Two-stage anti-interference online blood glucose monitoring method and system for hemodialysis machine

Abstract

The present application relates to hemodialysis and blood glucose monitoring technical field, disclose a two-stage anti-interference online blood glucose monitoring method and system for hemodialysis machine. Method includes in situ micro blood sample diversion processing, multi-source signal synchronous acquisition and pretreatment, two-stage anti-interference correction, blood glucose change trend prediction, hierarchical early warning and parameter linkage five steps;The system is composed of blood sample diversion processing, multi-source signal acquisition pretreatment, blood glucose detection anti-interference correction, blood glucose change trend prediction, hierarchical early warning and parameter linkage module in turn interact. The present application adopts two-stage anti-interference mechanism to make the detection error ≤±4.5%, introduces blood viscosity dynamic correction to realize non-invasive blood sample processing, blood sampling self-calibration gets rid of blood sampling dependence, multi-cycle prediction identifies blood glucose risk in advance, and realizes blood glucose active control through hemodialysis machine linkage, and through standard communication protocol, all kinds of hemodialysis machine is adapted, and blood glucose safety monitoring is realized throughout hemodialysis.

Description

Technical Field

[0001] This invention relates to the field of hemodialysis and blood glucose monitoring technology, specifically to a two-stage anti-interference online blood glucose monitoring method and system for hemodialysis machines, belonging to the field of online biochemical detection technology for hemodialysis equipment. Background Technology

[0002] Hemodialysis is a core maintenance treatment for patients with end-stage renal disease. During this treatment, the patient's blood composition changes continuously with the dialysis process. Combined with the dynamic adjustments to dialysate exchange and ultrafiltration, this can easily lead to drastic fluctuations in blood glucose levels. Asymptomatic hypoglycemia, in particular, can trigger serious cardiovascular events and even death. Therefore, continuous and accurate blood glucose monitoring throughout hemodialysis treatment is a clinical necessity. Current blood glucose monitoring technologies in hemodialysis settings are mainly divided into two categories: in vitro blood sampling and simple online monitoring. In vitro blood sampling is a traditional clinical method that requires medical staff to frequently collect venous or finger-prick blood from patients and perform testing using a biochemical analyzer. This method not only causes repeated physical discomfort to patients but also significantly increases the risk of infection during blood collection. Furthermore, it only obtains discrete blood glucose data and cannot achieve real-time continuous monitoring. While simple online monitoring technology attempts to integrate blood glucose sensors into the hemodialysis machine for in-situ blood sample testing, it often directly transplants the sensor structure and algorithms of ordinary blood glucose testing without adapting and optimizing for the specific conditions of hemodialysis, revealing many shortcomings in clinical applications.

[0003] The core defects of the existing hemodialysis blood glucose monitoring technology are firstly reflected in the serious lack of anti-interference ability. Macromolecular interfering substances such as heparin, creatinine, and urea present in the blood are prone to non-specific reactions with the detection electrode, generating false electrical signals, and the existing technology does not set a targeted physical anti-interference structure. At the same time, the dynamic changes in physical parameters such as blood temperature, pH, and flow rate during hemodialysis, as well as the fluctuations in working conditions such as blood pump speed, transmembrane pressure, and ultrafiltration rate, will directly affect the glucose-specific reaction rate and signal acquisition results. The existing technology lacks an effective parameter correction and working condition compensation mechanism, ultimately resulting in the detection accuracy far from meeting the clinical requirements. Secondly, the blood sample flow control of the existing technology does not conform to the actual working conditions of hemodialysis, ignoring the dynamic characteristic that blood viscosity will gradually increase with the treatment duration. Only a fixed flow parameter is used to extract blood samples, which is prone to flow control deviation, and even cause blood coagulation in the bypass pipeline, directly interrupting the monitoring process and unable to ensure the continuity of monitoring. Furthermore, all existing online monitoring technologies rely on blood sampling calibration. It is necessary to regularly collect the patient's blood sample to obtain the reference blood glucose value and manually calibrate the monitoring system, which not only increases the workload of medical staff but also further exacerbates the patient's pain and infection risk, lacking a blood sampling-free self-calibration mechanism suitable for the hemodialysis scenario. In addition, the existing technology can only achieve passive warning after the blood glucose concentration exceeds the threshold, unable to predict the blood glucose change trend in advance based on the blood glucose change rule and real-time working condition parameters. It is difficult for medical staff to formulate intervention strategies in advance; and there is no effective data interaction and parameter linkage mechanism between the monitoring system and the main control system of the hemodialysis machine, unable to automatically adjust treatment parameters such as dialysate glucose concentration and ultrafiltration rate according to the blood glucose status, and can only rely on manual intervention by medical staff, resulting in the problem of lagged intervention and unable to achieve the whole-process blood glucose safety control of hemodialysis. Finally, the existing simple online monitoring system has poor versatility and lacks a standardized communication protocol with different brands and models of hemodialysis machines, making it difficult to be applied on a large scale in clinical practice. These defects lead to the fact that the existing hemodialysis blood glucose monitoring technology cannot meet the whole-process blood glucose safety monitoring requirements of hemodialysis treatment in terms of detection accuracy, continuity, convenience, and clinical control ability, becoming the key problem restricting the improvement of the safety of hemodialysis treatment.

[0004] Da, the full English name is Dalton, Chinese name: Dalton; Modbus-RTU, the full English name is Modicon Bus-Remote Terminal Unit, Chinese name: Modbus bus in the remote terminal unit mode; RS485, the full English name is Recommended Standard 485, Chinese name: Recommended Standard 485; ISO, the full English name is International Organization for Standardization, Chinese name: International Organization for Standardization; UFR, the full English name is Ultrafiltration Rate, Chinese name: Ultrafiltration Rate; nA, the full English name is nanoampere, Chinese name: Nanoampere; mmHg, the full English name is millimeter of mercury, Chinese name: Millimeter of mercury; mL, the full English name is milliliter, Chinese name: Milliliter; min, the full English name is minute, Chinese name: Minute; h, the full English name is hour, Chinese name: Hour; μL, the full English name is microliter, Chinese name: Microliter; mm, the full English name is millimeter, Chinese name: Millimeter; Hz, the full English name is Hertz, Chinese name: Hertz; rpm, the full English name is revolutions per minute, Chinese name: Revolutions per minute; L, the full English name is liter, Chinese name: Liter; mmol, the full English name is millimole, Chinese name: Millimole. Summary of the Invention

[0005] The technical problems to be solved by the present invention at least include one of the following: overcoming the technical defects of the existing blood dialysis machine blood glucose monitoring technology with poor anti-interference ability and low detection accuracy, the blood sample flow control not being adapted to the blood dialysis working conditions and easily causing pipeline coagulation, relying on blood sampling calibration to increase the pain and infection risk of patients, being unable to predict the blood glucose change trend and having no effective linkage with the blood dialysis machine, only being able to give passive warnings and unable to achieve active blood glucose control, and having poor versatility and being difficult to be clinically applied on a large scale.

[0006] To solve the above technical problems, the present invention provides the following technical solutions.

[0007] A dual-level anti-interference online blood glucose monitoring method for hemodialysis machines is characterized by the following steps: Step 1, in-situ micro-blood sample diversion processing, enabling non-invasive extraction, detection, and reinfusion of micro-blood samples from the main pipeline of the hemodialysis machine, while preventing coagulation through pipeline maintenance; Step 2, multi-source signal synchronous acquisition and preprocessing, synchronously acquiring and eliminating noise from glucose-specific signals, blood physical parameters, and hemodialysis machine operating parameters; Step 3, dual-level anti-interference correction processing, eliminating detection deviations through a dual-level mechanism of molecular sieving physical anti-interference, parameter correction, and operating condition compensation, while achieving bloodless self-calibration and calculating the effective blood glucose concentration; Step 4, blood glucose change trend prediction processing, completing multi-cycle trend prediction based on historical blood glucose data, blood glucose change patterns during hemodialysis treatment, and real-time operating parameters, and classifying risk levels according to the predicted blood glucose values; Step 5, graded early warning and parameter linkage processing, achieving synchronous transmission of monitoring data and prediction results, issuing graded early warnings based on risk levels, and automatically adjusting hemodialysis machine treatment parameters according to the prediction results.

[0008] In a preferred embodiment of the present invention, step 1 includes sub-steps such as pre-filling and status checking of the bypass tubing, controlling the extraction of a small amount of blood from the arterial tubing, ensuring the blood sample remains stably within the detection chamber, returning the tested blood sample to the venous tubing, and periodically flushing the bypass tubing to prevent coagulation. A dynamic blood viscosity correction term and an anti-coagulation constraint term are introduced when extracting a small amount of blood, and the blood flow rate of the bypass branch is controlled to ≤0.3. mL / bad The filling time of the detection cavity is controlled within ≤2 seconds. s .

[0009] In a preferred embodiment of the present invention, the signals synchronously acquired in step 2 include glucose response current, blood temperature, and blood... pH Blood pressure, hemodialysis machine operating conditions, and blood flow rate were collected. After collection, the raw signals were sequentially processed by moving average smoothing filtering, minimum-maximum normalization, and time synchronization matching. The acquisition time deviation of different signal channels met the preset timing constraints.

[0010] In a preferred embodiment of the present invention, the molecular sieving physical interference resistance in step 3 adopts a molecular retention threshold of 300. Yes Nanofiltration membrane devices for molecular weight ≥300 Yes Interference substance rejection efficiency ≥95%; parameter correction includes temperature correction, pH The stepwise calibration of the calibration and flow field calibration, the working condition compensation including ultrafiltration interference compensation, dialysis exchange interference compensation, blood circuit state interference compensation, and the triggering conditions for bloodless self-calibration are any of the following: identification of the effective steady state interval, every preset time period, and blood glucose detection error exceeding the preset value.

[0011] In a preferred embodiment of the present invention, the multi-cycle trend prediction in step 4 includes short-cycle, medium-cycle, and long-cycle predictions, with the short-cycle being the prediction for the next 10 weeks. bad The medium-term cycle is the next 30 years. bad The "long period" refers to the remaining time of this hemodialysis treatment; the risk level is divided into low risk, medium risk, and high risk, with low risk defined as a blood glucose concentration of 3.9. mmol / L ~10.0 mmol / L Medium risk is defined as a blood glucose concentration of 3.0. mmol / L ~3.9 mmol / L Or 10.0 mmol / L ~16.7 mmol / L High risk is defined as a blood glucose concentration <3.0. mmol / L Or >16.7 mmol / L .

[0012] An online blood glucose monitoring system for hemodialysis machines based on dual-level anti-interference and dynamic prediction is characterized by comprising a blood sample diversion processing module for data transmission and command interaction, a multi-source signal acquisition and preprocessing module, a blood glucose detection anti-interference correction module, a blood glucose change trend prediction module, and a graded early warning and parameter linkage module. Each module independently completes its own preset function, and the overall system operates in an orderly manner through an internal data link.

[0013] In a preferred embodiment of the present invention, the blood sample diversion and processing module includes a bypass pipeline pre-filling unit, a micro-blood sample extraction unit, a blood sample stabilization and retention unit, a blood sample reinfusion unit, and a pipeline anti-coagulation flushing unit. The detection chamber is a cylindrical cavity with a medical heparin anti-coagulation coating sprayed on the inner wall. A molecular sieving physical anti-interference device is integrated at the inlet. The effective volume of the detection chamber is microliters.

[0014] In a preferred embodiment of the present invention, the multi-source signal acquisition and preprocessing module includes a signal acquisition triggering unit, a glucose signal acquisition unit, a blood physical parameter acquisition unit, a hemodialysis machine operating condition parameter acquisition unit, and a signal preprocessing unit. Modbus - RTU The standard communication protocol interacts with the main control system of the hemodialysis machine to synchronously collect real-time operating parameters of the hemodialysis machine.

[0015] In a preferred embodiment of the present invention, the blood glucose detection anti-interference correction module includes a physical anti-interference effect verification unit, a blood parameter correction unit, an operating condition interference compensation unit, a bloodless self-calibration unit, and a blood glucose concentration calculation and verification unit. Through a dual-level anti-interference mechanism, detection deviations are eliminated, and the blood glucose detection error is controlled within ±4.5%, which meets the requirements. ISO 15197 Clinical Standards.

[0016] In a preferred embodiment of the present invention, the graded early warning and parameter linkage module includes a data synchronization transmission unit, a graded early warning triggering unit, a treatment parameter adjustment unit, and an adjustment effect tracking unit, through which... RS The system uses dual 485+ Ethernet communication links to transmit data with the hemodialysis machine's main control system and the medical workstation. The treatment parameter adjustment unit primarily adjusts the dialysate glucose concentration and ultrafiltration rate, with the dialysate glucose concentration adjusted between 0 and 5.5%. mmol / L Within the clinical safety range.

[0017] Compared with existing hemodialysis blood glucose monitoring technologies, this invention has the following significant advantages: The dual-stage anti-interference mechanism significantly improves detection accuracy. The first stage uses a molecular cutoff threshold of 300. Yes The polyamide nanofiltration membrane molecular sieving device enables the separation of heparin, creatinine, and other molecules with a molecular weight ≥300. Yes The interference substance interception efficiency is ≥95%, blocking most interference substances at the physical level; the second level uses blood temperature, pH The stepwise correction of the flow field and the compensation for operating conditions of ultrafiltration, dialysis exchange, and blood circuit status eliminate residual interference and detection deviations caused by changes in operating conditions. Combined with the bloodless self-calibration mechanism, the blood glucose detection error is controlled within ±4.5%, which meets the requirements. ISO The 15197 clinical testing standard solves the core problem of large deviations in existing technologies.

[0018] The blood sample diversion process closely matches the actual working conditions of hemodialysis, enabling non-invasive continuous monitoring. When extracting small amounts of blood, a dynamic blood viscosity correction term and an anticoagulation constraint term are introduced to control the blood flow rate of the bypass branch to ≤0.3. mL / min The filling time of the detection cavity is controlled within ≤2 seconds. s It adapts to the dynamic changes in blood viscosity with treatment duration; at the same time, the bypass tubing is flushed with saline every 30 minutes to prevent coagulation, enabling non-invasive extraction, testing and reinfusion of trace blood samples from the hemodialysis machine's main tubing. The entire process does not require frequent blood collection, eliminating the patient's physical pain and the risk of blood collection infection, and ensuring the continuity of monitoring.

[0019] This system enables accurate multi-cycle prediction of blood glucose trends, transforming passive warnings into proactive identification. Based on clinical data from over 200 hemodialysis patients, it categorizes blood glucose change patterns and extracts core trend features by combining historical blood glucose data from the current treatment and real-time operating parameters. This allows for short-cycle (future 10-day) prediction. bad ), medium cycle (the next 30) bad The long-term (remaining time of this treatment) multi-cycle blood glucose trend prediction can identify the risk of abnormal blood glucose in advance, leaving sufficient time for medical staff to intervene, and solving the problem that existing technologies can only provide passive early warning.

[0020] A risk grading and early warning mechanism is established to link hemodialysis machine parameters, enabling proactive blood glucose management. Based on predicted blood glucose concentration, three risk levels—low, medium, and high—are identified, triggering corresponding graded early warning alerts. Simultaneously, treatment parameters such as dialysate glucose concentration and ultrafiltration rate are automatically adjusted based on blood glucose status, predicted trends, and risk level, with adjustments controlled within the range of 0–5.5%. mmol / L Within the clinical safety range, and by introducing an individual tolerance correction coefficient to take into account the differences among patients of different ages, it achieves an upgrade from passive early warning to active intervention, solving the problems of lack of linkage between existing technologies and hemodialysis machines and delayed intervention.

[0021] The system adopts a modular architecture design, with each module independently performing its function and being flexibly adaptable; through Modbus RTU , RS485+ Standard communication protocols such as Ethernet enable data exchange with the main control system of the hemodialysis machine and the medical workstation. It can be applied to hemodialysis machines of different brands and models, and has good prospects for large-scale clinical application. Moreover, the entire monitoring and intervention process does not require manual intervention, which greatly reduces the workload of medical staff.

[0022] The blood-free self-calibration mechanism improves the convenience of monitoring. Based on the steady-state characteristics of hemodialysis treatment and combined with physiological characteristics such as patient age and dialysis duration, it achieves blood-free self-calibration. The self-calibration trigger conditions are any one of the following: identification of the effective steady-state range, every 30 minutes, or blood glucose detection error exceeding 4.5%, thus eliminating the dependence on blood sampling calibration. Attached Figure Description

[0023] Figure 1 The flowchart illustrates the steps of the dual-level anti-interference online blood glucose monitoring method for hemodialysis machines provided by this invention. Figure 2 The core correction coefficient calibration curve provided for this invention; Figure 3 The individualized metabolic correction factor provided by this invention k meta Clinical statistical distribution chart; Figure 4 A statistical comparison chart of the classification of blood glucose change patterns during hemodialysis treatment provided by this invention; Figure 5 The correlation curve between the amount of glucose concentration adjustment in the dialysate and the risk level / rate of blood glucose change provided by the present invention; Figure 6 The diagram showing the relationship between ultrafiltration rate and dialysate glucose concentration during high-risk hypoglycemia provided by this invention; Figure 7 This is a schematic diagram of the module composition of the dual-level anti-interference online blood glucose monitoring system for hemodialysis machines provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0026] like Figure 1 As shown, in the first embodiment of the present invention, a two-stage anti-interference online blood glucose monitoring method for a hemodialysis machine is provided, the method comprising the following steps: Step 1: In-situ micro-blood sample diversion processing. This step is used to achieve non-invasive extraction, testing, and reinfusion of micro-blood samples from the main tubing of the hemodialysis machine, while preventing clotting through tubing maintenance. Step 1 further includes the following steps: Step 1-1: Bypass Tubing Pre-filling and Status Check. Before officially starting hemodialysis treatment, pre-fill the bypass branch with physiological saline (0.9% sodium chloride solution), the same pre-filling fluid used for hemodialysis, and control the pre-filling flow rate at 2. mL / min Precharge duration 1 bad Ensure that all pipeline cavities within the bypass branch are completely filled with pre-filled liquid, with no residual air. After pre-filling, detect the flow status using pressure sensors at both ends of the bypass branch. When the pressure difference between the inlet and outlet of the bypass branch is ≤ 5mmHg If the flow is deemed normal and there are no blockages or poor flow, the bypass branch enters standby mode; if the pressure difference is >5... mmHg This triggers a pipeline blockage warning, requiring manual inspection and recharging.

[0027] Steps 1-2: Controlling the extraction of a small amount of blood from the arterial tubing. Once hemodialysis treatment is officially started, the blood pump is running and the blood flow rate in the arterial tubing is stable. This should be done when the hemodialysis machine has been running for ≥30 minutes. badWhen the blood pump speed and transmembrane pressure fluctuation are ≤5%, the hemodialysis machine is determined to have entered a stable treatment state, triggering the bypass branch blood sample extraction action, opening the miniature solenoid valve at the bypass branch inlet, and allowing arterial blood to enter the bypass branch under the action of the blood pump pressure difference.

[0028] Steps 1-3: Allow the blood sample to remain stable within the detection chamber. After the detection chamber is filled, close the miniature solenoid valve at the bypass branch inlet. The blood flow within the detection chamber will stop, entering a stable state. Set the residence time to 15 seconds. s This duration is calibrated based on the reaction kinetics of glucose and an electrochemical sensor to ensure the full progress of the glucose-specific reaction; during the residence time, the blood temperature is monitored in real time by a micro-sensor integrated into the detection chamber. pH Pressure, when the temperature is 35~40 ℃ , pH 7.0~7.8, pressure 80~120 mmHg If the blood condition meets the testing requirements, the subsequent signal acquisition process is triggered; if the parameters are out of range, an abnormal blood sample status prompt is triggered, and the blood sample in the testing chamber is returned and re-extracted.

[0029] Steps 1-4: The blood sample after testing is reinfused into the venous line. After the blood sample signal is collected, the miniature solenoid valve at the outlet of the bypass branch is opened. The blood in the detection chamber is reinfused into the venous line under the pressure difference of the hemodialysis line. The reinfusion time is controlled to be 2 seconds. After the reinfusion is completed, the optical sensor in the detection chamber detects whether there is any blood sample residue in the chamber. When the light transmittance is ≥95%, it is determined that there is no blood sample residue. After confirming that there is no residue, the outlet miniature solenoid valve is closed, and the bypass branch returns to the standby state, waiting for the trigger signal for the next blood sample extraction.

[0030] Steps 1-5: Regularly flush the bypass tubing to prevent blood clotting. Throughout the hemodialysis treatment, flush the bypass tubing every 30 minutes using normal saline for hemodialysis at a flow rate of 5. mL / min , Rinse volume 3 mL This parameter was determined based on pipeline anticoagulation engineering experiments, ensuring that it avoids coagulation without causing abnormal electrolyte / volume status in the patient. During flushing, the miniature solenoid valves at the inlet and outlet of the bypass branch are opened, allowing the flushing fluid to flow through the entire bypass branch channel, carrying away residual trace amounts of blood / impurities. After flushing, the flow status check in step 1-1 is repeated to ensure that the bypass branch is in a working state throughout.

[0031] Before officially starting hemodialysis treatment, perform step 1-1, which involves pre-filling and status confirmation. The pre-filling fluid used is the same as that used in the hemodialysis machine. During pre-filling, it is necessary to confirm that all tubing cavities in the bypass branch are completely filled with the pre-filling fluid, with no residual air, to prevent the pre-filling fluid from entering the main hemodialysis line and affecting the pre-filling process of the hemodialysis machine. It is also confirmed that the bypass branch has normal flow and is free from blockages or obstructed flow. After pre-filling, the bypass branch enters a standby state, awaiting the trigger signal for the official start of hemodialysis treatment.

[0032] Once hemodialysis treatment is officially started, the blood pump of the hemodialysis machine begins to operate, and the blood in the arterial tubing flows steadily at the set flow rate. This is unrestricted, and occurs when the hemodialysis machine has been running for ≥30 minutes. bad When the blood pump speed and transmembrane pressure fluctuation are ≤5%, the hemodialysis machine is considered to have entered a stable treatment state, triggering the bypass branch blood sampling action. This opens the bypass branch inlet micro-solenoid valve, and after the hemodialysis machine enters a stable treatment state, the micro-sample sampling control operation in steps 1-2 is performed. At this time, the bypass branch blood sampling action is triggered, the bypass branch flow control structure opens, and the blood in the arterial line enters the bypass branch under the pressure difference generated by the blood pump.

[0033] Considering the dynamic changes in blood viscosity during hemodialysis with treatment duration, in a preferred embodiment of the present invention, a dynamic blood viscosity correction term and an anticoagulation constraint term are introduced when controlling the blood flow rate of the bypass branch: In the formula, Blood flow rate of bypass branch, in units of mL / bad Design constraints are ; The inner diameter of the bypass / branch road passage; This refers to the pressure difference between the arterial and venous lines of the hemodialysis machine. Blood dynamic viscosity (changes with treatment duration) To treat initial blood viscosity, Treatment duration, in units of h ); This refers to the length of the bypass / branch road passage. To prevent the coagulation correction factor from being used, this factor was calibrated based on coagulation tests of ex vivo blood samples from 50 hemodialysis patients, under the following conditions: temperature 37℃ and blood viscosity 4.0–5.0. mPa·s This is used to prevent blood samples from clotting due to prolonged retention and to achieve dynamic flow rate adaptation.

[0034] By incorporating dynamic changes in blood viscosity and anticoagulation constraints, the flow control deviation problem caused by neglecting changes in blood viscosity during treatment, as seen in existing technologies, is resolved, making it more closely aligned with actual hemodialysis conditions. The extracted blood enters the detection chamber along the bypass branch; based on the dynamic flow characteristics, the detection chamber filling time is corrected using the following methods: In the formula, For cavity filling time, the unit is... s Design constraints are ; To measure the effective volume of the cavity, the unit is... μL ; This is a blood viscosity correction factor used to compensate for the effect of changes in blood viscosity on the filling speed.

[0035] In one non-limiting embodiment, the detection cavity is a cylindrical cavity with an effective volume of The inner wall is coated with a medical heparin anticoagulant coating and installed in the middle section of the bypass branch. A molecular sieving physical anti-interference device is integrated at the inlet.

[0036] Once the blood filling chamber is complete, the blood sample stabilization process described in steps 1-3 is performed. At this time, the inlet flow control structure of the bypass branch closes, the blood flow in the detection chamber stops, and the blood enters a stable stabilization state. Optionally, the stabilization time can be set to 15 seconds. s The residence time is controlled according to preset requirements to ensure that the target substances in the blood can fully react, while avoiding excessive residence time that could lead to blood clotting or changes in blood composition. During the residence process, the basic physical parameters of the blood are monitored in real time to confirm that the blood state meets the detection requirements. Once the blood state is confirmed to be stable and meets the detection requirements, the subsequent signal acquisition process begins.

[0037] Once the signal acquisition process for this blood sample is complete, the blood sample recirculation process (steps 1-4) is executed. At this time, the outlet flow control structure of the bypass branch opens, and the blood in the detection chamber flows into the venous tubing of the hemodialysis machine along the outlet channel of the bypass branch under the pressure difference of the hemodialysis tubing. During the recirculation process, the blood recirculation rate is controlled to avoid affecting the blood flow in the venous tubing. Simultaneously, it is confirmed that all blood in the detection chamber has completely returned to the venous tubing, with no residual blood remaining in the detection chamber or bypass branch tubing, preventing coagulation or cross-contamination of subsequent samples due to residual blood. After complete blood recirculation, the outlet flow control structure closes, and the bypass branch returns to a standby state, awaiting the trigger signal for the next blood sample extraction.

[0038] Throughout the hemodialysis treatment, steps 1-5, involving anticoagulation and tubing maintenance, will be performed at set intervals. The flushing fluid used is the same as the dialysate or saline used in the hemodialysis treatment. During flushing, the inlet and outlet flow control structures of the bypass branch will be activated, allowing the flushing fluid to flow through the complete flow path of the bypass branch, carrying away any residual trace amounts of blood or impurities and preventing coagulation or blockage in the tubing.

[0039] Preferably, the duration and interval of the flushing operation are performed according to a set rule, which will not affect the normal operation of hemodialysis treatment, nor will it cause abnormal changes in the patient's electrolyte or volume status. At the same time, after each flush, the patency of the bypass branch is reconfirmed to ensure that the bypass branch is in a normal and working state throughout the entire hemodialysis treatment.

[0040] Step 2: Synchronous Acquisition and Preprocessing of Multi-Source Signals. This step is used for the synchronous acquisition and noise reduction of glucose-specific signals, blood physical parameters, and hemodialysis machine operating parameters.

[0041] Step 2 further includes the following steps: Step 2-1: Initiate multi-signal synchronous acquisition; after determining in Step 1-3 that the blood condition meets the detection requirements, a unified hardware trigger signal is issued to control the glucose response current, blood temperature, and blood... pH The system simultaneously collects data from six channels: blood pressure, hemodialysis machine status, and blood flow rate. During the data collection process, the operating status of each channel is monitored in real time. If the data loss rate of a channel exceeds 5% or the data exceeds a preset reasonable range, the collection channel is deemed abnormal, the current data collection process is immediately stopped, the blood sample diversion operation in step 1 is re-executed, and a channel abnormality warning is triggered.

[0042] Step 2-2: Acquire glucose-specific electrical signals; during the acquisition process, the electrochemical three-electrode sensor (working electrode is gold electrode, reference electrode is...) is used in the detection chamber. Ag / AgCl (The electrode and counter electrode are platinum electrodes) 0.6 V A constant operating voltage is used to ensure the stable progress of the glucose-specific redox reaction. The current signal generated by the reaction is collected, and this current signal is linearly correlated with the blood glucose concentration.

[0043] Steps 2-3: Synchronously collect blood temperature and pH parameters. Simultaneously collect basic blood physical parameters, including temperature, pH, flow rate, and pressure, along with the glucose-specific electrical signal. The collection frequency and cycle must be completely consistent with the glucose response signal to ensure that each glucose response signal corresponds to the blood physical parameters at the same moment. After collection, validity is assessed, and abnormal data exceeding the normal operating range of the hemodialysis machine are discarded. All operating parameters are matched with the glucose response signal and blood physical parameters according to the collection timestamp to ensure data temporal consistency. After collection, initial validity screening is performed, discarding data with temperatures <35℃ or >40℃, pH <7.0 or >7.8, and pressure <80℃. mmHg or >120 mmHg If the percentage of abnormal data is greater than 10%, the signal will be re-acquired.

[0044] Steps 2-4: Read the hemodialysis machine operating parameters. Modbus RTU The standard communication protocol interacts with the main control system of the hemodialysis machine to synchronously collect real-time operating parameters of the hemodialysis machine. The collection frequency is consistent with the glucose response signal, including blood pump speed, dialysate flow rate, ultrafiltration rate, transmembrane pressure, dialysate glucose concentration, and treatment duration.

[0045] Steps 2-5: Smooth and normalize the original signal.

[0046] Once the blood in the detection chamber reaches a stable state in step 1, and the blood condition is confirmed to meet the detection requirements, the acquisition triggering and synchronization control operation in step 2-1 is executed. The synchronous acquisition operation uses a unified trigger signal to control all data acquisition structures to start the acquisition process at the same time, ensuring that all acquired signals correspond to the blood condition and hemodialysis machine operating conditions at the same moment, avoiding time deviations between different signals that could affect subsequent correction and calculation accuracy.

[0047] Synchronous acquisition of all signal channels satisfies the following constraint: In the formula, , These represent the acquisition times for different signal channels, in units of... s ; The sampling frequency of the multi-source signal, in units of Hz The set value is 10. Hz The acquisition frequency is controlled according to the set requirements, enabling the complete acquisition of all signals within the set acquisition cycle, while ensuring that the acquired signals have sufficient time resolution to accurately reflect changes in blood status and hemodialysis machine operating conditions.

[0048] During the collection process, the working status of all collection channels will be monitored in real time. If an abnormality is detected in a collection channel, the current collection process will be stopped immediately, and the blood sample diversion operation in step 1 will be re-executed, or the corresponding abnormality prompt will be triggered to avoid the abnormal signals collected affecting the subsequent processing results.

[0049] After completing the acquisition triggering and synchronization control, the glucose response signal acquisition and processing operation in step 2-2 is executed. During the acquisition process, a stable set voltage is applied to ensure that the glucose-specific response can proceed stably, and the corresponding current signal is acquired simultaneously. The acquired current signal has a corresponding correlation with the glucose concentration in the blood.

[0050] During the data acquisition process, multiple sets of current signals are continuously acquired at set intervals, and the acquisition time for each set of signals is recorded simultaneously. This provides sufficient raw data for subsequent preprocessing. t The raw glucose response current collected at each time point is denoted as The unit is nA After the data acquisition is completed, a preliminary validity assessment will be performed on the multiple sets of current signals acquired, and abnormal signals that are significantly outside the reasonable range will be removed to ensure the validity of the original signals.

[0051] Simultaneously execute steps 2-3 for acquiring and processing blood basic physical parameter signals. The acquired parameters include blood temperature, pH, flow rate, and pressure. These parameters all affect the specific response to glucose and the signal acquisition results, and need to be acquired simultaneously for subsequent correction processing. t The blood temperature collected at any time is recorded as The unit is ℃; blood pH is recorded as ℃. Dimensionless; blood flow velocity is denoted as The unit is mm / s Blood pressure is recorded as The unit is mmHg .

[0052] Preferably, all basic physical parameter signals are synchronously acquired using the same acquisition frequency and period as the glucose response signals, ensuring that each set of glucose response signals has corresponding basic blood physical parameter signals at the same time. After acquisition, the acquired parameter signals are also preliminarily evaluated for validity, and abnormal signal data are removed to ensure the validity of the original data.

[0053] The hemodialysis machine operating parameter signal acquisition and processing operations in steps 2-4 are executed synchronously. Through data interaction with the hemodialysis machine's main control system, real-time operating parameters of the hemodialysis machine are collected, including the hemodialysis pump speed, dialysate flow rate, ultrafiltration rate, transmembrane pressure, dialysate glucose concentration, and treatment duration. These parameters reflect the real-time status of hemodialysis treatment, and changes in the status of hemodialysis treatment directly affect the patient's blood glucose metabolism and also influence the composition and physical state of the blood. Therefore, synchronous data acquisition is necessary for subsequent correction and trend prediction processing.

[0054] Will t The ultrafiltration rate collected at each time point is denoted as . The unit is mL / h The glucose concentration in the dialysate is denoted as The unit is mmol / L The acquisition frequency of the parameters is consistent with the acquisition frequency of the glucose response signal to ensure that each set of glucose response signals has corresponding hemodialysis machine operating parameters at the same time.

[0055] After obtaining the operating parameters, the validity of the parameters will be judged to confirm that the parameters are within the normal operating range of the hemodialysis machine. Abnormal parameter data will be eliminated. At the same time, the obtained operating parameters will be organized according to the set format and matched with other synchronously acquired signals to ensure the time consistency of all data.

[0056] After completing the acquisition and validity assessment of all raw signals, perform the multi-source acquisition signal preprocessing operation in steps 2-5. The preprocessing operation first smooths all signals using a moving average smoothing filter method. This method processes multiple sets of continuously acquired signal data to eliminate random noise generated during signal acquisition while preserving the inherent trend of the signal and avoiding the loss of effective signal features due to smoothing.

[0057] Preferably, the formula for calculating smoothing filtering is: In the formula, for t Glucose response current after time-smoothing processing; To smooth out half the window width; The sampling period is expressed in units of 10 ... s .

[0058] After smoothing, all signals are normalized using the min-max normalization method to convert signal data of different types and dimensions into a unified numerical range, eliminating dimensional differences between different signals and facilitating subsequent correction and calculation.

[0059] After normalization, all signals are time-synchronized to ensure that all signal data accurately correspond to the same acquisition time, avoiding time misalignment. Once matching is complete, all preprocessed signal data are consolidated into a unified dataset and transmitted to the subsequent calibration process. Simultaneously, the original and preprocessed signal data are temporarily stored for later backtracking and verification.

[0060] Step 3: Anti-interference processing for blood glucose detection is completed through physical sieving and parameter calibration. A two-stage mechanism—first-stage molecular sieving for physical interference and second-stage parameter calibration and operating condition compensation—eliminates detection deviations caused by interfering substances in the blood and changes in hemodialysis conditions. Simultaneously, it achieves self-calibration without blood collection, ultimately calculating an accurate blood glucose concentration and verifying its effectiveness.

[0061] Step 3 further includes the following steps: Step 3-1: Determine the actual effectiveness of physical anti-interference treatment by assessing signal baseline stability. The first level of physical anti-interference is achieved through a polyamide nanofiltration membrane molecular sieving device integrated into the detection chamber inlet. This device has a molecular rejection threshold of 300 Da, a membrane pore size of 2 nm, and is fabricated using a high-voltage electrospinning process. It allows glucose molecules (molecular weight 180 Da) to pass through smoothly while blocking molecules with a molecular weight ≥300 Da from the blood. Yes Interfering substances (heparin, creatinine, urea, etc.) are eliminated to avoid detection bias caused by non-specific reactions. The effectiveness of physical anti-interference is judged by the baseline stability of the glucose response signal after preprocessing. The standard deviation of the signal baseline is calculated. When the fluctuation coefficient = (standard deviation / signal mean) × 100% ≤ 5%, the physical anti-interference effect is considered to be satisfactory. If the fluctuation coefficient > 5%, it indicates that the interfering substances have not been effectively blocked, and the blood sample splitting operation in step 1 is repeated, triggering an abnormality prompt from the physical anti-interference device.

[0062] Step 3-2: Perform stepwise calibration of the glucose response signal based on multiple blood physical parameters. Blood temperature, pH, flow rate, and other physical parameters can affect the glucose-specific reaction rate and signal acquisition results. Using calibration coefficients calibrated in pre-experiments, calibration rules are established for each parameter to complete stepwise calibration. After stepwise calibration, a glucose response signal that eliminates the influence of blood physical parameters is obtained, which more accurately reflects blood glucose concentration.

[0063] Preferably, all correction factors were determined based on calibration experiments using ex vivo blood samples from ≥50 hemodialysis patients, with calibration conditions of 35-40°C. ℃ , pH 7.0~7.8, flow rate 0~5 mm / s The correction coefficients are obtained through linear fitting.

[0064] like Figure 2 As shown, Figure 2The core calibration curve contains three sub-plots, representing the temperature correction factor k. T Calibration curve pH Correction coefficient k pH Calibration curves and flow field correction coefficients k flow Calibration curve; This figure is based on blood samples from 50 hemodialysis patients taken outside the body within the clinical testing range (temperature 35~40°C). ℃ pH 7.0~7.8, flow rate 0~5 mm / s The calibration experiment confirmed that the glucose response current is related to blood temperature, pH Both flow rate and velocity showed significant linear correlation characteristics; the fitting formulas and value ranges of each correction coefficient were experimentally verified to be reliable and could be directly used as the quantitative basis for stepwise correction of blood physical parameters. Furthermore, those skilled in the art could repeat the parameter calibration based on the calibration curve.

[0065] Step 3-3: Compensate for residual interference based on hemodialysis conditions.

[0066] Steps 3-4: Non-invasive calibration based on steady-state signals. Based on the steady-state characteristics during hemodialysis treatment and combined with the patient's physiological characteristics, blood-free self-calibration is achieved, eliminating reliance on blood sampling calibration. The self-calibration formula incorporates physiological characteristics such as patient age and dialysis duration, overcoming calibration biases caused by individual physiological differences. The updated fitting parameters retain the correlation with individual physiological characteristics, take effect immediately, and are stored in the patient's local database for subsequent blood glucose concentration calculations. The self-calibration is triggered automatically when any one of the following conditions is met: each time a valid steady-state interval is identified, every 30 minutes, or the blood glucose detection error is >4.5%.

[0067] Steps 3-5: Calculate blood glucose concentration and verify the validity of the results.

[0068] After completing the preprocessing operation in step 2, the first step 3-1, the confirmation operation of the first-level physical anti-interference treatment effect, is performed. The first-level physical anti-interference treatment is achieved through a molecular sieving mechanism. By setting a molecular retention threshold, glucose molecules are allowed to pass through smoothly, while interfering substances in the blood with a molecular weight greater than glucose, including heparin, creatinine, and urea, are blocked. This prevents these interfering substances from causing non-specific reactions and generating additional electrical signals that could lead to deviations in the test results.

[0069] Preferably, the interception efficiency of the first-level physical interference suppression is calculated by the following formula: In the formula, To interfere with the efficiency of substance retention; The concentration of interfering substances on the transmission side; This represents the concentration of interfering substances on the feed side. The molecular rejection threshold is set to 300. Yes For molecular weight ≥300 YesInterfering substances, with a retention efficiency constraint of This achieves physical-level interference blocking.

[0070] Before proceeding with subsequent correction processing, the effectiveness of the first-level physical interference suppression needs to be confirmed. This confirmation process is accomplished by judging the baseline stability of the preprocessed glucose response signal. When the baseline fluctuation of the signal is within the set reasonable range, it indicates that the physical interference suppression processing has achieved the expected effect, most of the interfering substances have been effectively blocked, and the signal baseline is in a stable state, providing a reliable foundation for subsequent correction processing.

[0071] When the baseline fluctuation of the signal exceeds the set reasonable range, it indicates that the effect of the physical anti-interference processing has not met expectations. It is necessary to repeat the blood sample diversion operation in step 1, replace the blood sample with a new one for signal acquisition, or trigger the corresponding abnormal prompt to avoid abnormal signals affecting the subsequent correction results.

[0072] After confirming the effectiveness of the first level of physical anti-interference, step 3-2, the basic calibration processing based on blood physical parameters, is performed. Blood temperature, pH, flow rate, pressure, and other physical parameters all affect the specific response rate of glucose and thus the signal acquisition results. Even with the same glucose concentration, the acquired response signals will differ under different blood physical parameter conditions, requiring basic calibration processing to eliminate these differences. Basic calibration processing establishes corresponding calibration rules for each blood physical parameter using calibration coefficients calibrated in pre-experiments. Based on the actual values ​​of the parameters, the corresponding signal correction amount is calculated, completing the step-by-step calibration.

[0073] During the calibration process, the preprocessed glucose response signal is first calibrated based on the synchronously acquired blood temperature values ​​to eliminate the influence of temperature changes on the response signal. The formula for temperature calibration is as follows: In the formula, for t Glucose response current after temperature correction at any given time; The temperature correction factor is dimensionless and is used for pre-experiment calibration. This is a reference value for the core body temperature.

[0074] After temperature correction, the temperature-corrected signal is then subjected to pH correction based on the synchronously acquired blood pH values. This process eliminates the influence of pH changes on the response signal. The pH correction calculation formula is as follows: In the formula, for t The glucose response current after pH correction at any given time; The pH correction coefficient is dimensionless and determined in the preliminary experiment. For normal human blood pH Reference value.

[0075] After pH correction, flow field correction is performed on the signal based on the synchronously acquired blood flow velocity values ​​to eliminate the influence of changes in blood flow state on the response signal. The calculation formula for flow field correction is as follows: In the formula, for t The glucose response current after flow field correction at any given moment; The flow field correction coefficient is dimensionless and is used for pre-experiment calibration. To detect steady-state reference flow velocity.

[0076] After all basic calibration processes are completed, a glucose response signal corrected for blood physical parameters is obtained. This signal has eliminated the detection bias caused by changes in blood physical parameters and can more accurately reflect the glucose concentration in the blood.

[0077] After completing the basic calibration, step 3-3, the interference compensation processing based on the hemodialysis machine's operating parameters, is performed. During hemodialysis treatment, changes in the hemodialysis machine's operating conditions can cause changes in the composition of the patient's blood. Even after the first level of physical anti-interference processing, a small amount of interfering substances will still remain in the blood. The concentration of these interfering substances will change with the changes in the hemodialysis machine's operating conditions. At the same time, changes in the hemodialysis machine's operating conditions will also affect the patient's blood glucose metabolism rate, leading to changes in blood glucose concentration. Interference compensation processing is needed to eliminate the detection bias caused by these factors.

[0078] Interference compensation processing utilizes extensive clinical data analysis of hemodialysis treatment to establish corresponding compensation rules for each hemodialysis machine operating parameter. Taking into account individual patient metabolic differences, an innovative ultrafiltration interference compensation formula is designed, incorporating individual metabolic coefficients and treatment duration correction terms. The formula is as follows: In the formula, For t Glucose response current after ultrafiltration compensation at any time; Total blood volume of the patient, in units of mL ; Individual metabolic correction factor ( The value is the patient's age (in years). This is used to compensate for the differences in metabolic rate among patients of different ages and the interference caused by changes in metabolic state during treatment. In this way, individualized adaptation of ultrafiltration compensation can be achieved, solving the problem of compensation deviation caused by ignoring individual differences in existing technologies.

[0079] Figure 3 Individualized metabolic correction factor Clinical statistical distribution chart, Figure 3Based on clinical statistics from 100 hemodialysis patients with different dialysis durations (3 / 6 / 12 months), the individualized metabolic correction factor was determined. The variation pattern showed a negative correlation with patient age and a positive correlation with the duration of hemodialysis treatment; the fitting formula... It can accurately characterize this relationship, and the scatter plot distribution of different dialysis ages all fits the trend line, verifying the clinical rationality and individualized adaptability of the dual correction terms of age and treatment duration in the formula, and solving the compensation bias problem caused by neglecting individual metabolic differences in existing technologies.

[0080] After ultrafiltration compensation is completed, an innovative dialysis exchange interference compensation formula is designed based on the dynamic characteristics of transmembrane exchange, introducing the interference residual coefficient and the operating condition stability factor. The formula is as follows: In the formula, For t The glucose response current after constant dialysis exchange compensation is the final response signal after completing the two-stage anti-interference processing. The transmembrane glucose exchange coefficient, calibrated in a preliminary experiment, is dimensionless. This is the reference concentration of glucose in the dialysate, in units of... mmol / L ; The residual interference coefficient; This represents the interception efficiency of the first level of physical interference suppression. Through this method, the effect of physical interference suppression can be correlated with dialysis exchange compensation, achieving synergistic linkage between the two levels of interference suppression, improving compensation accuracy, and preventing residual interfering substances from affecting the detection results.

[0081] After dialysis exchange compensation is completed, blood path interference compensation is performed on the signal based on the synchronously acquired blood pump speed and transmembrane pressure values. Changes in blood pump speed and transmembrane pressure affect the flow state of blood within the dialyzer and the distribution of blood components, thus affecting the test results. Blood path compensation can calculate the detection deviation caused by changes in blood path state based on the blood pump speed and transmembrane pressure values, and compensate the response signal accordingly to eliminate interference caused by changes in blood path state.

[0082] After all interference compensation processing is completed, a glucose response signal with dual-level anti-interference processing is obtained. This signal has eliminated the trace interference remaining after physical obstruction and the detection deviation caused by changes in hemodialysis conditions, and can accurately reflect the glucose concentration in the blood.

[0083] After completing the two-level anti-interference processing, perform step 3-4, which is a blood-free self-calibration process based on steady-state characteristics.

[0084] Existing blood glucose monitoring technologies require frequent collection of patients' fingertip or venous blood. Calibrating the monitoring system with blood glucose values ​​obtained from reference testing methods increases patient discomfort and infection risks. This invention's bloodless self-calibration process eliminates the need for blood samples. Instead, it utilizes the steady-state characteristics of hemodialysis treatment, combined with the patient's physiological characteristics, to innovatively design a self-calibration formula for individualized self-calibration. The specific method is as follows: In the formula, The baseline blood glucose concentration corresponding to the steady-state range, in units of mmol / L ; , These are the concentration-current fitting parameters before calibration; This represents the mean of the response signal within the steady-state interval; , Correction factors for age and dialysis duration ( ) (Patient's dialysis duration, in months).

[0085] Because it incorporates physiological characteristics such as patient age and dialysis duration, it can overcome calibration deviations caused by individual physiological differences, improve self-calibration accuracy, and retain the advantage of bloodless operation.

[0086] The self-calibration process first determines the steady-state range during hemodialysis treatment. The steady-state range refers to the treatment phase in which the operating parameters of the hemodialysis machine remain stable and the patient's blood status also remains stable. Within the steady-state range, the patient's blood glucose metabolism is in a stable state, the blood glucose concentration will not fluctuate drastically, and the corresponding glucose response signal will also remain stable.

[0087] Schematic, the determination of the steady-state interval satisfies the following two conditions: In the formula, For interval The standard deviation of the internal response signal; For interval The mean of the internal response signal; This is the steady-state fluctuation threshold; This represents the maximum change in ultrafiltration rate within the interval. This represents the maximum change in glucose concentration in the dialysate within the specified interval.

[0088] When both conditions are met simultaneously, the decision interval is determined. This represents the steady-state range that can be used for calibration.

[0089] After determining the steady-state interval, multiple sets of glucose response signals processed with two levels of anti-interference were extracted within the steady-state interval, and the average value of these signals was calculated as the steady-state response benchmark value. Then, based on the pre-established baseline correspondence between the response signal and blood glucose concentration, and combined with the innovative self-calibration formula, the baseline blood glucose concentration corresponding to the steady-state range is calculated. At the same time, combined with the patient's basic information, it is confirmed that the baseline blood glucose concentration conforms to the patient's physiological change pattern.

[0090] After matching is complete, the obtained baseline blood glucose concentration is used to automatically calibrate the correspondence between the response signal and the blood glucose concentration, update the relevant parameters of the correspondence, and complete the bloodless self-calibration process. The updated fitting parameters are then used. , Preserving individual physiological characteristics ensures individualized accuracy in subsequent concentration calculations.

[0091] The self-calibration process is performed at set intervals throughout the entire hemodialysis treatment. Each time a new steady-state range is determined, a self-calibration is executed to ensure the monitoring system's correlation remains accurate and to prevent detection deviations caused by baseline drift. The entire self-calibration process does not require any blood samples from the patient; it is completed entirely through the steady-state characteristics of the hemodialysis treatment process, completely eliminating reliance on blood sampling calibration and reducing patient discomfort and infection risks.

[0092] After completing the self-calibration process, perform steps 3-5 to calculate and verify the validity of the corrected blood glucose concentration. First, based on the glucose response signal after dual-level anti-interference processing and the correspondence between the self-calibrated updated response signal and blood glucose concentration, calculate the blood glucose concentration value corresponding to the current blood sample. The calculation formula is as follows: In the formula, for t Final blood glucose concentration after time correction, in units of mmol / L ; , These are the fitted parameters after self-calibration and updating. The blood glucose concentration calculated using the above formula allows the detection error to be controlled within ±4.5%, which meets the requirements. ISO 15197 Clinical Standards.

[0093] After calculation, the obtained blood glucose concentration value is validated. Validation first determines whether the blood glucose concentration value is within a reasonable physiological range. When the value exceeds the normal physiological range for human blood glucose, all calibration processes need to be re-examined to confirm whether there are abnormal signals or incorrect calibration operations. Simultaneously, the blood glucose concentration value is recalculated to avoid erroneous test results. Secondly, the validity validation compares the currently calculated blood glucose concentration value with multiple previously collected blood glucose concentration values ​​to determine whether the trend of the value conforms to the physiological changes in human blood glucose and the blood glucose change patterns during hemodialysis treatment. When the value shows a sudden and drastic change, and the trend does not conform to the corresponding pattern, it indicates that the value may be abnormal, and the signal acquisition and calibration processes need to be repeated to confirm the accuracy of the value.

[0094] Validation also incorporates synchronously collected hemodialysis machine operating parameters to determine if changes in blood glucose concentration correspond to changes in the machine's operating conditions. If the hemodialysis machine's operating conditions remain relatively unchanged while the blood glucose concentration fluctuates drastically, it indicates a potential anomaly, requiring re-validation. Once the blood glucose concentration passes all validity validations, it is determined to be the accurate blood glucose concentration for the current sample. This value is then transmitted to the subsequent trend prediction process. Simultaneously, the value, along with the corresponding collection time and correction parameters, is temporarily stored for later review and analysis.

[0095] Step 4: Blood Glucose Trend Prediction and Processing. This step, based on historical blood glucose data from this treatment, the pattern of blood glucose changes during hemodialysis, and real-time operating parameters, completes multi-period trend prediction of blood glucose changes and classifies risk levels according to the predicted values. Step 4 further includes the following steps: Step 4-1: Construct a historical blood glucose dataset for this treatment. Integrate all validated blood glucose concentration values ​​from this hemodialysis treatment process, along with the corresponding acquisition time, blood physical parameters, and hemodialysis machine operating parameters for each value, and sort them in chronological order of acquisition time to form a continuous time series of blood glucose changes.

[0096] The dataset undergoes an integrity check. If the percentage of missing data points is ≤5%, it is supplemented using linear interpolation of adjacent data. If the percentage of missing data points is >5%, a data missing warning is triggered, and data is re-collected and supplemented. Preliminary analysis of the time-series blood glucose variation characteristics is performed, including the rate of increase, rate of decrease, and fluctuation amplitude. Simultaneously, this dataset is matched with the patient's past hemodialysis treatment history data (stored in a local database) to extract patterns in blood glucose changes during past treatments.

[0097] Step 4-2: Matching the blood glucose variation pattern throughout the treatment. Preferably, the blood glucose variation pattern during hemodialysis treatment is obtained based on statistical analysis of clinical data from ≥200 hemodialysis patients. Patients are categorized into different types according to their baseline characteristics. First, the baseline blood glucose variation pattern type is determined based on whether the patient has diabetes, the type of diabetes, the duration of dialysis, and the treatment plan. Then, the blood glucose time series of this treatment is compared with the standard pattern of the corresponding type to calculate the matching degree. When the matching degree is ≥85%, the treatment is considered to match the standard pattern. If the matching degree is <85%, the standard pattern is fine-tuned based on the dialysate glucose concentration, ultrafiltration protocol, and other operating parameters of this treatment to make the pattern more closely match the actual treatment.

[0098] Typical patterns include: slow decrease in blood glucose during routine hemodialysis in diabetic nephropathy patients, stable blood glucose during hemodialysis in non-diabetic patients, and slow increase in blood glucose under high dialysate glucose concentration. The standard patterns for each type are stored in the system in the form of curves, and the matching degree is calculated using the Pearson correlation coefficient.

[0099] Figure 4 This is a statistical comparison chart showing the different types of blood glucose changes during hemodialysis. The three statistical mean curves in the chart correspond to three typical patterns: slow decrease in blood glucose during routine hemodialysis in diabetic nephropathy patients, stable blood glucose during hemodialysis in non-diabetic patients, and slow increase in blood glucose under high dialysate glucose concentration. Figure 3 Based on clinical statistics of 60 hemodialysis patients with different baseline characteristics, it was confirmed that there are three typical and significant patterns of blood glucose changes during hemodialysis treatment: slow decline in routine hemodialysis in diabetic nephropathy patients, stable blood glucose in non-diabetic patients, and slow rise in blood glucose under high dialysate glucose concentration. The range of blood glucose change rate for each pattern is clear and the statistical average curve fit is high.

[0100] Step 4-3: Extract features such as the current rate of change in blood glucose. Extract core features that reflect the future trend of blood glucose changes to form a trend feature set. During the feature extraction process, correlate the fluctuation characteristics of hemodialysis conditions and introduce a condition fluctuation correction factor to ensure feature accuracy.

[0101] Step 4-4: Calculate future blood glucose levels over multiple time periods. Based on the matched patterns and trend characteristics of blood glucose changes, complete the short-term (future 10) calculation. bad ), medium cycle (the next 30) bad ), long-term (remaining time of this treatment) multi-cycle blood glucose prediction.

[0102] Steps 4-5: Classify blood glucose risk levels based on the predicted values.

[0103] After completing the calculation and verification of blood glucose concentration in step 3, the blood glucose historical dataset construction and organization operation in step 4-1 is performed first.

[0104] The historical dataset includes verified and accurate blood glucose concentration values ​​obtained from all previous collection cycles during this hemodialysis treatment, along with the corresponding collection time for each value, synchronously collected blood physical parameters, and hemodialysis machine operating parameters. During dataset construction, all blood glucose values ​​and their corresponding associated information are sorted according to the chronological order of collection time to form a continuous time series of blood glucose changes. Simultaneously, the data in the time series undergoes a completeness check to confirm that all data have corresponding collection times and associated information, and that no data items are missing.

[0105] For any missing data, it will be supplemented according to pre-defined rules based on adjacent blood glucose values ​​and corresponding changes in working conditions to ensure the continuity of the time series. After the dataset is processed, a preliminary analysis of blood glucose change trends will be conducted to extract the blood glucose change characteristics that have occurred during this hemodialysis treatment, including the rate of increase, rate of decrease, and amplitude of fluctuation, providing a basic reference for subsequent trend prediction. Simultaneously, the processed historical dataset will be matched with the patient's past hemodialysis treatment data to extract patterns of blood glucose changes during past treatments, providing more reference information for trend prediction.

[0106] After constructing and organizing the historical dataset, step 4-2, which involves matching the blood glucose change patterns throughout hemodialysis treatment, is executed. During hemodialysis treatment, the patient's blood glucose changes are not random but exhibit corresponding patterns as the treatment progresses. These patterns have been verified through extensive clinical data statistics and analysis. The trend prediction processing of this invention matches the blood glucose changes during the current treatment with the verified blood glucose change patterns during hemodialysis treatment to determine the type of blood glucose change pattern corresponding to the current treatment.

[0107] Pattern matching first determines the patient's baseline blood glucose variation pattern based on the patient's basic information, including whether they have diabetes, the type of diabetes, the duration of dialysis, and the treatment plan. Different types of patients have significantly different blood glucose variation patterns during hemodialysis treatment. For example, the blood glucose variation patterns of patients with diabetes are completely different from those of patients without diabetes.

[0108] Once the basic pattern type is determined, the time series of blood glucose changes collected during this treatment is compared with the standard blood glucose change pattern of the corresponding type. The degree of matching between the two is calculated. When the degree of matching reaches a set threshold, the blood glucose change pattern corresponding to this treatment can be determined. Simultaneously, the matching process also incorporates the operating parameters of the hemodialysis machine during this treatment, including dialysate glucose concentration and ultrafiltration protocol, to adjust the obtained pattern. This is because different treatment protocols can affect the patient's blood glucose change pattern, and the adjusted pattern can more accurately reflect the patient's blood glucose change trend during this treatment.

[0109] After matching the blood glucose change pattern, perform step 4-3 to extract trend features based on the current treatment status.

[0110] Trend characteristics refer to relevant features that can reflect the future trend of a patient's blood glucose, including the current blood glucose concentration, the rate and direction of blood glucose change during the current treatment, the current operating parameters of the hemodialysis machine, and the hemodialysis machine operating adjustment plan set in the future treatment cycle. These characteristics will have a direct impact on the patient's future blood glucose changes.

[0111] To illustrate, during the feature extraction process, blood glucose change data within a set time period prior to the current moment is first extracted from the historical dataset. Then, considering the fluctuation characteristics of hemodialysis operations, a condition fluctuation correction factor is introduced, as shown in the following formula: In the formula, for t The rate of change of blood glucose at time t, in units of mmol / ( L • bad ); For rate calculation window; This is the coefficient for the influence of operating condition fluctuations; For nearly 5 bad The standard deviation of the internal ultrafiltration rate is used to compensate for the impact of fluctuations in hemodialysis conditions on the rate of change in blood glucose.

[0112] Because it correlates the fluctuations in hemodialysis conditions with the rate of change in blood glucose, it can overcome the problem of rate calculation deviation caused by ignoring fluctuations in conditions in existing technologies, and improve the accuracy of trend extraction.

[0113] Secondly, extract all operating parameters of the hemodialysis machine at the current moment, as well as the pre-set adjustment plan for operating parameters of the hemodialysis machine within the future prediction period, including the adjustment of ultrafiltration rate and dialysate glucose concentration. These adjustments will directly affect the patient's future blood glucose changes.

[0114] Then, the patient's basic physiological information, as well as the blood glucose change characteristics during the same treatment stage and under the same working conditions in the past hemodialysis treatment, are extracted as reference features. After all features are extracted, the features are screened for effectiveness, and invalid features with little impact on the blood glucose change trend are removed, while core features with a greater impact on the blood glucose change trend are retained and organized into a trend feature set.

[0115] After completing the trend feature extraction, perform step 4-4 to calculate the multi-period prediction of blood glucose change trends.

[0116] Multi-cycle prediction refers to predicting the trend of blood glucose changes over different time periods in the future, including short-cycle prediction, medium-cycle prediction, and long-cycle prediction. The prediction results of different cycles can provide different references for subsequent risk warning and treatment intervention.

[0117] The predictive calculation is based on the previously matched blood glucose change patterns and the extracted trend feature set. In a preferred embodiment of the present invention, the multi-period predictive calculation method is as follows: In the formula, for t Always looking towards the future Predicted blood glucose concentration over time, in units of mmol / L ; To estimate the duration, the unit is... bad ; The working condition influence coefficient is dimensionless. To predict the amount of blood glucose change caused by changes in hemodialysis conditions during the cycle, the unit is... mmol / L .

[0118] First, a short-cycle prediction calculation is performed, with the prediction period being the next 10 minutes. During the calculation, based on the current blood glucose concentration value, combined with the current blood glucose change rate and the operating parameters of the hemodialysis machine in the next 10 minutes, the blood glucose concentration value corresponding to each time point in the next 10 minutes is calculated using the above formula according to the matched blood glucose change pattern, forming a short-cycle blood glucose change trend curve.

[0119] Short-term predictions are highly accurate and can accurately reflect a patient's blood glucose changes in a short period of time.

[0120] After the short-cycle prediction is completed, the medium-cycle prediction is calculated. The medium-cycle prediction period is the next 30 minutes. During the calculation, based on the results of the short-cycle prediction, combined with the hemodialysis machine's operating condition adjustment plan within the next 30 minutes and the patient's blood glucose change pattern, the blood glucose concentration value corresponding to each time point within the next 30 minutes is calculated using the above formula, forming the blood glucose change trend curve of the medium-cycle.

[0121] The results of mid-cycle prediction can reflect the patient's blood glucose change trend in advance, allowing sufficient time for early risk warning and intervention.

[0122] After the intermediate-cycle prediction is completed, the long-cycle prediction is calculated. The duration of the long-cycle prediction is the total remaining treatment time of this hemodialysis treatment. In the calculation process, based on the results of the intermediate-cycle prediction, combined with the complete treatment plan for the remaining time of this treatment, as well as the blood glucose change pattern of the patient throughout the past treatment, the blood glucose change trend curve for the remaining treatment time is calculated using the above formula.

[0123] The results of long-term predictions can reflect the changes in blood glucose levels throughout the patient's treatment, providing a reference for medical staff to adjust treatment plans.

[0124] After all the prediction calculations for each cycle are completed, the prediction results will be verified for reasonableness. This will determine whether the predicted blood glucose change trend conforms to the physiological change pattern of human blood glucose and the blood glucose change pattern of hemodialysis treatment. If the prediction results show obvious unreasonableness, the matching pattern and extracted features will be readjusted, and the prediction calculation will be performed again.

[0125] After completing the multi-cycle prediction calculation, perform the risk level classification operation of the prediction results in steps 4-5.

[0126] Risk levels are determined based on the predicted blood glucose concentration and the corresponding clinical risk level. The predicted results are categorized into different risk levels, each corresponding to different warning and intervention measures. The risk level classification criteria are based on clinical guidelines and extensive clinical data statistics. An illustrative classification rule is as follows: Low risk: The predicted blood glucose concentration meets the requirements. This range represents the clinically safe range for a patient's blood glucose concentration during hemodialysis treatment. Blood glucose levels within this safe range will not harm the patient's health and do not require additional intervention. When the predicted blood glucose concentration value remains within the safe range throughout the prediction period, the risk level is classified as low risk. Low risk indicates that the patient's blood glucose status is stable and there will be no abnormal blood glucose levels in the future.

[0127] Medium risk: The predicted blood glucose concentration meets the requirements. or This range is the clinical concern range for blood glucose concentration. If the blood glucose concentration exceeds the safe range but has not yet reached the abnormal threshold, it needs to be monitored by medical staff, but no emergency intervention is required. When the predicted blood glucose concentration value will enter the concern range in the future prediction period, but will not exceed the threshold, the risk level is classified as medium risk. Medium risk means that the patient's blood glucose will fluctuate to some extent in the future, requiring medical staff to monitor and prepare for intervention in advance.

[0128] High risk: The predicted blood glucose concentration meets the requirements. or This range represents the clinically abnormal range for blood glucose concentration. When blood glucose concentration exceeds the clinically set threshold, it can harm the patient's health and requires emergency intervention. If the predicted blood glucose concentration value is expected to enter the abnormal range within the future prediction period, the risk level is classified as high risk. High risk means that the patient will experience abnormal blood glucose in the future, requiring immediate warning and intervention to avoid serious health risks for the patient.

[0129] Once the risk level is determined, the predicted blood glucose trend curve, the corresponding blood glucose concentration value, and the determined risk level will be transmitted to the subsequent early warning process. At the same time, all prediction results will be stored for future retrospective analysis.

[0130] Step 5: Tiered Early Warning Processing. Step 5 enables the synchronous transmission of blood glucose monitoring data and prediction results, tiered early warning based on risk level, and automatic adjustment of hemodialysis machine treatment parameters, while tracking blood glucose changes after parameter adjustment. Step 5 further includes the following steps: Step 5-1: Send monitoring data to the hemodialysis machine and workstation. (Through...) RS485+ The dual Ethernet communication links synchronously transmit the following data to the hemodialysis machine's main control system and the hemodialysis treatment room's medical workstation: current effective blood glucose concentration, historical blood glucose change curve for this treatment, future multi-cycle blood glucose prediction trend curve, predicted blood glucose value, and risk level classification results.

[0131] Step 5-2: Issue tiered alerts based on risk levels. Based on the risk levels identified in Step 4-5, trigger different levels of alerts.

[0132] Step 5-3: Adjust treatment parameters according to the predicted results. The hemodialysis machine treatment parameters are automatically adjusted based on the patient's real-time blood glucose status, predicted trends, and risk level.

[0133] Optionally, it also includes step 5-4: tracking blood glucose changes after parameter adjustment.

[0134] After completing step 4 of predicting blood glucose change trends and classifying risk levels, step 5-1 is executed to synchronize the transmission of blood glucose monitoring data and prediction results.

[0135] The targets of synchronous transmission include the main control system of the hemodialysis machine and the medical workstations in the hemodialysis treatment room. The transmitted content includes the current verified accurate blood glucose concentration value, the historical blood glucose change curve during this treatment process, the predicted blood glucose change trend curve for different future periods, and the corresponding risk level classification results.

[0136] During transmission, data interaction is conducted through standard communication protocols to ensure the stability and accuracy of data transmission. At the same time, a corresponding timestamp is added to all transmitted data to ensure that the data received by the hemodialysis machine's main control system and the medical workstation completely corresponds to the actual collection time and the predicted time, avoiding time misalignment.

[0137] After data transmission is completed, a feedback signal will be received from the receiving end to confirm that the data has been fully received. If no feedback signal is received, data transmission will be re-initiated, and a corresponding communication error message will be triggered. This ensures that all blood glucose data and prediction results can be accurately synchronized to the corresponding receiving end, providing an accurate data foundation for subsequent early warning and coordinated adjustments.

[0138] After the data synchronization and transmission is completed, execute step 5-2, which involves triggering a graded early warning based on the risk level.

[0139] The tiered early warning system triggers different levels of warning prompts based on the risk levels identified in step 4. The different levels of warning prompts have different prompting methods and target audiences, ensuring that medical staff can obtain patients' blood glucose risk information in a timely and accurate manner.

[0140] When the risk level is low, a Level 1 warning is triggered. The Level 1 warning only displays the patient's blood glucose value and trend in real time on the operation interface of the hemodialysis machine and the interface of the medical staff workstation. It does not trigger any additional audio or visual prompts to avoid unnecessary interference with the normal work of medical staff.

[0141] When the risk level is medium, a level 2 warning is triggered. In addition to displaying blood glucose data and risk information on the interface, the level 2 warning will also trigger an alarm sound at the medical staff workstation and display a prominent attention prompt on the operation interface of the hemodialysis machine, reminding medical staff to pay attention to the patient's blood glucose changes and prepare for intervention in advance.

[0142] When the risk level is high, a Level 3 warning is triggered. In addition to displaying detailed risk information on the interface, the Level 3 warning will also trigger an audible and visual alarm on the hemodialysis machine and the medical staff workstation. The alarm volume and flashing frequency are set according to the standards of clinical emergency alarms, which can immediately attract the attention of medical staff. At the same time, the patient's blood glucose risk information will be sent to the mobile terminal of the medical staff responsible for the patient, ensuring that medical staff can obtain emergency risk information as soon as possible and deal with it in a timely manner.

[0143] All warning messages will be recorded in detail, including the trigger time, risk level, corresponding blood glucose level, and prediction result, to facilitate subsequent review and analysis.

[0144] After triggering the tiered early warning, step 5-3, which involves adjusting the hemodialysis machine treatment parameters based on the patient's blood glucose status, is executed. This coordinated adjustment refers to automatically adjusting the hemodialysis machine's treatment parameters based on the patient's real-time blood glucose levels and future blood glucose trends. This proactive intervention prevents abnormal blood glucose levels and represents an upgrade from passive early warning to active intervention. The parameters and adjustment ranges for this coordinated adjustment are determined based on the patient's blood glucose status and risk level, while also complying with clinical treatment guidelines. All adjustment rules have been validated by clinical experts to ensure the safety and effectiveness of the adjustments.

[0145] Taking into account risk level and individual differences, and introducing a risk level correction factor and an individual tolerance factor, the dialysate glucose concentration is calculated as follows: In the formula, This is the adjustment amount for the glucose concentration in the dialysate, in units of... mmol / L ; The safety adjustment factor for clinical validation is dimensionless. To anticipate the lead time; The risk level adjustment coefficients (low risk = 1.0, medium risk = 1.2, high risk = 1.5) are used to match the adjustment range with the risk level. This is an individual tolerance correction factor used to compensate for differences in glucose tolerance among patients of different ages.

[0146] When a patient's risk level is medium and blood sugar is predicted to decline and approach the threshold of hypoglycemia, the linkage adjustment will automatically increase the glucose concentration of the dialysate in the hemodialysis machine according to the above formula, increase the amount of glucose exchange between the blood and the dialysate, slow down the rate of decline in the patient's blood sugar, and prevent blood sugar from entering the abnormal range. At the same time, the adjustment range will be controlled within the clinically safe range and will not have an adverse effect on the patient's treatment.

[0147] When a patient's risk level is medium and blood sugar is predicted to rise, approaching the critical value of hyperglycemia, the linkage adjustment will automatically reduce the glucose concentration of the dialysate in the hemodialysis machine according to the above formula, reduce the glucose content in the blood, slow down the rate of increase in the patient's blood sugar, and prevent blood sugar from entering the abnormal range.

[0148] When a patient's risk level is high and it is predicted that the patient will experience hypoglycemia within a future prediction period, the linkage adjustment will immediately stop the ultrafiltration operation of the hemodialysis machine, satisfying the following constraint: Simultaneously, taking into account individual tolerance differences, an innovative formula for emergency adjustment and supplementation of glucose in cases of high-risk hypoglycemia was designed to quickly and accurately replenish glucose. In the formula, This refers to the emergency glucose concentration in dialysate during high-risk hypoglycemia, in units of... mmol / L ; To predict hypoglycemia concentration, the unit is... mmol / L .

[0149] When a patient's risk level is high and a severe case of hyperglycemia is anticipated, the linkage adjustment will immediately adjust the glucose concentration of the dialysate to the lowest safe value, while simultaneously adjusting the ultrafiltration rate to a clinically safe range. This helps lower the patient's blood glucose concentration and triggers an emergency alarm to alert medical staff for timely clinical intervention.

[0150] Figure 5 This is a graph showing the correlation between the amount of glucose concentration adjustment in the dialysate and the risk level / rate of blood glucose change. Figure 5 The correlation between dialysate glucose concentration adjustment and the rate of blood glucose change, as well as risk level, was intuitively quantified: the adjustment amount increased systematically with the increase of the rate of blood glucose decrease, and increased with the risk level (low, medium, high), and the risk level correction coefficient (…). k risk A gradient matching of adjustment magnitudes (e.g., 1.0 / 1.2 / 1.5) can be achieved, and the individual tolerance coefficient will have a reasonable influence on the adjustment amount; at the same time, all adjustment amounts are controlled within the range of 0~5.5. mmol / L Within the clinical safety range, the quantitative logic of individualized and risk-matching adjustment rules has been clarified.

[0151] Figure 6 This is a graph showing the relationship between ultrafiltration rate and dialysate glucose concentration during emergency adjustments in cases of high-risk hypoglycemia. Figure 6 The core principles of emergency intervention for high-risk hypoglycemia have been clarified: the lower the predicted hypoglycemia concentration, the higher the required emergency glucose concentration in the dialysate, and the point at which the ultrafiltration rate stops ( UFR=0 The core emergency intervention threshold for high-risk hypoglycemia is ); the individual tolerance coefficient of patients of different ages ( k toler The fitted curve and emergency formula will correspond to different emergency concentrations. C d,glu,emerg =5.5-(3.0- Ĉ glu )・ k toler It can accurately characterize the individualized emergency relationship; the results realize the precise intervention of ultrafiltration cessation and glucose concentration supplementation in high-risk hypoglycemia, and take into account the individual tolerance differences due to age, and can quickly and safely correct hypoglycemic trends.

[0152] All linked adjustments will display detailed adjustment information on the hemodialysis machine's operating interface, including the adjusted parameters, adjustment range, and reason for adjustment. All adjustments will also be recorded for medical staff to review and trace. All adjustments can be manually modified or stopped by medical staff at any time to ensure the safety of treatment.

[0153] After completing the coordinated adjustment of the hemodialysis machine treatment parameters, perform the post-adjustment treatment effect tracking and feedback operation in step 5-4. The purpose of tracking and feedback is to confirm the effect of the coordinated adjustment, determine whether the adjusted parameters can effectively control the patient's blood glucose changes, and optimize subsequent adjustment rules based on the tracking results to ensure the effectiveness of the coordinated adjustment.

[0154] The tracking process begins after parameter adjustments are completed. Following a set collection period, the patient's blood glucose concentration is continuously collected, recording changes in blood glucose levels after adjustment. Simultaneously, the operating parameters of the hemodialysis machine are collected to confirm that the adjusted parameters have taken effect. The tracking duration is the set effect verification period, comprehensively reflecting the impact of parameter adjustments on the patient's blood glucose changes. During the tracking process, the real-time blood glucose values ​​are compared with the predicted results before adjustment to determine whether the blood glucose trend has changed in the expected direction.

[0155] When blood glucose levels are controlled as expected and gradually return to a safe range, it indicates that the adjustment has achieved the desired effect. Continue to maintain the adjusted parameters and continuously monitor blood glucose levels. If the blood glucose trend does not change as expected and there is still a risk of entering an abnormal range, it indicates that the current adjustment magnitude is insufficient. The adjustment parameters need to be recalculated and a second adjustment performed. Simultaneously, the effect data of this adjustment should be fed back into the adjustment rules to optimize them and prevent similar situations from occurring in the future. If the blood glucose trend is over-adjusted, causing abnormal fluctuations in the opposite direction, it indicates that the current adjustment magnitude is too large. The parameters need to be immediately adjusted back to safe values, and corresponding prompts should be triggered to remind medical staff to check and handle the situation.

[0156] The following example illustrates the entire process of the online blood glucose monitoring method for hemodialysis machines based on dual-level anti-interference and dynamic prediction according to the present invention.

[0157] Patient basic information: Male, 55 years old, end-stage type 2 diabetic nephropathy, on regular hemodialysis for 6 months, with no serious complications; initial blood viscosity at treatment. η 0 = 4.2 mPa • s Total blood volume V blood =4800 mL .

[0158] Basic parameters for hemodialysis: Treatment duration: 4 hours; blood pump flow rate: 250. mL / bad dialysis fluid flow rate 500 mL / bad The initial glucose concentration in the dialysate was 1.5%. mmol / L Initial ultrafiltration rate 300 mL / h .

[0159] Monitoring hardware parameters: Inner diameter of bypass branch channel d =0.3 mm Channel length L =150 mm Effective volume of the detection cavity V chamber =5 μL Multi-source signal sampling frequency fs =10 Hz .

[0160] Step 1: In-situ micro-blood sample shunt treatment (30 minutes after treatment starts, hemodialysis enters a stable state).

[0161] 1-1 Before starting treatment, the bypass branch is pre-filled with the same normal saline as the hemodialysis machine to confirm that there are no residual air bubbles or blockages in the tubing lumen, and that the flow is normal, and the bypass is ready to work.

[0162] 1-2. Thirty minutes after the start of treatment, the hemodialysis machine entered a stable state, and the arterial pressure... P art =120 mmHg venous pressure P ven =30 mmHg Pressure difference ΔP =90 mmHg Treatment duration t treat =0.5 h .

[0163] Blood dynamic viscosity calculation: η ( t )= η 0・(1+0.02・ t treat =4.2 × (1 + 0.02 × 0.5) = 4.242 mPa • s Calculation of bypass branch blood flow (compliant with) Q bypass ≤0.3 mL / bad constraint): Qbypass =[ π • d 4 • ΔP ] / [128・ η ( t )・ L ]・(1- k coag • t treat Substituting the parameters, the calculated basic flow rate is 0.225. mL / bad Multiply by the anticoagulation correction term of 0.9925, and the final result is... Q bypass =0.223 mL / bad It meets the design constraints.

[0164] Calculation of filling time for detection chamber (compliant with) t fill ≤2 s constraint): t fill = V chamber / Q bypass • square ( η ( t ) / η 0); Substituting the parameters yields t fill ≈1.35 s It meets design constraints and completes blood sample extraction and testing chamber filling.

[0165] 1-3 After filling is complete, close the bypass inlet valve, check if the blood sample in the chamber stops flowing, and set the residence time to 15 seconds. s This ensures a full reaction between glucose and the detection electrode; the blood temperature is simultaneously monitored at 36.8℃ during the treatment process. pH 7.38, Pressure 102 mmHg Once the blood sample is confirmed to be stable, subsequent signal acquisition is triggered.

[0166] After signal acquisition 1-4 is completed, open the bypass outlet valve and use the arteriovenous pressure difference to control the flow of signal 5 into the detection chamber. μL The blood sample was returned to the intravenous tubing of the hemodialysis machine, with a return time of 2 minutes. s After confirming that there is no blood residue in the testing chamber and pipeline, close the outlet valve and return the bypass to the standby state.

[0167] 1-5 During the entire hemodialysis process, a saline flush should be performed every 30 minutes. After this blood sample collection is completed, use 3 mLThe bypass flow path was flushed with saline solution to remove trace amounts of residual impurities. After flushing, the flow of the tubing was reconfirmed to be normal, with no risk of coagulation or blockage.

[0168] Step 2: Synchronous acquisition and preprocessing of multi-source signals.

[0169] 2-1 After the blood sample stabilizes, a unified trigger signal is issued to control the synchronous acquisition of six signal channels, including glucose response current, blood physical parameters, and hemodialysis machine operating status. The acquisition cycle is 2. s Each channel collects 20 data points; the time deviation of all channel acquisitions is ≤0.05. s , conforms to | t i - t j |≤1 / (2 f s Synchronization constraints.

[0170] Apply 0.6 to 2-2. V Constant operating voltage, continuously collecting 20 sets of raw glucose response currents I raw ( t )(unit nA ): 24.2, 24.5, 24.1, 24.3, 24.4, 24.6, 24.2, 24.3, 24.5, 24.4, 24.3, 24.1, 24.4, 24.5, 24.3, 24.2, 24.4, 24.3, 24.5, 24.4. Initial effectiveness screening completed, no exceeding 20-30. nA All abnormal signals within a reasonable range are retained.

[0171] 2-3. Simultaneously collect basic blood parameters at the same time point. The mean of the valid data is: temperature. T ( t = 36.8℃, pH pH ( t =7.38, Flow velocity v ( t )=0 mm / s ,pressure P ( t )=102 mmHg All parameters were within physiologically reasonable ranges and were retained after the initial effectiveness screening.

[0172] 2-4 Connect to the main control system of the hemodialysis machine and synchronously collect real-time operating parameters: blood pump speed 2500 rpm 500 dialysis fluid flow rate mL / bad Ultrafiltration rate UFR ( t )=300 mL / h Transmembrane pressure 180 mmHg dialysis fluid glucose concentration C _ d , thank you ( t )=1.5 mmol / L Treatment duration: 0.5 seconds h All parameters are within the normal working range, and timestamp matching and processing are completed.

[0173] 2-5 Moving Average Smoothing Filter: Smoothing Window Half Width N =2, window size is 5 data points, formula is: Taking the 5th data point as an example, the smoothed value is (24.1+24.3+24.4+24.6+24.2) / 5=24.32 nA After smoothing all the data, the final average current after smoothing was 24.35. nA Eliminate random noise.

[0174] Min-Max Normalization: Convert all signals to the [0,1] interval, for example, the reasonable range for glucose current is 0-100. nA 24.35 nA After normalization, it is 0.2435; the reasonable temperature range is 35-40℃, and after normalization, it is 0.36 at 36.8℃, thus completing the dimensional unification of all signals.

[0175] Complete the time synchronization matching of all signals, organize them into a standardized dataset, and transmit it to the calibration processing flow.

[0176] Step 3: Two-level anti-interference correction processing.

[0177] 3-1 The first level of physical anti-interference adopts 300 Yes Molecular retention threshold, measured retention efficiency for interfering substances such as heparin and creatinine. η reject =96.2%, meeting the design constraint of ≥95%; the baseline standard deviation of the smoothed glucose signal is 0.14. nA The fluctuation coefficient is 0.57%, which is far below the reasonable threshold of 5%, confirming that the physical anti-interference effect meets the standard.

[0178] 3-2 Temperature Correction: T ref =37℃, pre-calibration temperature correction factor k T =0.02 / ℃, the formula is: I T ( t )= I smooth( t )・[1+ k T ・( T ( t )- T ref Substitute the values: I T =24.35×[1+0.02×(36.8-37)]=24.2526 nA .

[0179] pH correction: pH ref =7.4, pre-calibration pH Correction coefficient k pH =0.08 / pH The unit, the formula is: I pH ( t )= I T ( t )・[1+ k pH ・( pH ( t )- pH ref Substitute the values: I pH =24.2526×[1+0.08×(7.38-7.4)]≈24.2138 nA .

[0180] Flow field correction: v ref =0 mm / s Pre-calibrated flow field correction coefficients k flow =0.001 / ( mm / s The formula is: I flow ( t )= I pH ( t )・[1+ k flow ・( v ( t )- v ref Substitute the values: I flow =24.2138 × 1 = 24.2138 nA Complete the basic calibration.

[0181] 3-3 Ultrafiltration Interference Compensation: The formula is: ,in =300 mL / h ×0.5 h =150 mL Individual metabolic correction factor k meta ( t ) = 1 + 0.01 × 0.5 - 0.005 × 55 = 0.73; Substitute the value: I uf =24.2138×(1-(150 / 4800)×0.73) -1 ≈24.779 nA Dialysis exchange interference compensation: The formula is: I comp ( t )= I uf ( t )・[1+ k d ・( C d,glu ( t )- C d,ref )・(1- α • η reject )],in C d,ref =1.5 mmol / L Pre-calibration of transmembrane exchange coefficient k d =0.15 / ( mmol / L ), α =0.02, this time C d,glu ( t Consistent with the reference value, the correction factor is 1, and the final value is... I comp =24.779 nA This completes the dual-level anti-interference processing.

[0182] 3-4 Steady-state interval determination: During the 25-30 minute treatment interval, two steady-state conditions must be met: signal fluctuation coefficient ≤ 2% (0.48%). ΔUFR =0 mL / h ≤5 mL / h , ΔC d,glu =0 mmol / L ≤0.5 mmol / L This is determined to be an effective steady-state interval.

[0183] Self-calibration calculation: Extracting the mean steady-state response I comp,steady =24.78 nA The formula is: C glu,ref =( k 0+ k 1・ I comp,steady )・(1+ k age • old + k dial • t dial ), where the initial fitting parameters are pre-calibrated. k 0 = 0.25 k 1 = 0.22, k age =0.003 / year, k dial =0.002 / month, substituting the patient's age of 55 years and dialysis duration of 6 months, we get the individual correction term = 1 + 0.003 × 55 + 0.002 × 6 = 1.177, from which the baseline blood glucose concentration is calculated. C glu,ref =(0.25+0.22×24.78)×1.177≈6.71 mmol / L .

[0184] Fitting parameter update: Obtaining new parameters after self-calibration k 1'≈0.2607, k 0'=0.25, completing the bloodless self-calibration, eliminating the need to collect patient blood samples.

[0185] 3-5 Blood glucose concentration calculation: The formula is as follows C glu ( t )= k 1'・ I comp ( t )+ k 0', Substitute the value: C glu =0.2607×24.78+0.25≈6.71 mmol / L .

[0186] Validity verification: Values ​​range from 3.9 to 10.0. mmol / LThe physiologically reasonable range; compared with the previously collected 7.02 mmol / L 6.85 mmol / L A slow downward trend was observed, consistent with the typical blood glucose fluctuations during the initial stages of hemodialysis; the hemodialysis conditions remained stable, and the blood glucose trend matched the operating conditions, thus the verification was successful, confirming the final blood glucose concentration as 6.71. mmol / L .

[0187] Step 4: Predicting and Processing Blood Glucose Trends 4-1 Integrating this treatment 0 bad 10 bad 20 bad 30 bad The effective blood glucose level is 7.25. mmol / L 7.02 mmol / L 6.85 mmol / L 6.71 mmol / L The system matches corresponding timestamps, blood physical parameters, and hemodialysis operating parameters to form a continuous time series with no missing data; it also extracts the characteristics of slow blood glucose decline and matches the blood glucose change patterns of the patient's previous 6 hemodialysis treatments.

[0188] 4-2 Based on the patient's type 2 diabetic nephropathy, 6 months of dialysis experience, and basic information about the routine hemodialysis regimen, the basic pattern type was determined to be the slow decrease of blood glucose during routine hemodialysis in diabetic nephropathy patients. The 30-minute blood glucose time series was compared with the corresponding standard pattern, and the matching degree was 92.3%, reaching the set threshold of 85%. It was determined that the current treatment matched this pattern. Considering the work conditions and the lack of adjustment plan, no pattern correction was required.

[0189] 4-3 The formula for calculating the rate of change in blood glucose is: ;in Δt window =5 bad , C glu ( t -5 bad )=6.85 mmol / L , k fluc =0.05, nearly 5 bad Standard deviation of ultrafiltration rate σ UFR =0, substitute the value: v glu,change =(6.71-6.85) / 5×(1-0)=-0.028 mmol / ( L • bad Simultaneously extract current operating parameters, future treatment plans, and the patient's past blood glucose characteristics at the same stage, and screen core features to form a trend feature set.

[0190] 4-4 Multi-period prediction calculation of blood glucose change trends, the core prediction formula is: Ĉ glu ( t + Δt )= C glu ( t )+ v glu,change ( t )・ Δt + k treat • ΔC treat ( Δt ), where the pre-calibrated operating condition influence coefficient k treat =0.8, no operational condition adjustments expected in the next 30 minutes. ΔC treat =0, the formula simplifies to Ĉ glu =6.71-0.028× Δt .

[0191] Short-term forecast (next 10 minutes): Δt =10 bad Predicting blood sugar Ĉ glu =6.43 mmol / L It generates a continuous blood glucose change curve within 10 minutes, with an accuracy that meets clinical requirements.

[0192] Mid-cycle forecast (next 30 minutes): Δt =30 bad Predicting blood sugar Ĉ glu =5.87 mmol / L This generates a 30-minute blood glucose trend curve.

[0193] Long-term prediction (remaining 3.5 hours of treatment time): Based on the patient's past treatment patterns, the rate of blood glucose decline slows down after 2 hours of treatment, and the calculated blood glucose level at the end of treatment is approximately 4.1 mmol / L. mmol / L This generates a trend curve of blood glucose changes throughout the entire process, and all prediction results are verified to be reasonable according to physiological laws.

[0194] The lowest predicted blood glucose level for this multi-cycle period was 4.1. mmol / L The entire process meets the 3.9 standard. mmol / L ≤ Ĉ glu ≤10.0 mmol / L The safety zone requirements meet the low-risk level classification criteria and are therefore classified as low-risk.

[0195] Step 5: Tiered Early Warning Processing 5-1 The current blood glucose level is 6.71. mmol / L Historical blood glucose curves, multi-cycle prediction results, and low-risk level information are synchronously transmitted to the hemodialysis machine's main control system and medical workstation via a standard communication protocol. All data is timestamped, and the receiving end provides feedback confirmation, ensuring complete data transmission without loss.

[0196] 5-2 This case is low risk, triggering a Level 1 warning. Blood glucose data, change curves, and prediction results are only displayed in real time on the hemodialysis machine operation interface and the medical workstation interface. No audio or visual prompts are triggered to avoid interfering with normal clinical work.

[0197] Supplementary demonstration of a medium-risk scenario (120 minutes of treatment): At 120 minutes of treatment, the measured blood glucose level was 4.2. mmol / L The rate of change in blood glucose was -0.022. mmol / ( L • bad Predicted blood sugar level to be 3.54 mmol / L in 30 minutes. mmol / L The patient's blood sugar level meets the criteria for a medium-risk level, triggering a Level 2 warning: the interface displays a medium-risk warning, and the medical workstation triggers an alert sound to remind medical staff to pay attention to changes in the patient's blood sugar.

[0198] Supplementary high-risk scenario demonstration (180 minutes of treatment): At 180 minutes of treatment, the measured blood glucose level was 3.2. mmol / L The rate of change in blood glucose was -0.03. mmol / ( L • bad (The blood sugar level is predicted to be 2.9 mmol / L in 10 minutes.) mmol / L If the risk level is met, a Level 3 warning is triggered: the hemodialysis machine and the medical workstation simultaneously trigger an audible and visual alarm, and at the same time, the risk information is pushed to the mobile terminal of the responsible medical staff.

[0199] 5-3 Low-risk state: Blood glucose is predicted to be within a safe range throughout the treatment, and no adjustment of hemodialysis treatment parameters is required. Continue monitoring according to the original plan.

[0200] Medium-risk status (120 minutes of treatment): Dialysis fluid glucose concentration is adjusted accordingly. The clinical validation safety adjustment factor k adj =0.3, lead time for prediction Δt pre =30 bad Medium risk adjustment coefficient k risk =1.2, Individual tolerance coefficient k toler =1+0.01×55=1.55, substituting the values, we get... ΔC d,glu ≈0.37 mmol / L After adjustment, the glucose concentration of the dialysate is 1.5 + 0.37 = 1.87. mmol / L It conforms to [0, 5.5]. mmol / L The system automatically adjusts parameters to meet safety constraints, and displays the adjustment details and reasons on the interface.

[0201] High-risk condition (180 minutes of treatment): Immediately stop ultrafiltration. UFR new =0; simultaneously, the emergency glucose concentration in the dialysate is calculated using the emergency formula: C d,glu,emerg =5.5-(3.0- Ĉ glu )・ k toler Substituting the values ​​into the equations is worthwhile. C d,glu,emerg ≈5.35 mmol / L Immediately implement the adjustment and simultaneously trigger an emergency alarm.

[0202] After adjusting the risk parameters in section 5-4, the blood glucose level was 4.1 mmol / L for three consecutive collection cycles. mmol / L 4.05 mmol / L 4.02 mmol / L The rate of decrease in blood glucose was -0.022. mmol / ( L • bad It dropped to -0.006 mmol / ( L • bad The adjustment achieved the expected control effect, confirming its effectiveness. The current parameters will continue to be tracked. The data on the effect of this adjustment will be fed back to the adjustment rule base to complete the rule optimization.

[0203] In another embodiment of the present invention, a dual-level anti-interference online blood glucose monitoring system for hemodialysis machines is provided. This system is applied to the online blood glucose monitoring scenario of hemodialysis machines, and can realize in-situ, continuous and non-invasive monitoring of patients' blood glucose during hemodialysis treatment. At the same time, it can predict and classify the trend of blood glucose changes and provide early warning. It can also form data interaction and parameter adjustment with the main control system of the hemodialysis machine, providing technical support for the safe management of blood glucose in hemodialysis treatment.

[0204] like Figure 7 As shown, the system consists of a blood sample diversion and processing module, a multi-source signal acquisition and preprocessing module, a blood glucose detection anti-interference correction module, a blood glucose change trend prediction module, and a graded early warning and parameter linkage module.

[0205] The five modules described above sequentially transmit data and exchange commands. The output of the previous module serves as the input data for the next module. Each module independently completes its preset functions, while the overall system operates in an orderly manner through internal data links. The working parameters of all modules are calibrated based on clinical and engineering experiments of hemodialysis, closely matching the actual working conditions of hemodialysis treatment. The detection and prediction results meet the relevant standards of clinical diagnosis and treatment. The system can be adapted to different models of hemodialysis machines and achieves data interaction and parameter linkage through standard communication protocols, making it suitable for online blood glucose monitoring in various hemodialysis treatment scenarios.

[0206] The blood sample diversion and processing module is used to achieve non-invasive extraction, testing, and reinfusion of trace blood samples from the main circuit of the hemodialysis machine. At the same time, it prevents clotting in the bypass circuit through pipeline maintenance operations, providing trace blood samples that meet the testing requirements for subsequent signal acquisition. The workflow of this module covers the entire cycle of hemodialysis treatment, from pre-filling of the pipeline before treatment to blood sample extraction, testing, and reinfusion during treatment, and pipeline flushing and maintenance throughout the treatment process, all of which are completed independently by this module.

[0207] Optionally, the module includes a bypass pipeline pre-filling unit, a micro-blood sample extraction unit, a blood sample stabilization unit, a blood sample reinfusion unit, and a pipeline anti-coagulation flushing unit. Each unit is started sequentially according to a preset time sequence to form a closed-loop blood sample diversion operation process.

[0208] The bypass tubing pre-filling unit is operated before the formal start of hemodialysis treatment. It uses 0.9% sodium chloride saline, the same pre-filling fluid used in the hemodialysis machine, to pre-fill the bypass branch, with the pre-filling flow rate controlled at 2... mL / bad The pre-charge duration is set to 1 minute. bad Ensure that all pipeline cavities within the bypass branch are completely filled with pre-filled liquid, with no residual air. After pre-filling, the unit detects the flow status of the pipeline through pressure sensors at both ends of the bypass branch. When the pressure difference between the inlet and outlet of the bypass branch is less than or equal to 5... mmHgWhen the flow is determined to be normal, with no blockage or poor flow, the bypass branch enters standby mode. If the pressure difference is greater than 5... mmHg This unit will trigger a pipeline blockage warning, and the pre-charge operation will be re-executed after manual inspection.

[0209] The micro-blood sampling unit starts monitoring after the hemodialysis treatment officially begins. It detects when the hemodialysis machine's runtime is greater than or equal to 30 seconds. bad When the blood pump speed and transmembrane pressure fluctuation are less than or equal to 5%, the hemodialysis machine is considered to have entered a stable treatment state. This triggers the blood sample extraction action of the bypass branch, opening the miniature solenoid valve at the bypass branch inlet, allowing arterial blood to enter the bypass branch under the pressure difference of the blood pump. This unit incorporates a dynamic blood viscosity correction term and an anticoagulation constraint term when controlling the bypass branch blood sample flow rate. Based on the hemodialysis treatment duration, the dynamic dynamic viscosity of the blood is calculated in real time. Combined with parameters such as the bypass branch channel inner diameter, arteriovenous pressure difference, and channel length, the blood sample flow rate of the bypass branch that meets the design constraints is a blood sample flow rate not exceeding 0.3. mL / bad Simultaneously, the filling time of the detection chamber is adjusted according to the blood viscosity correction factor, with the design constraint that the filling time should not exceed 2 seconds. s This design closely reflects the dynamic changes in blood viscosity over treatment duration, preventing deviations in flow control. The detection chamber in this module is cylindrical, with an effective volume set to 5. μL The inner wall is coated with a medical heparin anticoagulant coating and installed in the middle section of the bypass branch. A molecular sieving physical anti-interference device is integrated at the inlet to provide a physical basis for subsequent anti-interference treatment.

[0210] The blood sample stabilization unit activates after the detection chamber is filled with blood sample, closing the bypass branch inlet micro solenoid valve to stop blood flow within the detection chamber and enter a stable stabilization state. The stabilization time is set to 15 seconds. s This duration is calibrated based on the reaction kinetics of glucose and an electrochemical sensor to ensure the glucose-specific reaction proceeds fully. During the residence time, the unit monitors the blood temperature in real time using a micro-sensor integrated into the detection chamber. pH Physical parameters such as pressure, when blood temperature is between 35 and 40°C, pH The value is between 7.0 and 7.8, and the pressure is between 80 and 120. mmHg When the blood condition is determined to meet the testing requirements, a trigger signal is sent to the multi-source signal acquisition and preprocessing module to start the subsequent signal acquisition process. If the above parameters exceed the preset range, the unit will trigger a blood sample status abnormality prompt, return the blood sample in the detection chamber and re-execute the blood sample extraction operation.

[0211] The blood sample reinfusion unit starts after receiving feedback from the multi-source signal acquisition and preprocessing module that the signal acquisition has been completed. It opens the miniature solenoid valve at the bypass branch outlet, allowing the blood in the detection chamber to be reinfused into the venous line under the pressure difference of the hemodialysis tubing. The reinfusion time is controlled to be 2 seconds. s After the reinfusion is completed, the optical sensor in the detection chamber detects whether there is any blood sample residue inside. When the light transmittance is greater than or equal to 95%, it is determined that there is no blood sample residue. After confirming that there is no residue, the outlet micro solenoid valve is closed, returning the bypass branch to the standby state, waiting for the trigger signal for the next blood sample extraction. This avoids coagulation caused by residual blood sample or cross-contamination of samples for subsequent testing. The tubing anti-coagulation flushing unit works continuously throughout the hemodialysis treatment, performing bypass tubing flushing operations every 30 minutes. The flushing fluid is normal saline for hemodialysis treatment, and the flushing flow rate is controlled at 5. mL / bad The flushing volume is set to 3. mL This parameter was determined based on pipeline anticoagulation engineering experiments, and it does not cause abnormalities in the patient's electrolyte or volume status while preventing coagulation.

[0212] During flushing, the unit opens the inlet and outlet micro solenoid valves of the bypass branch, allowing the flushing fluid to flow through the complete flow path of the bypass branch, carrying away any trace amounts of blood or impurities remaining in the tubing. After flushing is completed, the flow status detection function of the bypass tubing pre-filling unit is reactivated to confirm that the flow status of the bypass branch is normal, ensuring that the bypass branch is in a working state throughout the entire hemodialysis treatment.

[0213] The multi-source signal acquisition and preprocessing module is used to synchronously acquire and eliminate noise from glucose-specific signals, blood physical parameters, and hemodialysis machine operating parameters. After smoothing, normalizing, and time-synchronizing the acquired raw signals, a standardized dataset is formed and transmitted to the blood glucose detection anti-interference correction module. This module includes a signal acquisition trigger unit, a glucose signal acquisition unit, a blood physical parameter acquisition unit, a hemodialysis machine operating parameter acquisition unit, and a signal preprocessing unit. Each unit starts working synchronously after receiving the trigger signal from the blood sample diversion processing module, ensuring that all acquired signals correspond to the blood state and hemodialysis machine operating conditions at the same time.

[0214] The signal acquisition trigger unit is the control core of this module. When the blood sample diversion and processing module determines that the blood condition meets the detection requirements, it sends a unified hardware trigger signal to control the glucose response current, blood temperature, and blood... pH The system simultaneously starts collecting data from six signal channels: blood pressure, hemodialysis machine status, and blood flow rate.

[0215] During the data acquisition process, this unit monitors the working status of each channel in real time. When the data loss rate of a certain channel exceeds 5% or the data exceeds the preset reasonable range, the acquisition channel is determined to be abnormal. The current acquisition process is immediately stopped, and a command is sent to the blood sample splitting processing module to re-execute the blood sample splitting operation. At the same time, a channel abnormality prompt is triggered to prevent invalid data from abnormal channels from entering the subsequent processing flow. This unit sets strict timing constraints for the synchronous acquisition of all signal channels. The difference in acquisition time between different signal channels is no greater than half of the reciprocal of the sampling frequency, which is set to 10. Hz This ensures the consistency of data collection time across all channels, preventing time deviations from affecting subsequent correction and calculation results.

[0216] After receiving the trigger signal, the glucose signal acquisition unit applies a 0.6 μL pressure to the electrochemical three-electrode sensor in the detection chamber. V The sensor operates at a constant voltage; its working electrode is a gold electrode, and its reference electrode is... Ag / AgCl The electrodes are platinum electrodes, and the application of a constant voltage ensures the stable conduct of the glucose-specific redox reaction. This unit acquires the current signal generated during the reaction in real time. This current signal is linearly correlated with blood glucose concentration, providing a core basis for subsequent blood glucose concentration calculation. The unit continuously acquires multiple sets of current signals at a set sampling frequency, simultaneously recording the acquisition timestamp for each set of signals. After acquisition, the raw current signals undergo initial validity screening, removing abnormal signals that significantly exceed reasonable ranges to ensure the validity of the raw signals.

[0217] The blood physical parameter acquisition unit starts synchronously with the glucose signal acquisition unit. The acquired blood physical parameters include temperature, pH, flow rate, and pressure. The acquisition frequency and acquisition period are completely consistent with the glucose response signal, ensuring that each set of glucose response signals corresponds to the blood physical parameters at the same moment.

[0218] After data collection, this unit performs a validity check on the parameter data, discarding data with temperatures below 35℃ or above 40℃. pH Less than 7.0 or greater than 7.8, pressure less than 80 mmHg or greater than 120 mmHg If abnormal data is found, and the proportion of abnormal data is greater than 10%, an instruction is sent to the signal acquisition trigger unit to re-execute the signal acquisition operation. After the data is filtered, all physical parameters are matched with the glucose response signal according to the acquisition timestamp to further ensure the time consistency of the data.

[0219] The hemodialysis machine operating parameter acquisition unit obtains data through... Modbus - RTUThe standard communication protocol interacts with the main control system of the hemodialysis machine to synchronously collect real-time operating parameters of the hemodialysis machine. The collection frequency is consistent with the glucose response signal. The collected parameters include blood pump speed, dialysate flow rate, ultrafiltration rate, transmembrane pressure, dialysate glucose concentration, and treatment duration. These parameters can reflect the real-time status of hemodialysis treatment and provide data support for subsequent operating condition compensation and trend prediction.

[0220] After collecting the operating parameters, this unit judges the validity of the parameters, confirms that the parameters are within the normal operating range of the hemodialysis machine, and removes abnormal parameter data. At the same time, the acquired operating parameters are matched with the glucose response signal and blood physical parameters according to the collection timestamp to ensure the time synchronization of the three types of data.

[0221] The signal preprocessing unit is the core data processing unit of this module. After receiving valid data from the three acquisition units, it first smooths all signals using a moving average smoothing filter. By using a set smoothing window half-width, it calculates and processes multiple continuously acquired signal data sets to eliminate random noise generated during signal acquisition while preserving the signal's inherent trend and preventing the loss of effective signal characteristics during smoothing. After smoothing, all signals are normalized using a minimum-maximum normalization method to convert signal data of different types and dimensions to a uniform numerical range, eliminating dimensional differences between different signals and facilitating subsequent correction and calculation processing.

[0222] After normalization, a second time synchronization matching is performed on all signals, and the timestamps of each signal are checked one by one to confirm that all signal data accurately correspond to the same acquisition time and avoid time misalignment. After matching, all preprocessed signal data are organized into a unified standardized dataset and transmitted to the blood glucose detection anti-interference correction module through an internal data link. At the same time, the original signal and preprocessed signal data are temporarily stored to facilitate subsequent backtracking and verification.

[0223] The blood glucose detection anti-interference correction module is used to eliminate detection deviations caused by interfering substances in the blood and changes in hemodialysis conditions through a two-level anti-interference mechanism. At the same time, it can achieve blood glucose self-calibration without blood collection, and finally calculate the blood glucose concentration value that meets clinical requirements and complete the validity verification of the value. The blood glucose concentration value that has passed the verification will be transmitted as the core data to the blood glucose change trend prediction module.

[0224] The module's dual-level anti-interference mechanism includes a first-level molecular sieving physical anti-interference and a second-level parameter correction and operating condition compensation. The module is equipped with a physical anti-interference effect verification unit, a blood parameter correction unit, an operating condition interference compensation unit, a bloodless self-calibration unit, and a blood glucose concentration calculation and verification unit. Each unit performs operations sequentially to gradually complete the anti-interference correction and blood glucose concentration calculation. The processing results of each unit must meet the preset constraints before proceeding to the next unit's processing flow.

[0225] The physical anti-interference effect verification unit is the first processing unit of this module. The first level of physical anti-interference is achieved by the polyamide nanofiltration membrane molecular sieving device integrated into the inlet of the detection chamber. The molecular rejection threshold of this device is 300. Yes The membrane pore size is 2 nm Prepared using a high-voltage electrospinning process, it allows glucose molecules to pass through smoothly while blocking molecules with a molecular weight greater than or equal to 300 in the blood. Yes Interfering substances, including heparin, creatinine, and urea, are eliminated to avoid detection bias caused by non-specific reactions. This unit judges the actual effect of physical anti-interference treatment by the baseline stability of the glucose response signal after preprocessing, calculates the standard deviation of the signal baseline, and calculates the fluctuation coefficient by the ratio of the standard deviation to the signal mean. When the fluctuation coefficient is less than or equal to 5%, the physical anti-interference effect is deemed to be satisfactory, and the subsequent parameter calibration process can proceed. If the fluctuation coefficient is greater than 5%, it indicates that the interfering substances have not been effectively blocked. This unit will send a command to the blood sample splitting module to re-execute the blood sample splitting operation and trigger an abnormality prompt from the physical anti-interference device.

[0226] Simultaneously, this unit calculates the retention efficiency of interfering substances, which is the ratio of the interfering substance concentration on the permeate side to the interfering substance concentration on the feed side. For molecules with a molecular weight greater than or equal to 300... Yes To mitigate interference, a constraint is set on a rejection efficiency of no less than 95% to ensure the actual effectiveness of physical anti-interference.

[0227] The function of the blood parameter correction unit is based on the blood temperature, pH Multiple physical parameters, such as flow rate, are used to perform stepwise correction of the glucose response signal, eliminating the influence of changes in blood physical parameters on the glucose-specific response rate and signal acquisition results, so that the corrected signal can more accurately reflect the blood glucose concentration. All correction coefficients were determined based on in vitro blood sample calibration experiments of no less than 50 hemodialysis patients, with calibration conditions of 35 to 40°C. pH 7.0 to 7.8, flow rate 0 to 5 mm / s The corresponding correction coefficients are obtained through linear fitting. All calibration data and fitting curves are stored in the system database and can be retrieved and updated at any time.

[0228] The calibration operation of this unit is based on temperature, pH The flow rate was adjusted sequentially, starting with temperature correction. Using the core human body temperature of 37°C as a reference, and combining this with a pre-experimental calibrated temperature correction coefficient, the smoothed and normalized glucose response current was calculated to obtain the temperature-corrected current signal. Subsequently, pH correction was performed, using a normal human blood pH of 7.4. pH Using the reference value and the pH correction coefficient calibrated in the pre-experiment, the temperature-corrected current signal is calculated to obtain the pH-corrected current signal; finally, flow field correction is performed, with 0... mm / s Using the steady-state reference flow velocity as a benchmark and combining the flow field correction coefficient calibrated in the pre-experiment, the current signal after pH correction is calculated to obtain the glucose response signal after blood physical parameter correction. This signal has eliminated the detection deviation caused by blood physical parameters.

[0229] The interference compensation unit receives the processing results from the blood parameter correction unit and, based on the real-time operating parameters of the hemodialysis machine, compensates for interference in the signal after blood parameter correction. This eliminates detection deviations caused by changes in hemodialysis operating conditions and compensates for residual trace interference after physical anti-interference measures. The unit's compensation operation is divided into ultrafiltration interference compensation, dialysis exchange interference compensation, and blood circuit status interference compensation, sequentially processing compensation for different operating parameters. Each compensation step has corresponding calculation formulas and calibration coefficients, closely reflecting the actual operating conditions of hemodialysis treatment.

[0230] Ultrafiltration interference compensation is achieved through a dedicated ultrafiltration interference compensation formula, which incorporates the patient's total blood volume and individual metabolic correction coefficient. The individual metabolic correction coefficient is calculated based on the patient's age and the duration of hemodialysis treatment, compensating for the interference caused by differences in metabolic rates among patients of different ages and changes in metabolic state during treatment, thus achieving individualized adaptation of ultrafiltration compensation. Dialysis exchange interference compensation is designed based on the dynamic characteristics of transmembrane exchange, incorporating the transmembrane glucose exchange coefficient, the interference residual influence coefficient, and the physical anti-interference retention efficiency, linking the effect of physical anti-interference with dialysis exchange compensation, thereby improving the actual compensation effect. Blood circuit state interference compensation calculates the detection deviation caused by changes in blood circuit state based on the synchronously acquired blood pump speed and transmembrane pressure values, and performs corresponding compensation processing on the signal to eliminate the interference caused by changes in blood pump speed and transmembrane pressure.

[0231] After completing all the above compensation operations, the obtained current signal is the final response signal after completing the two-stage anti-interference processing. This signal has eliminated various detection deviations caused by interfering substances in the blood and changes in hemodialysis conditions. The bloodless self-calibration unit is the calibration core of this module. Based on the steady-state characteristics during hemodialysis treatment and combined with the patient's physiological characteristics, it achieves bloodless self-calibration, eliminating dependence on blood sampling calibration and overcoming calibration deviations caused by individual physiological differences.

[0232] The self-calibration formula incorporates the patient's age, dialysis history, and other physiological characteristics. The updated fitting parameters retain the correlation with individual physiological characteristics, take effect immediately, and are stored in the patient's local database for subsequent blood glucose concentration calculations. The self-calibration trigger conditions for this unit are: each time a valid steady-state interval is identified, every 30 minutes, and the blood glucose detection error is greater than 4.5%. Meeting any one of these conditions will automatically trigger the self-calibration operation.

[0233] This unit first determines the steady-state interval during hemodialysis treatment. Determining the steady-state interval requires meeting two conditions: first, the fluctuation coefficient of the response signal after two-stage anti-interference is less than or equal to 2% within the interval; second, the maximum change in ultrafiltration rate within the interval is no greater than 5%. mL / h Furthermore, the maximum change in glucose concentration in the dialysate should not exceed 0.5%. mmol / L When both conditions are met simultaneously, the interval is determined to be a valid steady-state interval suitable for calibration. After determining the valid steady-state interval, the mean value of the response signal within the interval is extracted as the steady-state response benchmark value. Combined with pre-established concentration-current fitting parameters, the benchmark blood glucose concentration corresponding to the steady-state interval is calculated using a self-calibration formula. Subsequently, the concentration-current fitting parameters are automatically updated using this benchmark blood glucose concentration, completing the bloodless self-calibration process. The entire process does not require collecting blood samples from patients, reducing patient discomfort and infection risks.

[0234] The blood glucose concentration calculation and verification unit first calculates the blood glucose concentration value corresponding to the current blood sample using a linear formula based on the glucose response signal after completing the dual-level anti-interference processing and the concentration-current fitting parameters updated by the bloodless self-calibration unit. The detection error of this calculation formula can be controlled within ±4.5%, which meets the requirements. ISO 15197 Clinical Standards.

[0235] After the calculation is completed, the unit will perform multi-dimensional validity verification of the blood glucose concentration value. First, it will determine whether the value is within the normal physiological range of blood glucose in the human body. If it is outside the range, all correction processes will be rechecked and recalculated. Second, the current value will be compared with multiple blood glucose concentration values ​​collected previously to determine whether the trend of the value changes in accordance with the physiological change law of blood glucose in the human body and the blood glucose change law during hemodialysis treatment. If there is a sudden and drastic change that does not conform to the law, the signal acquisition and correction process will be re-executed. Finally, combined with the synchronously acquired hemodialysis machine operating parameters, it will determine whether the change of blood glucose concentration value matches the change of hemodialysis machine operating conditions. If there is no significant change in operating conditions but the blood glucose value fluctuates drastically, the verification will be repeated.

[0236] Once the blood glucose concentration value passes all validity verifications, it is determined to be the accurate blood glucose concentration value of the current blood sample. The value is then transmitted to the blood glucose change trend prediction module. At the same time, the value, along with the corresponding collection time, correction parameters, and other information, is temporarily stored for subsequent backtracking and analysis.

[0237] The blood glucose trend prediction module, based on historical blood glucose data from the current treatment, blood glucose variation patterns during hemodialysis, and real-time operating parameters, performs multi-period trend predictions of blood glucose changes. It then categorizes risk levels based on the predicted blood glucose values, and the prediction results and risk levels are transmitted to the graded early warning and parameter linkage module. This module includes a historical blood glucose dataset construction unit, a blood glucose variation pattern matching unit, a trend feature extraction unit, a multi-period blood glucose prediction unit, and a risk level classification unit. Each unit performs operations sequentially, from basic data organization to clinical pattern matching, then to core feature extraction and multi-period prediction calculations, ultimately completing the risk level classification.

[0238] The blood glucose history dataset construction unit first integrates all blood glucose concentration values ​​that have been validated by the blood glucose detection anti-interference correction module during this hemodialysis treatment, as well as the collection time, blood physical parameters, and hemodialysis machine operating parameters corresponding to each value, and sorts them in chronological order of collection time to form a continuous blood glucose change time series.

[0239] This unit performs an integrity check on the constructed dataset. If the percentage of missing data points is less than or equal to 5%, it supplements the data using linear interpolation of adjacent data. If the percentage of missing data points is greater than 5%, a data missing warning is triggered, and a command is sent to the blood glucose detection anti-interference correction module to re-collect and supplement the data. After the data integrity check is completed, the characteristics of the blood glucose change time series are preliminarily analyzed, including the rate of increase, rate of decrease, and fluctuation amplitude. Simultaneously, the constructed dataset is matched with the patient's historical hemodialysis treatment data, which is stored in the patient's local database. By matching, the blood glucose change patterns during the patient's past treatments are extracted, providing individual reference for subsequent pattern matching.

[0240] The blood glucose change pattern matching unit is based on the blood glucose change pattern of hemodialysis treatment, which is obtained from the statistical analysis of clinical data of no less than 200 hemodialysis patients. It is divided into different types according to the basic characteristics of patients. Typical pattern types include slow decrease of blood glucose in routine hemodialysis for diabetic nephropathy patients, stable blood glucose in hemodialysis for non-diabetic patients, and slow increase of blood glucose under high dialysate glucose concentration. The standard pattern of each type is stored in the system in the form of curves and can be optimized according to the updates of clinical data.

[0241] This unit first determines the corresponding baseline blood glucose variation pattern based on the patient's basic characteristics, including whether they have diabetes, the type of diabetes, the duration of dialysis, and the treatment plan. Then, it compares the blood glucose variation time series of this treatment with the corresponding standard pattern, using the Pearson correlation coefficient to calculate the degree of match. When the degree of match is greater than or equal to 85%, the treatment is considered to match the standard pattern. If the degree of match is less than 85%, the standard pattern is fine-tuned based on the dialysate glucose concentration, ultrafiltration protocol, and other operating parameters of this treatment, making the adjusted pattern more closely reflect the actual situation of this treatment and providing an accurate basis for subsequent trend prediction.

[0242] The function of the trend feature extraction unit is to extract core features that reflect future blood glucose trends, forming a trend feature set. During feature extraction, the fluctuation characteristics of hemodialysis conditions are considered, and a condition fluctuation correction factor is introduced to ensure the practical reference value of the extracted features. This unit first extracts blood glucose change data from the historical blood glucose dataset for the five minutes preceding the current moment. It then calculates the rate of blood glucose change by combining this with the fluctuation characteristics of hemodialysis conditions. The calculation formula for the rate of blood glucose change incorporates the influence coefficient of condition fluctuation and the standard deviation of the ultrafiltration rate over the past five minutes to compensate for the bias caused by fluctuations in hemodialysis conditions in calculating the rate of blood glucose change. Next, it extracts all real-time operating parameters of the hemodialysis machine at the current moment, as well as the pre-set adjustment plans for operating parameters within the future prediction period, including adjustments to the ultrafiltration rate and dialysate glucose concentration. These parameter adjustments directly affect the patient's future blood glucose changes. Finally, it extracts the patient's basic physiological information and blood glucose change characteristics under the same treatment stage and operating conditions during past hemodialysis treatments as reference features.

[0243] After all features are extracted, the features are screened for effectiveness. Invalid features with little impact on blood glucose trend are removed, while core features with significant impact on blood glucose trend are retained and organized into a standardized trend feature set to provide feature basis for subsequent multi-period prediction.

[0244] The multi-cycle blood glucose prediction unit completes multi-cycle blood glucose prediction based on the matched blood glucose change pattern and the extracted trend feature set. The short cycle is the next 10 minutes, the medium cycle is the next 30 minutes, and the long cycle is the remaining time of the current treatment. The prediction results of different cycles correspond to different clinical application scenarios, providing a reference for blood glucose management for different durations.

[0245] The prediction calculation of this unit is based on a preset prediction formula, which includes the current blood glucose concentration, blood glucose change rate, operating condition influence coefficient, and the amount of blood glucose change caused by changes in hemodialysis operating conditions within the prediction period. If there is no adjustment of hemodialysis operating conditions within the prediction period, the amount of blood glucose change caused by changes in operating conditions is 0, and the formula is simplified accordingly. Short-term prediction is based on the current blood glucose concentration, combined with the current rate of blood glucose change and the operating parameters of the hemodialysis machine in the next 10 minutes. It calculates the blood glucose concentration at each time point within the next 10 minutes according to the matched blood glucose change pattern, forming a short-term blood glucose trend curve. The results of short-term prediction closely match the patient's short-term blood glucose changes. Medium-term prediction is based on the results of short-term prediction, combined with the hemodialysis machine's operating adjustment plan for the next 30 minutes and the patient's blood glucose change pattern, calculating the blood glucose concentration at each time point within the next 30 minutes, forming a medium-term blood glucose trend curve. This allows sufficient time for early risk warning and intervention. Long-term prediction is based on the results of medium-term prediction, combined with the complete treatment plan for the remaining time of this treatment and the patient's blood glucose change patterns throughout past treatments, calculating the blood glucose trend curve for the remaining treatment time.

[0246] After all the prediction calculations for all cycles are completed, the rationality of the prediction results is verified to determine whether the predicted blood glucose change trend conforms to the physiological change law of human blood glucose and the blood glucose change law of hemodialysis treatment. If the prediction results show obvious irrationality, the matching law and extracted features are readjusted, and the prediction calculation is performed again.

[0247] The risk grading unit classifies the predicted blood glucose levels based on the multi-cycle blood glucose prediction unit, combined with clinical guidelines and statistical standards derived from extensive clinical data. The risk levels are categorized into low, medium, and high risk, determined by the clinically safe, watch-area, and abnormal blood glucose concentration ranges: low risk is defined as a predicted blood glucose concentration of 3.9 mmol / L. mmol / L Up to 10.0 mmol / L The clinically safe range is defined as blood glucose levels within which there is no harm to the patient's health and no additional intervention is required; medium risk is defined as a predicted blood glucose concentration of 3.0. mmol / L Up to 3.9 mmol / L Or 10.0 mmol / L Up to 16.7 mmol / L The clinical attention range is defined as blood glucose levels exceeding the safe range but not reaching the abnormal threshold, requiring continuous monitoring by healthcare professionals; high risk is defined as a predicted blood glucose concentration less than 3.0 mmol / L. mmol / L Or greater than 16.7 mmol / L The clinically abnormal range of blood glucose levels indicates that blood glucose levels within this range can be harmful to the patient's health and require emergency intervention.

[0248] Based on all the predicted blood glucose values ​​over multiple cycles, this unit will determine whether the patient's blood glucose will fall into different ranges in the future, and then classify the corresponding risk levels. After the classification is completed, the multi-cycle blood glucose prediction curve, the predicted blood glucose value, and the risk level classification results will be transmitted to the graded early warning and parameter linkage module.

[0249] The graded early warning and parameter linkage module is used to realize the synchronous transmission of blood glucose monitoring data and prediction results, complete graded early warning based on risk level, and automatically adjust the treatment parameters of the hemodialysis machine according to the prediction results and risk level, and track blood glucose changes after parameter adjustment to verify the adjustment effect.

[0250] This module includes a data synchronization and transmission unit, a graded early warning triggering unit, a treatment parameter adjustment unit, and an adjustment effect tracking unit. Each unit performs operations sequentially to complete the entire process of early warning and parameter adjustment. All operations comply with clinical diagnosis and treatment guidelines, while retaining the manual operation permissions of medical staff to ensure the safety of treatment.

[0251] Data synchronization transmission unit through RS The dual communication links of RS485 and Ethernet synchronously transmit the current effective blood glucose concentration output by the blood glucose detection anti-interference correction module, the historical blood glucose change curve of this treatment constructed by the blood glucose change trend prediction module, the future multi-cycle blood glucose prediction trend curve, the predicted blood glucose value, and the risk level classification results to the main control system of the hemodialysis machine and the medical workstation in the hemodialysis treatment room.

[0252] When transmitting data, this unit adds a corresponding timestamp to all data to ensure that the data received by the hemodialysis machine's main control system and the medical workstation completely corresponds to the actual collection time and prediction time, avoiding time misalignment. After data transmission, it waits for feedback signals from the receiving end to confirm that the data has been completely received. If no feedback signal is received, the data transmission is re-initiated, and a communication error prompt is triggered at the same time to ensure that all blood glucose data and prediction results can be accurately synchronized to the receiving end.

[0253] The tiered early warning triggering unit triggers different levels of early warning prompts based on the risk level defined by the blood glucose change trend prediction module. Different levels of early warning prompts correspond to different prompting methods and recipients, ensuring that medical staff can obtain patients' blood glucose risk information in a timely and accurate manner.

[0254] When the risk level is low, a Level 1 warning is triggered. The patient's blood glucose level, trend, prediction result, and risk level are displayed in real time only on the hemodialysis machine's operating interface and the medical staff's workstation interface. No additional audio or visual prompts are triggered to avoid unnecessary interference with the normal work of medical staff. When the risk level is medium, a Level 2 warning is triggered. In addition to displaying the above blood glucose data and risk information on the interface, an audio prompt will be triggered on the medical staff's workstation. At the same time, a prominent attention prompt will be displayed on the hemodialysis machine's operating interface to remind medical staff to pay attention to the patient's blood glucose changes and prepare for intervention in advance. When the risk level is high, a Level 3 warning is triggered. In addition to displaying detailed risk information on the interface, an audio and visual alarm will be triggered simultaneously on the hemodialysis machine and the medical staff's workstation. The alarm volume and flashing frequency are set according to the standards of clinical emergency alarms to immediately attract the attention of medical staff. At the same time, the patient's blood glucose risk information is simultaneously sent to the mobile terminal of the medical staff responsible for the patient, ensuring that medical staff can obtain emergency risk information and deal with it in a timely manner. This unit records in detail the trigger time, risk level, corresponding blood glucose value, and prediction result of all warnings, and stores them in the system database for easy retrospective analysis.

[0255] The treatment parameter adjustment unit automatically adjusts the hemodialysis machine's treatment parameters based on the patient's real-time blood glucose status, predicted blood glucose change trends, and risk level. This proactive intervention prevents the patient's blood glucose from entering abnormal ranges. The unit primarily adjusts the dialysate glucose concentration and, under high-risk conditions, adjusts the ultrafiltration rate. All adjustment parameters and magnitudes comply with clinical treatment guidelines, and the adjustment rules have been validated by clinical experts to ensure safety.

[0256] This unit incorporates clinically validated safety adjustment factors, advance prediction amounts, risk level correction factors, and individual tolerance correction factors when calculating adjustments to dialysate glucose concentration. The risk level correction factor is set according to risk level: 1.0 for low risk, 1.2 for medium risk, and 1.5 for high risk, ensuring the adjustment range matches the risk level. The individual tolerance correction factor is calculated based on the patient's age to compensate for differences in glucose tolerance among patients of different ages, achieving individualized adjustments. Furthermore, the adjustment amount for dialysate glucose concentration is controlled within the range of 0 to 5.5. mmol / L Within the clinical safety range.

[0257] When a patient is at medium risk and their blood glucose is predicted to decrease and approach the hypoglycemic threshold, the dialysate glucose concentration is automatically increased according to a calculation formula to increase the glucose exchange between the blood and the dialysate, thus slowing the rate of blood glucose decline. When a patient is at medium risk and their blood glucose is predicted to increase and approach the hyperglycemic threshold, the dialysate glucose concentration is automatically decreased to reduce the glucose content in the blood and slow the rate of blood glucose rise. When a patient is at high risk and hypoglycemia is predicted, the ultrafiltration rate of the hemodialysis machine is immediately adjusted to 0, and an emergency glucose concentration in the dialysate is calculated using an emergency formula to quickly replenish glucose. This emergency formula incorporates an individual tolerance correction coefficient to account for tolerance differences among patients of different ages. When a patient is at high risk and severe hyperglycemia is predicted, the dialysate glucose concentration is immediately adjusted to the minimum safe value, and the ultrafiltration rate is adjusted to the clinically safe range to help the patient reduce the glucose concentration in their blood. This is combined with a three-level warning system to trigger an emergency alarm.

[0258] After completing parameter adjustments, this unit will display detailed adjustment information on the hemodialysis machine's operating interface, including the adjusted parameters, adjustment range, and adjustment reasons. It will also record all adjustment operations and store them in the system database for easy viewing and review by medical staff. Furthermore, all automatic adjustment operations can be manually modified or stopped by medical staff at any time to further ensure the safety of treatment.

[0259] The effect tracking unit starts tracking after the treatment parameters are adjusted. According to the system's preset blood sample collection cycle, it continuously collects the patient's blood glucose concentration value, records the changes in blood glucose after adjustment, and simultaneously collects the operating parameters of the hemodialysis machine to confirm that the adjusted parameters have taken effect normally.

[0260] The tracking duration is set as the effect verification cycle, which can fully reflect the impact of parameter adjustments on patients' blood glucose changes. During the tracking process, the real-time blood glucose values ​​are compared with the predicted results before the adjustment to determine whether the blood glucose trend changes in the expected direction. If the blood glucose trend is controlled as expected and gradually returns to the clinically safe range, it indicates that the parameter adjustment has achieved the expected effect. The control system maintains the adjusted parameters and continues to track blood glucose changes. If the blood glucose trend does not change as expected and there is still a risk of entering the abnormal range, it indicates that the current adjustment range is insufficient. The adjustment parameters are recalculated, triggering the treatment parameter adjustment unit to perform a second adjustment. At the same time, the effect data of this adjustment is fed back to the system's adjustment rule library for optimization of the adjustment rules. If the blood glucose trend is over-adjusted, causing blood glucose to fluctuate abnormally in the opposite direction, it indicates that the current adjustment range is too large. The hemodialysis machine parameters are immediately adjusted back to safe values, and a prompt is triggered to remind medical staff to check and handle the situation.

[0261] The tracking results of this unit will be continuously fed back to the adjustment rule base to continuously optimize the system's parameter adjustment rules, so that subsequent parameter adjustments are more in line with the individual patient's condition and the actual working conditions of hemodialysis treatment.

[0262] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0263] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A two-stage anti-interference online blood glucose monitoring method for hemodialysis machines, characterized in that, Includes the following steps: Step 1: In-situ micro-blood sample diversion processing. Micro-blood samples from the main pipeline of the hemodialysis machine are non-invasively extracted, tested, and reinfused, while maintaining the pipeline to prevent clotting. Step 2: Synchronous acquisition and preprocessing of multi-source signals, including simultaneous acquisition and noise reduction of glucose-specific signals, blood physical parameters, and hemodialysis machine operating parameters; Step 3: Dual-level anti-interference correction processing, eliminating detection bias through a dual-level mechanism of molecular sieving physical anti-interference, parameter correction, and operating condition compensation, while simultaneously achieving bloodless self-calibration to calculate the effective blood glucose concentration; Step 4: Blood glucose trend prediction processing, completing multi-cycle trend prediction based on historical blood glucose data, blood glucose change patterns during hemodialysis treatment, and real-time operating parameters, and classifying risk levels according to the predicted blood glucose values; Step 5: Graded early warning and parameter linkage processing, synchronously transmitting monitoring data and prediction results, issuing graded early warnings based on risk levels, and automatically adjusting hemodialysis machine treatment parameters according to the prediction results.

2. The dual-level anti-interference online blood glucose monitoring method for hemodialysis machines according to claim 1, characterized in that, Step 1 includes sub-steps such as pre-filling and status check of the bypass tubing, controlling the extraction of a small amount of blood from the arterial tubing, ensuring stable residence of the blood sample within the testing chamber, returning the tested blood sample to the venous tubing, and periodically flushing the bypass tubing to prevent coagulation. A dynamic blood viscosity correction term and an anti-coagulation constraint term are introduced when extracting a small amount of blood, and the blood flow rate of the bypass branch is controlled to ≤0.

3. mL / min The filling time of the detection cavity is controlled within ≤2 seconds. s .

3. The dual-level anti-interference online blood glucose monitoring method for hemodialysis machines according to claim 1, characterized in that, The signals synchronously acquired in step 2 include glucose response current, blood temperature, and blood... pH Blood pressure, hemodialysis machine operating conditions, and blood flow rate were collected. After collection, the raw signals were sequentially processed by moving average smoothing filtering, minimum-maximum normalization, and time synchronization matching. The acquisition time deviation of different signal channels met the preset timing constraints.

4. The dual-level anti-interference online blood glucose monitoring method for hemodialysis machines according to claim 1, characterized in that, In step 3, the molecular sieving physical interference resistance uses a molecular retention threshold of 300. Da Nanofiltration membrane devices for molecular weight ≥300 Da Interference substance rejection efficiency ≥95%; parameter correction includes temperature correction, pH The stepwise calibration of the calibration and flow field calibration, the working condition compensation including ultrafiltration interference compensation, dialysis exchange interference compensation, blood circuit state interference compensation, and the triggering conditions for bloodless self-calibration are any of the following: identification of the effective steady state interval, every preset time period, and blood glucose detection error exceeding the preset value.

5. The dual-level anti-interference online blood glucose monitoring method for hemodialysis machines according to claim 1, characterized in that, Step 4 involves multi-cycle trend prediction, including short-cycle, medium-cycle, and long-cycle predictions. The short-cycle prediction refers to the next 10... min The medium-term cycle is the next 30 years. min The long cycle refers to the remaining time of this hemodialysis treatment; Risk levels are categorized as low, medium, and high, with low risk defined as a blood glucose concentration of 3.

9. mmol / L ~10.0 mmol / L Medium risk is defined as a blood glucose concentration of 3.

0. mmol / L ~3.9 mmol / L Or 10.0 mmol / L ~16.7 mmol / L High risk is defined as a blood glucose concentration <3.

0. mmol / L Or >16.7 mmol / L .

6. A dual-level anti-interference online blood glucose monitoring system for hemodialysis machines, characterized in that, The system includes a blood sample diversion processing module that sequentially transmits data and interacts with commands, a multi-source signal acquisition and preprocessing module, a blood glucose detection anti-interference correction module, a blood glucose change trend prediction module, and a graded early warning and parameter linkage module. Each module independently completes its preset functions, and the overall system operates in an orderly manner through internal data links.

7. The dual-level anti-interference online blood glucose monitoring system for a hemodialysis machine according to claim 6, characterized in that, The blood sample diversion and processing module includes a bypass pipeline pre-filling unit, a micro-blood sample extraction unit, a blood sample stabilization and retention unit, a blood sample reinfusion unit, and a pipeline anti-coagulation flushing unit. The detection chamber is a cylindrical cavity with a medical heparin anti-coagulation coating sprayed on the inner wall. A molecular sieving physical anti-interference device is integrated at the inlet. The effective volume of the detection chamber is microliters.

8. The dual-level anti-interference online blood glucose monitoring system for a hemodialysis machine according to claim 6, characterized in that, The multi-source signal acquisition and preprocessing module includes a signal acquisition triggering unit, a glucose signal acquisition unit, a blood physical parameter acquisition unit, a hemodialysis machine operating condition parameter acquisition unit, and a signal preprocessing unit. Modbus - RTU The standard communication protocol interacts with the main control system of the hemodialysis machine to synchronously collect real-time operating parameters of the hemodialysis machine.

9. The dual-level anti-interference online blood glucose monitoring system for a hemodialysis machine according to claim 6, characterized in that, The blood glucose detection anti-interference correction module includes a physical anti-interference effect verification unit, a blood parameter correction unit, an operating condition interference compensation unit, a bloodless self-calibration unit, and a blood glucose concentration calculation and verification unit. It eliminates detection deviations through a two-level anti-interference mechanism, controlling the blood glucose detection error within ±4.5%, which meets the requirements. ISO 15197 Clinical Standards.

10. The dual-level anti-interference online blood glucose monitoring system for a hemodialysis machine according to claim 6, characterized in that, The graded early warning and parameter linkage module includes a data synchronization and transmission unit, a graded early warning triggering unit, a treatment parameter adjustment unit, and an adjustment effect tracking unit, which... RS The system uses dual 485+ Ethernet communication links to transmit data with the hemodialysis machine's main control system and the medical workstation. The treatment parameter adjustment unit primarily adjusts the dialysate glucose concentration and ultrafiltration rate, with the dialysate glucose concentration adjusted between 0 and 5.5%. mmol / L Within the clinical safety range.

Images