WO2008094077A2 - Method for determining the glucose concentration in the human blood - Google Patents

Abstract

A determination method of human blood glucose level (BGL) relates to the field of medical engineering. More particularly, this non-invasive method relates to diagnostics and continuous BGL monitoring. Electric transmitting functions of skin-deep and under layer tissue are continuous measured by two tetra-electrode sensors located at the same extremity place of human body. First sensor is located along major blood vessels of extremity, and second sensor is located normally to a first sensor. Measurements of two sensors process by mathematic model calibrated preliminary. The model is calibrated by the comparison between invasive BGL measurements and BGL values calculated by method proposed. Thereafter BGL data are restored from author dependence derived. Method proposed permits to restore BGL by measurements of two sensors without additional information. Also this method improves an accuracy of BGL determination.

Description

 The method for determining the concentration of glucose in human blood

The invention relates to medicine, namely to methods for diagnosis and continuous monitoring of the state of glucose in human blood.

There are several analogues of the present invention based on the relationship between blood glucose and electrical characteristics of the skin or tissue. For example, a utility model is known (IPC 6 A 61 V 5/00 98104833/20 dated 03/19/98), entitled "Device for monitoring the state of a patient with diabetes", authors Esikov B. B., Petrov SV., Strelnikov AC, Khmelevsky B.I., Antonov E.V. It has a capacitive sensor and an indicator microammeter, according to the testimony of which they judge the insertion capacity and the concentration of sugar in the blood. The main disadvantage of this utility model is its applicability to a narrow class of diabetics, as well as the low accuracy of glucose determination.

Also known is the RF Patent for the invention JV "2073242 (MKI 6 G 01 N 33/48), entitled" A method for indicating blood sugar and a device for its implementation ". In it, the level of sugar in the blood is judged by the change in capacity due to the proportional change in the dielectric constant of the blood. A change in capacitance affects the oscillation frequency of the resonant circuit. At the same time, the current through the indicator changes, "the readings of which in accordance with the calibrated curve (indicator scale) will reflect the patient's blood sugar level." The main disadvantage of this invention is also the applicability for a narrow class of diabetics as well as lack of accuracy in determining glucose.

Also known is a method for determining a method for determining the concentration of glucose in human blood according to patent N ° 2230485 IPC 7 A61B5 / 053, which consists in the fact that at the preparatory stage of implementation The method simultaneously measures the total electrical resistance of the skin and the concentration of glucose in the blood by an invasive method, and these measurements are performed both with increasing and decreasing glucose concentration, while the coefficients ao and a _{b are} determined, while the coefficient a _b also depends on the direction of the concentration change blood glucose. To determine the concentration of glucose in the blood both at the calibration stage and at the stage of using the method, coefficient ai can be taken as a constant value at intervals of monotony of change in glucose concentration. Therefore, it is necessary to know whether the concentration of glucose increases or decreases in a given period of time. This qualitative information can be obtained, for example, by changing the shape of the electrocardiogram, which is associated with a change in the level of glucose in the blood, or using any other external source of information.

The main disadvantage of this method is the need to obtain additional external information about the direction of change in the concentration of glucose in the blood of a person, as well as the insufficient accuracy of the method when using small sensors, which is explained by the practical impossibility of measuring impedances for them.

The claimed method has the task of determining the concentration of glucose in human blood in a non-invasive way without involving additional information, using small sensors.

The task is achieved as follows: measure the electrical transfer function (PFT), which is taken as the ratio of the output voltage, that is, the voltage on the tissue area through which the electric current passes during the measurement, to the input current coming from the alternating current electric generator, by two pairs of four-electrode sensors mounted on the surface of the human body, the first pair they are fixed along the main blood vessels, mainly the limbs, and the second pair of electrodes is fixed in the same place orthogonally to the first, the electrical transfer functions of not only the surface of the skin, but also of the subcutaneous tissues are continuously measured, then the measurements of the four-electrode sensors are processed using a pre-calibrated mathematical model, and the model is calibrated by comparing the results of the proposed method for determining glucose in human blood and any other known method of determination glucose in human blood, after which the concentration of glucose in human blood is calculated by the formula:

G _n = G ₀₀ + G _n ., Exp [- a ₀ β _n . _I (t _p - t _p . _L ) J + {a _l [L _p ^LF - P _] , L _p ^ULF - P, ₂ L _p ^HF J + + a ₂ fK _p ^LF -P ₂₁ K _p ^ULF -P ₂₂ K _p ^HF ] exp [-a ₀ (β _p ., T _p -β _p .it _p .,)]} (T _p - t _p .,) / 2; β _p = l / [l + a ₃ \ G _p , - G _O o \] where G _n = G (ty) is the determined value of the concentration of glucose in the blood in ^! a given point in time t _p

- G _0O - the level of glucose concentration in the blood of the human body ';'^' corresponding to the homeostasis of this organism in relation to the concentration of glucose in the blood,

- a ₀ - coefficient characterizing the rate of glucose utilization in the body of a particular person, taking into account the variability of external factors,

- ¹ CIi - coefficient characterizing the relationship between the concentration of glucose in the blood of a particular person and the values of the transfer functions of tissues, measured along the main blood vessels at different frequencies,

- a ₂ - coefficient characterizing the relationship between the concentration of glucose in the blood of a particular person and the values of the transfer functions of tissues, measured in the orthogonal direction to the first pair of four-electrode sensors at different frequencies,

- a ₃ - coefficient taking into account the facilitated transfer of glucose,

- P p, P p, P 2i, P 22 are the coefficients of the model. s - L _n , L _n ^LF , L _n ^HF - inverse values to the electrical transfer functions of tissues, measured along the main blood vessels at an ultra-low frequency from 0.01 Hz to 5 Hz, a low frequency from 10 kHz to 60 kHz and a high frequency from 100 kHz to 10 MHz at time t _p , and K _n ^ULF , K _n ^LF , K _n ^HF are the reciprocal of the electrical transfer functions of tissues, measured at ultra-low, low and high frequencies in the orthogonal direction to the first pair of four-electrode sensors at time t _p discrete sampling, starting from zero at tо = 0, wherein n is an integer (n = 1, 2,...), while the ao, aj, a ₂ , a ₃ , Pc, Pn, Pc, P ₂₂ values are determined at the preparatory stage of calibration, while the calibration time sufficient is selected so that multidirectional changes in the concentration of glucose in the blood can be recorded.

As PFT, the modulus of the electrical transfer function of tissues, that is, its absolute value, can be measured. ' ^®

They can also measure the active component of the electrical transfer function.

Can also measure the reactive component of the electrical transfer function.

Can also measure the ratio of reactive to active components of the electrical transfer function.

For an insulin-dependent patient, the value of GQ _{O is} assumed to be zero.

The technical result of the proposed method is to determine the concentration of glucose in human blood, based entirely on the measurement system adopted in the method, without involving additional information, while increasing the accuracy of determining the concentration of glucose in the blood. In FIG. 1 shows a flowchart of a method in a calibration step.

In FIG. 2 shows a flowchart of a method during the step of using the method (calculating glucose from the measurement results).

In FIG. Figure 3 shows the calibration results for one of the type II diabetics (non-insulin-dependent).

In FIG. Figure 4 shows a graph of the time variation of the measured tissue transfer functions (TFT) at low and high frequencies during the restoration of glucose in the blood of an organism of a second-type diabetic.

On Fig 5 presents a graph of the time change of the measured PFT at an ultra-low frequency, with the restoration of glucose in the blood of a second-type diabetic.

In Fig 6. presents a graph of the change in time of the estimated concentration of glucose in the blood for a diabetic of the second kind. Dots indicate glucose control values measured by the invasive method ^' ; * -

In FIG. 7 shows calibration results for a healthy person.

In FIG. Figure 8 shows a graph of the time variation of the measured PFT of the tissues at low and high frequencies during the restoration of glucose in the blood of the body.

Fig. 9 is a graph of the time variation of the measured PFT at an ultra-low frequency during restoration of glucose in the blood of a healthy volunteer.

Figure 10 shows a graph of the time variation of the estimated blood glucose concentration for a healthy volunteer. Dots indicate glucose control values measured in an invasive manner.

The method for determining the concentration of glucose in human blood is as follows: at the preliminary stage of the calibration of the mathematical model of the method (see Fig. 1), fix 1, for example, four-electrode sensors according to the scheme with two external and two internal (measuring) electrodes, moreover, an external probing electric current is supplied to external electrodes spaced at a distance of 20-50 mm, and the smaller linear size of the sensor is determined by the uncertainty of localization of the main blood vessels under the sensor, as well as by the increase in the specific contribution of glucose transfer from the main vessels in the transverse direction. The increase in the linear dimensions of the sensor is limited by the possibility of its use on the wrist or other area of the limbs. Then provide simultaneous measurements of blood glucose concentration by the invasive method 2, and measurement of electrical tissue transfer functions (TFT) in a limited area of limb 3, for example, the wrist, in two orthogonal directions (along the arm and in the transverse direction) at three different frequencies. The ratio of the output voltage, that is, the voltage at the tissue site through which the electric current in the measurement process, to the input current coming from the alternating current electric generator is taken as PFT. In the method, PFT is a complex function of time. These measurements can be carried out by any suitable method, for example, according to the usual four-electrode circuit with two external and two internal (measuring) electrodes, and at the same time, alternating electric current of three different frequencies in the ultra-low ranges is supplied to the external electrodes, spaced at a distance of 20-50 mm frequency (VLF - from 0.01 Hz to 5 Hz), low frequency (LF - from 5 kHz to 50 kHz), high frequency (HF - from 200 kHz to YmHz).

In the inventive method for measuring PFT simultaneously two identical sensors are used, the first of which is located along the arm, and the second in the transverse direction. This allows you to simultaneously measure the modules of the electric PFT for ultra-low, low and high frequencies along the arm, respectively designated t / ⁷ ^, lf ^{^ h} , lf ^1h , and PFT in the transverse direction, denoted respectively by F ^ULF , F ^{11 '} , F " ^F.

The values of K ^ULF = 1 / ^, K ^LF = 1 / ^, K ^ИF = 1 / F " ^F , inverse to the PFT measured in the transverse direction at ultra-low, low and high frequencies, respectively, according to our studies, correlate with the liquid content in the dermis, the content of extracellular fluid and the total fluid content in the tissue under the measuring electrodes, while the contribution to the fluid content is made mainly by capillary blood transfer, without the contribution of blood transfer through large vessels.

The values L ^ULF = 1 / U ^ULF , L ^LF = 1 / U ^LF , L ^HF = l / lf ^F , inverse to PFT measured in the direction along the arm at ultra-low, low and high frequencies, respectively, according to our studies, correlate with the content fluid in the dermis, the content of extracellular fluid and the content of total fluid in the tissue under the measuring electrodes. In this case, capillary blood transfer and blood transfer along the main vessels make a contribution to the fluid content. ^{v dc}

The experimental results of the restoration of blood glucose concentration confirm the findings of our studies.

Simultaneous measurements of PFT in two orthogonal directions make it possible to avoid the negative influence of the vascular component of the blood on the measurement results and the results of glucose recovery, and to determine the direction of change in glucose in the blood.

The concentration of glucose in the blood is related to the fluid content in the tissues, which fall into the area covered by the lines of force of the propagating current from the sensor electrodes. The values that are inverse to the measured PFT are proportional to the fluid content in the tissue, as our studies show, confirmed by the results of the restoration of the glucose concentration in the blood.

The concentration of glucose in human blood at a given time is determined by the recurrence ratio of 4, linking it with the inverse of the measured modules of the electrical transfer functions of the tissue, and the glucose concentration at the previous time:

P₁₂L_п ^HF]+G _n = G ₀₀ + G _n ., Expf-

P ₁₂ L _p ^HF ] +

+ a ₂ [K _p ^LF ~ P ₂₁ K _p ^ULF -P ₂₂ K _p ^HF J exp [-a ₀ (β _p ., t _p -β _p ., t _p .,)]}} (t _p - t _p j / 2; β _n = l / [l + a ₃ \ G _n .j - G _O o U; (1)

where G _n = G (t _p ) is the determined value of the concentration of glucose in the blood at a given time t _p ,

- G ₀₀ - the level of glucose concentration in the blood of the human body, corresponding to the homeostasis of this organism in relation to the concentration of glucose in the blood,

- a ₀ is a coefficient characterizing the rate of glucose utilization in the body of a particular person, taking into account the variability of external factors, both in the presence of insulin produced by the human body (in healthy, prediabetics, and type II diabetics), and in the presence of external insulin (long and short ) in diabetics of the first kind.

- a _\ is a coefficient characterizing the relationship between the concentration of glucose in the blood of a particular person and the transmission ratios of the skin and tissues, measured along the human limb at different frequencies,

- a ₂ - coefficient characterizing the relationship between the concentration of glucose in the blood of a particular person and the transmission ratios of the skin and tissues, measured in the transverse direction to the human limb at different frequencies,

- a ₃ - coefficient taking into account the facilitated transfer of glucose,

- P l, P _{] 2} , P ₂₁ , P ₂₂ - model coefficients, determined at the preparatory stage of calibration, -L _p ^ULF , L _n ^LF , L _n ^HF are the reciprocal of the electrical transfer functions of the tissue, measured along the human limb at ultra-low, low and high frequencies at time t _p , and K _n , K _n ^L , K _n ^HF are the reciprocal of the electrical transfer functions of the tissue, measured at ultralow, low and high frequencies in the transverse direction to the limb at the time t _{p of} taking discrete samples, starting from zero at to = 0, while n is an integer (n = 1 , 2,...).

The mentioned values of a ₀ , a _} , ci ₂ , az, Pc, Pn, Pc, P 22 are determined by approximating the dependence of the concentration of glucose in blood 5 obtained by the invasive method 3 on the above calculated dependence G _n = G (I _n ) calculated according to the formula (1), while the time T is chosen sufficient so that it is possible to record changes in the concentration of glucose in the blood associated with the natural daily cycle of its changes, or artificially caused, for example; nutrition, exercise, injection of glucose or insulin. The determination of the values a ₀ , Ci ₁ , α ₂ , <* З, Pn, Pn, Ргi, P 22 'is carried out by standard mathematical methods, for example, minimizing the nonlinear functional, for example, by the least squares method. As a result, a choice is made of such coefficients a ₀ , a _/ , a ₂ , b, P ιι, P 12, P 2i for the P ₂₂ model, which best approximate the results of invasive measurements (approximating the experimental curve) obtained during the calibration process. Such a procedure for selecting model parameters can be implemented using standard software on the MATLAB platform. For example, the above procedure for determining the model parameters can be performed using the "survefit" procedure located in MATLAB. During the calibration process, the number of invasive points is no less than 3–4 times the number of determined model parameters.

At the glucose recovery stage (see Fig. 2), 6 sensors are fixed in the same way as at the calibration stage, 7 PFTs are measured in discrete time points, after which 8 glucose in a person’s blood is calculated according to the ratio (1) with the coefficients determined at the calibration stage. Then, the calculated glucose concentration is displayed 9, for example, on a liquid crystal indicator.

The inventive method was tested experimentally on five first-type diabetics, two second-type diabetics and three healthy ones. Moreover, one test day for each subject was used to calibrate the method. The remaining test days (from one to three for different subjects) were used to restore blood glucose. Tests on test days took from two to four hours, with changes in glucose taking place in different directions (downward and upward). At the same time, invasive measurements of blood glucose were performed using Assu-Chek Assive instruments for control. More information on the device device can be found at httr: // www. assu-chek.sh / ru / rewrite / content / ru RU / 11.30.10: 10 / article / ACCM heperal article 237. &^' htm.

The results of some of these experiments are shown in FIG. 3 - 10.

Below are the test results for one of the type II diabetics (non-insulin-dependent). In FIG. 3 shows the calibration results, while the calibration coefficients of the model have the following values: a ₀ = 0.0015; U ₁ = 15.03; a ₂ = - 0.206; a ₃ = 0.5; P ₁₁ = 9821; P ₁₂ = 0; P ₂₁ = 20070; P ₂₂ = - 2.05. The points on the graph correspond to blood glucose values measured using an invasive device. A continuous line shows the results of the calculation method.

In FIG. Figures 4 and 5 show graphs of measured PFT for the second day of testing, during which glucose was restored in the blood of a diabetic. On .. 6 presents the results of the restoration of glucose in

Yu blood according to the formula (1) obtained using calibration coefficients and measured transfer functions of tissues.

In FIG. 7 shows the results of calibration on a test day for one of the healthy subjects, while the calibration coefficients are determined: a ₀ = 0.0002; a, = 45; a ₂ = - 20; a ₃ = 0.4; P ₁₁ = 50; P _n = 0; P ₂ / = 5.014; P ₂₂ = 0.502.

In FIG. 8-10 show the results of restoring glucose concentration on a test day according to formula (1) obtained using measured tissue transfer functions and calibration factors.

Thus, the technical result of the proposed method is achieved, namely, the determination of glucose concentration in human blood is completely based on the measurement system adopted in the method, without involving additional information, with improved accuracy in determining blood glucose concentration.

Claims

Claim 1. The method of determining the concentration of glucose in the blood of a person, characterized in that the electrical characteristics of the current flowing through the tissue are measured by means of four electrode sensors mounted on the surface of the human body, then the measurements of the four electrode sensors are processed by a pre-calibrated mathematical model by comparing the results of the proposed method for determining glucose in human blood and any other known method for determining glucose in human blood, after which they determine glucose concentration in human blood, characterized in that two pairs of four-electrode sensors are fixed on the surface of the human body, the first pair being fixed along the main blood vessels, mainly limbs, and the second pair of electrodes is fixed in the same place orthogonally to the first, the electrical transfer functions are not continuously measured only the surface of the skin, but also of the subcutaneous tissues, which are taken as the ratio of the output voltage, that is, the voltage on the tissue site through which passes the electric current in the measurement process, to the input current coming from the generator of electric alternating current, then the concentration of glucose in the blood of a person is calculated by the formula: G _n = G ₀₀ + G _n ., exp-a _o β _p _, (t _p - t _n _,)] + {a, [L _n ^LF - P ,, L _n ^ULF - P _{] 2} L _n ^HF J + + a ₂ [K _n ^LF -P ₂ , K _n ^ULF -P ₂₂ K _n ^HF ] exp [-ao (β _p ., t _p -β _p ., t _p .,)]} (t _p - t _p _,) / 2; β _p = l / [l + ^Ci ₃ ] G _n ^. , - G ₀₀ U, where G _n = G (I _n ) is the determined value of the concentration of glucose in the blood at a given time t _p , - Goo - the level of glucose concentration in the blood of the human body, corresponding to the homeostasis of this organism in relation to the concentration of glucose in the blood, - a ₀ - coefficient characterizing the rate of glucose utilization in the body of a particular person, taking into account the variability of external factors, - P] - coefficient characterizing the relationship between the concentration of glucose in the blood of a particular person and the values of the transfer functions of tissues, measured along the main blood vessels at different frequencies, - a ₂ - coefficient characterizing the relationship between the concentration of glucose in the blood of a particular person and the values of the transfer functions of tissues, measured in the orthogonal direction to the first pair of four-electrode sensors at different frequencies, - a ₃ - coefficient taking into account the facilitated transfer of glucose, - P]], P] ₂ , P _2I , P ₂₂ - model coefficients, - L _p ^ULF , L _n ^LF , L _n ^HF are the reciprocal of the electrical transfer functions of tissues, measured along the main blood vessels at an ultra-low frequency from 0.01 Hz to 5 Hz, a low frequency from 10 kHz to 60 kHz, and a high frequency from 100 kHz to ^{10 MHz} at time t _p , and K _n ^υLF , K _n ^LF , K _n ^HF are the reciprocal of the electrical transfer functions of tissues, measured at ultra-low, low and high frequencies in the orthogonal direction to the first pair of four-electrode sensors at the time time t _p removal of discrete samples, starting from zero p when t0 = 0, at that n is an integer (n = 1, 2, ...), while the above-mentioned quantities aη, П], a ₂ , аз, Рц, P п, P ₂ i, P _{22 are} determined on the preparatory stage of calibration using standard mathematical methods of minimizing the nonlinear functional, while the calibration time is chosen sufficient so that multidirectional changes in the concentration of glucose in the blood can be recorded; 2. The method according to p. 1, characterized in that they measure the module of the electrical transfer function; 3. The method according to p. 1, characterized in that they measure the active component of the electrical transfer function; 4. The method according to p. 1, characterized in that they measure the reactive electric transfer function; 5. The method according to p. 1, characterized in that they measure the ratio of reactive to active components of the electrical transfer function; 6. The method according to any one of paragraphs. 1-5, characterized in that for an insulin-dependent patient, the value of G _O o is taken equal to zero;