WO2008099356A2

WO2008099356A2 - Drug delivery device

Info

Publication number: WO2008099356A2
Application number: PCT/IB2008/050526
Authority: WO
Inventors: Hendrika C. Krijnsen; Anna-Maria Janner; Ventzeslav P. Iordanov; Michel P. B. Van Bruggen
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2007-02-15
Filing date: 2008-02-13
Publication date: 2008-08-21
Also published as: WO2008099356A3

Abstract

The invention relates to a method and a device(10) for supplying a substance (e.g. a drug) to a physiological system(1) (e.g. a patient) during a predetermined time period (e.g. 24 hours). The amounts(u) of substance that have to be administered are determined in such a way that the predicted concentration of the substance, which is predicted with a pharmacokinetic model, matches a target concentration(c tar ) within the predetermined time period according to some optimization criterion. The optimization may particularly minimize the overall difference between target and predicted concentration during the time period. In a preferred embodiment, parameters of the pharmacokinetic model are individually adjusted to the associated patient.

Description

Drug delivery device

FIELD OF THE INVENTION

The invention relates to a drug delivery device and a method for supplying a substance to a physiological system, particularly to a patient, during a predetermined time period.

BACKGROUND OF THE INVENTION

The US 2006/0009729 Al discloses a device for delivering a medical fluid continuously with a controlled rate. The device comprises inter-alia a patient model that estimates the current concentration of the medical fluid in the patient based on the previously applied dosages. The difference between the estimated concentration and a target concentration is then provided as input to a controller. The device further comprises a second patient model that extrapolates the concentration of the medical fluid. If this model predicts that a threshold concentration will be exceeded, further dosage can immediately be stopped as a safety measure.

SUMMARY OF THE INVENTION Based on this background it was an object of the present invention to provide means for an improved delivery of substances, for example drugs, to a physiological system.

This object is achieved by a drug delivery device according to claim 1 and a method according to claim 6. Preferred embodiments are disclosed in the dependent claims.

The drug delivery device according to the present invention serves for supplying a substance, for example a medical drug, to a physiological system during a predetermined time period. The physiological system may particularly be a human or animal body or some in-vitro system, and the predetermined time period during which the substance is applied is typically longer than about one minute, preferably longer than 10 minutes, most preferably longer than 30 minutes. In many cases the predetermined time period is just one cycle of a rhythm with repeated drug applications. In cases of a chronic drug application, the considered time period is usually 24 hours and repeated every day. The drug delivery device comprises the following components: a) A "target module" for providing a "target concentration" of the substance in the physiological system over the predetermined time period. The target module may typically be realized by some data processing unit in combination with appropriate software. In the most simple case, the target concentration may be a constant value.

For longer time periods, the target concentration is however typically a curve that varies according to some rhythm of the physiological system. Moreover, the target concentration may be determined in a feedforward manner, for example be taken from a lookup table, or in a feedback manner comprising the processing of measured physiological parameters and/or external inputs of the patient and/or physician.

b) A "modeling module" for determining at least one amount of substance that is to be delivered to the physiological system during the predetermined time period. This determination is done with the help of a pharmacokinetic model of the physiological system in such a way that the "predicted concentration" of the substance in the physiological system, which is predicted by said pharmacokinetic model, optimally matches the aforementioned "target concentration" (that is provided by the target module) with respect to a given optimization criterion. The pharmacokinetic model may be selected from a wide variety of models known in the art according to the required accuracy and the available processing power. A pharmacokinetic model that is suited for modeling the concentration of a substance in the blood volume of a patient is for example a one-, two- or three-compartment model as it is described in literature (e.g. Clinical Pharmacokinetics, First edition, Edited by Soraya Dhillon, ISBN 978 0 85369 571 4 (2006)).

c) An applicator for delivering within the predetermined time period the aforementioned determined amount of substance to the physiological system, wherein said substance is usually taken from some storage within the applicator. The applicator may both be designed to provide discrete deliveries of substance at certain time points as well as to provide a continuous delivery of substance with some rate, i.e. amount per time. It may be realized by any of the devices known for this purpose, for example by a battery powered infusion pump.

The drug delivery device has the advantage that the target concentration is optimally met for a whole given predetermined time period, i.e. that the system looks ahead into the future and does not only base its current action on data of the system history. Thus it is for example possible to apply an amount of substance that currently implies some difference between the desired and the actual concentration of substance but that in some time, e.g. minutes or even hours later, avoids the occurrence of much larger differences.

It was already mentioned that the applicator may deliver the substance continuously during the predetermined time period or, in the opposite extreme, in a one-shot manner all at once at one point in time. In a preferred embodiment, the applicator is designed to deliver a plurality of amounts of substance, which were determined with the help of the pharmacokinetic model, during the predetermined time period at temporal distances of at least one minute.

In a further development of the invention, the drug delivery device comprises a sensor for measuring some physiological parameter of the associated physiological system, e.g. the actual concentration of the applied substance. This measurement may for example be used as input in a feedback control loop that provides a higher accuracy than a solely feedforward control.

The drug delivery device may in the aforementioned case further comprise an estimator for estimating a parameter of the pharmacokinetic model based on the measured physiological parameter. Thus the measured concentration of some test substance in the blood of a patient may for example be used to determine the volume of distribution that is needed for an accurate pharmacokinetic model.

The drug delivery device is preferably designed as an implantable device. It may for example be provided in a casing with a physiologically compatible surface, a storage for the substance to be supplied, and/or a battery for power supply. An implantable drug delivery device may particularly be used for a chronic supply of drugs. In the case of a 24-hours-a-day delivery, the global optimization of drug concentration provided by the delivery device is of particular benefit.

The drug delivery device may optionally comprise a receiver for receiving external signals, particularly a receiver for receiving wireless signals. Thus an input for example of the patient or the medical staff can be given to the target module to establish or alter the dosage of a drug.

The invention further relates to a method for supplying a substance to a physiological system during a predetermined time period, said method comprising the following steps: a) Providing a target concentration of the substance over the time period. b) Determining at least one amount of substance to be delivered within the time period to the physiological system such that the predicted concentration of the substance in the physiological system, which is calculated with a given pharmacokinetic model of the physiological system, matches the target concentration according to a given optimization criterion. c) Delivering within the time period said determined amount of substance to the physiological system.

The method comprises in general form the steps that can be executed with a drug delivery device of the kind described above. Therefore, reference is made to the preceding description for more information on the details, advantages and improvements of that method.

The substance may be delivered continuously during the predetermined time period or, in the opposite extreme, in a one-shot manner all at once at one point in time. In a preferred embodiment, a plurality of amounts of substance, which were determined with the help of the pharmacokinetic model, is delivered during the predetermined time period at temporal distances of at least one minute. In the case of a chronic application, the temporal distance between the delivery of two amounts of substance typically ranges from about 1 minute to four hours (depending on the half- life of the applied drug, wherein drugs with small elimination half-lives need more frequent deliveries than drugs with long half-lives).

In the following, various embodiments of the invention will be described that apply both to the drug delivery device as well as to the method.

The optimization criterion that is applied for the determination of the amount(s) of substance to be delivered preferably comprises the minimization of the distance between predicted concentration and target concentration of the substance at a given set of time points during the predetermined time period, wherein said distance is calculated by some suitable measure (e.g. absolute or squared value of difference between predicted and target concentration). The minimization may be done globally, i.e. the (optionally weighted) sum of all differences at all considered time points is minimized; or the minimization may be done on a sequential basis, i.e. the difference is minimized individually for each considered time point. In the latter case, the amount of substance administered at some point in time may for example be determined such that the difference between predicted and target concentration is minimal (ideally zero) at the next administration time.

In another preferred embodiment of the invention, the pharmacokinetic model comprises at least one of the following parameters or settings (limitations) of the physiological system and/or of the supplied substance: the bioavailability, the clearance, the lag time, the therapeutic window, the toxicity threshold, the elimination half-life, the volume(s) of distribution, the distribution rate(s), the absorption rate(s), the activity rhythm, the dosing frequency, the trough level, the remaining activity, and the temperature. Thus the principal factors that affect the observed concentration of a substance in the physiological system can be taken into account.

The pharmacokinetic model may use parameters of the physiological system as such, i.e. typical or averaged values that are best fits to a large number of similar physiological systems. Preferably, the pharmacokinetic model comprises however at least one individual parameter of the associated physiological system, wherein said individual value is for example supplied by a measurement. Thus the control action can be fine-tuned to the situation of an individual patient, allowing a more efficient and improved therapy.

At least one physiological parameter of the associated physiological system may optionally be measured, e.g. the actual concentration of the applied substance. This measurement may for example be used as input in a feedback control loop that provides a higher accuracy than a solely feedforward control.

In a further development of the aforementioned approach, a parameter of the pharmacokinetic model is estimated based on the measured physiological parameter. Thus the measured concentration of some test substance in the blood of a patient may for example be used to determine the volume of distribution that is needed for an accurate pharmacokinetic model. The drug delivery device is preferably implanted and may particularly be used for a chronic supply of drugs.

BRIEF DESCRIPTION OF THE DRAWINGS These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. These embodiments will be described by way of example with the help of the accompanying drawings in which:

Figure 1 shows a principal sketch of an implanted drug delivery device according to the present invention;

Figure 2 is a diagram with exemplary curves of a target drug concentration, the

(predicted) actual drug concentration, and the administered amounts of drug when it is tried to locally meet the target concentration as well as possible; Figure 3 is a similar diagram as Figure 2 when it is tried to globally meet the target concentration as well as possible.

Like reference numbers in the Figures refer to identical or similar components.

DETAILED DESCRIPTION OF EMBODIMENTS Many known implantable drug-delivery products providing constant (continuous) release of medication are designed such that every patient receives the same amount of medication per time unit, independent of body weight, drug distribution volume, elimination half-life and other relevant parameters. Other products are implantable (battery) powered infusion pumps. Some of these pumps can be programmed by the physician/nurse (and often provide constant flow rate) and/or can be adjusted within limits by the patient him/herself based on experienced symptoms.

Preprogrammed delivery systems cannot be halted to (temporarily) stop release and cannot be adjusted in case the patient requires an adjusted amount of medication, for example because of a change in body weight, or observed side effects. Further, delivery is mainly such that a constant level of drug is provided: there is no adaptation to symptoms changing upon time, mimicking of physiological patterns, or changes in perception of symptoms. Also, the patient may desire to change dosage, for example if he would in a haemophilia A treatment like to temporarily increase the blood FVIII level when going sporting. All of these aspects may require adjustment of medication, which is however not an option in preprogrammed drug delivery products.

Systems with an external feedback allow the administration of an additional dose on top of the amount of drug already present. This can lead to more medication than required, which results in a faster emptying of the drug reservoir, higher expenses, and harm for the patient. Moreover, the delivery follows the experience of symptoms instead of anticipating them.

Known systems also have the disadvantage that they do not take "inter-individual properties" such as bioavailability and volumes of distribution into account, so that the amount of medication is not optimized for each patient. This means that some patients get on average too much (resulting in higher costs for the drug and an earlier need to refill or replace the system), or patients get on average too little.

Another aspect is that (clinical) tests have a relative large noise on their results as they are carried out with only a fixed amount of drug for many patients. If "inter- individual properties" could be taken into account, the noise could be reduced leading sooner to statistically significant results.

Based on the above considerations, a drug delivery device and method are proposed here that anticipate the symptoms and thus can avoid the occurrence or the level of occurrence of disease symptoms. Anticipation of the symptoms can be achieved by making use of a (pharmacokinetic) model that contains data on the patient, his/her pharmacokinetic characteristics on the relevant drug, half-life of that drug in the patient, half-lives in blood and/or other relevant tissue (cells, spine, ...). Such a model and device may further enable a better comparison of clinical outcome, as aspects like bioavailability, elimination half- life, volume of distribution, etc. are not considered in current drug delivery systems. Further, (clinical) tests do not take into account these inter-individual parameters causing noise in the test results. The accuracy of (clinical) tests can therefore be enhanced by first determining the pharmacokinetic characteristics of each individual patient before the real/actual test starts, which then uses the determined pharmacokinetic parameters in the drug delivery profile of the drug delivery system. In such a way the blood drug profile can for example be kept constant (or according to a certain desired pattern) and not the amount of drug administered. Therefore, it is proposed here to use a pharmacokinetic algorithm in combination with a drug delivery device.

Figure 1 shows in a principal sketch the layout of an implantable drug delivery device 10 according to the present invention. The device is housed in a physiologically compatible casing and provided with some data processing means (e.g. a microcontroller) together with appropriate storages (e.g. ROM, PROM, EEPROM etc.) and software. The modules shown in the Figure are in first place conceptually different entities and may in practice for example be realized by different software procedures running on the same hardware.

The drug delivery device 10 comprises the following components:

A "target module" 11 that provides a target concentration c_tar of the substance (drug) to be supplied to the blood volume 1 of a patient. This target concentration is at least valid for the time interval until the next drug amount is administered; preferably, it is valid for a larger time interval, e.g. for one day (24 hours). The target module 11 optionally comprises a receiver 12 for wirelessly receiving external input, for example a request of the patient for a higher or reduced dosage.

A "modeling module" 14 which is provided with the target concentration c_tar by the target module 11 and which comprises a pharmacokinetic model of the patient. With the help of this model, the modeling module calculates the amount u of drug that has to be supplied at the next time point of administration.

An applicator 13 for delivering the determined amount u of drug to the vessel system 1. The applicator 13 typically comprises some storage for the drug and some pumping mechanism as they are known to a person skilled in the art.

An optional sensor 15 that is in contact with the physiological system and can determine some physiological parameter. Thus the actual blood drug concentration c_m or the drug level in other relevant tissue may be measured making use of tags like optical tags or magnetic tags, i.e. labels adhered to the drug (or a part of it) for measuring with a suitable sensor the level of drugs in the blood/plasma. The measured concentrations c_m can for example be used by the modeling module 14 to indirectly determine relevant parameters such as the distribution volume, half-lives etc. Moreover, the measured values c_m may be used by the target module 11 to adapt the provided target values c_tar.

The drug delivery device 10 may comprise further components like a battery or some receiver for wirelessly transmitted external power that are generally known and not depicted in the Figure.

By combining a desired blood drug level profile c_tar with a pharmacokinetic model containing pharmacokinetic information about the patient, an elaborated feedforward system can be created. This system is particularly of impact or relevance for patients that experience their symptoms according to some rhythmic pattern. This pattern may be a daily rhythm, an infradian rhythm, or an ultradian rhythm.

Parameters for the pharmacokinetic model like the one that is described below could be: bioavailability of the individual patient (as a function of delivery location, time, etc.); - lag time until the maximum plasma / blood concentration is reached (sometimes referred to as T_max); half- life of medication of individual patient; vo lume(s) of distribution (A); drug absorption rate; - remaining activity of the drug (over time, interaction with body fluids, ...); temperature (measured e.g. by the drug delivery device).

The pharmacokinetic model has to estimate drug dosage within certain limitations / setting. These could be: therapeutic window of the drug; this window may differ from patient to patient and may depend on e.g. time of day, or be a function of other medications administered; toxicity threshold; activity/rhythm of the patient; used optimization model; dosing frequency; - trough level (i.e. minimum allowable drug level).

A feedforward control that can be executed by the drug delivery device 10 may consist of the following steps: 1. Estimate or set the desired blood drug level c_tar(t), at least for the next period, but preferably for some larger time interval into the future.

2. Set the administration frequency, i.e. the time intervals t_α between two administrations.

3. At each time point U = i-t_α (i = 0; 1; 2; ...) the drug concentration is incremented by Ac₁ from an initial concentration C₁ = C(I₁). This increment Ac₁ is calculated for a one- compartment model such that the concentration c(t) = (C₁ +Ac₁) • exp(-k(t-t_!)) equals the given target concentration c_tar(t₁₊i) at the next time point t = t₁₊₁ :

c(t₁₊i) = (C₁ +Ac₁) •

= (C₁ +Ac₁) • exp(-kt_α)

= C_tar(t₁₊l)

→ Ac₁ = c_tar(t₁₊i) • exp(kt_α) - C₁

with k being the elimination half-life constant of the drug in the patient (units: 1/h). In other words: the drug level increases above the target levels at times to reach the target blood drug level at the next time period i+1, just prior to any new administration - this is indicated in Fig 2 by the sudden increases above the target line, which exponentially decay to reach the target blood drug level just prior to the next sudden increase.

More information on pharmacokinetic models that could be applied here can be found on www.rxkinetics.com/pktutorial.

IF the calculated difference Ac₁ is zero or smaller than zero, no drug has to be administered, and the concentration at t₁₊i is given by

C_1+I = c(t₁₊i) = C₁ • exp(-kt_α) .

ELSE IF the calculated difference Ac₁ is positive, the following amount u of drug is administered (with A being the distribution volume): u(t^ = A-Ac₁ .

The concentration at t₁₊i is in this case given by

C₁₊I = c(t₁₊i) = (C₁ +Ac₁) • exp(-kt_α). Figure 2 shows a first diagram with simulation results for the above described approach. The diagram shows

- the administered amounts u of drug (triangles) with the corresponding

(dimensionless) scale on the right vertical axis;

the target concentration c_tar of the drug with the corresponding (dimensionless) scale on the left vertical axis;

the predicted actual concentration c of the drug with the corresponding (dimensionless) scale also on the left vertical axis.

The time intervals between two administrations are set to 0.5 hours. Further parameters of the applied pharmacokinetic model are as follows: volume of distribution: A = 140 dL; patient weight: 70 kg; k = 0.06 1/h; bioavailability: 100%; time lag T_max: 0 h; drug absorption time: 0 h; drug distribution time: 0 h;

Instead of mimicking as closely as possible the set or required drug profile c_tar by the real drug profile c, one can also try and mimick certain time intervals during the day. In Figure 2, the target concentration c_tar was mimicked for the first part and because of the relatively high half-life - compared to the set drug profile -, the drug level stays above the set drug profile. It depends on the drug and on the type of symptoms whether one wishes to mimick the peaks up (as in Figure 2) or one wishes for example to optimize the entire 24h rhythm. In the latter case more rigorous calculations have to be performed for example based on least-squares optimization (minimization of the difference between real and target drug profile). The order of the steps described above also changes. The aim is to have the smallest difference of the whole curves c_tar and c in the diagram. It starts with setting a number of administration amounts (u's), based upon which the differences in concentrations are calculated. These differences are then minimized for the entire interval of interest (e.g. 24 h). Exemplary curves resulting from this approach are shown in Figure 3.

It should be noted that the optimization procedure does affect the amount of drug administered. Thus if it is not really necessary to follow the peak of the set drug profile, but a procedure such as the least square optimization is fine, then less drug can be delivered. A saving of drug means a saving of money and a delay for the refill or re-implantation of a new device.

Moreover, it has to be taken into account that some drugs have a bioavailability smaller than 100%, and that (quite often) increase in blood level may not be instantaneous (i.e. there is a lag time period). Both these aspects can be taken into account in the model calculations.

The pharmacokinetic model can be based on group data. Preferably it is however based on data of the individual patient under treatment.

The model is usually generated off-line. A preferred embodiment of the invention is that the actual blood drug level (or drug level in other relevant tissue) is measured in-line making use of tags, such as for example optical tags or magnetic tags, to indirectly determine relevant parameters such as distribution volume, half- life etc.

The described invention is applicable in the area of drug delivery devices, containing at least parts of a device that delivers medication based on the use of battery power. It relates especially to implantable drug delivery devices but may also relate to the other drug delivery approaches, such as a pill or capsule containing electronics, or transdermal drug delivery devices.

Finally it is pointed out that in the present application the term "comprising" does not exclude other elements or steps, that "a" or "an" does not exclude a plurality, and that a single processor or other unit may fulfill the functions of several means. The invention resides in each and every novel characteristic feature and each and every combination of characteristic features. Moreover, reference signs in the claims shall not be construed as limiting their scope.

Claims

CLAIMS:

1. A drug delivery device (10) for supplying a substance to a physiological system (1) during a predetermined time period, comprising a) a target module (11) for providing a target concentration (c_tar) of the substance in the time period; b) a modeling module (14) for determining at least one amount (u) of substance to be delivered within the time period to the physiological system (1) such that the predicted concentration (c) of the substance in the physiological system (1), which is calculated with a pharmacokinetic model of the physiological system (1), matches the target concentration (c_tar) according to a given optimization criterion; c) an applicator (13) for delivering within the time period said amount (u) of substance to the physiological system (1).

2. The device according to claim 1, characterized in that the drug delivery device (10) comprises a sensor (15) for measuring a physiological parameter of the physiological system (1).

3. The device according to claim 2, characterized in that the drug delivery device (10) comprises an estimator for estimating a parameter of the pharmacokinetic model based on the measured physiological parameter.

4. The device according to claim 1, characterized in that the drug delivery device (10) is designed as an implantable device (10).

5. The device according to claim 1, characterized in that the drug delivery device (10) comprises a receiver (12) for receiving external signals.

6. A method for supplying with a drug delivery device (10) a substance to a physiological system (1) during a predetermined time period, comprising the following steps: a) providing a target concentration (c_tar) of the substance in the time period; b) determining at least one amount (u) of substance to be delivered within the time period to the physiological system (1) such that the predicted concentration (c) of the substance in the physiological system (1), which is calculated with a pharmacokinetic model of the physiological system (1), matches the target concentration (c_tar) according to a given optimization criterion; c) delivering within the time period said determined amount (u) of substance to the physiological system (1).

7. The method according to claim 6, characterized in that a plurality of determined amounts (u) of substance is delivered during the predetermined time period with temporal distances of at least one minute.

8. The method according to claim 6, characterized in that the optimization criterion comprises the overall or the sequential minimization of the distances between predicted concentration (c) and target concentration (c_tar) of the substance at given time points.

9. The method according to claim 6, characterized in that the pharmacokinetic model comprises at least one of the following parameters/settings of the physiological system (1) and/or of the substance: bioavailability, clearance, lag time, therapeutic window, toxicity threshold, elimination half-life, volume(s) of distribution, distribution rate(s), the absorption rate(s), activity rhythm, dosing frequency, trough level, remaining activity, and temperature.

10. The method according to claim 6, characterized in that the pharmacokinetic model comprises at least one parameter that is individually adjusted to the physiological system (1).

11. The method according to claim 6, characterized in that a physiological parameter of the physiological system (1) is measured.

12. The method according to claim 11, characterized in that a parameter of the pharmacokinetic model is estimated based on the measured physiological parameter.