Noninvasive cardiac output function detection method and device

Abstract

The invention relates to a non-invasive cardiac output function detection method and device, and belongs to the technical field of biological measurement. According to the method, on the basis of pulse wave related data, electrocardiowave related data and biological impedance and biological reactance related data which are obtained through detection, the bioelectrical impedance method is adopted to calculate and obtain the cardiac output of the bioelectrical impedance method, and the pulse wave method is adopted to calculate and obtain the cardiac output of the pulse wave method. And according to the impedance signal confidence coefficient and the pulse wave signal confidence coefficient, carrying out weighted average on the cardiac output of the bioelectrical impedance method and the cardiac output of the pulse wave method, and taking the weighted average as a cardiac output detection result. The detection device comprises a pulse wave detection system, an electrocardiogram detection system, a biological impedance and reactance detection system and a host used for data processing and result display. The reliability of cardiac output detection is remarkably improved while portability is guaranteed, and accurate, continuous and portable detection of non-invasive cardiac output is achieved.

Description

Technical Field

[0001] The present invention relates to a non-invasive cardiac output function detection method and device, belonging to the field of biological measurement technology. Background Art

[0002] Cardiac output is the total volume of blood ejected into the artery by a single ventricle per minute, denoted by CO (Cardiac Output). The volume of blood ejected by a single ventricle during each systolic phase is the stroke volume, denoted by SV (Stroke Volume). Cardiac output can indicate the heart function status of the human body. Related hemodynamic parameters show the peripheral circulation function, can give early warnings when problems occur in the circulatory and metabolic functions, can comprehensively evaluate the heart function and circulatory function of the human body, diagnose the heart function status of patients in all aspects, guide doctors in clinical medication for patients, evaluate the treatment effect of patients, and improve the monitoring and diagnosis and treatment levels. Therefore, detecting hemodynamic parameters such as cardiac output and providing an intuitive parameter basis for the diagnosis and treatment of clinical cardiovascular patients has important clinical significance.

[0003] Current detection methods for cardiac output mainly include invasive methods, minimally invasive methods, and non-invasive methods. Among them, the typical detection method for invasive cardiac output, which is regarded as the gold standard for output measurement, is the thermodilution method. Thermodilution (TD) is carried out through a pulmonary artery catheter (PAC). A large dose of ice-cold liquid (usually normal saline) is injected through the proximal hole (opening of the right atrium), and the thermosensitive resistor at the tip of the PAC located in the center of the pulmonary artery records the generated thermodilution curve, and then the cardiac output (CO) is calculated according to the Stewart-Hamilton equation method. However, due to the invasiveness of this method, high cost, and the need for professional operation of complex catheters, continuous measurement cannot be achieved, and the risk of complications (such as pulmonary artery rupture, arrhythmia, etc.) is high, so its clinical application has great limitations.

[0004] Typical minimally invasive cardiac output detection methods are transesophageal echocardiography Doppler monitoring (ODM) and transpulmonary thermodilution technology (TPTD). The ODM technology is to place an ultrasonic Doppler probe in the middle of the esophagus, and the tip of the probe points backward to the aorta to obtain the Doppler velocity-time waveform of the descending artery blood flow, and then measure and obtain hemodynamic parameters; however, the ODM detection method also has great limitations. For example, the measurement of cardiac output in patients with aortic lesions or severe arterial blood pressure changes is not accurate enough. Operations such as electrocautery during surgery or other factors that cause blood flow changes will affect the measurement accuracy, and the measurement will cause nausea and retching discomfort to conscious patients.

[0005] Transpulmonary single indicator dilution method (TPTD) is based on the Stewart-Hamilton equation method to indirectly measure CO. The PiCCO system is the first commercial device based on transpulmonary thermodilution method. Due to the less invasive nature of TPTD technology, it has been developed into the gold standard for verifying other minimally invasive and unvalidated CO monitoring technologies. The limitation of this method is that it cannot achieve continuous monitoring of cardiac output, and recalibration is required after changes in preload, afterload, and myocardial contractility.

[0006] Typical detection methods for non-invasive cardiac output are bioelectrical impedance method (mainly thoracic impedance method) and pulse wave method. The thoracic impedance method obtains the differential value of impedance change through amplitude modulation (AM) or frequency modulation (FM), analyzes the ejection time (LVET) by combining with the electrocardiogram waveform, and obtains the cardiac output through the Kubicek formula. This method is safe, non-invasive, simple to operate, and low-cost, so it has been favored by the medical industry in recent years. However, it has poor anti-motion interference, is easily affected by factors such as respiration, and is affected by the position of electrode patches, resulting in low measurement accuracy, which greatly limits its clinical application.

[0007] The pulse wave method obtains the waveform characteristic quantity K by calculating the change in the area under the pulse wave pressure curve, and combines the measured blood pressure value to obtain the cardiac output. This method has high accuracy, real-time dynamic detection, convenient operation, and diverse applicable scenarios, so it conforms to the current trend of mobile and home medical care. However, its disadvantages are that it relies on individual calibration, is greatly affected by vascular conditions, and is easily interfered by local blood flow. Summary of the Invention

[0008] The object of the present invention is to achieve non-invasive, accurate, and continuous monitoring of cardiac output.

[0009] The technical solution of the present invention is: a non-invasive cardiac output function detection method, based on the detected pulse wave-related data, electrocardiogram-related data, and bioelectrical impedance and bioelectrical reactance-related data, calculates the cardiac output by bioelectrical impedance method (the cardiac output calculated by bioelectrical impedance method), calculates the cardiac output by pulse wave method (the cardiac output calculated by pulse wave method), and performs weighted averaging on the cardiac output by bioelectrical impedance method and the cardiac output by pulse wave method according to the impedance signal confidence level (the confidence level involved in the bioelectrical impedance method) and the pulse wave signal confidence level (the confidence level involved in the pulse wave method), and uses this as the cardiac output detection result (the final cardiac output).

[0010] Under the background of the existing technology, the bioelectrical impedance method can adopt the thoracic impedance method, and the pulse wave method can adopt the brachial artery pulse wave method.

[0011] Further, the weight coefficient of the cardiac output measured by bioelectrical impedance method is: confidence level of impedance signal (bioelectrical impedance signal) / (confidence level of impedance signal + confidence level of pulse wave signal); the weight coefficient of the cardiac output measured by pulse wave method is: confidence level of pulse wave signal / (confidence level of impedance signal + confidence level of pulse wave signal).

[0012] The weighted average formula can be expressed as: CO final = w TEB ×CO TEB + w brachial ×CO brachial , where, CO final is the detection result of cardiac output, CO TEB and CO brachial are the cardiac output measured by bioelectrical impedance method and the cardiac output measured by pulse wave method respectively, w TEB and w brachial are the weighted coefficients of the cardiac output measured by bioelectrical impedance method and the cardiac output measured by pulse wave method respectively, where, w TEB =C TEB / ( C TEB + C brachial ), w brachial = C brachial / ( C TEB + C brachial ), C TEB and C brachial are the confidence levels of impedance signal and pulse wave signal respectively.

[0013] Preferably, during continuous monitoring, the weight coefficients of the cardiac output measured by bioelectrical impedance method and the weight coefficients of the cardiac output measured by pulse wave method are dynamically adjusted according to the real-time confidence levels of impedance signal and pulse wave signal.

[0014] Preferably, when any confidence level (confidence level of impedance signal or confidence level of pulse wave signal) is lower than the set threshold (lower limit, for example, 0.3), the cardiac output calculated based on the other method is used as the detection result of cardiac output, and an alarm is given for the corresponding confidence level lower than the threshold.

[0015] Preferably, the following formula (impedance signal confidence level model) is used to calculate the confidence level of impedance signal: C TEB =a×C resp +b×C period +c×C baseline , where, C TEB is the confidence level of impedance signal, C resp 、C period and Cbaseline They are the respiration interference suppression ratio, the periodicity of the impedance waveform, and the baseline stability of the impedance waveform, respectively. a, b, and c are the corresponding coefficients.

[0016] Furthermore, the respiration interference suppression ratio, the periodicity of the impedance waveform, and the baseline stability are calculated respectively using the following formulas: C resp = 1 - energy in the respiration frequency band / total energy; C period = 1 / (1 + standard deviation of the intervals between adjacent wave peaks); C baseline = 1 / average value of the slopes of the baseline.

[0017] Preferably, the following formula (pulse wave signal confidence model) is used to calculate the confidence of the pulse wave signal: C brachial = d × C SNR + e × C morph + f × C cuff + g × C PWV , where C brachial is the confidence of the pulse wave signal, C SNR , C morph , C cuff and C PWV are the proportion of the energy of the effective signal, the waveform shape stability, the cuff pressure stability, and the vascular quality rationality respectively, and d, e, f, and g are the corresponding coefficients.

[0018] Furthermore, the proportion of the energy of the effective signal, the waveform shape stability, the cuff pressure stability, and the vascular quality rationality are calculated respectively using the following formulas: C SNR = energy in the heart rate frequency band / total energy; C morph =(R1 + R2 + … + RN) / N, where R1, R2, …, RN represent the correlation coefficient R of the 1st, 2nd, …, Nth adjacent waveforms respectively, and N is the number of the correlation coefficients of the adjacent waveforms; , , where AI = height of the reflected wave / central arterial pulse pressure.

[0019] Non-invasive cardiac output function detection device, including a pulse wave detection system (or signal acquisition device), an electrocardiogram detection system, a bioimpedance and reactance detection system, and a host (data processing and display system). The pulse wave detection system, the electrocardiogram detection system, and the bioimpedance and reactance detection system are used to perform synchronous detections, respectively obtaining corresponding pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bio-reactance-related data. The host receives the pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bio-reactance-related data from the pulse wave detection system, the electrocardiogram detection system, and the bioimpedance and reactance detection system respectively, and uses any non-invasive cardiac output function detection method disclosed in the present invention to process the pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bio-reactance-related data, generating and outputting the cardiac output detection result.

[0020] Preferably, the host is provided with a human-computer interaction device and a storage device. The human-computer interaction device is used for human-computer interaction, and is provided with a display screen for displaying the cardiac output detection result (which can also be used for displaying other data). The storage device is used for data storage and model storage.

[0021] The beneficial effects of the present invention are as follows: The present invention combines the cardiac output detected by two methods, namely the bioelectrical impedance method and the pulse wave method, avoiding the deviation caused by the influence of respective interference factors in a single detection method. Moreover, a confidence calculation model for the bioelectrical impedance method is constructed by using the respiratory interference suppression ratio, impedance waveform periodicity, and baseline stability. In view of the characteristics of the interference factors involved in the two detection methods, a confidence calculation model for the pulse wave method is constructed by using the energy ratio of the effective signal, waveform shape stability, cuff pressure stability, and vascular quality rationality. The confidence levels of the bioelectrical impedance method and the pulse wave method are reasonably set, and through dynamic weighting of each confidence level, the accuracy of the detection result is further improved.

[0022] While ensuring portability, the present invention significantly improves the reliability of cardiac output detection, achieving accurate, continuous, and portable detection of non-invasive cardiac output. Description of the Drawings

[0023] Figure 1 is an example of the detection device of the present invention and its usage method; Figure 2 is an example of the detection data processing flow of the present invention. Detailed Embodiments

[0024] See Figure 1 and Figure 2, the present invention proposes a non-invasive cardiac output measurement system that combines arterial pulse waves, electrocardiogram waveforms, and bioimpedance and bioelectrical reactance (ICG). It simultaneously collects arterial pulse wave signals, electrocardiogram signals, and impedance signals, calculates the cardiac output based on the impedance signal according to the impedance signal and through the Kubicek formula and CO = SV * HR, and calculates the cardiac output based on the pulse wave signal according to the arterial pulse wave signal and the K-value method formula and CO = SV * HR; then calculates the confidence level of the impedance signal and the confidence level of the pulse wave signal respectively, and finally dynamically adjusts the weights according to the real-time confidence level and calculates the final cardiac output based on the weights.

[0025] I. Calculating Stroke Volume Based on Bioimpedance and Bioelectrical Reactance 1) The system is built-in with two high-precision and low-drift standard resistors, with resistance values of 25Ω and 150Ω respectively (covering the range of human chest impedance measurement under an alternating current excitation current of 62.5KHz). When the device starts, it automatically measures the calibration resistors respectively and automatically adjusts the calibration coefficient, and then measures the standard impedance again. This process continues until the measurement error of the standard resistor is within ±0.1Ω. 2) A low-pass filter is used to obtain the patient's basal impedance value. 3) The stroke volume SV1 is obtained according to the following formula: SV = ρ × (L / Z0)2 × (dZ / dt)max × LVET.

[0026] Where, SV is the stroke volume, LVET is the left ventricular ejection time, ρ is the blood resistivity (usually taking a fixed value of 135 Ω·cm), L is the distance between the two measurement electrodes, Z0 is the thoracic basal impedance, and (dZ / dt) max is the maximum value of the time derivative of △Z; 4) According to the formula: CO = SV * HR, the cardiac output is calculated (denoted as CO TEB ).

[0027] II. Calculating Stroke Volume Based on Brachial Artery Pulse Wave 1) First, use a cuff blood pressure monitor to measure the systolic pressure Ps and diastolic pressure Pd. 2) Then inflate to an appropriate pressure and non-invasively measure the pulse pressure P(t) using a pulse sensor. 3) Finally, the stroke volume SV2 is obtained according to the following formula: SV = (0.283 / K 2 ) × T × (P s - P d ), K = (P m - P d ) / (P s - P d ); Among them, P s is the systolic blood pressure, P d is the diastolic blood pressure, P(t) is the pulse pressure, T is each pulse cycle, P m is the mean arterial pressure, which is the average value of the integral of P(t) over time T.

[0028] The K value only depends on the area under the pulse wave curve. 4) According to the formula: CO = SV * HR, the cardiac output is calculated (denoted as CO brachial ).

[0029] III. Confidence algorithm A fusion system of two methods is constructed by dynamically weighting the confidence of bioelectrical impedance method and pulse wave method for non-invasive cardiac output detection. A confidence calculation model of bioelectrical impedance method is constructed by using the respiration interference suppression ratio, impedance waveform periodicity, and baseline stability; a confidence calculation model of pulse wave method is constructed by using the energy ratio of effective signals, waveform morphological stability, cuff pressure stability, and vascular quality rationality.

[0030] I. Confidence design and calculation of bioelectrical impedance method 1) Signal quality indicators of bioelectrical impedance method i) Respiration interference suppression ratio (C resp ): C resp = 1 - energy in respiration frequency band / total energy.

[0031] Calculate the energy ratio in the respiration frequency band (0.1 - 0.3 Hz). The higher the ratio, the greater the signal interference from respiration.

[0032] ii) Impedance waveform periodicity (C period ): C period = 1 / (1 + standard deviation of adjacent peak intervals).

[0033] Calculate the standard deviation of the intervals between two adjacent peaks within a certain sampling time t (range 0 - 100 s). The larger the standard deviation, the worse the impedance waveform periodicity.

[0034] iii) Baseline stability (C baseline ): C baseline = 1 / average value of the slope of the baseline.

[0035] Calculate the slope of the baseline change within a certain sampling time t1 (range 0 - 100 s). The larger the average value of the slope, the worse the baseline stability.

[0036] 2) Confidence calculation: C TEB = a × C resp + b × Cperiod + c × C baseline 。

[0037] a, b, and c are the corresponding coefficients, which can be determined based on experience, adjusted or screened through experiments. When necessary, it can also be determined (inverted) by combining standard detection methods and using neural networks or curve fitting with sufficient experimental data. Generally, the value range can be: 0 < a < 1, 0 < b < 1, 0 < c < 1, and a + b + c ≤ 1. For example, a = 0.4, b = 0.3, c = 0.3.

[0038] According to the above formula, 0 < C TEB < 1.

[0039] II. Confidence Design and Calculation of Pulse Wave Method 1) Pulse Wave Signal Quality Index i) Proportion of Effective Signal Energy (C SNR ): C SNR = Energy in Heart Rate Band / Total Energy.

[0040] Calculate the proportion of the energy of the effective component (such as the heart rate band) in the pulse wave signal. The higher the signal-to-noise ratio, the larger C SNR is.

[0041] ii) Waveform Morphology Stability (C morph ): C morph = (R1 + R2 + … + RN) / N.

[0042] Calculate the average value of the correlation coefficient R of adjacent waveforms within a certain sampling time t2 (ranging from 0 - 100 s). The more correlated the adjacent waveforms, the larger C morph is. Here, R1, R2, …, RN represent the correlation coefficients R of the 1st, 2nd, …, Nth adjacent waveforms respectively, and N is the number of correlation coefficients of adjacent waveforms.

[0043] iii) Cuff Pressure Stability (C cuff ): Detect the pressure fluctuation range. If the pressure fluctuation < 5 mmHg, then C cuff = 1; if 5 mmHg < pressure fluctuation < 10 mmHg, then C cuff = 0.5; if the pressure fluctuation > 10 mmHg, then C cuff = 0.

[0044] iv) Rationality of Vascular Quality (C AI ): The state of the patient's vascular quality directly affects the accuracy of the pulse wave signal. By calculating the enhancement effect (AI) of the reflected wave on the systolic blood pressure after superimposing with the forward wave, the rationality of the vascular quality is reflected. AI = Reflection Wave Height (AP) / Central Arterial Pulse Pressure (CPP). If 0.75 < AI < 0.85, then CAI = 1; if 0.85 < AI < 0.90, then C AI = 0.5; if AI > 0.90, then C AI = 0.

[0045] 2) Confidence calculation: C brachial = d × C SNR + e × C morph + f × C cuff + g × C PWV .

[0046] d, e, f, and g are the corresponding coefficients, which can be determined based on experience and adjusted or screened through experiments. When necessary, they can also be determined (inverted) by combining standard detection methods and using neural networks or curve fitting with sufficient experimental data. Usually, the value ranges can be: 0 < d < 1, 0 < e < 1, 0 < f < 1, 0 < g < 1, and d + e + f + g ≤ 1). For example: d = 0.3, e = 0.3, f = 0.2, g = 0.2.

[0047] According to the above formula, 0 < C brachial < 1.

[0048] IV. Dynamic weighting algorithm 1) Weight assignment Dynamically adjust the weight according to the real-time confidence: w TEB = C TEB / ( C TEB + C brachial ), w brachial = C brachial / ( C TEB + C brachial ).

[0049] 2) Calculation of the final cardiac output (CO final ): CO final = w TEB × CO TEB + w brachial × CO brachial .

[0050] When the confidence of a certain method is lower than the threshold (C < 0.3), the result of only another method can be used, and a warning is issued to prompt the user to re-measure.

[0051] Existing detection devices using thoracic impedance method and pulse wave method can be adopted for detection. With the support of software, the detection data are analyzed and processed. The corresponding cardiac output is calculated respectively according to the thoracic impedance method and the pulse wave method. The confidence levels of the two methods are calculated according to the confidence models (confidence calculation methods) of the two detection methods. The detection results (cardiac output) of the two methods are weighted and averaged using corresponding formulas according to the confidence levels, and this is used as the final cardiac output (cardiac output detection result). During continuous monitoring, dynamic weighting is implemented according to the real-time confidence level.

[0052] Each of the preferred and optional technical means disclosed in the present invention can be arbitrarily combined to form several different specific embodiments, unless otherwise specified or one preferred or optional technical means is a further limitation of another technical means.

Claims

1. A non-invasive cardiac output function detection method, based on the detected pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bioelectrical reactance-related data, calculates the cardiac output by bioelectrical impedance method to obtain the cardiac output by bioelectrical impedance method, calculates the cardiac output by pulse wave method to obtain the cardiac output by pulse wave method, and weights and averages the cardiac output by bioelectrical impedance method and the cardiac output by pulse wave method according to the impedance signal confidence level and the pulse wave signal confidence level, and uses this as the cardiac output detection result.

2. The non-invasive cardiac output function detection method according to claim 1, wherein The weight coefficient of the cardiac output by bioelectrical impedance method is: impedance signal confidence level / (sum of impedance signal confidence level and pulse wave signal confidence level); the weight coefficient of the cardiac output by pulse wave method is: pulse wave signal confidence level / (sum of impedance signal confidence level and pulse wave signal confidence level).

3. The non-invasive cardiac output function detection method according to claim 2, wherein During continuous monitoring, the weight coefficients of the cardiac output by bioelectrical impedance method and the weight coefficient of the cardiac output by pulse wave method are dynamically adjusted according to the real-time impedance signal confidence level and pulse wave signal confidence level.

4. The non-invasive cardiac output function detection method according to claim 1, characterized in that When any confidence level is lower than the set threshold, the cardiac output obtained by another method is used as the cardiac output detection result, and an alarm for the confidence level lower than the threshold is given accordingly.

5. The non-invasive cardiac output function detection method according to any one of claims 1-4, characterized in that The impedance signal confidence level is calculated by the following formula: C TEB = a × C resp + b × C period + c × C baseline , Among them, C TEB is the impedance signal confidence level, and C resp , C period and C baseline are the respiration interference suppression ratio, the impedance waveform periodicity, and the baseline stability of the impedance waveform respectively, and a, b, and c are the corresponding coefficients.

6. The non-invasive cardiac output function detection method according to claim 5, characterized in that The respiration interference suppression ratio, impedance waveform periodicity, and baseline stability are calculated respectively by the following formulas: C resp = 1 - Energy in the respiratory frequency band / Total energy; C period = 1 / (1 + standard deviation of adjacent peak intervals); C baseline = 1 / average slope of the baseline.

7. The non-invasive cardiac output function detection method according to any one of claims 1-4, characterized in that The pulse wave signal confidence level is calculated by the following formula: C brachial = d × C SNR + e × C morph + f × C cuff + g × C PWV , Among them, C brachial is the confidence level of the pulse wave signal, C SNR , C morph , C cuff and C PWV are the proportion of effective signal energy, waveform morphology stability, cuff pressure stability, and vascular quality rationality respectively, and d, e, f, and g are the corresponding coefficients.

8. The non-invasive cardiac output function detection method according to claim 7, wherein The effective signal energy ratio, waveform shape stability, cuff pressure stability, and vascular quality rationality are calculated respectively by the following formulas: C SNR = Heart rate band energy / Total energy; C morph =(R1 + R2 + … + RN) / N, where R1, R2, …, RN respectively represent the correlation coefficients R of the 1st, 2nd, …, Nth adjacent waveforms, and N is the number of correlation coefficients of the adjacent waveforms; ， ， Wherein, AI = reflected wave height / central arterial pulse pressure.

9. Non-invasive cardiac output function detection device, characterized in that It includes a pulse wave detection system, an electrocardiogram detection system, a bioimpedance and bioelectrical reactance detection system, and a host computer. The pulse wave detection system, the electrocardiogram detection system, and the bioimpedance and bioelectrical reactance detection system are used to synchronously perform detections, and respectively obtain the corresponding pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bioelectrical reactance-related data. The host computer receives the pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bioelectrical reactance-related data respectively from the pulse wave detection system, the electrocardiogram detection system, and the bioimpedance and bioelectrical reactance detection system, and processes the pulse wave-related data, electrocardiogram wave-related data, and bioimpedance and bioelectrical reactance-related data by using the non-invasive cardiac output function detection method described in any one of claims 1-8, and generates and outputs the cardiac output detection result.

10. The non-invasive cardiac output function detection device according to claim 9, characterized in that The host computer is provided with a man-machine interaction device and a storage device. The man-machine interaction device is used for man-machine interaction, and is provided with a display screen for displaying the cardiac output detection result. The storage device is used for data storage and model storage.

Images