Continuous analyte monitoring sensor calibration and measurement by connection function

Abstract

A method of determining glucose values during continuous glucose monitoring (CGM) measurements includes providing a CGM device including a sensor, a memory, and a processor; applying a constant voltage potential to the sensor; measuring a primary current signal resulting from the constant voltage potential and storing the measured primary current signal in the memory; applying a probe potential modulation sequence to the sensor; measuring a probe potential modulation current signal resulting from the probe potential modulation sequence and storing the measured probe potential modulation current signal in the memory; determining an initial glucose concentration based on a conversion function value and the primary current signal; determining a connection function based on the primary current signal and a plurality of the probe potential modulation current signals; and determining a final glucose concentration based on the initial glucose concentration and a connection function value. Other aspects are provided.

Description

[0001] This application claims U.S. Provisional Patent Application No. 63 / 061,135, filed August 4, 2020, entitled "CONTINUOUS ANALYTE MONITORING SENSOR CALIBRATION AND MEASUREMENTS BY A CONNECTION FUNCTION"; U.S. Provisional Patent Application No. 63 / 061,152, filed August 4, 2020, entitled "NON-STEADY-STATE DETERMINATION OF ANALYTE CONCENTRATION FOR CONTINUOUSGLUCOSE MONITORING BY POTENTIAL MODULATION"; and U.S. Provisional Patent Application No. 63 / 061,152, filed August 4, 2020, entitled "EXTRACTING PARAMETERS FOR ANALYTE CONCENTRATION". The benefits of U.S. Provisional Patent Application No. 63 / 061,157 entitled “Biosensor with membrane structure for steady-state and non-steady-state conditions for determining analyte concentrations”, filed August 4, 2020, and U.S. Provisional Patent Application No. 63 / 061,167 entitled “Biosensor with membrane structure for steady-state and non-steady-state conditions for determining analyte concentrations”, each disclosure of which is hereby incorporated herein by reference in its entirety for all purposes. Technical Field

[0002] This disclosure relates to continuous sensor monitoring of analytes in body fluids. Background Technology

[0003] Continuous analyte sensing in in vivo or in vitro samples, such as continuous glucose monitoring (CGM), has become routine sensing in the medical device field, and more specifically, in diabetes care. For biosensors that use discrete sensing to measure analytes in whole blood samples, such as by pricking a finger to obtain a blood sample, the temperature of the sample and the hematocrit of the blood sample are likely to be the main sources of error. However, for sensors deployed in non-whole blood environments with relatively constant temperatures, such as those used in continuous in vivo sensing operations, other sensor error sources may exist.

[0004] Therefore, there is a need for improved devices and methods for determining glucose levels using CGM sensors. Summary of the Invention

[0005] In some embodiments, a method for determining a glucose value during continuous glucose monitoring (CGM) measurement includes: providing a CGM device comprising a sensor, a memory, and a processor; applying a constant voltage potential to the sensor; measuring a primary current signal generated by the constant voltage potential and storing the measured primary current signal in the memory; applying a probe potential modulation sequence to the sensor; measuring a probe potential modulation current signal generated by the probe potential modulation sequence and storing the measured probe potential modulation current signal in the memory; determining an initial glucose concentration based on a transition function and the primary current signal; determining a connection function value based on the primary current signal and a plurality of the probe potential modulation current signals; and determining a final glucose concentration based on the initial glucose concentration and the connection function value.

[0006] In some embodiments, a continuous glucose monitoring (CGM) device includes: a wearable portion having: a sensor configured to generate a current signal from interstitial fluid; a processor; a memory coupled to the processor; and a transmitter circuitry coupled to the processor. The memory includes a connection function based on a primary current signal generated by applying a constant voltage potential to a reference sensor, and a plurality of probe potential modulated current signals generated by applying a probe potential modulation sequence between measurements of the primary current signal. The memory includes computer program code stored therein, which, when executed by the processor, causes the CGM device to: measure and store the primary current signal using the sensor and the memory of the wearable portion; measure and store a plurality of probe potential modulated current signals associated with the primary current signal; determine an initial glucose concentration based on a conversion function and the primary current signal; determine a connection function value based on the primary current signal and the plurality of probe potential modulated current signals; and determine a final glucose concentration based on the initial glucose concentration and the connection function value.

[0007] Other aspects, features, and advantages of this disclosure will become apparent from the detailed description and illustrations of the following numerous exemplary embodiments and implementations, including the best mode contemplated for carrying out the invention. Other and different embodiments of this disclosure are also possible, and modifications may be made to various aspects thereof, all without departing from the scope of the invention. For example, although the following description relates to continuous glucose monitoring, the apparatus, systems, and methods described below can be readily adapted to monitor other analytes, such as cholesterol, lactate, uric acid, alcohol, etc., in other continuous analyte monitoring systems. Attached Figure Description

[0008] The accompanying drawings described below are for illustrative purposes and are not necessarily drawn to scale. The drawings are not intended to limit the scope of this disclosure in any way.

[0009] Figure 1 The in vivo sensor sensitivity relative to the in vitro sensor sensitivity is shown according to the embodiments provided herein.

[0010] Figure 2A A linearity test of the working electrode (Iw) current over 15 days is demonstrated according to the embodiments provided herein.

[0011] Figure 2B The linearity of the working electrode (Iw) current for glucose solutions of 50, 100, 200, 300, and 450 mg / dL according to the embodiments provided herein is demonstrated.

[0012] Figure 3A The applied voltage and current sampling timing is shown in an example probe potential modulation (PPM) cycle according to the embodiments provided herein.

[0013] Figure 3B An example of a constant voltage application according to the embodiments provided herein, and an example of timing for current measurement of individual primary data points, are shown.

[0014] Figure 3C The output current (time-current curve) for the first four cycles of primary data points according to the embodiments provided herein is shown, as well as the current generated by... Figure 3B constant applied voltage and Figure 3A The PPM current generated by the PPM cycle.

[0015] Figure 3D and 3E The embodiments provided herein are shown respectively. Figure 3C The superimposed PPM working electrode current Iw and PPM blank electrode current Ib of the first four PPM cycles are shown.

[0016] Figure 3F Example timing and marking of the PPM current measured in a cycle according to the embodiments provided herein are shown.

[0017] Figure 4A The working electrode current (minus background current) of 10 sensors according to embodiments provided herein is shown in relation to a reference glucose concentration.

[0018] Figure 4B A graph showing the distribution of the relative error (deviation %) of the original glucose concentration with unit calibration according to the embodiments provided herein is presented.

[0019] Figure 4C This document demonstrates the use of a connection function with uniform calibration according to embodiments provided herein. Figure 4B The output (final) glucose concentration.

[0020] Figure 4D This document illustrates embodiments of the invention provided herein. Figure 4A G of data 原始 and G 连接 Comparison of the percentage deviation.

[0021] Figure 5A This demonstrates the effect of employing a connection function with unit calibration according to the embodiments provided herein. Figure 2B The output glucose concentration data.

[0022] Figure 5BThis demonstrates how, according to the embodiments provided herein, after employing a join function, [the following occurs]... Figure 4A G of data 原始 and G 连接 Comparison of the percentage deviation.

[0023] Figure 6 The document presents a comparison of the relative error of glucose %-ΔG / G using a single prediction equation relative to a transition function with uniform calibration, followed by a connection function for the same dataset, according to embodiments provided herein.

[0024] Figure 7A and 7B An example sensor response population is shown according to the embodiments provided herein.

[0025] Figure 7C and 7D This paper demonstrates how, according to the embodiments provided herein, a wide range of sensor responses can be subdivided into two subsets of responses: an upper limit and a lower limit.

[0026] Figures 8A-8D The embodiments provided herein demonstrate the output glucose concentrations relative to a reference glucose concentration from a connection function with uniform calibration on days 1, 3, 7, and 14, respectively.

[0027] Figure 9A and 9B The initial response of the CGM sensor in a linear test starting at 50 mg / dL and 100 mg / dL, respectively, according to the embodiments provided herein, is shown, as well as the linear glucose output using a connection function with uniform calibration.

[0028] Figure 10 An example method for determining glucose values ​​during continuous glucose monitoring measurements is shown according to embodiments provided herein.

[0029] Figure 11A A high-level block diagram of an example CGM device according to embodiments provided herein is shown.

[0030] Figure 11B A high-level block diagram of another example CGM device according to the embodiments provided herein is shown.

[0031] Figure 12 This is a side view of an example glucose sensor according to the embodiments provided herein.

[0032] Figure 13A and 13BThe combined in vivo and in vitro responses, with and without subtracting current Ib from the blank electrode, are shown according to the embodiments provided herein, from CGM clinical studies and laboratory tests of multiple sensors, respectively.

[0033] Figure 13C and 13D The in vivo and in vitro data according to the embodiments provided herein are shown to have responses from their respective lower and upper boundaries from each study. Detailed Implementation

[0034] Reference will now be made in detail to exemplary embodiments of this disclosure, which are illustrated in the accompanying drawings. Where appropriate, the same reference numerals will be used throughout the drawings to refer to the same or similar parts. Unless otherwise specifically indicated, features of the various embodiments described herein can be combined with each other.

[0035] The terms “voltage,” “potential,” and “voltage potential” are used interchangeably herein. Similarly, “current,” “signal,” and “current signal” are used interchangeably herein, such as in “continuous analyte monitoring” and “continuous analyte sensing.” As used herein, probe potential modulation (PPM) refers to the intentional, periodic alteration of a constant voltage potential applied to other aspects of the sensor during continuous analyte sensing, such as applying a probe potential step, pulse, or other potential modulation to the sensor.

[0036] Primary data point or primary current refers to the measured value of the current signal generated during continuous analyte sensing in response to an analyte at a constant voltage potential applied to the sensor. Probe potential modulation (PPM) current refers to the measured value of the current signal generated during continuous analyte sensing in response to probe potential modulation applied to the sensor. Reference sensor refers to the sensor used to generate primary data point and PPM current in response to a reference glucose concentration, such as that expressed by a blood glucose meter (BGM) reading (e.g., primary current and PPM current measured for the purpose of determining a prediction equation, such as a conversion function and connection function subsequently stored in the continuous analyte monitoring (CAM) device and used to determine the analyte concentration during continuous analyte sensing).

[0037] Similarly, a reference sensor data point refers to a reference sensor reading at a time that closely corresponds to the signal time of the sensor during continuous operation. For example, a reference data point can be obtained directly as the concentration of a reference analyte solution prepared by gravimetric analysis and validated by a reference sensor / instrument such as a YSI glucose analyzer (YSI Incorporated of Yellow Springs, Ohio), Contour NEXTOne (Ascensia Diabetes Care US, Inc. of Parsippany, New Jersey), and / or similar instruments. In vitro studies involving linearity studies are performed by exposing the continuous analyte sensor to the reference solution. In another instance, reference sensor data points can be obtained from readings of the reference sensor during periodic in vivo measurements of the target analyte via venous blood draws or finger prick sampling.

[0038] Uniform calibration refers to a calibration mode in which only one calibration sensitivity, or one subset of calibration sensitivities, is always applied to all sensors. Under uniform calibration, in-situ manual pointer puncture calibration or calibration using sensor codes can be minimized or eliminated.

[0039] A connection function is a function established from in vitro tests of multiple sensors / sensitivity to capture the range of variation in sensor sensitivity, or a function established from a combination of in vitro and in vivo data to capture the range of variation in sensor sensitivity.

[0040] For sensors deployed in non-whole blood environments with relatively constant temperatures, such as those used in continuous in vivo sensing operations, sensor error can be related to both short-term and long-term sensor sensitivity, as well as subsequent calibration methods. Several challenges / problems exist associated with such continuous sensing operations: (1) long break-in (warm-up) times; (2) factory or field calibration; and (3) sensitivity variations during continuous sensing operations. These problems / problems appear to be related to sensor sensitivity as represented by initial decay (break-in / warm-up time), sensitivity variations due to sensor environmental sensitivity during sensor production, and the environmental / conditions of subsequent sensor deployment.

[0041] According to one or more embodiments of this disclosure, the apparatus and method are operable to detect initial start conditions for continuous sensor operation for a sample analyte, and to detect sensor conditions at any point thereafter during continuous sensing operation of the sensor.

[0042] The embodiments described herein include systems and methods for applying probe potential modulation over a constant voltage applied to other aspects of an analyte sensor. The terms voltage, potential, and voltage potential are used interchangeably herein.

[0043] Methods are provided for developing parameters for a prediction equation that can be used to continuously and accurately determine analyte concentrations from an analyte sensor. In some embodiments, as described below, the prediction equation may include a transformation function and a connection function. Furthermore, methods and apparatus are provided for determining analyte concentrations using a probe potential modulation (PPM) self-supplied signal (e.g., a working electrode current generated by applying probe potential modulation). Such methods and apparatus can allow analyte concentration determination while (1) overcoming the influence of different background interference signals, (2) eliminating or removing the influence of different sensor sensitivities, (3) shortening the warm-up time at the start of a (long-term) continuous monitoring process, and / or (4) correcting for changes in sensor sensitivity during continuous monitoring. Reference is made below. Figure 1-13D These and other embodiments are described.

[0044] For continuous glucose monitoring (CGM) biosensors that typically operate under a constant applied voltage, the current from the mediator is continuously measured due to the enzymatic oxidation of the target analyte, glucose. In practice, the current is usually measured or sensed every 3 to 15 minutes, or at another regular interval, although it is referred to as continuous. There is an initial break-in period when the CGM sensor is first inserted / implanted into the user, which can last from 30 minutes to several hours. Once the CGM sensor has been broken in, its sensitivity may still change for various reasons. Therefore, it is necessary to sense the sensor's operating conditions during its initial period and after the break-in period to identify any changes in its sensitivity.

[0045] After the CGM sensor is subcutaneously inserted / implanted into the user, operation of the sensor begins with the applied voltage E0. The applied voltage E0 is typically located at a point on the redox plateau of the mediator. For the natural mediator of oxygen-glucose oxidase, in a medium with a chloride concentration of approximately 100-150 mM, the oxidation plateau of hydrogen peroxide (H₂O₂, the oxidation product of the enzyme reaction) relative to an Ag / AgCl reference electrode is in the range of approximately 0.5 to 0.8 V. The operating potential of the glucose sensor can be set at 0.55-0.7 V, which is within the plateau region.

[0046] The embodiments described herein employ probe potential modulation as a periodic perturbation of a constant voltage potential applied to other aspects of the working electrode of a subcutaneous biosensor during continuous sensing operation (e.g., for monitoring biological sample analytes, such as glucose). During continuous sensing operation, such as continuous glucose monitoring, the sensor working electrode current is typically sampled every 3–15 minutes (or at some other frequency) for glucose value determination. These current measurements represent primary currents and / or primary data points used for analyte determination during continuous sensing operation. In some embodiments, after each primary current measurement, periodic cycling of probe potential modulation may be employed, such that a set of self-supplied currents is accompanied by each primary data point containing information about the sensor / electrode state and / or condition.

[0047] Probe potential modulation can include one or more potential steps that differ from the constant voltage potential typically used during continuous analyte monitoring. For example, probe potential modulation can include a first potential step above or below a constant voltage potential, a first potential step above or below a constant voltage potential and then returning to a constant voltage potential, a series of potential steps above and / or below a constant voltage potential, voltage steps, voltage pulses, pulses of the same or different durations, square waves, sine waves, triangular waves, or any other potential modulation.

[0048] As described, conventional biosensors for continuous analyte sensing operate by applying a constant potential to the working electrode (WE) of the sensor. Under these conditions, the current from the WE is recorded periodically (e.g., every 3–15 minutes or at some other time interval). In this way, the current generated by the biosensor is attributed only to changes in analyte concentration, not to changes in the applied potential. That is, non-steady-state currents associated with applying different potentials are avoided or minimized. While this method simplifies continuous sensing operation, the current signal from the data stream from which a constant potential is applied to the sensor provides minimal information about the sensor's state / condition. In other words, the sensor current signal from which a constant potential is applied to the sensor provides little information related to issues associated with long-term continuous monitoring of the sensor, such as batch-to-batch sensitivity variations, long warm-up times due to initial signal attenuation, changes in sensor sensitivity during long-term monitoring, and the effects of varying background interference signals.

[0049] The embodiments described herein include systems and methods for applying probe potential modulation over a constant voltage applied to other aspects of an analyte sensor. Methods are provided for developing parameters for a prediction equation, which may include a transformation function and a connection function, and which can be used to continuously and accurately determine analyte concentrations from an analyte sensor. For example, in some embodiments, a transformation function may be applied to a raw glucose signal (e.g., working electrode current or working electrode current minus background current) to determine an initial glucose value, and a connection function may then be applied to the initial glucose value to determine a final glucose value.

[0050] Subcutaneously implanted continuous glucose monitoring (CGM) sensors require timely calibration against reference glucose levels. Typically, the calibration process involves obtaining a blood glucose monitoring (BGM) reading, or capillary glucose value, from a finger prick glucose measurement and inputting the BGM value into the CGM device to set the CGM sensor's calibration point for the next operating cycle. This calibration process is usually performed once daily, or at least once daily with a finger prick glucose measurement, as the sensitivity of the CGM sensor can vary daily. This is an inconvenient but necessary step to ensure the accuracy of the CGM sensor system.

[0051] To improve usability and minimize the number of manual finger-prick BGM tests, a logical approach would be to correlate the in vitro and in vivo sensitivities of individual sensors one by one. However, in practice, the correlation between the in vitro and in vivo sensitivities of individual sensors is difficult to observe and complex. This is not surprising, as sensor sensitivity varies over time and under different conditions during CGM. Furthermore, the sample types for in vitro and in vivo tests differ. For in vitro tests, the sample type is a glucose solution in buffer, or at most, simulated interstitial fluid (ISF). For in vivo tests, the sample type is interstitial fluid referencing capillary glucose. A more robust correlation between in vitro and in vivo glucose measurements is desirable to keep the number of manual finger-prick BGM measurements used for calibration minimal.

[0052] Embodiments of this disclosure provide a method for uniform calibration and for establishing a connection between in vitro and in vivo analytes for use in subcutaneously implanted (under the skin of a subject) continuous monitoring sensors without additional calibration. For example, in some embodiments, this disclosure provides a connection between in vitro and in vivo glucose for subcutaneously implanted continuous monitoring sensors.

[0053] In one or more embodiments, a method is provided for determining continuous glucose values ​​by uniformly calibrating sensors with linear and nonlinear responses, sensors with normal hydration membrane conditions, sensors with rapidly changing membrane and / or environmental conditions, and / or sensors with unbalanced enzyme-mediator conditions. In some embodiments, a method is provided for determining glucose concentration for a reference capillary glucose level using a simple conversion function, a connection function, and, in some embodiments, a final adjustment function. In yet other embodiments, adjustments to the connection function relating to in vivo sensor behavior are embedded in a process of deriving the connection function by combining in vitro and in vivo data. Further, the biosensor system for continuous glucose monitoring is provided with a membrane-covered biosensor and means operating under probe potential modulation as described herein.

[0054] As described above, uniform calibration is highly desirable and can be achieved by projecting in vivo sensitivity from the in vitro sensitivity of individual sensors, thereby reducing / eliminating the need for finger-prick calibration during CGM procedures. In this pursuit of a one-to-one correspondence between in vitro and in vivo sensitivity, individual sensor sensitivities from in vitro release tests are tracked in clinical studies to establish a correlation with initial in vivo sensitivity. It has been found that such a one-to-one correlation between in vitro and in vivo sensitivity does not actually exist (R...). 2 (Value is 0.04), such as Figure 1 As shown in the figure, the in vivo sensor sensitivity is compared to the in vitro sensor sensitivity according to the embodiments provided herein.

[0055] Even with a finite R 2 Values ​​(e.g., 0.5) exhibit some correlation, and changes in sensitivity over time still render the one-to-one correspondence of sensitivity from in vitro to in vivo unpredictable. Figure 2A and 2B In the study, we can see such sensitivity variations in six in vitro linear tests over 15 different days. Figure 2A Linearity testing of the working electrode (Iw) current over 15 days was demonstrated, and Figure 2B The linearity of the Iw current of glucose solutions of 50, 100, 200, 300 and 450 mg / dL according to the embodiments provided herein is demonstrated.

[0056] The sensitivity of subsequent linear tests is 35–45% higher than that of the tests conducted on day 1. Therefore, even though the initial in vivo sensitivity can be predicted by the in vitro sensitivity, subsequent changes in sensitivity will still ensure in-situ manual finger puncture calibration during continuous monitoring.

[0057] The two examples above demonstrate the desire to establish a unified calibration method that provides continuous analyte sensing results through a reliable and predictable connection from in vitro glucose to in vivo glucose, preferably without manual finger-prick calibration during continuous glucose monitoring. The embodiments described herein include systems and methods for connecting in vitro glucose to in vivo glucose by applying a probe potential modulation over a constant voltage applied to other aspects of the analyte sensor. Methods are provided for developing parameters for prediction equations (e.g., transition functions and / or connection functions) that can be used to continuously and accurately determine analyte concentrations from the analyte sensor.

[0058] In some embodiments, the CGM sensor may comprise four electrodes, including a working electrode, a blank electrode (or background electrode), a reference electrode, and a counter electrode enclosed in a membrane structure. The working electrode is adhered to and covered by a cross-linking enzyme layer, such as glucose oxidase, which is not part of the outer membrane structure and is used to catalyze the oxidation of the target analyte, glucose. The blank electrode provides a background signal (background current Ib) from all oxidizable chemicals at the same operating potential as the working electrode. Therefore, the difference current (Iw–Ib) between the two electrode currents Iw and Ib serves as the analyte response signal, effectively removing background / interference signal components by subtracting Ib. The reference electrode provides a potential reference for both the working and blank electrodes, while the counter electrode is used to perform all reverse redox reactions of both the working and blank electrodes to complete the electrochemical reaction. In some embodiments, when the membrane structure provides the function of repelling interfering substances, the current signal Iw from the working electrode can be used alone. In yet another embodiment, the function of repelling interference signals is embedded in the algorithm using the PPM signal and the primary signal from the working electrode current Iw, without subtracting the background current signal Ib.

[0059] In one or more embodiments, a probe potential modulation (PPM) sequence generated from the CGM transmitter can be simultaneously applied to the working and blank electrodes of the CGM sensor, and the Iw and Ib currents can be sampled. For example, the Iw and / or Ib currents can be sampled for each primary data point in a repeating cycle every 3 minutes, and sampled once every 2 seconds during the PPM cycle, thereby generating the Iw and Ib primary currents and the Iw and Ib PPM currents. Other sampling rates for the primary and / or PPM currents can be used. In one or more embodiments, the ratio of the PPM time (the total time of the PPM sequence) to the overall cycle time (in this case, 3 minutes) can be between about 0.05 and 0.5, and in some embodiments, about 0.1 to 0.2. Other ratios of the PPM time to the overall cycle time can be used.

[0060] Figure 3AThe applied voltage and current sampling timing is illustrated in an example PPM cycle or sequence according to the embodiments provided herein. Figure 3A In the example PPM sequence, there are six voltage potential steps 1-6. Other numbers, values, or types of voltage potential changes can be used. Figure 3B An example of a constant voltage application according to the embodiments provided herein is shown, as well as an example of timing for current measurement of individual primary data points (e.g., every 3-5 minutes). Figure 3B The square above illustrates the timing for sampling primary current signals (such as Iw and Ib) under a constant applied voltage. The first four sampling cycles are labeled 302, 304, 306, and 308, respectively. The figure is based on an embodiment provided herein. Figure 3C The output current (time-current curve) for the first four cycles of the primary data points is shown, along with the current generated by... Figure 3B constant applied voltage and Figure 3A The PPM currents generated by the PPM cycle (labeled as 302', 304', 306' and 308', respectively). Figure 3D and 3E This document illustrates embodiments of the invention provided herein. Figure 3C The superimposed PPM working electrode current Iw and PPM blank electrode current Ib of the first four PPM cycles are shown. Figure 3D and 3E The numbered circles in the text correspond to Figure 3B The example PPM sequence contains six voltage potential steps. Finally, Figure 3F Example timing and marking of the PPM current measured in a cycle according to the embodiments provided herein are shown. Figure 3F The numbered circles in the text correspond to Figure 3B The example PPM sequence contains six voltage potential steps.

[0061] In some embodiments, the PPM sequence in the first cycle 302' triggers current recording 2 seconds after the CGM system starts the sensor. That is, the PPM sequence is applied before recording the first primary data point at 3 minutes. Figure 3D and 3E As shown, both the PPM Iw and Ib currents in the first cycle are significantly higher than the PPM currents in subsequent cycles. Furthermore, compared to other PPM currents under non-start conditions, the currents at steps 1 and 5 in the first cycle are disproportionately higher than the currents at other PPM steps. This behavior serves as a good indicator of the start-up conditions of the CGM sensor.

[0062] To investigate the link between in vitro and in vivo glucose, a series of linear tests from different CGM sensors were performed at various points throughout the 15-day period. Target glucose concentrations in the glucose solution were 50, 100, 200, 300, and 450 mg / dL, with acetaminophen added at a concentration of 0.2 mg / dL as a substitute interfering substance for the normal interference level in in vitro testing. Using the current from the blank electrode representing the interference signal, the current difference (Iw–Ib) was used as the glucose response signal, which was plotted on... Figure 4A In this context, the current from all sensors used in the linearity test is used. Specifically, Figure 4A The composite CGM sensor response is demonstrated by in vitro linearity testing of different sensor sensitivities from various sensors (e.g., 10 sensors) according to embodiments provided herein.

[0063] like Figure 4A As shown, the glucose response signals (primary data points) from all sensors tested at different time points during 15 days of continuous in vitro monitoring are distributed over a wide range. The current response range is concentrated at (Iw-Ib) = 0.0965*G, crossing the zero point. ref At the regression line, the response population is defined by two lines ranging from approximately 0.0667 to 0.1778 nA / mg / dL. It can be seen that in the combined response, the upper line has a slope approximately three times that of the lower line. Figure 2B The response can be seen in one of the sensors tested with an initial low response, which was subsequently increased during later tests.

[0064] Considering the uncertainty in establishing a one-to-one correlation between in vitro and in vivo sensitivity, this paper discloses a method for establishing a connection between in vitro and in vivo glucose by applying a unified "transformation function" to data from a wide range of sensor responses, followed by applying a "connection function" to reduce glucose error to a narrow band. The unified transformation function calculates the original or "initial" glucose value G. 原始 = f(signal), where signal is the measured current signal, and f can be a linear or nonlinear function. When the conversion function f is nonlinear, sensitivity or response slope should not be applied (as described below).

[0065] In its simplest form, the unified transfer function can be a linear relationship between the measured current signal and a reference glucose level obtained from in vitro test data. For example, the unified transfer function could be the glucose signal (e.g., Iw-Ib), the slope, and the reference glucose level. ref Linear relationship between them:

[0066] (1) Signal = Slope * G ref

[0067] Make

[0068] (2)G ref =Signal / Slope

[0069] Where the slope represents the composite slope (slope) 复合 This is also known as the unified composite slope, selected based on in vitro sensor data, as described below. The above relationship can then be used to calculate the initial or raw glucose concentration G during CGM. 原始 :

[0070] (3)G 原始 =Signal / Slope 复合

[0071] In some embodiments, a nonlinear transformation function, such as a polynomial, may be used instead of a linear transformation function (e.g., to better fit the sensor’s changing response).

[0072] Now for reference Figure 4B-4D Describe other details, among which Figure 4B The original glucose concentration G according to the embodiments provided herein is shown. 原始 = (Iw – Ib) / S 复合 The distribution of relative error (deviation%). Figure 4C This document demonstrates the use of join functions from [the document's] embodiments. Figure 4B The output of glucose. Figure 4D This document illustrates embodiments of the invention provided herein. Figure 4A G of data 原始 and G 连接 Comparison of the percentage deviation.

[0073] Figure 4B The example shows a slope of 0.1333. 复合 Calculated G 原始 The percentage distribution of the deviation. This composite slope was pre-selected to be greater than 0.0965 starting from the center of the data group, such as... Figure 4A As shown (e.g., based on linear regression), and based on considerations of providing overlapping coverage between two subsets of the entire response population according to the sensor's manufacturing specifications. Used to calculate G 原始 The uniform composite slope makes the deviation % value more dispersed because there is no one-to-one corresponding slope to calculate glucose for each sensor, nor a separate slope for subsequent responses during the 15-day monitoring period. However, if the connection function is applied to the individual errors (deviation % = 100% * ΔG / G = 100% * (G... 原始 –G ref ) / G refThis allows for narrow band glucose delivery, and a single conversion simplifies in vitro-to-in vivo connections without calibration. This connection function is based on ΔG... 原始 / G ref The value is derived from the PPM parameter. This reduction is achieved by using the PPM parameter from G. 原始 The error band of the data is defined by the connection function, which is called the connection function that establishes the connection from external to internal sensing without calibration. This means that it adapts to all responses of the sensor to a narrow error band.

[0074] A connection function is referred to as a wide-range connection from in vitro glucose to in vivo glucose when it provides predicted in vivo glucose values ​​to a narrow error band without calibration. In this context, no one-to-one correspondence is sought between in vitro sensitivity and in vivo sensitivity (for a particular sensor). Instead, the connection function provides glucose values ​​from the sensor over a large sensitivity range, provided the sensor is responsive to glucose. The response can be linear or non-linear.

[0075] Utilizing the rich information about the CGM sensor derived from the PPM current, this connection function is derived from the PPM current and related parameters (in some cases, including the primary current, which is responsive to a constant applied voltage). Each response data point in the periodic cycle is then converted to a glucose value G via a composite conversion function. 原始 At that time, there exists an associated error or deviation %ΔG / G 原始 =(G 原始 –G ref ) / G ref By setting G 连接 =G ref ,but:

[0076] (4)G 连接 =G 原始 / (1+ΔG / G 原始 ) = G 原始 / (1+join function)

[0077] Where the connection function = ΔG / G 原始 =f(PPM parameters). One way to derive the connection function is by using the relative error ΔG / G 原始 The objective was set as a multivariate regression with input parameters derived from PPM parameters. The application of the PPM method in CGM sensor systems and the resulting PPM current... Figure 4C and 4D As shown in [the image]. Figure 4C In this context, output glucose refers to the final glucose concentration G after applying the join function. 连接 .

[0078] In summary, in some embodiments, the primary current i10 (e.g., Iw-Ib) can be used as part of a conversion function to convert the original current signal into the original or initial glucose value G. 原始 For example, G 原始 It can be calculated as follows:

[0079] (5) G 原始 = (Iw-Ib) / 0.1333

[0080] G can be used 原始 Other relationships with the primary current.

[0081] Once we know G 原始 Then, a connection function can be used to calculate the compensated or final glucose signal or concentration G. comp (also known as G) 连接 For example, a steady-state signal (primary current i10) and a non-steady-state signal (PPM signal) can be used as input parameters, using the relative error ΔG / G. 原始 As the target of multivariate regression, the connection function is derived from the in vitro data.

[0082] In one embodiment, the connection function is G 连接 =G 原始 / (1+connection function) is provided, where the connection function = f(PPM parameters) is derived from multiple regression, making the error deviating from the composite transformation function, such as the slope, the result. 复合 This is reduced / minimized to produce glucose values ​​within a narrow error band. In another embodiment, by... Ref For multiple regression with PPM input parameters set as the regression objective, the join function is simply the predictive equation. An example join function CF is provided below. It should be understood that other quantities and / or types of terms may be used.

[0083] (6)ΔG / G 原始 =-5.838261-3.511979*z53+6.5e-6*GR6-0.005973*GR53+0.012064*Gz61-0.00 5874*Gy52-0.038797*Gy43-14.75114*R63R51+0.385802*R64R43-6.134046*R6 5R52+0.059922*R51R43-0.009478*GR54R42-31.22696*z61z32+7.036651*z63z 42-5.153158*z64z42-10.93096*z65z54-1.981203*z51z31+1.578947*z51z21……

[0084] Where GR6 = G 原始 *R6, Gz61=G 原始 *z61,Gy52=G 原始 *y52, and R63R51=R63 / R51, R64R43=R64 / R43, z64z42=z64 / z42, etc., as further described below. Many other PPM currents and / or parameters can be used.

[0085] Example PPM current and PPM parameters

[0086] Example PPM current marking, such as Figure 3F As shown, the first number in the ixy format represents the potential step number (1-6), while the second number indicates which sampling current (1-3) within the potential step is used. For example, i10 represents the primary current; i11 is the first of the three currents recorded in step 1, and i13 is the third of the three currents recorded in step 1. Similarly, i63 is the third of the three currents recorded in step 6. Additional parameters are provided below.

[0087] Rx parameters: These parameters are generally given by dividing the final current within a step by the first current. For example, R1 = i13 / i11, R2 = i23 / i21, R3 = i33 / i31, R4 = i43 / i41, R5 = i53 / i51, and R6 = i63 / i61.

[0088] Xij parameters: The general format for this type of parameter is given by dividing the end current of the next potential step by the end current of the previous step. For example, parameter x61 is determined by i63 / i13, where i63 is the end current of potential step 6 out of the three currents recorded for each step, and i13 is the end current of step 1. For example, x61=i63 / i13, x62=i63 / i23, x63=i63 / i33, x64=i63 / i43, x65=i63 / i53, x51=i53 / i13, x52=i53 / i23, x53=i 53 / i33, x54=i53 / i43, x41=i43 / i13, x42=i43 / i23, x43=i43 / i33, X31=i33 / i13, x32=i33 / i23, and X21=i23 / i13.

[0089] Yij parameters: The general format for this type of parameter is given by dividing the end current of the next step by the first current of the previous step. For example, parameter y61 is determined by i63 / i11, where i63 is the end current of step 6 out of the three currents recorded for each step, and i11 is the first current of step 1. For example, y61 = i63 / i11, y62 = i63 / i21, y63 = i63 / i31, y64 = i63 / i41, y65 = i63 / i51, y51 = i53 / i11, y52 = i53 / i21, y53 = i53 / i31, y54 = i53 / i41, y41 = i43 / i11, y42 = i43 / i21, y43 = i43 / i31, y31 = i33 / i11, y32 = i33 / i21, and y21 = i23 / i11.

[0090] Zij parameter: The general format of this type of parameter is given by dividing the first current of the subsequent step by the end current of the previous step. For example, parameter z61 is determined by i61 / i13, where i61 is the first current of step 6 out of the three currents recorded for each step, and i13 is the end current of step 1. For example, z61 = i61 / i13, z62 = i61 / i23, z63 = i61 / i33, z64 = i61 / i43, z65 = i61 / i53, z51 = i51 / i13, z52 = i51 / i23, z53 = i51 / i33, z54 = i51 / i43, z41 = i41 / i13, z42 = i41 / i23, z43 = i41 / i33, z31 = i31 / i13, z32 = i31 / i23, and z21 = i21 / i13.

[0091] Wij parameter: The general format of this type of parameter is given by dividing the neutral current of the next step by the neutral current of the previous step. For example, parameter w61 is determined by i62 / i12, where i61 is the neutral current of step 6 out of the three currents recorded for each step, and i12 is the end current of step 1. For example, w61 = i62 / i12, z62 = i62 / i22, z63 = i62 / i32, z64 = i62 / i42, z65 = i62 / i52, w51 = i52 / i12, z52 = i52 / i22, z53 = i52 / i32, z54 = i52 / i42, w41 = i42 / i12, z42 = i42 / i22, z43 = i42 / i32, w31 = i32 / i12, z32 = i32 / i22, and w21 = i22 / i12.

[0092] Other PPM parameters may include normalized PPM currents ni11 = i11 / i10, ni12 = i12 / i10, ..., ni63 = i63 / i10, relative differences d11 = (i11–i12) / i10, d12 = (i12–i13) / i10, d21 = (i21–i22) / i10, d22 = (i22–i23) / i10, ..., d61 = (i61–i62) / i10 and d62 = (i62–i63) / i10, average currents for each PPM potential step av1 ​​= (i11+i12+i13) / 3, av2 = (i21+i22+i23) / 3, ... and their ratio av12 = av1 / av2, etc.

[0093] Other types of parameters can also be used, such as PPM current difference or relative difference carrying equivalent or similar information. To demonstrate the feasibility of overcoming issues such as different sensor sensitivities, initial warm-up time, long-term sensitivity variations, and different background signals caused by varying amounts of interfering substances, the above parameters, along with their temperature cross-terms, can be used as inputs to a simple form of multivariate regression. Additional terms / parameters can be provided in the regression.

[0094] Using the join function method, Figure 5A It shows Figure 2B The non-uniform response shown is compressed into essentially a single output glucose response line, without considering individual response sensitivity. This means that as long as the sensor produces a glucose-responsive signal within its response range, future sensitivity and its variations are irrelevant. This effectively provides glucose determination without requiring sensor calibration. This method is further applied to... Figure 4A The dataset shown represents a large in vitro linearity test of 10 CGM sensors. Figure 4B In the middle, a uniform transformation slope is used. 复合 The resulting broadly distributed error, expressed as a percentage of deviation, may produce a narrow error band without requiring additional calibration using a connection function, such as... Figure 5B As shown. Due to the use of a preset slope. 复合 To calculate the initial glucose concentration G 原始 Therefore, as long as the sensor response is within the target response range, the connection function can significantly reduce the error of all data points without further calibration, regardless of the sensor response. Figure 4A As shown, the upper boundary of the response is almost three times the lower boundary of the response range.

[0095] The example embodiments provided herein describe the use of a transformation function to determine an initial glucose concentration (e.g., based on a composite slope) and a link function developed using multivariate regression to determine a final glucose concentration from the initial glucose concentration. Alternatively, multivariate regression can be used to develop a single predictive equation without using separate transformation and link functions. Example predictive equations are described, for example, in U.S. Patent Application No. 16 / 782,974, filed February 5, 2020, entitled “Apparatus and Methods of Probing Sensor Operation and / or Determining Analyte Values ​​During Continuous Analyte Sensing,” which is hereby incorporated herein by reference in its entirety for all purposes.

[0096] In Table 1 below, the initial transformation function G is... 原始 The results and the connection function G 连接 The results are consistent with those of a single prediction equation G. 预测 The results were compared (as described in the prediction equations in previously incorporated U.S. Patent Application No. 16 / 782,974). Table 1 shows the following main advantages of the connection function of the PPM method. For G 原始 The CGM signal (primary data point) passes through a single slope of the sensor. 复合 =0.1333 is converted to glucose value, and the sensor has a wide response range that varies continuously over time. The results are widely distributed, with individual glucose errors deviating significantly from those caused by the composite slope. 复合 The described central behavior (MARD = 23.4%, and only 43.1% of the data within the ±20% error margin) was observed. Using a single prediction equation, error spread was reduced by 50%, and the MARD value decreased by over 60% to 7.79%, with the data population remaining within the ±20% error margin by approximately 91.3%. If the slope... 复合 If the same dataset is processed by a connection function instead of a transformation function with a conversion function of 0.1333, then the error will be from G. 原始 Further reduction, even compared to G 预测 There has been some improvement, with MARD decreasing from 23.4% to 5.55%, and the data populations approaching 94.1% within ±15% error margins and approaching 98.1% within ±20% error margins.

[0097] Table 1: G 原始 G 预测 G 连接 Comparison of glucose calculations

[0098]

[0099] Although both the single prediction equation method and the link function method are derived from multiple regression of PPM parameters, the regression ΔG / G will be used to obtain the link function. 原始 Using the relative error as the target reduces the error significantly more than a single prediction equation method in the low glucose range. This can be achieved... Figure 6 As seen in the diagram, the deviation percentage from a single prediction equation and a connection function is used to compare the values, where the hollow circle (○) represents G. 预测 The deviation is %, while the solid triangle (▲) indicates the deviation from G. 连接 The deviation percentage. Since the CGM sensor responds to both the glucose concentration in the primary data point and the PPM current, the connection function and the composite transformation function such as (Iw–Ib) / slope. 复合 The coupling provides glucose determination with a narrow error band. Therefore, it offers a wide range of connections between in vitro and in vivo glucose testing.

[0100] During the sensor manufacturing process, a wide range of sensor responses may be encountered in response to the release specifications. Figure 7A and 7B An example sensor response population according to the embodiments provided herein is shown. The lower and upper boundaries are approximately described by two lines y = 0.066x and y = 0.475x. For the entire response range, the slope... 复合 A simple linear line with a correlation coefficient of 0.1471 characterizes the center of the in vitro response of the CGM sensor. The correlation coefficient R0 2 =0.5677 indicates that, due to the wide response range of the sensor, the signal (Iw–Ib) and the reference glucose value G... Ref There is only a moderate correlation between them. For example... Figure 7A and 7B As shown, for the population, a portion of the response is within the upper bound, and a portion of the response is within the lower bound (indicated by the black triangle ▲). Figure 7B (Response within the upper limit of the slope). 复合 A single transformation function can be used to transform a signal into G. 原始 *(Iw–Ib) / Slope 复合 This is followed by a single connection function that reduces the original deviation % value to a narrow band.

[0101] Given the wide response range of the CGM sensor, the expected production response range can also be subdivided into more than one subset of responses. Figure 7C and 7D This document demonstrates how, according to the embodiments provided herein, a wide range of sensor responses can be subdivided into two subsets: an upper limit and a lower limit. Figures 7A-7DIn the diagram, the hollow circle (○) represents the entire response range (e.g., ...). Figure 7A and 7B As shown), while solid triangles (▲) represent subsets of the upper or lower response limits (e.g., ...). Figure 7C and 7D (As shown).

[0102] Figure 7C The content can be found Figure 4A As seen, it has an extended y-axis, with the lower and upper boundaries defined by two lines approximately ranging from 0.0667 to 0.1778 nA / mg / dL. For the lower-bound response subset, the correlation coefficient of the linear regression of the response increases from 0.5677 (the entire response population) to 0.8355. For the upper-bound response subset, the correlation coefficient of the linear regression of the response increases from 0.5677 to 0.725. Figure 7D If a polynomial is used to describe the central response of the upper bound, the correlation coefficient further increases to 0.8213. In this case, the transformation function is no longer related to the slope. 复合 A simple linear relationship. However, it is still applicable through the initial transformation function ΔG / G. 原始 =f(PPM parameters) is the connection function that connects the initial error, such that G 连接 =G 原始 / (1+connection function). In each case of the response subset, the response function signal = f(glucose) with an increased correlation coefficient helps to reduce the initial error ΔG / G. 原始 The connection functions further reduce the initial error, thus providing the CGM sensor with significantly better glucose concentration determination than a large subset of responses. The two connection functions are designed to have overlapping coverage of the edge responses between the two response subsets. For each of the two response subsets, a wide-range connection from in vitro to in vivo glucose is provided, where no further calibration is required to produce glucose values ​​within a narrow error band. Therefore, multiple transformation functions and / or multiple connection functions can be used.

[0103] Linearizing the glucose output from the nonlinear signal response is another advantage of the connection function. Figure 7BIn the upper response region, the glucose response signal is inherently nonlinear, which may be due to the unbalanced enzyme-mediator conditions of the biosensor. In biosensors using glucose oxidase (GO) as the enzyme, mediator oxygen in and around the tissue-surrounded sensor can become limited, especially for sensors with very high response sensitivity. This can be considered an unbalanced enzyme-mediator condition where, at medium to high glucose concentrations, the oxidative state of the enzyme cannot be completely and timely regenerated by mediator oxygen. Subsequently, a nonlinear response curve is obtained. In addition to adapting to a wide range of sensor responses with respect to changes in sensor sensitivity, a connection function can be used to generate a linearized glucose output value from the nonlinear response. Figures 8A-8D This illustrates such a transformation of the nonlinear response for generating linear output glucose values ​​from a CGM sensor, where hollow circles (○) represent the original primary data points, and solid rhombuses (◆) represent the output glucose values ​​transformed from the primary data points via a connection function. By doing so, the connection function adapts not only to different sensor sensitivities but also to the nonlinear response at different times during continuous in vitro glucose sensing. Figures 8A-8D The data are shown on days 1, 3, 7, and 14, respectively, and these adaptations were performed without further calibration.

[0104] Operating under rapidly changing membrane conditions / environments and reporting accurate glucose values ​​offers further advantages, provided by the use of connection functions without the need for further calibration. Rapidly changing membrane conditions and / or environments are primarily reflected in the time-varying output signal. An example of such conditions is the sensor response immediately after the sensor is immersed in the in vitro test solution, such as... Figure 9A (using 50 mg / dL glucose solution) and Figure 9B (Using 100 mg / dL glucose solution) as shown, or the sensor response immediately after subcutaneous insertion into the skin. In both in vitro and in vivo cases, relative to its initial dry state relative to surrounding solvent molecules and / or tissue, the sensor membrane will undergo rapid changes in its structure, including enzymes and the outer membrane. For Figure 9A and 9B In examples such as rehydration, this change is often referred to as rehydration and is most pronounced in the change in initial response, with signal attenuation lasting 30-40 minutes, even at constant glucose concentrations. Such subtle changes during rehydration are also reflected in the PPM current, and the circuit is incorporated into the regression of the connection function. The connection function derived from the PPM parameters can eliminate / minimize such attenuation effects during the initial state of continuous glucose sensing. The results can be found in... Figure 9A and 9BAs seen in the glucose concentration plot, the initial attenuation portion is effectively removed by applying the connection function to the current response. This occurs without the need for additional calibration during continuous glucose sensing.

[0105] Another method for determining glucose from capillary glucose references can be provided via a final adjustment function after establishing a wide range of connections (e.g., using transition and connection functions as described above). This adjustment can be implicitly implemented by combining in vitro and in vivo data. Given the discrete nature of in vitro testing, where only discrete levels of analyte concentration are used, combining data from in vitro and in vivo studies adapts to different information sources embedded in data from different test settings. The goal is to provide accurate predictions of in vivo analyte concentrations. Therefore, in vitro tests can be designed to provide different combinations and variations of sensors under different conditions, such as alternating high and low oxygen conditions. Figure 13A and 13B Examples of combining in vitro and in vivo data with and without background subtraction of the working electrode current are provided according to the embodiments provided herein. Figure 13C and 13D The embodiments provided herein further separate the datasets from in vivo and in vitro tests, illustrating the range of sensitivity variations from different sensors and under different conditions. In this example, the in vitro data were collected at room temperature. Under these lower temperature conditions, the response is relatively lower compared to the in vivo data (collected at approximately 32°C). This temperature effect is compensated for by an implicit implementation using a connection function derived from a regression process with data from different temperatures. This adjustment function can account for interstitial fluid hysteresis, even if the in vitro-to-in vivo connection is performed without calibration, for example, using relevant data from clinical studies.

[0106] Table 2 below shows the data results before (G-raw) and after (G-final) the application of a join function that adjusts a combined dataset of in vitro data from multiple sensors and in vivo data from 7-day CGM operations from multiple sensors. Uniform calibration slope 复合The mean deviation % for the in vivo data population is 0.15385; that is, G-raw = (Iw – Ib) / 0.15385. The mean deviation % for the in vivo data is slightly negative, but the mean deviation % for the in vitro data is essentially negative. This large negative mean deviation % partially reflects the temperature effect of the in vitro data collected at approximately 22–25 °C, while the target temperature for the uniform calibration slope is 32 °C under subcutaneous conditions. After compensation through a connection function with implicit adjustment, the mean deviation % values ​​for the in vivo, in vitro, and combined data are virtually zero. Furthermore, the MARD % values ​​are significantly reduced compared to the G-raw results. This demonstrates the effectiveness of combining uniform calibration with the connection function to overcome the differences in sensitivity between different sensors and their variations over time.

[0107] Table 2: G 原始 and G 最终 Comparison of glucose calculations

[0108]

[0109] In summary, the use of probe potential modulation (PPM) as described herein provides sufficient self-sufficiency to accommodate sensitivity differences between different sensor batches, sensitivity variations throughout the continuous monitoring period, background variations due to varying levels of interference, and the nonlinear effects of the glucose signal that appear immediately after insertion and activation (providing a shortened warm-up time). This can be achieved via PPM current and without the need for factory and / or field calibration.

[0110] Figure 10 An example method 1000 for determining glucose values ​​during continuous glucose monitoring measurements, according to embodiments provided herein, is illustrated. References Figure 10 Method 1000 includes, in block 1002, providing a CGM device comprising a sensor, memory, and processor (e.g., Figure 11A-11B The CGM device 1100 or 1150). In block 1004, method 1000 includes applying a constant voltage potential to the sensor (e.g., Figure 3B In block 1006, method 1000 includes measuring a primary current signal generated by a constant voltage potential and storing the measured primary current signal in a memory. In block 1008, method 1000 includes applying a probe potential modulation sequence to a sensor (e.g., E0 in the context of E0). Figure 3AThe method 1000 includes measuring a probe potential modulation current signal generated by the probe potential modulation sequence in block 1010, and storing the measured probe potential modulation current signal in memory. The method 1000 further includes: determining an initial glucose concentration based on a transition function and a primary current signal in block 1012; determining a connection function value based on the primary current signal and multiple probe potential modulation current signals in block 1014; and determining a final glucose concentration based on the initial glucose concentration and the connection function value in block 1016. The final glucose concentration can be transmitted to the user (e.g., via...). Figure 11A Or the 11B monitor 1117 or 1122).

[0111] In some embodiments, the PPM cycle or sequence is designed to take no more than half the time of the primary data cycle (e.g., 3-5 minutes) to allow sufficient time for a constant voltage to be applied to the working electrode to restore steady-state conditions before recording the next primary data point. In some embodiments, the PPM cycle may be approximately 1 to 90 seconds, or no more than 50% of a typical 180-second primary data cycle.

[0112] In one or more embodiments, the PPM cycle can be about 10–40 seconds, and / or include more than one modulation potential step near the redox plateau of the mediator. In some embodiments, the PPM sequence can be about 10–20% of a conventional primary data point cycle. For example, when a conventional primary data point cycle is 180 seconds (3 minutes), a 36-second PPM cycle is 20% of the primary data point cycle. The remaining time of the primary data cycle allows the steady-state conditions to recover under a constant applied voltage. For potential steps in the PPM cycle, the duration is transient, making the boundary conditions of the measurable substance resulting from these potential steps non-steady-state. Therefore, in some embodiments, each potential step can be about 1–15 seconds, in other embodiments about 3–10 seconds, and in still other embodiments about 4–6 seconds.

[0113] In some embodiments, the probe potential modulation can step into the potential region of non-diffusion-limited redox conditions, or the kinetic region of the mediator (meaning the output current depends on the applied voltage, where the higher the applied voltage, the greater the output current generated from the electrode). For example... Figure 3A Steps 2 and 3 are two potential steps in the dynamic region of the dielectric that generate unsteady output current from the electrodes. When the potential steps reverse, the applied voltage recovers to the same amplitude. Figure 3A (Step 4 and Step 5) to detect the unsteady output current from the electrodes.

[0114] Different implementations can be adopted, accompanied by unsteady-state conditions. For example, the unsteady-state conditions can also be achieved by directly reaching the target potential in one step. Figure 3A The step 2) and return to the starting potential ( Figure 3A The first step in the probe potential is followed by a second probe potential step, which directly reaches different potentials in dynamic regions with different unsteady conditions. Figure 3A The step 3 in the middle) and then directly return to the starting potential ( Figure 3A The step (1 or 6) is used for detection. The purpose is to modulate the applied potential, thereby creating an alternation of steady-state and non-steady-state conditions for the measurable substance at the electrode surface, from which the signal from the non-steady-state can be used to determine the analyte concentration.

[0115] Figure 11A A high-level block diagram of an example CGM device 1100 according to embodiments provided herein is shown. Although not shown in... Figure 11A As shown, but it should be understood, various electronic components and / or circuits are configured to be coupled to a power source, such as, but not limited to, a battery. The CGM device 1100 includes a bias circuit 1102, which can be configured to be coupled to the CGM sensor 1104. The bias circuit 1102 can be configured to apply a bias voltage, such as a continuous DC bias, to the analyte-containing fluid through the CGM sensor 1104. In this example embodiment, the analyte-containing fluid can be a human serous molecule, and the bias voltage can be applied to one or more electrodes 1105 of the CGM sensor 1104 (e.g., working electrode, background electrode, etc.).

[0116] The bias circuit 1102 can also be configured to apply a bias to the CGM sensor 1104 such as Figure 3A The probe potential modulation sequence shown is another probe potential modulation sequence. For example, the probe potential modulation sequence can be applied for each primary data point, as referenced above. Figures 3A-3F As described. For example, a probe potential modulation sequence can be applied before, after, or both before and after measuring primary data points.

[0117] In some embodiments, the CGM sensor 1104 may comprise two electrodes, and a bias voltage and probe potential modulation (PPM) may be applied across these electrodes. In such cases, the current through the CGM sensor 1104 can be measured. In other embodiments, the CGM sensor 1104 may comprise three electrodes, such as a working electrode, a counter electrode, and a reference electrode. In such cases, for example, a bias voltage and probe potential modulation may be applied between the working electrode and the reference electrode, and the current through the working electrode can be measured. The CGM sensor 1104 contains chemicals that react with a glucose-containing solution in a reduction-oxidation reaction, said chemicals affecting the concentration of charge carriers and the time-dependent impedance of the CGM sensor 1104. Example chemicals include glucose oxidase, glucose dehydrogenase, etc. In some embodiments, mediators such as ferricyanide or ferrocene may be used.

[0118] For example, the continuous bias voltage generated and / or applied by the bias circuit 1102 relative to the reference electrode can be in the range of about 0.1 to 1 volt. Other bias voltages can be used. An example probe potential modulation value has been described previously.

[0119] In response to probe potential modulation and a constant bias voltage, the probe potential modulated (PPM) current and the non-probe potential modulated (NPPM) current in the analyte-containing fluid can be transmitted from the CGM sensor 1104 to the current measurement (I0). 测量 Circuit 1106 (also referred to as a current sensing circuit system). The current measurement circuit 1106 can be configured to sense and / or record a current measurement signal having a magnitude indicating the amount of current transmitted from the CGM sensor 1104 (e.g., using a suitable current-to-voltage converter (CVC)). In some embodiments, the current measurement circuit 1106 may include a resistor having a known nominal value and a known nominal accuracy (e.g., 0.1% to 5%, or even less than 0.1% in some embodiments), through which the current transmitted from the CGM sensor 1104 passes. The voltage generated across the resistor in the current measurement circuit 1106 represents the magnitude of the current and may be referred to as the current measurement signal (or the original glucose signal). 原始 ).

[0120] In some embodiments, sampling circuit 1108 may be coupled to current measurement circuit 1106 and may be configured to sample the current measurement signal and generate digitized time-domain sampled data (e.g., a digitized glucose signal) representing the current measurement signal. For example, sampling circuit 1108 may be any suitable A / D converter circuit configured to receive the current measurement signal as an analog signal and convert it into a digital signal with a desired number of bits as output. In some embodiments, the number of bits output by sampling circuit 1108 may be sixteen, but more or fewer bits may be used in other embodiments. In some embodiments, sampling circuit 1108 may sample the current measurement signal at a sampling rate in the range of approximately 10 samples per second to 1000 samples per second. Faster or slower sampling rates may be used. For example, a sampling rate such as approximately 10 kHz to 100 kHz may be used and downsampled to further reduce the signal-to-noise ratio. Any suitable sampling circuit system may be employed.

[0121] Still referencing Figure 11A The processor 1110 may be coupled to the sampling circuit 1108 and may be further coupled to the memory 1112. In some embodiments, the processor 1110 and the sampling circuit 1108 are configured to communicate directly with each other via a wired path (e.g., via a serial or parallel connection). In other embodiments, the coupling between the processor 1110 and the sampling circuit 1108 may be implemented via the memory 1112. In this embodiment, the sampling circuit 1108 writes digital data to the memory 1112, and the processor 1110 reads the digital data from the memory 1112.

[0122] Memory 1112 may store one or more equations 1114 (e.g., prediction equations such as transition functions and connection functions, single prediction equations, multiple transition and / or connection functions, etc.), such as one or more connection functions, for determining glucose values ​​based on primary data point (NPPM current) and probe potential modulation (PPM) current (from current measurement circuit 1106 and / or sampling circuit 1108). For example, in some embodiments, two or more prediction equations may be stored in memory 1112, each equation for a different segment (time period) of data collected by CGM. In some embodiments, memory 1112 may contain prediction equations (e.g., connection functions) based on a primary current signal generated by applying a constant voltage potential to a reference sensor, and multiple probe potential modulation current signals generated by applying a probe potential modulation sequence between primary current signal measurements.

[0123] The memory 1112 may also store multiple instructions therein. In various embodiments, the processor 1110 may be a computing resource, such as, but not limited to, a microprocessor, a microcontroller, an embedded microcontroller, a digital signal processor (DSP), a field-programmable gate array (FPGA) configured to execute as a microcontroller, etc.

[0124] In some embodiments, a plurality of instructions stored in memory 1112 may include instructions that, when executed by processor 1110, cause processor 1110 to perform the following operations: (a) cause CGM device 1100 (via bias circuit 1102, CGM sensor 1104, current measurement circuit 1106, and / or sampling circuit 1108) to measure current signals (e.g., primary current signals and probe potential modulated current signals) from interstitial fluid; (b) store the current signals in memory 1112; (c) calculate prediction equation (e.g., connection function) parameters, such as the ratio (and / or other relationships) of currents from different pulses, voltage steps, or other voltage changes within the probe potential modulation sequence; (d) use the calculated prediction equation (e.g., connection function) parameters to calculate glucose values ​​(e.g., concentrations) using prediction equations (e.g., transformation functions incorporating connection functions); and / or (e) transmit glucose values ​​to a user.

[0125] Memory 1112 can be any suitable type of memory, such as, but not limited to, one or more of volatile and / or non-volatile memory. Volatile memory can include, but is not limited to, static random access memory (SRAM) or dynamic random access memory (DRAM). Non-volatile memory can include, but is not limited to, electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory (e.g., EEPROM of one type in either non-configurable or non-configurable, and / or stacked or planar arrangement and / or single-level cell (SLC), multi-level cell (MLC), or combined SLC / MLC arrangement), resistive memory, filamentary memory, metal-oxide memory, phase-change memory (e.g., chalcogenide memory), or magnetic memory. For example, memory 1112 can be packaged as a single chip or multiple chips. In some embodiments, memory 1112 can be embedded in an integrated circuit, such as, for example, an application-specific integrated circuit (ASIC), along with one or more other circuits.

[0126] As described above, memory 1112 may have a plurality of instructions stored therein, which, when executed by processor 1110, cause processor 1110 to perform various actions specified by one or more of the stored instructions. Memory 1112 may further have portions reserved for one or more "scratchpad" storage areas, which can be used by processor 1110 in response to the execution of one or more of the instructions for read or write operations.

[0127] exist Figure 11A In some embodiments, the bias circuit 1102, CGM sensor 1104, current measurement circuit 1106, sampling circuit 1108, processor 1110, and memory 1112 containing prediction equation 1114 may be housed within the wearable sensor portion 1116 of the CGM device 1100. In some embodiments, the wearable sensor portion 1116 may include a display 1117 for displaying information such as glucose concentration information (e.g., without the use of external devices). The display 1117 may be any suitable type of human-perceptible display, such as, but not limited to, liquid crystal display (LCD), light-emitting diode (LED) display, or organic light-emitting diode (OLED) display.

[0128] Still referencing Figure 11A The CGM device 1100 may further include a portable user device portion 1118. A processor 1120 and a display 1122 may be housed within the portable user device portion 1118. The display 1122 may be coupled to the processor 1120. The processor 1120 may control text or images displayed by the display 1122. The wearable sensor portion 1116 and the portable user device portion 1118 may be communicatively coupled. In some embodiments, for example, the communicative coupling of the wearable sensor portion 1116 and the portable user device portion 1118 may be via wireless communication through transmitter circuitry and / or receiver circuitry, such as the transmit / receive circuitry TxRx 1124a in the wearable sensor portion 1116 and the transmit / receive circuitry TxRx 1124b in the portable user device 1118. Such wireless communication may be conducted in any suitable manner, including, but not limited to, standard-based communication protocols, such as... Communication protocols are used. In various embodiments, wireless communication between the wearable sensor portion 1116 and the portable user device portion 1118 may alternatively be conducted via near field communication (NFC), radio frequency (RF) communication, infrared (IR) communication, or optical communication. In some embodiments, the wearable sensor portion 1116 and the portable user device portion 1118 may be connected via one or more wires.

[0129] Display 1122 can be any suitable type of human-perceptible display, such as, but not limited to, liquid crystal display (LCD), light-emitting diode (LED) display, or organic light-emitting diode (OLED) display.

[0130] Now for reference Figure 11B An example CGM device 1150 is shown, which is similar to Figure 11A The illustrated embodiments differ in component division. In the CGM device 1150, the wearable sensor portion 1116 includes a bias circuit 1102 coupled to the CGM sensor 1104, and a current measurement circuit 1106 coupled to the CGM sensor 1104. The portable user device portion 1118 of the CGM device 1150 includes a sampling circuit 1108 coupled to the processor 1120, and a display 1122 coupled to the processor 1120. The processor 1120 is further coupled to a memory 1112, which may contain a prediction equation 1114 stored therein. In some embodiments, the processor 1120 in the CGM device 1150 may also perform, for example, operations by... Figure 11A The processor 1110 of the CGM device 1100 performs the functions previously described. The wearable sensor portion 1116 of the CGM device 1150 can be compared to... Figure 11A The CGM device 1100 is smaller and lighter, and therefore less invasive, because the sampling circuitry 1108, processor 1110, memory 1112, etc., are not included. Other component configurations can be used. For example, as Figure 11B In a variant of the CGM device 1150, the sampling circuit 1108 can be retained on the wearable sensor section 1116 (so that the portable user device 1118 receives the digitized glucose signal from the wearable sensor section 1116).

[0131] Figure 12This is a side view of an example glucose sensor 1104 according to embodiments provided herein. In some embodiments, the glucose sensor 1104 may include a working electrode 1202, a reference electrode 1204, a counter electrode 1206, and a background electrode 1208. The working electrode may include a conductive layer coated with a chemical that reacts with a glucose-containing solution in a reduction-oxidation reaction (the chemical affects the concentration of charge carriers and the time-dependent impedance of the CGM sensor 1104). In some embodiments, the working electrode 1202 may be formed of platinum or a roughened platinum surface. Other working electrode materials may be used. Example chemical catalysts (e.g., enzymes) for the working electrode 1202 include glucose oxidase, glucose dehydrogenase, etc. For example, the enzyme component may be immobilized to the electrode surface by a crosslinking agent such as glutaraldehyde. An outer membrane layer may be applied to the enzyme layer to protect the entire internal assembly containing the electrode and the enzyme layer. In some embodiments, a mediator such as ferricyanide or ferrocene may be used. Other chemical catalysts and / or mediators may be used.

[0132] In some embodiments, the reference electrode 1204 may be formed of Ag / AgCl. The counter electrode 1206 and / or the background electrode 1208 may be formed of a suitable conductor such as platinum, gold, palladium, etc. Other materials may be used for the reference electrode, counter electrode, and / or background electrode. In some embodiments, the background electrode 1208 may be the same as the working electrode 1202, but without a chemical catalyst. The counter electrode 1206 may be isolated from other electrodes by an insulating layer 1210 (e.g., polyimide or another suitable material).

[0133] While the description primarily pertains to the determination of glucose concentration during continuous glucose monitoring, it should be understood that the embodiments described herein can be used in conjunction with other continuous analyte monitoring systems (e.g., cholesterol, lactate, uric acid, alcohol, or other analyte monitoring systems). For example, one or more predictive equations, such as one or more transition functions and / or connection functions, can be developed for any analyte to be monitored by using probe potential modulation of the output current and its associated cross terms.

[0134] The foregoing description discloses exemplary embodiments of the present disclosure. Modifications to the apparatus and methods disclosed above that fall within the scope of this disclosure will be readily apparent to those skilled in the art. Therefore, although the present disclosure has been disclosed in conjunction with exemplary embodiments, it should be understood that other embodiments may also fall within the scope of the present disclosure as defined by the appended claims.

Claims

1. A method for determining glucose values ​​during continuous glucose monitoring (CGM) measurements to adapt sensor sensitivity to a sensor without field calibration, the method comprising: A CGM device is provided, comprising a sensor, a memory, and a processor, wherein the memory contains connection functions and conversion functions; A constant voltage potential is applied to the sensor; The primary current signal generated by the constant voltage potential is measured, and the primary current signal is stored in the memory; A probe potential modulation sequence is applied to the sensor, wherein applying the probe potential modulation sequence includes sequentially applying a plurality of voltage potentials, each of the plurality of voltage potentials being applied at different times during the continuous glucose monitoring, wherein probe potential modulation is used to provide information to adapt to sensitivity changes within the continuous monitoring period; Measure multiple probe potential modulation current signals generated by the probe potential modulation sequence, and store the multiple probe potential modulation current signals in the memory; The system alternates between steady-state conditions associated with the constant voltage potential and non-steady-state conditions associated with the probe potential modulation sequence, wherein each measured primary current signal during the steady-state conditions is accompanied by a set of associated current signals measured during the subsequent non-steady-state conditions. The initial glucose concentration value is determined based on the conversion function and the primary current signal; A connection function value is determined based on the primary current signal and the plurality of probe potential modulated current signals, wherein the connection function value is determined at least in part based on the ratio of a first probe potential modulated current signal to a second probe potential modulated current signal of the plurality of probe potential modulated current signals, wherein the second probe potential modulated current signal is a first current signal measured at a first different time period at the first voltage potential of the plurality of voltage potentials, wherein the first probe potential modulated current signal is a termination current signal measured at a second different time period at the second voltage potential of the plurality of voltage potentials, wherein the first different time period occurs before the second different time period; and The final glucose concentration value is determined based on the initial glucose concentration value and the connection function value.

2. The method according to claim 1, wherein applying the probe potential modulation sequence includes applying a first voltage potential greater than the constant voltage potential, a second voltage potential less than the constant voltage potential, a third voltage potential less than the second voltage potential, and a fourth voltage potential greater than the third voltage potential.

3. The method according to claim 1, wherein determining the connection function value based on the plurality of probe potential modulated current signals includes determining the connection function value based on the ratio of the plurality of probe potential modulated current signals.

4. The method of claim 2, wherein determining the connection function value based on the plurality of probe potential modulated current signals comprises determining the connection function value based on the ratio of the plurality of probe potential modulated current signals of different potential voltage steps within the probe potential modulation sequence.

5. The method according to claim 1, wherein the primary current signal and the plurality of probe potential modulated current signals are working electrode current signals.

6. The method of claim 1, wherein the primary current signal is measured every 3 to 15 minutes.

7. The method of claim 4, wherein the probe potential modulation sequence comprises four or more voltage steps.

8. The method of claim 1, wherein the conversion function includes a slope determined based on in vitro data from multiple sensors.

9. The method of claim 8, wherein the slope is determined based on the in vitro working electrode current relative to reference glucose data.

10. The method of claim 9, wherein the external data is derived from the primary current signal.

11. The method of claim 1, wherein the connection function value is calculated based on the plurality of probe potential modulated current signals measured for the reference CGM sensor in response to the probe potential modulation sequence applied to the reference CGM sensor before or after measuring the primary current signal for the reference CGM sensor.

12. A non-transitory computer-readable medium containing computer-executable instructions, which, when executed by a processor, perform a method for determining glucose values ​​during continuous glucose monitoring (CGM) measurements to adapt sensor sensitivity to a sensor without field calibration, the method comprising: A constant voltage potential is applied to the sensor; The primary current signal generated by the constant voltage potential is measured and the primary current signal is stored. A probe potential modulation sequence is applied to the sensor, wherein applying the probe potential modulation sequence includes sequentially applying a plurality of voltage potentials, each of the plurality of voltage potentials being applied at different times during the continuous glucose monitoring, wherein probe potential modulation is used to provide information to adapt to sensitivity changes within the continuous monitoring period; Measure and store multiple probe potential modulation current signals generated by the probe potential modulation sequence; The system alternates between steady-state conditions associated with the constant voltage potential and non-steady-state conditions associated with the probe potential modulation sequence, wherein each measured primary current signal during the steady-state conditions is accompanied by a set of associated current signals measured during the subsequent non-steady-state conditions. The initial glucose concentration value is determined based on the conversion function and the primary current signal; A connection function value is determined based on the primary current signal and the plurality of probe potential modulated current signals, wherein the connection function value is determined at least in part based on the ratio of a first probe potential modulated current signal to a second probe potential modulated current signal of the plurality of probe potential modulated current signals, wherein the second probe potential modulated current signal is a first current signal measured at a first different time period at the first voltage potential of the plurality of voltage potentials, wherein the first probe potential modulated current signal is a termination current signal measured at a second different time period at the second voltage potential of the plurality of voltage potentials, wherein the first different time period occurs before the second different time period; and The final glucose concentration value is determined based on the initial glucose concentration value.

13. The computer-readable medium of claim 12, wherein applying the probe potential modulation sequence to the sensor comprises applying a first voltage potential greater than the constant voltage potential, a second voltage potential less than the constant voltage potential, a third voltage potential less than the second voltage potential, and a fourth voltage potential greater than the third voltage potential.

14. The computer-readable medium according to claim 13, in, The final glucose concentration value is determined at least in part based on the connection function value.

15. The computer-readable medium of claim 14, wherein the connection function value is based on the ratio of the plurality of probe potential modulated current signals of different potential voltage steps within the probe potential modulation sequence.

16. The computer-readable medium of claim 12, wherein the primary current signal and the plurality of probe potential modulated current signals are working electrode current signals.

17. The computer-readable medium according to claim 12, in, Measuring the primary current signal includes generating a digitized current signal from the primary current signal, and The measurement of the plurality of probe potential modulation current signals includes generating a plurality of digital probe potential modulation current signals from the plurality of probe potential modulation current signals.

18. The computer-readable medium of claim 12, further comprising transmitting information representing the final glucose concentration value to a portable user device for presentation to a user of the portable user device.

19. The computer-readable medium of claim 12, wherein the conversion function includes a slope determined based on in vitro data from a plurality of sensors.

20. The computer-readable medium of claim 19, wherein the slope is determined based on the in vitro working electrode current relative to reference glucose data.

21. The computer-readable medium of claim 20, wherein the external data originates from the primary current signal.

Images