A method and system for monitoring the value of blood glucose of a subject

Abstract

Methods and systems for monitoring the value of blood glucose of a subject are described, the method including the steps of: continuously obtaining ECG parameters of the subject including heart rate, heart rate variability, QT interval and Tp / Rp ratio; inputting the obtained ECG parameters into a learning algorithm which processes the parameters; and outputting a continuously updated output value which is representative of the value of blood glucose of the subject.

Description

[0001] A METHOD AND SYSTEM FOR MONITORING THE VALUE OF BLOOD

[0002] GLUCOSE OF A SUBJECT

[0003] Technical Field

[0004] The present invention relates to systems and methods for monitoring the value of blood glucose of a subject.

[0005] Background to the Invention

[0006] Patients with type 1 diabetes (T1D) and a percentage of patients with type 2 diabetes (T2D) depend on external insulin to maintain normal blood glucose levels (BGL) because their bodies produce little or no insulin. Achieving good glycemic control is a challenge faced by both patients and their caregivers. T1D are affected by the daily fluctuation of blood glucose concentration. Their blood glucose levels can be out of the normal scale at any time, ranging from low blood glucose (hypoglycemia) to high blood glucose (hyperglycemia). Both types of dysglycemia can cause severe illness.

[0007] Hypoglycemia is defined as a low blood glucose level (usually less than 3.9 mmol / L). This condition is a dangerous complication of insulin and sulphonylureas in diabetes treatment. The average T1D patients suffers thousands of episodes of symptomatic hypoglycemia over a lifetime of diabetes (Cryer 2008). The fear of hypoglycemia affects the daily routines of both patients and their carers and contributes a significant factor in the failure of achieving satisfactory glycemic targets for the patients (Barnard 2010). Hypoglycemia manifests both autonomic symptoms and neuroglycopenic symptoms. Hypoglcyemia is one of the complications most feared by patients, on a par with blindness and renal failure.

[0008] On the other hand, hyperglycemia was associated with cognitive dysfunction in T1D children (Davis 1996). A decrease in mental performance was observed when T1D children experiences hyperglycemic episodes > 22.2 mmol / L. With blood glucose levels >15 mmol / L, adults with T1D had a significant longer time in performing cognitive tests (Cox 2005). Hyperglycemia can progress to a life-threatening condition called diabetic ketoacidosis (DKA) (Trachtenbarg 2005). The onset of hyperglycemia could be defined as blood glucose levels of 8.3 mml / L (150 mg / dl). Hyperglycemic episodes (BGL > 8.3 mml / L) were associated with increased mortality in pediatric intensive care units (Hirsberg 2008).

[0009] Current technologies used for diabetes diagnostic testing and self-monitoring are known. However, devices which require a blood sample are unsatisfactory in that the sample is painful to obtain, and continuous monitoring is not possible. Dysglycemia can be monitored by frequently taking blood samples. With the emergence of minimally invasive technology, continuous glucose monitoring (CGM) devices measure the interstitial fluid using a sensor. Although the performance of these devices has been improved, there exists a time delay between CGM measures and actual blood glucose levels. These devices also require frequent calibration and changing sensors. For example, the Dexcom G6 system has a sensor survival rate of 87% in 10 days (Wadwa 2018) and a lag time of 13 min was reported during exercise (Guillot 2020).

[0010] There remains a need to provide improved systems and methods for monitoring the value of blood glucose of a subject.

[0011] Summary of the Invention

[0012] In a first aspect the present invention provides a method of monitoring the value of blood glucose of a subject including the steps of: continuously obtaining ECG parameters of the subject including heart rate, heart rate variability, QT interval and Tp / Rp ratio; inputting the obtained ECG parameters into a learning algorithm which processes the parameters; and outputting a continuously updated output value which is representative of the value of blood glucose of the subject.

[0013] The method may further include the step of activating an alarm condition if the value of blood glucose falls below a pre-determined threshold. The method may further include the step of activating an alarm condition if the value of blood glucose rises above a pre-determined threshold.

[0014] In a second aspect the invention provides a system for monitoring the value of blood glucose of a subject including: an ECG sensor which is arranged to detect ECG signals; means for determining ECG parameters from the ECG signals, wherein the parameters include heart rate, heart rate variability, QT interval and Tp / Rp ratio; and a computing device which is arranged to run a learning algorithm which is arranged to receive the ECG parameters and process the parameters to output a continuously updated output value which is representative of the value of blood glucose of the subject.

[0015] The learning algorithm may be run on a computing device and the ECG sensor communicates with the computing device by wireless communication.

[0016] The computing device may include a display and the continuously updated value is represented on the display.

[0017] The system may be arranged to activate an alarm condition if the value of blood glucose falls below a pre-determined threshold.

[0018] The system may be arranged to activate an alarm condition if the value of blood glucose rises above a pre-determined threshold.

[0019] Brief Description of the Drawings

[0020] An embodiment of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

[0021] Figure 1 shows a system according to the invention in association with a subject;

[0022] Figure 2 is a rear view of the sensor of figure 1;

[0023] Figure 3 is a side view of the sensor of figure 1; and Figure 4 is a schematic view of the system of figure 1.

[0024] Detailed Description of the Preferred Embodiment

[0025] Referring to figure 1, a system 10 is shown for monitoring the value of blood glucose of a subject 100 including: a sensor module 20 which is arranged to be worn by being adhered to the skin of the subject on their chest in the region of their heart; and a computing device in the form of smart phone 40 which receives and processes information received from the sensor module 20.

[0026] Referring to figures 2 and 3, sensor module includes a replaceable adhesive electrode pad 22 which attaches to a transmitter module 26. The adhesive electrode pad has a self adhesive surface and three electrodes 24a, 24b and 24c are provided on the surface. The adhesive pad includes a tab 23 which enables the user to correctly orient the pad at the time of adhering it to their skin.

[0027] The rear face of the electrode pad includes three metal snap connectors 25 a, 25b and 25c which engage with corresponding recesses in the rear face of the transmitter module 26. The snap connectors 25a, 25b and 25c physically attach the adhesive pad 22 to the transmitter module 26 and also electrically connect the electrodes to circuity inside the transmitter module 26.

[0028] To apply the device, the subject attaches an adhesive electrode pad 22 to the transmitter 26, then removes a peel away film to reveal the self adhesive surface. The subject then applies the sensor module to the skin of their chest, with the tab 23 pointing upwards.

[0029] The sensor module 20 produces ECG waveforms 20 from the output of the electrodes 24a, 24b, 24c. The waveforms are transmitted over a wireless communication channel in the form of a bluetooth connection from the sensor to a bluetooth receiver inside smart phone 40.

[0030] Referring to figure 4, the system includes means for determining ECG parameters from the ECG waveforms 30 in the form of segmentation block 32 and feature extractor block 44. These are processing blocks implemented in software running on smartphone 40.

[0031] The determined parameters include:

[0032] • heart rate

[0033] • heart rate variability

[0034] • QT interval; and

[0035] • Tp / Rp ratio

[0036] Smartphone 40 is also programmed to run a learning algorithm in the form of a hybrid Al algorithm 50 which receives the extracted ECG parameters and processes the parameters to generate a continuously updated output value which is representative of the value of blood glucose of the subject.

[0037] The output value is displayed on the display screen 60 of the phone 40. The output value is also logged in the memory of the phone and can be viewed as a historical chart.

[0038] Output alarms are configured in the phone to alert the user if the determined value of blood glucose falls outside of pre-determined user defined upper or lower threshold levels. These thresholds are set to alert the user if they are determined to have an abnormally high or low blood glucose value.

[0039] The ECG is achieved by placing three Ag-AgCl electrodes in a Lead II configuration on the person’s chest. The signal obtained from the electrodes is then amplified using an instrumentation amplifier with gain of 10 and CMRR > 100 dB at 100 Hz. This feeds through a highpass filter with cut-off frequency of 0.5 Hz. A second stage non-inverting amplifier is added to provide a gain of 100. To obtain a reliable heart rate and heart rate variability of the patient, a bandpass filter may be used, to detect the QRS complex of the ECG signal. In particular, SDNN the standard deviation of all normal RR intervals, is the most commonly used time- domain measure of heart rate variability (HRV). The QT interval and TpTe interval, on the other hand are clinical parameters which can be derived from the ECG signals. QT measurement requires the identification of the start of QRS complex and the end of the T wave. The intersection of the isoelectric line and a tangent to the T wave can be used to measure the QT interval. Tp / Rp is the ratio of the peak value of the T wave and the R wave.

[0040] For determining blood glucose levels using a combination of ECG parameters (heart rate, QT interval, Tp / Rp ratio, heart rate variability), a hybrid neuro estimator 50 is used. This estimator is embedded in a software app of a smart phone 40. This hybrid Al estimator 50 uses a combination of Bayesian neural network or deep learning model combined with regression techniques. Essentially, it is trained using a learning algorithm in which synaptic strengths are systematically approximate the blood glucose status given by the available quantitative data. The system uses real-time learning so as to be able to adapt to the physiological signals of individual subjects.

[0041] Essentially, ECG data are acquired wirelessly to the smart phone from the sensors and transmitter in real-time. From ECG data, heart rate, heart rate variability, QT interval and Tp / Rp parameters are calculated accordingly. These parameters are fed into a hybrid artificial intelligence (Al) framework which is based on a Bayesian neural network. Bayesian neural networks is a practical and powerful means to improve the generalization of neural networks. Its learning is performed by considering Gaussian probability distributions of the weight which can give the best generalization. In particular, the weight w in network X are adjusted to their most probable values given the training data D. Specifically, the posterior distribution of the weights can be computed using Bayes’ rule as follows.

[0042] Briefly, after the network training is completed, the values of parameters ygand ygare then used to compute the log evidence of network , having M hidden nodes as follows. where S is the cost function, parameters ygand ggare hyperparameters, Wgis the number of weights and biases in group g , and Q is set to be io3. The best network will be selected with the highest log evidence.

[0043] This hybrid Al framework is based on three bands of blood glucose values (hypoglycemia: BGL < 3.9 mmol / L, euglycemia: BGL from >3.9 to <8.3 mmol / L, and hyperglycemia: BGL > 8.3 mmol / L). For each band, a Bayesian neural network is designed, and the overall value of this hybrid Al algorithm is calculated from the combined regression of these three bands.

[0044] In a recent study led by the inventor which involved 17 sessions with 10 T1D adolescents, the inventor has found that hypoglycemia episodes are correlated significantly with increased corrected QT interval (QTc) (p<0.05) and with decreased Tp / Rp ratio (p<0.05). Futhermore, the inventor has found that hyperglycemia episodes are correlated significantly with increased heart rate (p<0.01) and with decreased heart rate variability (p=0.13). The learning algorithm is based on identifying changes in these parameters over time to determine an ongoing value of blood glucose of the subject.

[0045] During testing, it has been found that the overall values of blood glucose calculated using a prototype system according to the invention correlate significantly with actual blood glucose values (p < 0.001) measured from blood samples.

[0046] A subject can wear the device for as long as is required or desired. Periodically, they may change the adhesive electrode pad and recharge or replace the battery in the sensor module.

[0047] Communication between the sensors and a processor system or smart phone is via a wireless system. Radio frequency transmitters and receivers or Bluetooth communication may be used. The alarm may be of any convenient type, and might comprise a simple radio alarm, a signal transmitted to a monitor station or a smart phone, a tactile device, via the Internet, or the like.

[0048] It is also preferred that data transmitted from the sensors will be continuously logged. The system may be interfaced with a PC which will continuously log the relevant data using a data management system.

[0049] Clearly the invention can vary from that described herein without departing from the scope of the invention. In particular, the hybrid Al algorithm need not be of the type described herein, but any learning algorithm that is able to provide substantially real-time analysis of multiple data streams in the manner described herein could be used.

[0050] It can be seen that embodiments of the invention provide at least one of the following advantages:

[0051] • Value of blood glucose can be continually monitored completely non-invasively with no need for taking blood samples or analysing intersitial fluid.

[0052] • Sensor module can be easily fitted and removed by the end user.

[0053] • Module interacts with users existing smartphone with no need for proprietary hardware.

[0054] Any reference to prior art contained herein is not to be taken as an admission that the information is common general knowledge, unless otherwise indicated.

[0055] Finally, it is to be appreciated that various alterations or additions may be made to the parts previously described without departing from the spirit or ambit of the present invention.

Claims

CLAIMS:

1. A method of monitoring the value of blood glucose of a subject including the steps of: continuously obtaining ECG parameters of the subject including heart rate, heart rate variability, QT interval and Tp / Rp ratio; inputting the obtained ECG parameters into a learning algorithm which processes the parameters; and outputting a continuously updated output value which is representative of the value of blood glucose of the subject.

2. A method according to claim 1 further including the step of activating an alarm condition if the value of blood glucose falls below a pre -determined threshold.

3. A method according to claim 1 further including the step of activating an alarm condition if the value of blood glucose rises above a pre-determined threshold.

4. A system for monitoring the value of blood glucose of a subject including: an ECG sensor which is arranged to detect ECG signals; means for determining ECG parameters from the ECG signals, wherein the parameters include heart rate, heart rate variability, QT interval and Tp / Rp ratio; and a computing device which is arranged to run a learning algorithm which is arranged to receive the ECG parameters and process the parameters to output a continuously updated output value which is representative of the value of blood glucose of the subject.

5. A system according to claim 4 wherein the learning algorithm is run on a computing device and the ECG sensor communicates with the computing device by wireless communication.

6. A system according to claim 5 wherein the computing device includes a display and the continuously updated value is represented on the display.

7. A system according to any one of claims 4 to 6 which is arranged to activate an alarm condition if the value of blood glucose falls below a pre-determined threshold.

8. A system according to any one of claims 4 to 6 which is arranged to activate an alarm condition if the value of blood glucose rises above a pre-determinedthreshold.

Images