DEVICE, SYSTEM, AND METHOD FOR CALIBRATING A BLOOD PRESSURE SURROGATE FOR USE IN MONITORING A SUBJECT'S BLOOD PRESSURE - Patent application

Abstract

The present invention relates to a device, system and method for calibrating a blood pressure surrogate for use in monitoring the blood pressure of a subject. The device comprises a BP input 51, a sensor input 52 and a processing unit 53, the processing unit being configured to calculate a pulse-related value from a first and a second time-dependent sensor signal using a first feature of a first time-dependent sensor signal and a second feature of a second time-dependent sensor signal during cuff inflation, the pulse-related value being a value of a Pulse Arrival Time PAT or a value of a Pulse Transit Time PTT, to select pairs of pulse-related values ​​and corresponding cuff pressure values ​​to calculate calibration parameters of a BP surrogate, the pairs comprising a pulse-related value and a corresponding time-related cuff pressure value, only those pairs of pulse-related values ​​and corresponding cuff pressure values ​​for which one or more predetermined conditions regarding BP and / or cuff pressure are met are selected, and to calculate the calibration parameters of the BP surrogate from the BP measurements and the selected pairs of pulse-related values ​​and corresponding cuff pressure values.

Description

[Technical field]

[0001] The present invention relates to devices, systems and methods for calibrating blood pressure (BP) surrogates for use in monitoring the blood pressure of a subject. [Background technology]

[0002] Hemodynamic measurements in acute care generally tend to be more continuous, less invasive and less intrusive. Blood pressure (BP) is the fundamental hemodynamic parameter used in all hospital settings to assess the health status of the patient. Arterial blood pressure (ABP) is an important physiological parameter involved in medical diagnosis, prevention as well as treatment guidance. Invasive measurements - the gold standard - allow continuous measurements with the highest accuracy, but can only be performed by trained medical personnel and are mostly applied in acute settings where real-time alarms and very close monitoring are required.

[0003] The established way to measure blood pressure non-invasively is with an upper arm cuff (NIBP = non-invasive blood pressure). Although practical, NIBP measurements are only intermittent and therefore must be repeated (usually automatically) for monitoring applications, e.g. every few minutes during surgery or typically every 15 minutes in the ICU. Most NIBP measurements in patient monitoring are performed with standard cuff-based ABP measurements using automated oscillometry that allows intermittent measurements. There is an unmet need for continuous and at the same time highly accurate non-invasive blood pressure measurement technology. The availability of continuous non-invasive blood pressure information between blood pressure measurement intervals is a key differentiator in order to avoid missing hypotensive (or hypertensive) episodes that are associated with adverse patient outcomes.

[0004] An elegant way to achieve this is to utilize continuous physiological signals already available in acute care settings and infer continuous blood pressure information from them (a surrogate parameter for continuous BP), thus minimizing additional costs and leaving clinical work flow unaffected.

[0005] For continuous NIBP measurement, some dedicated devices exist, e.g. finger cuff devices, but these have not yet been widely adopted in hospitals due to various issues related to lack of accuracy and robustness, but also due to the additional costs and negative impact on clinical workflow.

[0006] The approach of continuous non-invasive blood pressure measurement is based on BP surrogates, which are well-known parameters such as pulse transit time (PTT) and pulse arrival time (PAT), pulse wave velocity, or a combination of these. These parameters can be used to estimate blood pressure non-invasively and continuously without using external pressure other than for reference or initialization purposes, often called calibration. Calibration is a non-trivial issue for practical applications in clinical and home environments.

[0007] US Patent Application Publication No. 2010 / 160798A1 discloses a technique for continuous measurement of BP based on PTT and without the need for any external calibration. This technique is implemented using a body-worn monitor that measures BP and other vital signs and transmits them wirelessly to a remote monitor. A network of body-worn sensors, typically placed on the patient's right arm and chest, is connected to the monitor to measure time-dependent ECG, PPG, accelerometer, and pressure waveforms. The sensors may include a cuff featuring an inflatable bladder coupled to a pressure sensor, three or more electrical sensors (e.g., electrodes), three or more accelerometers, a temperature sensor, and an optical sensor (e.g., a light source and a photodiode) attached to the patient's thumb. During a controlled reduction in transmural pressure under the cuff, a dedicated PTT change is obtained. Summary of the Invention [Problem to be solved by the invention]

[0008] It is an object of the present invention to provide devices, systems and methods that can further improve the accuracy of continuous, non-invasive BP measurements using BP surrogates. [Means for solving the problem]

[0009] In one aspect of the invention, there is provided a device for calibrating a BP surrogate for use in monitoring blood pressure in a subject, the device comprising: a BP input configured to obtain a value of a time-dependent cuff pressure during inflation of a cuff of a pressure delivery system attached to a body part of a subject and to obtain a BP measurement; a sensor input configured to obtain a first time-dependent sensor signal and a second time-dependent sensor signal related to the subject's heart rate, the first time-dependent sensor signal and the second time-dependent sensor signal being measured at different locations on the subject's body during inflation of the cuff; Processing unit and The processing unit comprises: - calculating, during inflation of the cuff, a value related to a pulse from the first and second time dependent sensor signals using a first feature of the first time dependent sensor signal and a second feature of the second time dependent sensor signal, the value related to a pulse being a value of a pulse arrival time PAT or a value of a pulse transit time PTT; - selecting pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating a calibration parameter of a surrogate of BP, the pairs including a pulse-related value and a corresponding time-related cuff pressure value, and only those pairs of pulse-related values ​​and corresponding cuff pressure values ​​for which one or more predetermined conditions regarding BP and / or cuff pressure are satisfied are selected; - Calculating BP surrogate calibration parameters from the BP measurements and selected pairs of pulse-related values ​​and corresponding cuff pressure values. It is configured as follows.

[0010] In a further aspect of the present invention, there is provided a system for calibrating a BP surrogate for use in monitoring blood pressure of a subject, the system comprising: a pressure delivery system comprising a cuff configured to be attached to a body part of a subject and to deliver pressure to the body part of the subject by inflating the cuff; a pressure sensor configured to obtain a value of a time-dependent cuff pressure during inflation of a cuff of a pressure delivery system attached to a body part of a subject, and to obtain or estimate a BP measurement; a first sensor attached to a first portion of the subject's body and configured to obtain a first time-dependent sensor signal related to the subject's heart rate during inflation of the cuff; a second sensor attached to a second portion of the subject's body and configured to obtain a second time-dependent sensor signal related to the subject's heart rate during inflation of the cuff; and the device for calculating calibration parameters for calibrating the BP surrogate from the BP measurements, the time-dependent cuff pressure values, and the first and second time-dependent sensor signals; Equipped with.

[0011] In yet another aspect of the present invention, there is provided a corresponding method and a computer program having program code means which, when the computer program is run on a computer, causes the computer to perform the steps of the methods disclosed herein, and a non-transitory computer readable recording medium having stored therein a computer program product which, when run by a processor, causes the computer to perform the methods disclosed herein.

[0012] Preferred embodiments of the invention are defined in the dependent claims. It is to be understood that the claimed methods, systems, computer programs and media have similar and / or identical preferred embodiments to the claimed devices, in particular those defined in the dependent claims and disclosed herein.

[0013] The present invention allows for a reliable and personalized calibration procedure (sometimes called a start or initialization procedure) by exploiting that cuff inflation induces a well-defined and associated change in blood pressure surrogate. Insight into the physiological effects during cuff inflation has resulted in an optimized method and procedure used to calculate one or more calibration parameters. A reliable calibrated blood pressure surrogate allows for highly accurate and non-invasive tracking of blood pressure. That is, the use of a calibrated BP surrogate allows for the conversion of an initially undetermined / undefined set of parameters into a meaningful clinical parameter blood pressure.

[0014] The present invention is based on the finding that the change in PAT (or PTT) during cuff inflation is distorted by dynamic filling effects due to venous occlusion in the lower arm (or more generally in the peripheral region of the subject's body part on which the cuff is placed). This finding has an impact on the manner in which the calibration parameter(s) are obtained. For example, parametric regression, which is one of the preferred options for calculating the calibration parameter(s), is affected, which needs to be addressed in order to estimate the sensitivity parameter. This effect has been observed and interpreted from invasively measured BP signals obtained peripheral to the cuff, e.g., the lower arm if the cuff is placed on the upper arm, or the lower leg if the cuff is placed on the upper leg.

[0015] According to the invention, one or more conditions (or criteria) are used to select pairs of pulse-related values ​​and corresponding cuff pressure values ​​that are subsequently used to calculate the calibration parameters of a BP surrogate for estimating BP. Such conditions relate to BP and / or cuff pressure, i.e., depending on BP and / or cuff pressure, it is determined whether a pair of pulse-related values ​​and corresponding cuff pressure values ​​is used for the calculation of the calibration parameters of a BP surrogate. There are several embodiments of such conditions, as will be explained in detail below.

[0016] The processing unit is configured according to one embodiment to calculate the calibration parameters of the BP surrogate using only pairs of pulse-related values ​​and corresponding cuff pressure values ​​calculated from the first and second time-dependent sensor signals measured during partial inflation of the cuff. It has been found that at high cuff pressures, due to pulse distortion, no reliable pulse signal appears at the peripheral part of the body part on which the cuff is placed. This has a negative effect on the calculation of the calibration parameters if pairs of values ​​calculated from sensor signals measured during full or almost full inflation of the cuff are used. In this embodiment, such negative effects are avoided since only pulse signals with a defined relationship to BP are taken into account, which leads to an improved calibration accuracy and therefore ultimately to an improved determination of BP by using a calibrated BP surrogate.

[0017] According to another embodiment, the processing unit is configured to calculate the calibration parameters of the BP surrogate using only pairs of pulse-related values ​​and corresponding cuff pressure values ​​for which the peripheral BP is substantially constant, in particular with a variation of less than 10% or less than 5%. Other reasonable values ​​may be used as well. As explained above, in this way a similar effect can be obtained, but different conditions are used to select the value pairs for the calculation.

[0018] The processing unit is thereby configured to determine whether the peripheral BP is substantially constant based on the first and / or second time-dependent sensor signals. For this purpose, for example, a period during which the amplitude of the second time-dependent signal is substantially constant, in particular with a fluctuation of less than 10% or less than 5% (or any other reasonable threshold, predefined or set by the user) with respect to a previously obtained average value, is determined. This previously obtained average value refers to a previous average value of the second time-dependent signal.

[0019] The processing unit is further configured to calculate the BP surrogate calibration parameters using only pairs of pulse-related values ​​and corresponding cuff pressure values, where the cuff pressure value is below the diastolic BP of the subject, in particular the most recently measured diastolic BP of the subject, below the mean BP of the subject, in particular the most recently measured mean BP of the subject, or below a set cuff pressure threshold, which allows a rather simple yet efficient implementation of the disclosed solution.

[0020] In another embodiment, the processing unit is configured to calculate the calibration parameters of the BP surrogate using only pairs of pulse-related values ​​and corresponding cuff pressure values, the cuff pressure values ​​being in a range between a set minimum cuff pressure and a set maximum cuff pressure, in particular the minimum cuff pressure being set in a range of 10 to 30 mmHg and the maximum cuff pressure being set in a range of 40 to 90 mmHg (or set in a range defined by other reasonable values), which again allows a rather simple yet efficient implementation of the disclosed solution.

[0021] In yet further embodiments, the processing unit is configured to calculate the calibration parameters of the BP surrogate using only those pairs of pulse-related values ​​and corresponding cuff pressure values ​​for which the pulse-related value is substantially constant, in particular fluctuates by less than 10% (or any other reasonable value), or until the slope of the curve of the pulse-related value over time exceeds a threshold value. Such embodiments utilize other conditions that are easily implemented.

[0022] The processing unit is further configured to calculate a control signal for controlling the pressure delivery system to inflate a cuff of the pressure delivery system, in particular to fully inflate the cuff to obtain the BP measurement and to only partially inflate the cuff to obtain the first time-dependent sensor signal and the second time-dependent sensor signal. The device can thereby actively control the inflation of the cuff as required for the BP measurement or for obtaining the sensor signals used to determine the calibration parameters. The BP measurements used for calibration (i.e. for calculating the calibration parameters) are obtained through BP measurements by the BP measuring device.

[0023] In a preferred embodiment, the first time-dependent sensor signal is an ECG signal and / or the second time-dependent sensor signal is a photoplethysmography PPG signal, in particular a contact PPG signal or a remotely acquired PPG signal (e.g. acquired by a camera used in remote PPG). Other signals related to the subject's heart rate and allowing the calculation of PAT and / or PTT can additionally or alternatively be used. For example, instead of a PPG signal, a bioimpedance or cuff signal is used. ECG and PPG signals are commonly known signals, and PAT and / or PTT can be calculated from the ECG and PPG signals in a commonly known manner that can also be used in the device and method of the present invention.

[0024] The processing unit is further preferably configured in one embodiment to calculate the calibration parameters using regression. The calibration parameters are calculated, for example, as the slope of the dependence of the relationship of blood pressure, which is a function of the blood pressure surrogate at zero cuff pressure. The regression derives a parameterized functional relationship between the transmural blood pressure and the BP surrogate, which serves as a transfer function for estimating the blood pressure from the BP surrogate when the cuff is not inflated (zero cuff pressure). Further details are described, for example, in "Surrogate based continuous noninvasive blood pressure measurement", Pielmus, AG, Muhlsteff, J., Bresch, E., Glos, M., Jungen, C., Mieke, S., Zaunseder, S., Biomedical Engineering / Biomedizinische Technik, 66(3), pp. 231-245, 2021.

[0025] The processing unit is still further configured to use the calculated BP surrogate and the first and second time-dependent sensor signals measured at different locations on the subject's body when no pressure is delivered to the subject's body part to determine a value of BP when no pressure is delivered to the subject's body part by the pressure delivery system. This allows an otherwise unobtrusive determination of the subject's BP without inflating a cuff, i.e., the subject's BP can be continuously monitored without inflating a cuff to bother or injure the patient.

[0026] The claimed system further comprises a control unit, which may be part of the device or may be external, configured to control the pressure delivery system, for example based on control signals provided by the device or calculated by the system.

[0027] The present invention is also used in combination with one or more of the following embodiments.

[0028] The processing unit, in one embodiment, comprises: - calculating a control signal to control the pressure delivery system to repeatedly inflate a cuff of the pressure delivery system to a cuff pressure that is partially inflated and is less than the systolic BP of the subject; - calculating a pulse-related value from the first and second time dependent sensor signals measured during repeated partial inflations of the cuff using a first feature of the first time dependent sensor signal and a second feature of the second time dependent sensor signal, the pulse-related value being a value of a pulse arrival time PAT or a value of a pulse transit time PTT; - configured to calculate BP surrogate calibration parameters from the BP measurements and pairs of pulse-related values ​​and corresponding cuff pressure values ​​calculated from first and second time-dependent sensor signals measured during repeated partial inflations of the cuff, the pairs including a pulse-related value and a corresponding time-related cuff pressure value.

[0029] According to this embodiment, only low cuff pressure levels are used for calibration, i.e., the cuff is only partially inflated for data acquisition. Repeated sampling at such appropriate cuff pressure levels can be performed in a short enough period so that the BP level is stable and there are no changes that would distort the calibration process. Furthermore, repeated inflation ramps (with or without pauses between ramps) are possible due to the low maximum pressure. There is no need to wait for tissues and arteries to recover from high pressurization (i.e., there is no hysteresis effect).

[0030] Lower cuff pressure levels, especially those below the diastolic pressure, are less intrusive to the patient, especially an awake patient, and result in less discomfort and stress on the skin than a ramp of full inflation to above the systolic level. Another advantage is that it places no extra requirements on the pressure delivery system of the NIBP device.

[0031] The BP measurements used for calibration (ie, to calculate the calibration parameters) are obtained through BP measurements with a BP measuring device.

[0032] There are various embodiments of the method of controlling the pressure delivery system during repeated partial inflations, all of which attempt to further avoid the adverse dynamic filling effects mentioned above, thus contributing to improved accuracy of the calibration, and thus ultimately improved BP determination, by using a calibrated BP surrogate.

[0033] The processing unit is, in one embodiment, configured to calculate a control signal for controlling the pressure delivery system to inflate the cuff of the pressure delivery system at partial inflation during periods when the peripheral BP is substantially constant, in particular fluctuating less than 10 percent or less than 5 percent. Other reasonable values ​​may be used as well.

[0034] The processing unit is thereby configured to determine whether the peripheral BP is substantially constant based on the first and / or second time-dependent sensor signal, in particular by determining a time period during which the amplitude of the second time-dependent signal is substantially constant, in particular fluctuating less than 10 percent or less than 5 percent (or any other reasonable threshold, predefined or user set) with respect to a previously obtained average value, where the previously obtained average value refers to a previous average value of the second time-dependent signal.

[0035] The processing unit is according to another embodiment configured to calculate a control signal for controlling the pressure delivery system to inflate, with partial inflation, the cuff of the pressure delivery system to the diastolic BP of the subject, in particular the most recently measured diastolic BP of the subject, to the mean BP of the subject, in particular the most recently measured mean BP of the subject, or to a set threshold BP, thereby allowing a rather simple yet efficient implementation of the claimed and disclosed solution.

[0036] The processing unit is configured to calculate a control signal for controlling the pressure delivery system to inflate the cuff of the pressure delivery system at partial inflation to a cuff pressure at which the PAT exceeds a PAT threshold, in particular an absolute PAT threshold, or a relative PAT threshold compared to the baseline PAT. The processing unit is thereby configured to calculate a baseline PAT by averaging PAT measurements over a period of time, in particular a period in the range of 10 seconds to 5 minutes (other reasonable values ​​are possible as well). A reasonable threshold for the relative change, i.e. PAT compared to the baseline PAT, is for example in the range of 10 to 60 milliseconds, e.g. 30 milliseconds.

[0037] In another embodiment, the processing unit is configured to calculate a control signal for controlling the pressure delivery system to inflate the cuff of the pressure delivery system with partial inflation to a cuff pressure where the amplitude of the second time-dependent signal exceeds an amplitude threshold, in particular an absolute amplitude threshold or a relative amplitude threshold. Such a criterion can be easily set by the user or can be predefined. A reasonable threshold for the relative change is, for example, within the range of 5 to 20%, e.g. 10%.

[0038] The processing unit is configured to calculate a control signal for controlling the pressure delivery system to inflate the cuff of the pressure delivery system with partial inflation at a predetermined inflation rate or at an inflation rate that varies depending on one or more of the subject's heart rate, the subject's diastolic BP, the subject's mean BP, and the subject's systolic BP.

[0039] The processing unit may, in a practical embodiment, - calculating a control signal for controlling the pressure delivery system to deflate a cuff of the pressure delivery system after each repetition of partial inflation; - i) for each iteration, calculating, using a respective regression, one or more calibration parameters from pairs of pulse-related values ​​and corresponding cuff pressure values ​​obtained in each iteration, and averaging each calibration parameter calculated over two or more iterations to obtain one or more average calibration parameters for use by the BP surrogate; or - ii) calculating, in a single regression, one or more calibration parameters from pairs of pulse-related values ​​and corresponding cuff pressure values ​​obtained in two or more iterations; It is configured as follows.

[0040] Both options i) and ii) may lead to improved accuracy depending on the circumstances and application of the disclosed device and method. Note that the applied cuff pressure causes a change in the transmural pressure below the cuff.

[0041] The processing unit is, in another embodiment, configured to calculate a control signal for controlling the pressure delivery system to inflate a cuff of the pressure delivery system at full inflation above the subject's systolic BP before and / or after one or more partial inflation repetitions to obtain a time-dependent BP reference measurement for use in monitoring the subject's BP using a BP surrogate. The last measured BP value is used, among other things, to calculate a calibration parameter.

[0042] The processing unit is further configured to control the number of partial inflation iterations to be performed based on a comparison of the regression error to a regression error threshold. The processing unit is further configured to control whether and when to perform a full inflation according to a fixed or variable schedule, or if one or more calibration parameters have changed significantly since the last calculation, particularly by more than 10 percent (or more than 15 or 20 percent). Thus, a trade-off is made between additional measurement time / effort and improved accuracy of the calibration.

[0043] The processing unit is still further configured to determine a value of BP when no pressure is delivered to the subject's body part by the pressure delivery system through the use of the calculated BP surrogate and the first and second time-dependent sensor signals measured at various locations of the subject's body when no pressure is delivered to the subject's body part. This allows an otherwise unobtrusive determination of the subject's BP without inflating a cuff, i.e., the subject's BP can be continuously monitored without inflating a cuff to annoy or injure the patient.

[0044] In a preferred embodiment, the first time-dependent sensor signal is an ECG signal and / or the second time-dependent sensor signal is a photoplethysmography PPG signal, in particular a contact PPG signal or a remote PPG signal. Other signals related to the subject's heart rate and allowing the calculation of the PAT and / or PTT can additionally or alternatively be used. ECG and PPG signals are commonly known signals, and the PAT and / or PTT can be calculated from the ECG and PPG signals in commonly known manners that can also be used in the disclosed devices and methods.

[0045] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. [Brief description of the drawings]

[0046] [Figure 1] 1 is a schematic diagram of one embodiment of the overall concept of the present invention; [Diagram 2] FIG. 1 is a diagram of an ECG signal and a PPG signal with PAT indicated. [Diagram 3] FIG. 1 illustrates robust versus non-robust regression for linear models. [Figure 4] FIG. 13 is an example of an increase in PAT caused during cuff inflation. [Diagram 5] FIG. 1 illustrates the overall concept of calibration. [Figure 6] FIG. 1 is a schematic diagram of the setup used in the general concept for acquiring an invasively measured BP signal acquired peripherally relative to the cuff. [Figure 7] FIG. 6 is a diagram of a signal measured using the setup shown in FIG. 5. [Figure 8] FIG. 13 shows an analysis of the inflation process, which affects blood pressure measured peripherally behind the cuff. [Figure 9] FIG. 13. Transmural blood pressure versus measured PAT. [Figure 10] FIG. 1 is a diagram of one embodiment of a system according to the present invention. [Figure 11] FIG. 1 is a diagram of one embodiment of a device according to the invention. [Figure 12] FIG. 2 is a diagram of an embodiment of the method according to the present invention. [Figure 13] FIG. 13 is a plot of PAT values ​​versus Pcuff values ​​during the entire cuff inflation process. [Figure 14] FIG. 13 illustrates another embodiment of a method of using the present invention. [Figure 15] FIG. 13 shows that only the Pcuff / PAT value pair is used unless the PPG signal features indicate pulse pressure changes in the blood pressure of the distal arm. [Figure 16] FIG. 13 illustrates another embodiment of a method of using the present invention. [Figure 17] FIG. 13 shows an example of fitting a parameterized model to measurement data pairs obtained during a full expansion process and during partial expansion. [Figure 18] FIG. 4 is a diagram of another embodiment of the method according to the invention. [Figure 19] FIG. 1 illustrates one embodiment of a cuff pressure profile that includes two partial inflations followed by a full inflation. [Figure 20] FIG. 1 illustrates one embodiment of a cuff pressure profile including a full inflation followed by two partial inflations. [Figure 21] FIG. 13 illustrates one embodiment of a cuff pressure profile that provides continuous tracking of the need for recalibration through repeated partial inflations. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0047] The present invention relates to the development of ABP measurement techniques that allow accurate and continuous blood pressure monitoring. BP tracking during intervals of standard cuff-based, non-invasive, intermittent blood pressure measurements is based on recalibrated / initialized blood pressure surrogates. Calibration is performed using the NIBP measurement itself (systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean blood pressure (MBP)) and the inflation / deflation process of the NIBP measurement, which induces controlled BP-related disturbances in the surrogate signal. The surrogate is obtained from, or using a combination of features of, a signal related to the subject's heart rate, such as an electrocardiogram (ECG) or photoplethysmogram (PPG). These signals are routinely acquired at such sites.

[0048] Since the sensor embodiment is small, lightweight, and provides a comfortable, nearly continuous, non-invasive measurement of ABP, the present invention focuses on a method of making pulse wave velocity a conduit to otherwise unobtrusive ABP measurements. An embodiment of this concept is shown diagrammatically in FIG. 1. In this embodiment, ECG and PPG signals are obtained (block 100), e.g., received or retrieved from the respective sensors (which obtain these signals) or a buffer or memory (where these signals are buffered or stored), and used to continuously determine a BP surrogate (block 101). This BP surrogate can then be used as an indicator of a change in BP (block 102). From time to time, or upon indication of a significant change in BP (used as a trigger), the NIBP cuff is activated (block 103), a recalibration of the BP surrogate is performed, and a BP reading is obtained by an oscillometric technique, e.g., a recognized standard oscillometric technique. The BP determination obtained by use of the cuff inflation and by use of the BP surrogate is reported (block 104).

[0049] This concept does not require additional sensors but can use available sensors, can be implemented with software alone, and has a low barrier to clinical acceptance. Advantages include earlier detection of BP changes allowing earlier intervention, reduced complications due to reduced use of A-lines, and better tracking of BP responses leading to improved drug titration.

[0050] An embodiment of the present approach refers to the use of PAT, a BP surrogate. PAT is defined as the time interval between the peak of the R wave of the ECG and the onset of the pulse of peripheral plethysmography. PAT has been intensively studied as a surrogate measurement of blood pressure and vascular stiffness. PAT is the sum of the pre-ejection time (PEP) and the PTT. PEP refers to the time required for isovolumic ventricular contraction until the aortic valve opening (AVO), while PTT is the natural transit time of the pressure pulse over a long, non-uniform vascular path along the arterial wall. Only the PTT is related to the propagation of the arterial wave, which varies with pressure, modeled by the Moens-Korteweg equation. PEP is a variable additive delay that is sensitive to stress, emotion, and physical effort. The main advantage of PAT vs. PTT is that only one transducer (e.g., the PPG sensor), which is sensitive to location, needs to be precisely placed.

[0051] FIG. 2 shows a diagram of an ECG signal 110 and a PPG signal 111 (adapted from "An Optimization Study of Estimating Blood Pressure Models Based on Pulse Arrival Time for Continuous Monitoring", Shao, J. et al.), where the PAT is shown as the time delay between the R-peak of the ECG signal and a feature point of the peripheral PPG signal. In this example, the maximum derivative (Dmax) is selected as the feature point.

[0052] A system that uses PAT to continuously estimate BP between NIBP measurements does not require any additional devices (hardware), assuming time-synchronized PPG and ECG sensors (and / or corresponding signals) are available. This system therefore represents a software-only approach to continuous BP estimation in clinical practice.

[0053] Changes in PAT are inversely proportional to changes in BP. An increase in BP increases the transmural pressure of the arteries, leading to a decrease in the compliance of the arterial wall, resulting in a faster propagation time of the blood pulse wave and therefore a shorter time for the pulse wave to reach the periphery. Conversely, a decrease in BP increases the PAT. The exact transfer function from BP to PAT depends on many factors, including the patient's individual arterial properties and geometry, the cardiac pre-ejection period (PEP), and more. Furthermore, the transfer function is not static but changes over time as the patient's hemodynamic state, i.e., arterial properties (e.g., arterial smooth muscle or PEP), change.

[0054] However, the transfer functions of BP and PAT can be modeled by fairly simple models (e.g. proportional linear, logarithmic, reciprocal, inverse squared, etc.) with, for example, two or three parameters. For example, if we consider a simple linear model (BP=m1·PAT+m2), the two patient-specific parameters are the offset parameter m2 and the sensitivity parameter (slope) m1. In order to actually apply the transfer function to continuously estimate blood pressure from continuous PAT values, the patient-specific parameters need to be "learned". This is called calibration. Furthermore, the changes in these parameters are preferably tracked, i.e. the calibration is preferably repeated when the model parameters (i.e. the calibration parameters) are changing. This is called recalibration.

[0055] Model calibration is typically performed by parametric regression. That is, when multiple (BP, PAT) data pairs are available, a model curve can be fitted to the data points by applying some optimality criterion (such as least squares), and optimal values ​​of the model parameters are obtained based on the given data. The larger the range of BP and PAT values, the more appropriate (more accurate) the parameters will be determined.

[0056] FIG. 3 shows an illustration of robust versus non-robust regression for linear models, specifically a diagram of a linear model of a surrogate fitted to blood pressure and surrogate data pairs. Two parameters, offset and slope (or sensitivity parameter), are determined by regression. The regression is not robust in the diagram shown in FIG. 3A because the data pair values ​​(denoted by x) are too similar. The regression shown in the diagram in FIG. 3B, in contrast, is very robust because the available data pair values ​​are spread over a wider range.

[0057] In practice, achieving robust calibration of the model is a key challenge to ensure that BP surrogates can be used for otherwise unobtrusive determination and monitoring of a subject's BP. Regression using multiple NIBP and PAT measurements taken over a short period of time has the problem that the measured NIBP and PAT values ​​are too similar, resulting in a poor curve fit. To improve the curve fit, the patient's blood pressure level needs to be altered, which is usually impossible to do without changing the stiffness of the arterial walls, which would be harmful to the patient and impractical.

[0058] Furthermore, the increase in PAT can be measured in relation to the applied cuff pressure, since the transmural pressure in the arterial section below the cuff decreases. This is illustrated in FIG. 4, which shows an example diagram of an increase in PAT 120 caused during cuff inflation 121. The cuff pressure changes the arterial transmural pressure. The value of PAT is obtained beat by beat in synchronism with the cuff pressure. This sequence is applied more precisely to the arterial section above the cuff. To estimate BP, a scaling factor needs to be introduced to scale the entire arterial section where the distal pulse sensor is located.

[0059] Using the acquired PAT paired with synchronously acquired cuff pressure and BP values, a PAT / BP sensitivity parameter S can be derived, which allows tracking of BP. This is illustrated in FIG. 5, which shows a schematic diagram illustrating the overall concept of calibration. During cuff inflation, an increase in PAT occurs, from which, together with the synchronously acquired cuff pressure (using oscillometry) and the derived BP, a sensitivity parameter S for tracking BP is derived. The subject-specific scaling factor is represented in the formula as k. Two examples of mathematical models using the estimated sensitivity parameter S for tracking BP are shown and are used according to embodiments of the present disclosure.

[0060] FIG. 6 shows a schematic diagram of an exemplary setup used in the general concept for acquiring an invasively measured BP signal acquired distal to the cuff. It should be noted that invasive BP measurements are primarily shown here to clarify the BP changes distal to the cuff. New insights and understandings in physiological processes that influence the calibration process and underlying assumptions for a robust and accurate regression of the sensitivity parameter S have been discovered. In particular, it has been discovered that the change in PAT during cuff inflation is distorted by dynamic filling effects in the lower arm due to venous occlusion (which affects arterial blood pressure and causes adverse effects) and the capacity of veins to store blood in the arm region. This finding impacts the parameter regression and needs to be addressed in order to reliably infer the sensitivity parameter S.

[0061] Figure 7 shows diagrams of signals measured using the setup shown in Figure 6. Figure 7A shows the ECG signal. Figure 7B shows the cuff signal. Figure 7C shows the radial BP / A-line (representing the invasive BP signal). Figure 7D shows the PPG signal.

[0062] The current calibration technique (applying cuff pressure and varying the PAT under the cuff) is subject to nonlinear dynamic filling effects in the peripheral arm arterial tree. The observed process can be subdivided into different stages depending on their impact on the mathematical modeling functions and the included / excluded results of the data pairs. Figure 8 shows the analysis of the inflation process with respect to the blood pressure measured peripherally behind the cuff. Different processes can be identified that affect the PAT. The basic assumption for a successful calibration that does not rely on complex model assumptions is a constant peripheral blood pressure in the distal arm, which is considered to be present only at lower cuff pressures.

[0063] These findings are consistent with the transmural blood pressure (SBP-P cuff , on the y-axis) versus the measured PAT (on the x-axis). A typical PAT increase during cuff inflation is observed and at least three zones A, B, and C can be distinguished. In zone A, the arm is compressed with a lower cuff pressure, resulting in a large change in arm volume. The peripheral BP is at this stage unchanged by the cuff inflation. In zone B, the arterial peripheral blood pressure is still not affected by the cuff inflation. In zone C, a nonlinear dynamic process is observed in the peripheral blood pressure signal. Zone C should not be included in the calibration procedure as it is related to the cuff pressure area and thus the peripheral arm blood pressure changes.

[0064] However, the regression model assumes a constant BP for a particular instant in the pulse phase (eg, systolic, mean, diastolic) with a given transmural pressure during inflation.

[0065] Transmural pressure is P trans (t)=P arterial (t)-P cuff(t). The key condition for reliable regression during cuff inflation / deflation is the time-dependent P trans (t) is P cuff (t) and (for a particular instant of the pulse) P arterial = constant. This suggests that for reliable regression, the internal arterial pressure must not be affected by the inflation of the cuff (as explained in more detail in Figure 9).

[0066] In one exemplary regression technique described in US2010 / 160798A1, the value of PAT is included only when the cuff pressure is higher than DBP. However, as shown in FIG. 9 and indicated as Zone C, at such cuff pressure, the blood pressure in the distal arm changes relative to the BP at zero cuff pressure and also changes with cuff inflation. The calibration concept disclosed in US2010 / 160798A1 should therefore be improved, if at all applicable, within the scope of this disclosure. Considering the findings of the present invention, at cuff pressures higher than diastolic blood pressure, a complex nonlinear regression model is applied that depends on vascular resistance, the effect of arterial occlusion, venous obstruction, and the effect of pulse distortion. Usually, such parameters are not available. To address these challenges, a different strategy is needed for the calibration method.

[0067] The impact on blood pressure monitoring is that inaccurate blood pressure calibration results in unreliable BP estimation leading to false alarms or late and delayed interventions that impact relevant patient outcomes. The present invention deals with a solution that, in contrast to the current state of the art, includes PAT as calibration data only when well-defined criteria are valid.

[0068] FIG. 10 shows a diagram of one embodiment of a system 1 for calibrating a BP surrogate for use in monitoring a subject's blood pressure according to the present invention.

[0069] The system 1 comprises a pressure delivery system 10 comprising a cuff 11 configured to be attached to a body part of a subject and to deliver pressure to the body part of the subject by inflating the cuff. Such pressure delivery systems 10, comprising a cuff 11 that can be attached to a limb of the subject, specifically the upper arm, upper leg, or wrist, are generally known in the art of NIBP measurement. The pressure delivery system generally comprises a pressure generating unit, e.g. a pump or pressure accumulator, configured to inflate the cuff 11, a valve configured to deflate the cuff 11, and a processor configured to control the pressure generating unit and the valve and to determine the subject's blood pressure based on the measured cuff pressure. A user interface, e.g. including one or more of a display, a keypad, a speaker, and a touchpad, is provided to emit the measured BP value, e.g. in a visible and / or audible form.

[0070] The system 1 further comprises a pressure sensor 20 configured to obtain a value of the time-dependent cuff pressure during inflation of a cuff of a pressure delivery system attached to the subject's body part to obtain a BP measurement. The pressure sensor 20 is typically not directly attached to the subject's body part, but rather the cuff 11 is attached to the body part and the pressure sensor 20 is attached to an air tube from the cuff 11 (in the pressure delivery system). Thus, the pressure sensor 20, for example in the form of a pressure transducer, is integrated within the cuff and realised with a conventional BP sensor. Other means of applying external pressure and detecting cuff pressure, such as a shell cuff design, are possible as well.

[0071] The system 1 further comprises a first sensor 30 attached to a first part of the subject's body, the first sensor 30 being configured to acquire a first time-dependent sensor signal related to the subject's heartbeat during inflation of the cuff. The first sensor 30 is an ECG sensor for acquiring an ECG signal. The first part is for example the subject's chest or torso, where ECG electrodes are preferably attached for the ECG measurement.

[0072] The system 1 further comprises a second sensor 40 attached to a second part of the subject's body and configured to acquire a second time-dependent sensor signal related to the subject's heartbeat, which acquires a signal changed by the inflation of the cuff. The second sensor 40 is a PPG sensor, such as a contact PPG sensor (such as a pulse oximetry sensor with one or more LEDs emitting light in the visible and / or infrared range, e.g. red and infrared light), or a remote PPG sensor (such as a camera or photodetector, as known in the art, for example, from "Remote plethysmographic imaging using ambient light", Verkruysse et al., Optics Express, 16(26), pp. 21434-21445, Dec. 22, 2008). PPG generally refers to the optical measurement of volume changes of an organ or body part, specifically the detection of volume changes due to cardiovascular pulse waves traveling through the subject's body with each heartbeat. The radiation reflected from or transmitted through the subject's skin area can be detected. The second site is, for example, the subject's hand or finger on which the contact PPG sensor is attached or monitored by a remote PPG sensor. The second site is generally a site distal to the first site. For example, if the cuff is attached to the left upper arm, the second site is the left hand or finger on which the second sensor is attached or monitored by the second sensor.

[0073] The first and second signals are acquired and sampled generally synchronously. Signals other than ECG and PPG signals that are related to the subject's heart rate and allow calculation of PAT and / or PTT may additionally or alternatively be used, such as sensors that detect heart sounds, sensors that detect chest vibrations, etc.

[0074] The system 1 further comprises a device 50 as disclosed herein and described below for calculating calibration parameters for calibrating a BP surrogate from the BP measurements, the time-dependent cuff pressure values, and the first and second time-dependent sensor signals. The device 50 is further optionally configured to determine and monitor the subject's BP using the BP surrogate.

[0075] The system 1 optionally further comprises an output interface 60 configured to emit any determined information, such as the subject's BP value determined and monitored by use of a BP surrogate. The output interface 60 is generally any means for outputting information in visual or audible form, such as in the form of text, images or diagrams, sound or spoken word, etc. The output interface 60 can be, for example, a display, a speaker, a touch screen, a computer monitor, a smartphone or tablet screen, etc.

[0076] The system 1 optionally further comprises a control unit 70 configured to control the pressure delivery system 10, for example based on control signals provided by the device 50 or calculated in the system 1. The control unit 70 may be part of the device 50 or may be external.

[0077] FIG. 11 shows a diagram of one embodiment of a device 50 for calibrating a BP surrogate for use in monitoring the blood pressure of a subject in accordance with the present invention.

[0078] The device 50 comprises a BP input 51 configured to acquire a value of a time-dependent cuff pressure during inflation of a cuff 11 of a pressure delivery system attached to a body part of the subject, to acquire a BP measurement. The device 50 further comprises a sensor input 52 configured to acquire a first time-dependent sensor signal and a second time-dependent sensor signal related to the subject's heartbeat and measured at different parts of the subject's body during inflation of the cuff. The BP input 51 and the sensor input 52 are directly coupled or connected to the cuff 11 and the first and second sensors 30, 40, or acquire (i.e. retrieve or receive) these signals from a storage device, buffer, network, bus, etc. The inputs 51 and 52 are thus communication or data interfaces (wired or wireless), such as, for example, a Bluetooth interface, a WiFi interface, a LAN interface, an HDMI interface, a direct cable connection, or any other suitable interface allowing signal transmission to the device 50.

[0079] The device 50 further comprises a processing unit 53. The processing unit 53 is any kind of means configured to process the signals and determine calibration parameters for calibrating the BP surrogate. The processing unit is further configured to determine and monitor the subject's BP using the BP surrogate. The processing unit is implemented in software and / or hardware, e.g. a programmed processor, computer or app on a user device, such as a smartphone, smartwatch, tablet, laptop, PC, workstation, etc.

[0080] The device 50 further comprises an output 54 configured to output any determined information. The output 54 is generally any interface that provides the determined information, e.g., transmits the determined information to another device or provides it for retrieval by another device (e.g., a smartphone, computer, tablet, etc.). The output is thus generally any (wired or wireless) communication or data interface.

[0081] Figure 12 shows a diagram of one embodiment of a method 200 according to the invention. The steps of the method 200 are performed by the device 50, the main steps of the method being performed by the processing unit 53. The method is implemented as a computer program running on a computer or processor.

[0082] In a first step 201, a time-dependent cuff pressure value is obtained during inflation of a cuff of a pressure delivery system attached to a body part of a subject. In a second step 202, a BP measurement is obtained. In a third step 203, a first time-dependent sensor signal and a second time-dependent sensor signal related to the subject's heart rate and measured at different parts of the subject's body during inflation of the cuff are obtained. It should be noted that the order of steps 201 to 203 does not indicate a chronological order and these steps may be performed in any chronological order. Preferably, steps 201 and 203 are performed simultaneously and step 202 is performed before or after steps 201 and 203.

[0083] In a fourth step 204, while the cuff is inflated, a pulse-related value is calculated from the first and second time-dependent sensor signals using a first feature of the first time-dependent sensor signal and a second feature of the second time-dependent sensor signal, the pulse-related value being a PAT value or a PTT value. The PAT or PTT value is extracted from the first and second signals (e.g., ECG and PPG signals) using a suitable feature extraction technique. Other features such as heart rate, morphological features of the PPG, etc. may also be derived from a single signal.

[0084] In a fifth step 205, calibration parameters of a BP surrogate are calculated from the BP measurements and pairs of pulse-related values ​​and corresponding cuff pressure values, including pulse-related values ​​and corresponding time-related cuff pressure values ​​(specifically values ​​at the same time). Only those pairs of pulse-related values ​​and corresponding cuff pressure values ​​for which one or more predefined conditions regarding BP and / or cuff pressure are met are used for the calculation of the BP surrogate. There are various options for such conditions (or criteria), which are explained below.

[0085] Thus, in accordance with a first main aspect of the present invention, a BP surrogate can be calibrated and then BP can be tracked continuously. Criteria are used in embodiments that include a range of cuff pressures during inflation to reliably calibrate a BP surrogate from a pair of cuff pressure / BP surrogate data, where the transmural pressure is P cuff >0P arterial But, P cuff P at =0 arterial PAT is well defined as a function of applied cuff pressure that remains unchanged compared to . Calibration traditionally required full inflation of the cuff. When calibrating a PAT-based surrogate from continuous inflation-based NIBP measurements, as used in the embodiment, only PAT values ​​resulting from cuff pressure levels given by well defined criteria that satisfy physiological constraints are utilized. Thus, according to one embodiment, certain data (data in zone C in FIG. 9 ) are discarded. Features of the PPG signal are used as indicators of violations of assumptions underlying the applied regression model, e.g., of changes in pulse pressure in peripheral arterial segments. Features of the oscillation envelope are used as indicators of violations of assumptions underlying the applied regression model, e.g., of blood pressure ranges. Thus, a reliable calibration of PAT, a surrogate of the patient's BP pressure, characterizing the cuff inflation process using dedicated procedures and methods is possible.

[0086] In the following description of the BP calibration method by regression of the pair of BP surrogate / cuff pressure, PAT will be used as the BP surrogate. One embodiment is based on including data regarding the PAT data versus the cuff pressure value only for the cuff pressure during a period where the peripheral blood pressure (i.e., the cuff pressure behind the cuff on the arm when the cuff is placed on the upper arm) is constant, for example, with a variation of less than 5 or 10%.

[0087] In a complete cuff inflation with BP estimation, for the BP estimation of SBP, DBP, and MBP, oscillometry is used and typically requires a cuff inflation beyond the systolic BP. Using various criteria, pairs of PAT values and cuff pressure values can be selected for use in determining the calibration parameters. For example, one or both of the following criteria are used. - Criterion 1: P cuff <For DBP or MBP only, the value of PAT / P cuff - Criterion 2: p min <P cuff <For DBP, for example, when p min = 20 mmHg, the value of PAT / P cuff p min can be defined with respect to the observed effect of tissue compression (Figure 7). At cuff pressures where the volume of the tissue changes significantly, it may become more difficult to reliably define the transmural pressure across the artery.

[0088] In the figure shown in Figure 13, the data acquired during the complete inflation process of the cuff is shown as points in the graph of PAT vs P cuff The vertical line DBP shows up to which pair of values of the PAT value (i.e., the value related to the pulse) and the cuff pressure value is used to determine the calibration parameter. cuff

[0089] ​​​14 shows a flow diagram of one embodiment of a method 300 for using this aspect of the invention. In a first step 301, a BP measurement is initiated. In a second step 302, the PAT / P cuff A complete set of data pairs is acquired. In a third step 303, BP values ​​such as systolic, diastolic and mean blood pressure are determined in a standard manner. In a fourth step 304, the PAT / P cuff The data range of the data pairs is restricted according to a predefined criterion, for example according to criterion 1 or criterion 2. In a fifth step 305, a regression is performed on the acquired (and restricted) data pairs. Finally, in a sixth step 306, BP is estimated from the calibrated PAT data.

[0090] In another embodiment, the cuff pressure is used, for example during periods when the SBP in the distal arm (when the cuff is attached to the upper arm) is kept constant. FIG. 15 shows the PPG signal characteristics during periods when the SBP in the distal arm does not show changes in pulse pressure. cuff 15A and 15B show the behavior of the PPG signal during cuff inflation. FIG. 15B shows the increase in PAT during inflation. FIG. 15C shows the cuff pressure during inflation. The change in PPG amplitude indicates the change in the peripheral arterial pulse pressure PP, and the maximum PAT / PAT value pair to be included in the regression / calibration is calculated. cuff The amplitude of the PPG is one example, where the onset of a decrease in the amplitude of the PPG indicates that the maximum pressure regression data should be used.

[0091] 16 shows a flow diagram of one embodiment of a method 400 of using this aspect of the invention, specifically the use of PPG amplitude change as a criterion. In a first step 401, a BP measurement is started. In a second step 402, the PAT / P is measured during the cuff inflation process for periods when the PPG amplitude is constant (or changes by less than a predefined percentage, e.g., 5 or 10%). cuffA set of data pairs is acquired. In a third step 403, a regression is performed on the acquired data, whereby, for example, the BP from the previous full inflation is used in the regression. Finally, in a fourth step 404, the BP is estimated from the calibrated PAT data.

[0092] As an alternative or in addition to the above mentioned criteria, other criteria can be used. Exemplary criteria are: i) the number of vertices during the expansion, e.g. PAT-PAT ref < PAT / P if max cuff ii) a definition of the maximum change in PAT, including the value; and ii) an observation of the behavior of the change in PAT, e.g., the slope of the PAT.

[0093] The regression of model parameters is generally based on multiple data pairs, i.e., measured PAT values ​​and corresponding cuff pressure values. This is illustrated in the diagram shown in FIG. 17, which shows an example of fitting a parameterized model to measured data pairs (cuff pressure, PAT) obtained during cuff inflation. FIG. 17A shows a diagram of full inflation, while FIG. 17B shows a diagram where only low cuff pressure levels are used for regression, resulting in a less robust regression. PAT values ​​can be measured beat-by-beat, as explained above. Each time a new pulse is detected at the peripheral sensor, the corresponding PAT can be calculated and the corresponding cuff pressure is measured. In other words, the number of data pairs that can be used for parameter regression is equal to the number of pulses detected during cuff inflation.

[0094] However, the measured PAT value is subject to measurement error. Therefore, the more data pairs available to perform the actual regression, the better the quality (i.e., accuracy) of the parameter regression. Therefore, as described in one embodiment above, using only the lower part of the curve, i.e., PAT and cuff pressure data pairs up to a certain cuff pressure level, for the parameter regression provides further improvement in terms of the quality of the parameter regression. For example, if only the lower part of the curve is used, the overall number of data pairs available for the parameter regression is reduced. Furthermore, the signal-to-noise ratio of the regression curve is reduced. The absolute error in measuring the PAT is independent of the applied cuff pressure. However, the signal component desired for the parameter regression is not the absolute value of PAT, but the change in PAT from the baseline PAT during cuff inflation. The baseline PAT is the PAT without the applied cuff pressure. With this in mind, if only low cuff pressure levels are considered, the relative PAT measurement error increases since the change in PAT is smaller at low cuff pressure levels.

[0095] Thus, according to a second main aspect of the invention, a system and a device are used that comprises generally the same units as those described above with respect to the system and device shown in figures 10 and 11. However, the processing unit 53 and the method executed by the processing unit 53 are arranged differently. Figure 18 shows a diagram of an embodiment of a method 500 according to the invention. The steps of the method 500 are executed by the device 50, the main steps of the method being executed by the processing unit 53. The method is implemented as a computer program executed on a computer or processor.

[0096] In a first step 501, a time-dependent cuff pressure value is obtained during inflation of a cuff of a pressure delivery system attached to a body part of a subject. In a second step 502, a BP measurement is obtained. In a third step 503, a first time-dependent sensor signal and a second time-dependent sensor signal related to the subject's heart rate and measured at different parts of the subject's body during inflation of the cuff are obtained.

[0097] In a fourth step 504, a control signal for controlling the pressure delivery system is calculated to repeatedly inflate the cuff of the pressure delivery system with partial inflation to a cuff pressure below the subject's deflated BP.

[0098] In a fifth step 505, a pulse-related value is calculated from the first and second time-dependent sensor signals measured during repeated partial inflations of the cuff using the first feature of the first time-dependent sensor signal and the second feature of the second time-dependent sensor signal, the pulse-related value being a PAT value or a PTT value.

[0099] In a sixth step 506, BP surrogate calibration parameters are calculated from the BP measurements and pairs of pulse-related values ​​and corresponding cuff pressure values ​​calculated from the first and second time-dependent sensor signals measured during repeated partial inflations of the cuff, the pairs including a pulse-related value and a corresponding contemporaneous cuff pressure value (or at least a time-related value).

[0100] Thus, in one exemplary embodiment, the pump control is configured to repeatedly apply short increasing pressure profiles (pressure ramps) ranging from a cuff pressure of 0 mmHg to approximately the diastolic pressure level or until a change in pulse amplitude in the PPG signal is detected, and record the resulting cuff pressure and PAT. The process involves performing a parameter regression for each applied pressure ramp and averaging the resulting regression parameters to improve overall calibration accuracy.

[0101] The embodiment described with reference to Fig. 18 is based on the idea that only low cuff pressure levels should be used for calibration. Repeated sampling at such appropriate cuff pressure levels can be performed in a short enough period so that the BP level is stable and there are no changes that distort the calibration process. Furthermore, inflation ramps can be repeated even without pauses between ramps because the maximum pressure is low. There is no need to wait for tissues and arteries to recover from high pressurization (i.e., there is no hysteresis effect).

[0102] Another advantage is that cuff pressure levels below the diastolic pressure are less intrusive to the patient, especially an awake patient, resulting in less discomfort and stress on the skin than a ramp to full inflation to above the systolic level.

[0103] Yet another advantage of such an embodiment is that it places no extra requirements on the pressure pump integrated into the NIBP device: alternative proposals to achieve the same effect with a single gentle pressure ramp from 0 mmHg to the diastolic blood pressure level would require a pump capable of delivering very low airflow rates at low pressure levels, which represents a significant challenge for the pump design and is not supported by standard pumps currently used in NIBP devices.

[0104] This concept applies not only to systems that use PAT measured with ECG and PPG sensors, but more generally to systems that use alternative methods of measuring pulse propagation delay (e.g., based on differential pulse wave transit time) and to systems that include a pressurizable cuff that is added within the propagation path of the pulse wave being considered.

[0105] In the following, an embodiment of a second aspect is described for achieving improved calibration of a BP surrogate by repeated partial inflations, in which the PAT value is estimated beat-by-beat using PPG and ECG sensors.

[0106] PAT baseline value PAT base is obtained, in one embodiment, by averaging PAT measurements over a particular period of time. The PAT measurement period can be predefined, fixed, or user-configurable, e.g., 20 seconds, 15 seconds, or other sufficiently long period that sufficiently suppresses undesirable fluctuations caused by, e.g., respiration (typical frequencies 0.1-0.5 Hz), Mayer waves (typical frequencies 0.1-0.5 Hz), or other measurement noise.

[0107] Activating the pump of the NIBP device provides a first (partial) inflation slope to a specific sub-systolic level. The slope of the first inflation may be determined based on the diastolic pressure (known from the last NIBP measurement), a specified sub-diastolic level, a specific fixed cuff pressure level (e.g., 60 mmHg or 40 mmHg), or the PAT. base It may be a pressure level exceeding a certain threshold compared to PAT or a pressure level at which a predefined amplitude change in the amplitude of the PPG is detected. The inflation rate is set to a fixed inflation rate, e.g. 10 mmHg / sec, or is otherwise selected to ensure a certain minimum heart rate within the inflation slope, e.g. depending on the heart rate. During inflation of the cuff, the cuff pressure P cuff is measured. Baseline PAT base The difference in PAT compared to is designated as ΔPAT and is calculated relative to the PAT value during cuff inflation.

[0108] Measured data pairs (ΔPAT, P cuff ) to find the unknown model parameters m1, m2, ~m of the parameterized model function f. N A parameter regression (curve fitting) is performed, and the model function f models the relationship between the change in PAT and the change in blood pressure ΔBP around the operating point. During cuff inflation, the under-cuff length L cuff The blood pressure in the arterial part of cuff The change in blood pressure is ΔBP=-P cuff (This explanation applies equally to the methods explained above.) Therefore, the regression equation is given by: ΔPAT=L cuff f(-P cuff ;m1, m2, ~m N )

[0109] As an example, in a simple linear proportional model, the number of unknown model parameters is two (f = m1 · ΔBP + m2). The preferred criterion for parameter regression is least squares error, although parameter regression may also be performed according to another error measure, for example least absolute error.

[0110] To improve the parameter regression, the cuff is deflated, another partial inflation is performed, and another parameter regression is performed. This can be repeated multiple times. In a preferred embodiment, one additional partial inflation is performed. The resulting model parameter regression values ​​are then averaged, thus improving the accuracy of the model parameter calibration. The degree of variation in the resulting regression values ​​can be used to determine how many partial inflation iterations to perform. For example, if the resulting regression values ​​after two partial inflations are closely aligned, no further partial inflations are necessary.

[0111] To calibrate the resulting continuous BP estimator, NIBP measurements were taken and a blood pressure reference measurement, BP ref This is achieved by continuing the final partial inflation beyond the systolic cuff pressure level so that a standard inflation-based oscillometric blood pressure measurement can be calculated. In other words, the slope of the final inflation is the slope of the "full" cuff inflation. FIG. 19 shows a diagram illustrating the proposed cuff pressure profile in the example of performing two additional partial inflations (before full inflation) to improve the calibration of the model parameters. Finally, the calibrated PAT-based continuous blood pressure estimator is obtained as follows:

number

[0112] These steps are typically repeated from time to time to recalibrate the continuous blood pressure estimator. Recalibration can occur according to some periodic schedule or when significant changes in blood pressure occur. In general, the proposed pressure profile of the type shown in FIG. 19, i.e., a pressure profile that includes repeated partial inflations, can be performed whenever an NIBP measurement is automatically or manually triggered.

[0113] In an alternative embodiment, partial inflations are added but not prepended. Figure 20 shows a diagram illustrating the proposed cuff pressure profile in the example of two additional partial inflations (after full inflation). This profile has the advantage of being able to display the NIBP value earlier compared to the previous embodiment of the pressure profile illustrated in Figure 19. However, the time interval between the first and second parameter calibrations is longer. If the patient's hemodynamic state is not constant during this time, i.e., if the model parameters are changing, averaging over multiple calibration events will provide less improvement in accuracy.

[0114] In another alternative embodiment, the improvement of the parametric regression is not obtained by averaging the regression parameters from multiple partial dilations, but by averaging all the data pairs (ΔPAT, P cuff ) for the regression error threshold. This option provides improved accuracy by allowing more regression data points to be used in a single regression. Furthermore, the number of partial dilations to be performed is not pre-determined, but rather determined through a threshold comparison of the regression error. If the regression error falls below that threshold, no further partial dilations are necessary.

[0115] In yet another alternative embodiment, the continuous blood pressure estimator further allows for only partial inflations, for example every 10 minutes, after it has been calibrated by model parameter regression and NIBP reference measurement. This is done as explained above. A "full calibration" with NIBP reference measurement results in the model parameters m1 to m2 being inflated by the model parameter regression compared to the results of the parameter regression following the last "full calibration". N This is only done if a significant change in the current continuous blood pressure estimator occurs. If no significant change is detected, the parameters of the current continuous blood pressure estimator are not updated and full inflation for NIBP measurement is not performed. Figure 21 shows a diagram of a pressure profile that continuously tracks the need for recalibration through repeated partial inflations. The advantage of this concept is that it may operate unobtrusively for long periods of time. This is advantageous for patients who are automatically NIBP monitored for several days, for example supporting the healing process of patients in the ICU through improved quality of sleep at night.

[0116] The present invention is used in a wide range of applications, e.g., in monitoring sites in hospitals, where NIBP, ECG, and PPG sensors are already available and in use. These sensors are used practically in all sites, e.g., in OR and ICU, including specific sites in general wards, as well as in specialized wards, such as neurology. More specifically, beat-to-beat arterial blood pressure can be measured non-invasively. Beat-to-beat blood pressure monitoring can provide early detection of hypotensive events associated with adverse patient outcomes. Furthermore, early warning of hemodynamic instability can be provided. The disclosed device and method can also be used for long-term monitoring, e.g., 24-hour BP monitoring, or BP monitoring during nighttime sleep.

[0117] While the invention has been illustrated and described in detail in the drawings and the foregoing description, such illustration and description are to be considered as illustrative or representative and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

[0118] In the claims, the word "comprising" does not exclude other elements or steps, nor does the word "a" or "an" exclude a plurality. A single element or other unit fulfills the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0119] The computer program may be stored / distributed on a suitable non-transitory medium, such as an optical storage medium or a solid-state medium, provided integrally with or as part of other hardware, or may be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.

[0120] Any reference signs in the claims should not be construed as limiting the scope.

[0121] Further embodiments of the systems, devices and methods described in the claims are configured as follows. 1. A device as disclosed herein, comprising: - calculating a control signal to control a pressure delivery system (10) to repeatedly inflate a cuff of the pressure delivery system, with partial inflation, to a cuff pressure below the subject's systolic BP; - calculating a value related to a pulse from first and second time dependent sensor signals measured during repeated partial inflations of the cuff using a first feature of the first time dependent sensor signal and a second feature of the second time dependent sensor signal, the value related to a pulse being a value of a pulse arrival time PAT or a value of a pulse transit time PTT; - calculating BP surrogate calibration parameters from the BP measurements and pairs of pulse-related values ​​and corresponding cuff pressure values, the pairs being calculated from first and second time-dependent sensor signals measured during repeated partial inflations of the cuff, the pairs being configured to include a pulse-related value and a corresponding time-related cuff pressure value. 2. The device as defined in embodiment 1, The processing unit (53) is configured to calculate a control signal to control the pressure delivery system (10) to inflate the cuff of the pressure delivery system with partial inflation during periods when the peripheral BP is substantially constant, specifically varying by less than 10 percent or less than 5 percent. 3. The device as defined in embodiment 2, The processing unit (53) is configured to determine whether the peripheral BP is substantially constant based on the first and / or second time-dependent sensor signals, in particular by determining a time period during which the amplitude of the second time-dependent signal is substantially constant, in particular fluctuating by less than 10 percent or less than 5 percent relative to a previously obtained average value. 4. The device as defined in embodiment 1, The processing unit (53) is configured to calculate a control signal to control the pressure delivery system (10) to inflate the cuff of the pressure delivery system with partial inflation to the subject's diastolic BP, in particular the subject's most recently measured diastolic BP, to the subject's average BP, in particular the subject's most recently measured average BP, or to a set threshold cuff pressure. 5. The device defined in embodiment 1, The processing unit (53) is configured to calculate a control signal to control the pressure delivery system (10) to partially inflate the cuff of the pressure delivery system to a cuff pressure at which the PAT exceeds a PAT threshold, specifically an absolute PAT threshold or a relative PAT threshold compared to a baseline PAT. 6. The device defined in embodiment 5, The processing unit (53) is configured to calculate a baseline PAT by averaging the PAT measurements over a period of time, specifically a period ranging from 10 seconds to 5 minutes. 7. The device defined in embodiment 1, The processing unit (53) is configured to calculate a control signal to control the pressure delivery system (10) to partially inflate the cuff of the pressure delivery system to a cuff pressure at which the amplitude of the second time-dependent signal exceeds an amplitude threshold, specifically an absolute amplitude threshold or a relative amplitude threshold. 8. The device defined in any one of embodiments 1 to 7, The processing unit (53) is configured to calculate a control signal to control the pressure delivery system (10) to inflate the cuff of the pressure delivery system with partial inflation at a predetermined inflation rate or at an inflation rate that is dependent on one or more of the subject's heart rate, the subject's diastolic BP, the subject's mean BP, and the subject's systolic BP. 9. The device defined in any one of embodiments 1 to 8, The processing unit (53) - calculating a control signal for controlling the pressure delivery system (10) to deflate the cuff of the pressure delivery system after each partial inflation; - i) for each iteration, calculating, using a separate regression, one or more calibration parameters from pairs of pulse-related values ​​and corresponding cuff pressure values ​​obtained in each iteration, and averaging each calibration parameter calculated in two or more iterations to obtain one or more average calibration parameters for use in a surrogate for BP; or - ii) calculating, in a single regression, one or more calibration parameters from pairs of pulse-related values ​​and corresponding cuff pressure values ​​obtained in two or more repetitions; It is configured as follows. 10. The device defined in any one of embodiments 1 to 9, The processing unit (53) is configured to calculate a control signal to control the pressure delivery system (10) to inflate the cuff of the pressure delivery system at full inflation above the subject's systolic BP before and / or after one or more repeated partial inflations, and to obtain a time-dependent BP reference measurement for use in monitoring the subject's BP using the BP surrogate. 11. The device defined in any one of embodiments 1 to 10, The processing unit (53) is configured to control the number of partial inflation iterations to perform based on a comparison of the regression error to a regression error threshold, and / or to control whether and when to perform a full inflation according to a fixed or variable schedule, or if one or more calibration parameters have changed substantially since the last calculation, particularly if they have changed by more than 10 percent. 12. The device defined in any one of embodiments 1 to 11, The processing unit (53) is configured to determine a value of BP using the calculated BP surrogate and the measured first and second time-dependent sensor signals measured at various parts of the subject's body when no pressure is delivered to the subject's body part by the pressure delivery system. 13. A system for calibrating a surrogate for blood pressure (BP) for use in monitoring a subject's blood pressure, the system comprising: A pressure delivery system (10) comprising a cuff (11) configured to be attached to a body part of a subject and to deliver pressure to the body part of the subject by inflating the cuff; a pressure sensor (20) configured to acquire a time-dependent cuff pressure value during inflation of a cuff of a pressure delivery system attached to a body part of a subject, and to acquire or estimate a BP measurement; a first sensor (30) attached to a first portion of the subject's body and configured to obtain a first time-dependent sensor signal related to the subject's heart rate while the cuff is inflated; a second sensor (40) attached to a second portion of the subject's body and configured to obtain a second time-dependent sensor signal related to the subject's heart rate while the cuff is inflated; A device (50) as disclosed herein for calculating one or more calibration parameters for calibrating a BP surrogate from BP measurements, time-dependent cuff pressure values, and first and second time-dependent sensor signals. Equipped with: 14. The method disclosed herein, comprising: - calculating a control signal to control a pressure delivery system (10) to repeatedly inflate a cuff of the pressure delivery system, with partial inflation, to a cuff pressure that is less than the systolic BP of the subject; - calculating a pulse-related value from first and second time-dependent sensor signals measured during repeated partial inflations of the cuff using a first feature of the first time-dependent sensor signal and a second feature of the second time-dependent sensor signal, the pulse-related value being a value of the pulse arrival time PAT or a value of the pulse transit time PTT; - calculating calibration parameters of a BP surrogate from the BP measurements and pairs of pulse-related values ​​and corresponding cuff pressure values, the pairs being calculated from first and second time-dependent sensor signals measured during repeated partial inflations of the cuff, the pairs including a pulse-related value and a corresponding, time-related, contemporaneous value of the cuff pressure. The method comprising: 15. A computer program comprising program code means which, when said computer program is executed on a computer, causes the computer to carry out the steps of the method defined in embodiment 14.

Claims

1. 1. A device for calibrating a blood pressure (BP) surrogate for use in monitoring the blood pressure of a subject, the device comprising: a BP input for obtaining a time-dependent cuff pressure value during inflation of a cuff of a pressure delivery system attached to the subject's body part and for obtaining a BP measurement; a sensor input configured to acquire first and second time-dependent sensor signals measured at different locations on the subject's body during inflation of the cuff, the first and second time-dependent sensor signals being related to the subject's heart rate; Processing unit and Equipped with the processing unit calculates, during inflation of the cuff, a pulse-related value from the first and second time-dependent sensor signals using a first feature of the first time-dependent sensor signal and a second feature of the second time-dependent sensor signal, the pulse-related value being a pulse arrival time value PAT or a pulse transit time value PTT; the processing unit selects pairs of the pulse-related value and the corresponding cuff pressure value for calculating calibration parameters of the BP surrogate, the pairs including the pulse-related value and the corresponding time-dependent cuff pressure value, and only the pairs of the pulse-related value and the corresponding cuff pressure value for which one or more predetermined conditions regarding the BP and / or the cuff pressure are satisfied are selected; the processing unit calculates the calibration parameters of the BP surrogate from the BP measurements and the selected pairs of pulse-related values ​​and corresponding cuff pressure values. device.

2. the processing unit uses only pairs of pulse-related values ​​and corresponding cuff pressure values ​​calculated from the first and second time-dependent sensor signals measured during partial inflation of the cuff to calculate calibration parameters of a BP surrogate. The device of claim 1 .

3. the processing unit uses only pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating the calibration parameters of the BP surrogate where peripheral BP is substantially constant, in particular where the fluctuation is less than 10% or less than 5%.

3. A device according to claim 1 or 2.

4. the processing unit determines, based on the first and / or second time-dependent sensor signals, whether the peripheral BP is substantially constant, in particular by determining a time period during which the amplitude of the second time-dependent signal is substantially constant, in particular fluctuating by less than 10% or less than 5% relative to a previously obtained average value. The device of claim 3.

5. the processing unit uses only pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating the calibration parameters of the BP surrogate, where the cuff pressure value is below the diastolic BP of the subject, in particular below the most recent measured diastolic BP of the subject, or below the mean BP of the subject, in particular below the most recent measured mean BP of the subject, or below a set cuff pressure threshold. The device of claim 1 .

6. the processing unit uses only pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating calibration parameters of the BP surrogate, the cuff pressure values ​​being in a range between a set minimum cuff pressure and a set maximum cuff pressure, in particular, the minimum cuff pressure being set in a range of 10 to 30 mmHg and the maximum cuff pressure being set in a range of 40 to 90 mmHg. The device of claim 1 .

7. the processing unit uses only pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating the calibration parameters of the BP surrogate, where the pulse-related values ​​are substantially constant, in particular, fluctuating by less than 10%. The device of claim 1 .

8. the processing unit uses only pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating calibration parameters of the BP surrogate until a slope of the pulse-related value curve over time exceeds a threshold. The device of claim 1 .

9. the processing unit calculates a control signal for controlling a pressure delivery system to inflate a cuff of the pressure delivery system, in particular to fully inflate the cuff for obtaining the BP measurement and to only partially inflate the cuff for obtaining the first time-dependent sensor signal and the second time-dependent sensor signal. The device of claim 1 .

10. the first time-dependent sensor signal is an ECG signal and / or the second time-dependent sensor signal is a photoplethysmography PPG signal, in particular a contact PPG signal or a remote PPG signal; The device of claim 1 .

11. the processing unit to calculate a calibration parameter using regression, in particular, to calculate the calibration parameter as a slope of the dependence of the relationship of blood pressure as a function of the BP surrogate for zero cuff pressure; The device of claim 1 .

12. the processing unit determines a value of BP when no pressure is delivered to the subject's body part by the pressure delivery system using the calculated BP surrogate and the measured first and second time-dependent sensor signals measured at various locations on the subject's body when no pressure is delivered to the subject's body part. The device of claim 1 .

13. 1. A system for calibrating a blood pressure surrogate for use in monitoring a subject's blood pressure, the system comprising: a pressure delivery system comprising a cuff adapted to be attached to a body part of the subject, the cuff delivering pressure to the body part of the subject by inflating the cuff; a pressure sensor that acquires a time-dependent cuff pressure value during inflation of the cuff of the pressure delivery system attached to the subject's body part to acquire or estimate a BP measurement; a first sensor attached to a first portion of the subject's body to obtain a first time-dependent sensor signal related to the subject's heart rate during inflation of the cuff; a second sensor attached to a second portion of the subject's body and configured to acquire a second time-dependent sensor signal related to the subject's heart rate during inflation of the cuff; 10. The device of claim 1 for calculating calibration parameters for calibrating a BP surrogate from the BP measurements, the time-dependent cuff pressure values, and the first and second time-dependent sensor signals. A system comprising:

14. 1. A method for calibrating a blood pressure surrogate for use in monitoring a subject's blood pressure, said method comprising: obtaining a time-dependent cuff pressure value during inflation of a cuff of a pressure delivery system attached to a body part of the subject; obtaining a BP measurement; obtaining first and second time-dependent sensor signals measured at different locations on the subject's body during inflation of the cuff and related to the subject's heart rate; calculating, during inflation of the cuff, a pulse-related value from the first and second time-dependent sensor signals using a first feature of the first time-dependent sensor signal and a second feature of the second time-dependent sensor signal, the pulse-related value being a value of a pulse arrival time PAT or a value of a pulse transit time PTT; selecting pairs of pulse-related values ​​and corresponding cuff pressure values ​​for calculating calibration parameters of the BP surrogate, the pairs including the pulse-related values ​​and corresponding time-related cuff pressure values, and selecting only those pairs of pulse-related values ​​and corresponding cuff pressure values ​​for which one or more predetermined conditions regarding BP and / or cuff pressure are satisfied; calculating calibration parameters of a BP surrogate from the BP measurements and the selected pairs of pulse-related values ​​and corresponding cuff pressure values; A method comprising:

15. A computer program comprising program code means for causing a computer to carry out the steps of the method according to claim 14 when the computer program is run on said computer.

Images