HK40059990B - Method for determining a current glucose value in a transported fluid - Google Patents

Description

Methods for determining the current glucose level in the transport fluid

Technical Field

The present invention relates to a method for preferably continuously determining the current glucose level in the transport fluid of an organism, particularly in the blood.

The present invention also relates to an apparatus for preferably continuously determining the current glucose level in the transport fluid of an organism, particularly in the blood.

The present invention also relates to an assessment device that preferably continuously determines the current glucose level in the transport fluid of an organism, particularly in the blood.

The present invention also relates to an intangible machine-readable medium for storing instructions that, when executed on a computer, cause an execution method to determine, preferably continuously, the current glucose level in the transport fluid of an organism, particularly in the blood.

Although the present invention can be generally applied to any method for determining the current glucose level in a transport fluid, it will be described with reference to blood glucose concentration in a living organism.

Background Technology

To determine blood glucose concentration (BG) in organisms, particularly the human body, systems for continuous glucose monitoring, also known as CGM (Continuous Glucose Monitoring), are known. In CGM systems, interstitial blood glucose concentration (IG) is typically automated, for example, measured every one to five minutes. Diabetic patients particularly benefit from CGM systems because measurements can be performed at a significantly higher frequency compared to self-monitoring processes (also known as self-monitoring processes) where patients manually determine their blood glucose levels four to ten times daily. This allows for automated assessment and alerts, especially when the patient is asleep, which helps prevent critical health conditions.

On the one hand, known CGM systems are based on electrochemical processes. For example, such a CGM system is described in WO 2006/017358A1. Furthermore, optical CGM systems are known, for example, from DE 102015101847B4, in which fluorescence dependent on glucose levels is used, and this literature is incorporated herein by reference. Both CGM systems measure interstitial tissue glucose concentrations.

Furthermore, it is known that tissue glucose concentration or interstitial glucose (IG) concentration deviates from blood glucose concentration (hereinafter referred to as BG). This deviation is particularly significant after a strong influence on blood glucose levels, such as through the ingestion of food or nutrients or the supply of insulin, as described in the non-patent literature “Time lag of glucose from intravascular to interstitial compartment in humans” (Diabetes (2013): DB-131132) by Basu, Ananda, et al. This deviation is caused by diffusion processes in the tissues surrounding the blood, which causes IG levels to be delayed in time and follow BG levels in an unspecified manner, as described, for example, in the non-patent literature “Subcutaneous glucose predicts plasma glucose independent of insulin: implications for continuous monitoring” (American Journal of Physiology-Endocrinology and Metabolism 277.3(1999): E561-E571) by Rebrin, Kerstin et al.

Calibration of CGM systems that manually determine blood glucose concentrations (e.g., by drawing a drop of blood from a finger and determining the glucose concentration in the drop using an external measuring device) can lead to significant inaccuracies due to the muffling and time delay between the two glucose concentrations, one in the blood BG and the other in the peripheral tissue (IG).

However, to achieve accurate calibration of the CGM system, the aforementioned difference between tissue glucose concentration and blood glucose concentration must be considered or at least assessed. Various methods are known for this purpose. According to the non-patent literature “Delays in minimally invasive continuous glucose monitoring devices: a review of current technology” (Journal of Diabetes Science and Technology 3.5 (2009): 1207-1214) by Keenan, D. Barry et al., the use of time-delayed glucose signals for calibration is known. Furthermore, according to the non-patent literature “The extended Kalman filter for continuous Glukose monitoring” by Knobbe, Edward J., and Bruce Buckingham (Diabetes Technology & Therapeutics 7.1 (2005): 15-27), it is known how to use a Kalman filter to compensate for the muffling and time delay of glucose diffusion between blood and tissues.

Summary of the Invention

Therefore, an object of the present invention is to provide an apparatus and an evaluation device that can more accurately determine glucose levels, particularly blood glucose levels, while offering greater flexibility, especially when considering additional parameters and simpler implementation. Another object of the present invention is to provide an alternative method, an alternative apparatus, and an alternative evaluation device. Yet another object of the present invention is to provide a method, apparatus, and evaluation device that better determine blood glucose concentration in an organism based on measurements of interstitial tissue glucose levels.

In one embodiment, the present invention achieves the above objective through a method for preferably continuously measuring the current glucose level in the transport fluid of an organism, particularly in the blood, the method comprising the following steps:

a) Use a sensor device to determine a series of measurements in the tissue surrounding the delivery fluid, including at least two time-separated measurements of tissue glucose levels.

b) Determine tissue glucose levels based on a sensor model using a given series of measurements, whereby the sensor model correlates the sensor measurements with tissue glucose levels while taking into account measurement noise.

c) Provide a state transition model, wherein, by means of the state transition model, at least one glucose level in the delivery fluid is correlated with a determined tissue glucose level while considering process noise, and

d) Determine the current glucose level based on the provided state transition model and the established tissue glucose level.

Among them, at least step d), and especially steps b)-d), are performed using at least one rolling time-domain estimation method.

In another embodiment, the present invention achieves the above objective by means of a device that preferably continuously determines the current glucose level in the transport fluid of an organism, particularly in the blood, said device being preferably adapted to perform the method according to the invention, said device comprising...

A sensor device, particularly for measuring fluorescence in tissue surrounding a delivery fluid via an optical fiber probe, is designed to determine a series of measurements, including at least two temporally separated measurements of the tissue surrounding the delivery fluid.

A device is provided, designed to provide a state transition model and a sensor model, wherein, by means of the state transition model, at least one glucose level in the delivery fluid is correlated with a determined tissue glucose level while taking process noise into account, and wherein, by means of the sensor model, the measured values of the sensor device are correlated with the tissue glucose level while taking measurement noise into account.

The assessment device is designed to determine tissue glucose levels based on a sensor model using a range of provided measurements, and to determine the current glucose level using a rolling time-domain estimation method based on a provided state transition model and the determined tissue glucose levels.

In another embodiment, the present invention achieves the above objective through an evaluation device for preferably continuously determining the current glucose level in the transport fluid of an organism, particularly in the blood, preferably suitable for performing the method according to the invention, the evaluation device comprising:

At least one interface for connecting a sensor device to provide a series of measurements, including at least two time-separated measurements of tissue glucose levels surrounding the delivery fluid.

At least one memory is provided for storing a state transition model and a sensor model, wherein, while considering at least one process noise level, at least one glucose level in the transport fluid is correlated with a tissue glucose level by means of the state transition model, and wherein, while considering at least one measurement noise, the measured values of the sensor device are correlated with the tissue glucose level by means of the sensor model.

The computing device is designed to determine tissue glucose levels based on a sensor model, using a range of provided measurements, and to determine the current glucose level using at least one rolling time-domain estimation method based on a stored state transition model and the determined tissue glucose level.

In another embodiment, the invention achieves the above objective through an intangible machine-readable medium for storing instructions that, when executed on a computer, execute a method preferably adapted to perform the method according to the invention, thereby continuously determining, in particular, the current glucose level in the transport fluid of an organism, especially in the blood, the method comprising the following steps:

a) Use a sensor device to determine a series of measurements in the tissue surrounding the delivery fluid, including at least two time-separated measurements of tissue glucose levels.

b) Determine tissue glucose levels based on a sensor model using a given series of measurements, where the sensor model correlates the sensor measurements with tissue glucose levels while taking into account measurement noise.

c) Provide a state transition model, wherein, by means of the state transition model, at least one glucose level in the delivery fluid is correlated with a determined tissue glucose level while considering process noise, and

d) Determine the current glucose level based on the provided state transition model and the established tissue glucose level.

Among them, at least step d), and especially steps b)-d), are performed using at least one rolling time-domain estimation method.

In other words, a method for determining blood glucose concentration in an organism is proposed. This method comprises the following steps:

In a first method step, a series of measurements of tissue glucose levels in an organism's tissue are obtained via sensors, these measurements having at least two sensor measurements spaced apart over time. In another step, a sensor model is provided that correlates the sensor measurements with the tissue glucose levels, and a state transition model is provided that includes a model correlating the tissue glucose levels and blood glucose values. In yet another method step, the blood glucose levels of the organism are quantified using the sensor model and the state transition model associated with the sensor measurements, wherein the use of a rolling time-domain estimation method is necessary.

Rolling time-domain estimation (MHE) is known in principle for evaluating measurement signals that exist as a series of statistical values. The applicant's research demonstrates that, in contrast to methods used to date, the rolling time-domain estimation method offers significant advantages for evaluating sensor measurements used to quantify blood glucose levels in terms of estimation accuracy, flexibility related to model assumptions, and speed of providing blood glucose levels. Preferably, the rolling time-domain estimation method produces a current blood glucose concentration and a retrospective blood glucose concentration, wherein, preferably, at least one previous blood glucose concentration (the blood glucose concentration previously determined during the measurement process (concentration process)) is considered in determining the retrospective blood glucose concentration. Thus, the retrospective blood glucose concentration particularly enables a better reconstruction of the current blood glucose measurement signal solely into the current blood glucose concentration.

Here, the evaluation device can be a computer, an integrated circuit, or a similar device, for example, specifically constructed for the optimized calculation of the trace of a matrix. According to one embodiment of the invention, the device and/or evaluation device can be designed as a portable device with an independent power source, allowing for efficient operation, such as a battery or similar device, thereby also keeping the energy depletion potential for performing the method low, so that the battery can operate for as long as possible, which enhances the user experience. For this purpose, energy-efficient processors, wiring, circuits, interfaces, especially wireless interfaces, can be used in particular. Here, the execution of the method can be matched in terms of its parameters, for example, with the following device, such as the evaluation device, in the evaluation time domain and/or noise time domain (this will be described later) to achieve sufficient accuracy on the one hand and a long operating time on the other.

One of the advantages is that it provides greater efficiency in terms of time and computational resources for estimating current glucose levels in transport fluids, particularly blood. Furthermore, it offers significantly increased flexibility compared to known methods by eliminating limitations imposed on certain sensor models and/or state transition models. Another advantage is that it not only increases the accuracy of current glucose levels but also improves the accuracy of previous glucose levels.

Other features, advantages and embodiments of the present invention will be described below or disclosed therein.

According to a preferred embodiment, a rolling time-domain estimation method is performed to provide the current glucose level in step d), and applied to previously provided glucose measurements and at least one previous tissue glucose measurement. In particular, this provides an efficient determination of the current glucose level based on the glucose level of previous measurements.

According to another preferred embodiment, the sensor model is provided as a linear function of the measured value and the tissue glucose level. This allows for the particularly efficient and rapid calculation of interstitial tissue glucose levels based on a series of measurements with sufficient accuracy.

According to another preferred embodiment, the sensor model is provided in the form of a nonlinear function between the measured value and the tissue glucose level. Here, for example, the following sensor model can be provided, where y represents the measured value, IG represents the glucose concentration, and a, b, c or A, b, c represent sensor parameters:

y = c - a * b / (IG + b)

Alternatively, the following nonlinear sensor model can be used:

y = (A*b + C*IG) / (IG + b)

Its advantage is that the calculated interstitial tissue glucose level is more accurate based on a series of measurements, especially through sensor devices with affinity binding sensors or optical sensors.

According to another preferred embodiment of the invention, the time domain value for the rolling time-domain estimation method used to provide the current glucose level is selected to be less than or equal to 10. This enables the calculation of interstitial tissue glucose levels with sufficient accuracy, particularly efficiently and rapidly, based on a series of measurements.

According to another preferred embodiment, the variance of measurement noise and/or process noise is estimated, particularly at least periodically. This allows for the simple and rapid provision of noise measurements, and thereby provides an overall accurate determination of the current glucose level.

According to another preferred embodiment, the variance of the measurement noise and/or the variance of the process noise are estimated, or in particular interpolated and/or weighted, preferably using exponential smoothing. In this case, any noise level that changes over time can be matched or updated, which further improves the overall accuracy of determining the current glucose level.

According to another preferred embodiment, measurements that are only partially used to calculate estimates of measurement noise and/or process noise can be temporarily stored, and the stored measurements can be used to interpolate necessary measurements that have not yet been temporarily stored. Thus, for example, for determining the level of measurement noise and/or process noise, the necessary and computationally intensive measurements can be temporarily stored at least partially, making them available for subsequent measurements. This generally reduces the computational load required to determine the current glucose level without significantly reducing its accuracy.

According to another preferred embodiment, some previous measurements are selected that are greater than the time-domain measurements of the rolling time-domain estimation method, particularly at least twice as large, preferably at least five times as large, thereby determining the accuracy of the estimation of process noise and/or measurement noise associated with the current glucose level and improving the overall accuracy of the determination or quantification of the current glucose level.

According to another preferred embodiment, the variance of measurement noise and/or process noise is regularly adjusted based on the sum of the number of previous measurements and the time domain of the rolling time-domain estimation method, wherein the previous measurements are used to calculate the estimates of measurement noise and/or process noise. This ensures that efficient adjustments to noise measurements at any given time are performed at regular intervals, which on the one hand to achieve sufficient accuracy for the current glucose level, and on the other hand to prevent unnecessary adjustments or updates that do not increase or significantly increase the accuracy of the current glucose level.

According to another preferred embodiment, the provided measurements are filtered by a filtering function, thereby suppressing errors, particularly measurement errors, by the sensor device. With the aid of the filtering function, erroneous measurements, such as sensor errors or outliers in the measurements, can be easily selected; this means that these erroneous measurements are not considered when further calculating the current glucose level.

According to another preferred embodiment, the measurement noise values are weighted using a filtering function. This prevents both underestimation and overestimation of the measurement noise level: underestimation leads to extremely erroneous signals or measurements, while overestimation results in an overly smooth measurement process across a series of measurements. In summary, this further improves accuracy.

According to another preferred embodiment, in order to determine the error of the sensor device, the current tissue glucose level and/or the gradient of increase in the current tissue glucose level are evaluated. Alternatively or additionally, this can also be performed by assessing the gradient of increase in the current glucose level and/or the current glucose level of the delivery fluid. This provides a simple, reliable, and efficient error identification for the sensor device.

According to another preferred embodiment, measurements provided below a preset low threshold level and/or above a preset high threshold level are filtered out by a filtering function. Specifically, the low and high threshold levels correspond to physiological limits. Preferably, the low threshold level is between 10-50 mg/dL, particularly 30 mg/dL, and the high threshold level is between 100-600 mg/dL, preferably 450 mg/dL. With the aid of appropriate threshold levels, erroneous sensor measurements are advantageously filtered out by a weighting matrix, preferably a diagonal weighting matrix, for further calculations. Erroneous sensor measurements include glucose measurements in the delivery fluid, particularly blood, and tissue glucose measurements. Advantageously, the weighting matrix operates in such a way that erroneous measurements of the sensor device are weighted by a factor of 0, while all other measurements are weighted by a factor of 1. Erroneous measurements, such as outliers in sensor measurements, are determined by the absolute glucose concentration in the delivery fluid, particularly blood, and its rate of change or gradient. In the first case, it is preferable to introduce physiological limits for blood glucose concentration, wherein the blood glucose concentration range is assumed to be 10 mg/dL-600 mg/dL, preferably 20 mg/dL-500 mg/dL, and most preferably 30 mg/dL-450 mg/dL. Measurements outside these physiological limits are weighted as sensor measurements with an error factor of 0, and the gradient of glucose concentration in the delivery fluid, particularly in the blood, or the rate of change of glucose concentration in the delivery fluid, particularly in the blood, can also be determined, and its level can be compared with the physiologically actual rate of change. Therefore, the quantified rate of change of glucose concentration in the delivery fluid, particularly in the blood, is a value from 0.1 mg/dL/min to 15 mg/dL/min, preferably from 0.5 mg/dL/min to 10 mg/dL/min, and most preferably from 1 mg/dL/min to 3 mg/dL/min.

According to another preferred embodiment, calibration of the sensor device measurements is performed after step d). Therefore, uncalibrated tissue glucose measurements can be used, with the advantage that calibration need not be performed before, but rather after, the rolling time-domain estimation method. Another advantage of using uncalibrated tissue glucose measurements is that they exhibit a high correlation with self-monitored blood glucose concentrations, and thus, for example, the parameters of the sensor model can be advantageously determined more simply and accurately.

According to another preferred embodiment, the state transition model includes a diffusion model for time-dependent modeling of the diffusion process of glucose from the transport fluid to surrounding tissues. The diffusion model, particularly one based on the diffusion constant, provides a simple and computationally inefficient modeling of the decay and time delay between glucose levels in the transport fluid, especially in the blood, and between glucose levels in the tissues.

According to another preferred embodiment, the parameters of the sensor model and/or the state transition parameters of the state transition model are estimated and/or updated at least periodically. The advantage of doing so is that, overall, the accuracy used to determine the current glucose level is improved; similarly, the parameters of the model in question can be flexibly adjusted to change the environment or influences.

Other important features and advantages of the invention are apparent from the dependent claims, the drawings, and the description of the corresponding drawings using the drawings.

It should be understood that, without departing from the scope of the invention, the above-described features and the features to be described can be used not only in the combinations indicated respectively, but also in other combinations or individually.

Attached Figure Description

Preferred designs and embodiments of the invention are shown in the accompanying drawings and further explained in the following description. All remodeling steps, including equations, assumptions, solution processes, etc., can be used independently without departing from the scope of the invention.

Figure 1 illustrates the steps of a method according to an embodiment of the present invention in graphical form;

Figure 2 illustrates the steps of a method according to an embodiment of the present invention in graphical form; and

Figure 3 illustrates, in graphical form, a comparison between the method according to an embodiment of the present invention and a known method.

Claims (29)

1. A method for continuously determining the current glucose level in a transport fluid of an organism, the method comprising the steps of: a) Using a sensor device, a series of measurements is determined, said series of measurements including at least two time-spaced measurements of tissue glucose levels in the tissue surrounding the delivery fluid. b) Using the given series of measurements, based on a sensor model, determine the tissue glucose level, wherein, with the aid of the sensor model, the measurements of the sensor device are correlated to the tissue glucose level while taking into account measurement noise. c) Provide a state transition model, wherein, by means of the state transition model, while taking process noise into account, at least one glucose level in the delivery fluid is associated with a determined tissue glucose level, and d) Determine the current glucose level based on the provided state transition model and the established tissue glucose level. Among them, at least step d is performed using at least one rolling time-domain estimation method. 2. The method according to claim 1, characterized in that the sensor model is provided in the form of a linear or nonlinear function between the measured value and the tissue glucose level. 3. The method according to claim 1, wherein the time domain value used to provide the rolling time domain estimation method for the current glucose level is selected to be less than or equal to 10. 4. The method according to any one of claims 1-3, characterized in that, at least regularly, the variance of the measurement noise and/or the variance of the process noise are estimated. Specifically, exponential smoothing is used to estimate the variance of measurement noise and/or process noise based on at least one previous value. 5. The method according to claim 4, characterized in that it is capable of temporarily storing measured values that are only partially used to calculate the estimation of the measurement noise and/or the process noise, and capable of interpolating the measured values that are not temporarily stored and are needed using the stored measured values. 6. The method according to claim 4, characterized in that the number of previous measurements is selected, which is greater than the number of time-domain measurements of the rolling time-domain estimation method. 7. The method according to claim 6, characterized in that the variance of the measurement noise and/or the variance of the process noise are regularly adjusted based on the level of the sum of the time domain of the rolling time-domain estimation method and the number of previous measurements, wherein the previous measurements are measurements used to calculate the estimate of the measurement noise and/or the process noise. 8. The method according to claim 1, characterized in that the measured value provided by the filtering function is filtered by means of the filtering function, wherein the error of the sensor device is suppressed by means of the filtering function. 9. The method according to claim 8, characterized in that the gradient of the increase in the current tissue glucose level and/or the current tissue glucose level are evaluated to determine the error of the sensor device. 10. The method according to claim 8 or 9, characterized in that the filtering function filters out measurement values provided below a preset low threshold level and/or above a preset high threshold level, the low threshold level and the high threshold level corresponding to physiological limits. 11. The method according to any one of claims 1-3 and 5-9, characterized in that, after performing step d), the calibration of the measured values of the sensor device is performed. 12. The method according to any one of claims 1-3 and 5-9, characterized in that the state transition model comprises a diffusion model for time-dependent modeling of the diffusion process of glucose from the transport fluid to the surrounding tissue. and/or At least regularly estimate and/or update the sensor model parameters of the sensor model and/or the state transition parameters of the state transition model. 13. The method according to claim 1, wherein the delivery fluid is blood. 14. The method according to claim 1, characterized in that steps b) to d) are performed using the at least one rolling time-domain estimation method. 15. The method according to claim 1, wherein the rolling time-domain estimation method is performed to provide the current glucose level in step d) and is applied to previously provided glucose measurements and at least one previous tissue glucose measurement. 16. The method according to claim 4, characterized in that, at least regularly, interpolation and/or weighting are performed on the variance of the measurement noise and/or the variance of the process noise. Herein, exponential smoothing is used to interpolate and/or weight the variance of the measurement noise and/or the variance of the process noise based on at least one previous value. 17. The method according to claim 6, characterized in that the number of previous measurements is selected, which is at least twice the number of time-domain measurements of the rolling time-domain estimation method. 18. The method according to claim 17, characterized in that the number of previous measurements selected is at least five times greater than the number of time-domain measurements of the rolling time-domain estimation method. 19. The method according to claim 8, wherein the measurement error of the sensor device is suppressed by means of a filtering function. 20. The method according to claim 8, characterized in that the measured values are weighted by means of the filtering function. 21. The method of claim 10, wherein the low threshold level is between 10 mg/dL and 50 mg/dL, and the high threshold level is between 100 mg/dL and 600 mg/dL. 22. The method according to claim 21, wherein the low threshold level is 30 mg/dL and the high threshold level is 450 mg/dL. 23. An apparatus for continuously determining the current glucose level in a transport fluid of an organism, said apparatus being adapted to perform the method according to any one of claims 1-22, said apparatus comprising: A sensor device for measuring fluorescence in tissue surrounding a delivery fluid via an optical fiber probe, the sensor device being designed to determine a series of measurements, including at least two time-separated measurements of the tissue surrounding the delivery fluid. A providing device is designed to provide a state transition model and a sensor model, wherein, by means of the state transition model, at least one glucose level in the delivery fluid is correlated to a determined tissue glucose level while taking process noise into account, and wherein, by means of the sensor model, the measurements of the sensor device are correlated to the tissue glucose level while taking measurement noise into account. An evaluation device is designed to determine the tissue glucose level based on the sensor model and using the provided series of measurements, and to determine the current glucose level using a rolling time-domain estimation method based on the provided state transition model and the determined tissue glucose level. 24. The apparatus according to claim 23, wherein the delivery fluid is blood. 25. An assessment apparatus for continuously determining the current glucose level in a transport fluid of an organism, said assessment apparatus being adapted to perform the method according to any one of claims 1-22, said assessment apparatus comprising: At least one interface for connecting to a sensor device for providing a series of measurements, said series of measurements including at least two time-separated measurements of tissue glucose levels surrounding the delivery fluid. At least one memory is provided for storing a state transition model and a sensor model, wherein, while considering at least one process noise level, at least one glucose level in the delivery fluid is correlated to the tissue glucose level by means of the state transition model, and wherein, while considering at least one measurement noise, the measurement value of the sensor device is correlated to the tissue glucose level by means of the sensor model. A computing device is designed to determine the tissue glucose level based on the sensor model and using the provided series of measurements, and to determine the current glucose level using at least one rolling time-domain estimation method based on the stored state transition model and the determined tissue glucose level. 26. The evaluation apparatus according to claim 25, wherein the delivery fluid is blood. 27. A non-volatile machine-readable medium for storing instructions, which, when executed on a computer, cause a method for continuously determining a current glucose level in a delivery fluid of an organism to be performed, adapted to perform the method according to any one of claims 1-22, the method comprising the steps of: a) Using a sensor device, a series of measurements is determined, said series of measurements including at least two time-separated measurements of tissue glucose levels in the tissue surrounding the delivery fluid. b) Determine the tissue glucose level based on a sensor model using a given series of measurements, wherein the sensor model, by means of the sensor device, correlates the measurements to the tissue glucose level while taking into account measurement noise. c) Provide a state transition model, wherein, by means of the state transition model, while taking process noise into account, at least one glucose level in the delivery fluid is associated with a determined tissue glucose level, and d) Determine the current glucose level based on the provided state transition model and the determined tissue glucose level. Among them, at least step d is performed using at least one rolling time-domain estimation method. 28. The non-volatile machine-readable medium according to claim 27, wherein the delivery fluid is blood. 29. The non-volatile machine-readable medium according to claim 27, characterized in that steps b) to d) are performed using the at least one rolling time-domain estimation method.