary device according to claim 9, further comprising a dispensing control unit that controls the amount of drug administered to the amount calculated by the calculation unit.

11. A drug administration assistance system comprising the auxiliary device described in claim 7 and a drug dispenser for administering the drug to the patient.

12. This is a drug administration support program designed to assist in the administration of drugs continuously given to patients via intravenous injection. Computers, A multi-compartment model is used in which a brain compartment is added to a three-compartment model that includes a central compartment, a rapid equilibrium compartment, and a slow equilibrium compartment, and the distribution rate constant k from the central compartment to the brain compartment is used. 1e The distribution coefficient K corresponds to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state. p Based on the product of the following, the following program functions as a prediction unit to derive predicted values ​​of brain concentration: [Math 3] During the ceremony, k 1e is the distribution rate constant from the central compartment to the brain compartment. Kp is the partition coefficient corresponding to the ratio of the concentration of the drug in the brain to the concentration in the plasma under steady state. C central is the plasma concentration of the drug in the central compartment. k e0' is the rate constant of disappearance from the brain compartment, and C brain represents the brain concentration of the drug in the brain compartment.

Images