METHOD AND DEVICE FOR DETERMINING THE CENTRAL SYSTOLIC BLOOD PRESSURE

Abstract

The invention relates to a method for determining the central systolic blood pressure of a patient, comprising the following steps: determining a peripheral, in particular brachial, blood pressure curve, applying a moving average filter to the blood pressure curve, applying an all-pass filter to the blood pressure curve, determining the central systolic blood pressure as the maximum of the all-pass-filtered and average-filtered blood pressure curves, as well as a device for performing this method.

Claims

1. A method for determining a patient's central systolic blood pressure, comprising the steps determining a peripheral, in particular brachial, blood pressure curve, applying a moving average filter to the blood pressure curve, applying an all-pass filter to the blood pressure curve, determining the central systolic blood pressure as the maximum of the all-pass-filtered and average-filtered blood pressure curve.

2. The method according to claim 1, characterized in that the peripheral blood pressure curve is determined in relative values, a systolic and a diastolic pressure of the patient are determined in absolute values and an absolute peripheral blood pressure curve is determined by calibration of the relative blood pressure curve with the systolic and diastolic blood pressure curves.

3. The method according to claim 1, characterized in that the patient's relative peripheral blood pressure curve, the systolic pressure, the diastolic pressure and/or the patient's pulse are all measured using a sphygmomanometer.

4. The method according to claim 1, characterized in that the moving average filter is applied in such a way that a fraction of the value of the scanning frequency (f.sub.S) is selected as a window width (N) of a time window of a group of measuring points of the peripheral blood pressure curve for measuring the peripheral blood pressure curve, wherein the denominator (n) for forming the fraction is between 3 and 6, in particular between 3.8 and 4.2.

5. The method according to claim 4, characterized in that the window width (N) with the width N=2m+1 is applied to the time axis around the respective measuring point of the measured peripheral blood pressure curve, wherein m is an integral variable for defining a number of measuring points.

6. The method according to claim 1, characterized in that the all-pass filter is applied by adding an offset to the average-filtered blood pressure curve.

7. The method according to claim 5, characterized in that the filtered blood pressure curve (y(k)) is determined according to a first filter equation (equation 1): $y\left(k\right)=A.Math.x\left(k\right)+B.Math.\frac{1}{2.Math..Math.m+1}.Math.\underset{i=-m}{\overset{m}{.Math.}}.Math..Math.x\left(k+i\right)$ or according to a second filter equation (equation 4): $y\left(k\right)=A.Math.x\left(k\right)+B.Math.\frac{1}{2.Math..Math.m+1}.Math.\underset{i=-m}{\overset{m}{.Math.}}.Math..Math.x\left(k+i-\right)$ where A is an all-pass coefficient and B is an average filter coefficient, where A+B=1, x(k) corresponds to the absolute peripheral blood pressure curve, k corresponds to the index of the measuring points of the peripheral blood pressure curve and i corresponds to the index of neighboring measuring points for determination of the average and also corresponds to a phase angle.

8. The method according to claim 7, characterized in that the all-pass coefficient (A), the average filter coefficient (B), the window width (N) and/or the denominator (n) are determined by the fact that a transfer function (H(f)) of a calibration transfer function that corresponds to the filter equation (equation 1) is approximated to a calibration transfer function.

9. The method according to claim 8, characterized in that a certain generalized transfer function determined on a patient population is used as the calibration transfer function.

10. The method according to claim 9, characterized in that the window width (N) is corrected with a pulse-dependent correction factor to yield a pulse-corrected window width (N.sub.1).

11. The method according to claim 10, characterized in that the pulse-corrected window width (N.sub.1) is calculated by a correction equation (equation 3), according to which: $N=\frac{.Math.f_s.Math.}{\mathrm{nk}}.Math.\frac{1}{{\mathrm{HR}}_{\mathrm{pat}}/{\mathrm{HR}}_{\mathrm{Ref}}}$ where f.sub.S corresponds to the sampling frequency in hertz, n is the denominator, k is a correction factor, HR.sub.Pat is the current heart rate of the patient and HR.sub.Ref is the reference heart rate.

12. A device for determining the central systolic blood pressure of a patient, comprising a blood pressure measuring device for determining a peripheral, in particular brachial, blood pressure curve and an evaluation unit which is configured for applying a moving average filter to the blood pressure curve, applying an all-pass filter to the blood pressure curve, determining the central systolic blood pressure as a maximum of the filtered blood pressure curve.

13. The device according to claim 12, characterized in that the blood pressure measurement device is suitable for determining the peripheral blood pressure curve as a relative value and for determining a systolic pressure and a diastolic pressure on a patient as absolute values, and the evaluation unit is configured to determine an absolute peripheral blood pressure curve by calibration of the relative blood pressure curve with the absolute systolic and diastolic pressure values.

14. The device according to claim 12, characterized in that the evaluation unit is configured to perform a method of determining a patient's central systolic blood pressure, comprising the steps determining a peripheral, in particular brachial, blood pressure curve, applying a moving average filter to the blood pressure curve, applying an all-pass filter to the blood pressure curve, and determining the central systolic blood pressure as the maximum of the all-pass-filtered and average-filtered blood pressure curve. wherein the moving average filter is applied in such a way that a fraction of the value of the scanning frequency (f.sub.S) is selected as a window width (N) of a time window of a group of measuring points of the peripheral blood pressure curve for measuring the peripheral blood pressure curve, wherein the denominator (n) for forming the fraction is between 3 and 6, in particular between 3.8 and 4.2

15. A blood treatment machine having an extracorporeal blood circulation for treatment of a patient's blood, comprising a device for determining the central systolic blood pressure of the patient according to claim 11.

Description

[0051] Additional advantages and embodiments of the invention are explained as an example on the basis of an exemplary embodiment with reference to the figures.

[0052] In these figures:

[0053] FIG. 1 shows schematically an arrangement of a device according to the invention for determining the central systolic blood pressure in combination with a blood treatment machine to which a patient is connected,

[0054] FIG. 2 shows a comparison of a transfer function, which is determined with a method according to the invention in comparison with a GTF (generalized transfer function) as a calibration transfer function, and

[0055] FIG. 3 shows a flowchart of an exemplary embodiment of the method according to the invention.

[0056] A blood treatment machine 1 according to the invention is suitable for performing dialysis, filtration and/or diafiltration, for example, in which a patient's blood is treated, sending it through bloodlines 2 into the blood treatment machine 1 and then back again. The blood treatment machine 1, as shown schematically in FIG. 1, has an evaluation unit 3, which in this exemplary embodiment is suitable for analyzing data that is directly associated with the blood treatment machine 1 and storing it as well as the data of a device 4 according to the invention for determining the central systolic blood pressure of the patient. With the evaluation unit 3, the machine 4 according to the invention determines a central systolic blood pressure as the maximum of a filtered blood pressure curve. The filtered blood pressure curve is obtained by the fact that the evaluation unit 3 applies an all-pass filter and a moving average filter to a peripheral blood pressure curve.

[0057] To do so, a peripheral blood pressure curve is obtained first. To do so, a sphygmomanometer having an inflatable cuff 6 that is placed on the patient's upper arm, is used as the blood pressure measuring device 5. Then a traditional oscillometric blood pressure measurement is performed and a diastolic blood pressure value, a systolic blood pressure value, an average blood pressure value and the pulse are obtained. The evaluation unit 3 is designed and configured to store these values. The scanning frequency f.sub.S, with which the measured blood pressure values are recorded and/or stored, amounts to 500 Hz in this exemplary embodiment. The evaluation unit 3 has a unit that is also known as a microcontroller and must be equipped with only a low storage and computation capacity. In this case, it is an LPC2368 with 16/32 bits.

[0058] After the traditional blood pressure measurement, the cuff 6 is pumped up to the diastolic blood pressure, and the cuff pressure is maintained for approximately 10 seconds. During this period of time, the pressure measurement device 5 detects pressure oscillations, which are also known as pulse curves.

[0059] These pulse curves are first freed of the steady component and high-frequency interference (>18 Hz) and then stored in the evaluation unit 3 for further processing.

[0060] Unsuitable pulse curves are sorted out from the recorded pulse curves in advance, where these pulse curves include obvious outliers, as well as pulse curves that have been falsified due to movements by the patient, for example.

[0061] Then evaluation unit 3 forms an average pulse curve, which represents a peripheral blood pressure curve with which the additional process steps according to the invention are carried out.

[0062] The evaluation unit 3 calibrates the peripheral blood pressure curve with the systolic blood pressure curve and the diastolic blood pressure curve that are measured in advance. Thus, an absolute peripheral blood pressure curve is obtained for the patient and is stored temporarily by the evaluation unit 3.

[0063] The absolute peripheral blood pressure curve x(k) is filtered using the filter equation (1)

[00006]$\begin{array}{cc}y\left(k\right)=A.Math.x\left(k\right)+B.Math.\frac{1}{2.Math..Math.m+1}.Math.\underset{i=-m}{\overset{m}{.Math.}}.Math..Math.x\left(k+i\right).& \left(1\right)\end{array}$

[0064] The filtered blood pressure curve y(k) is determined by the all-pass filter component A x(k), where A is an all-pass coefficient, and the average filter component is applied to the peripheral blood pressure curve, where B is the average coefficient and 2m+1 is the window width N of the time window, which is applied to the moving average value formation. It has been found that the window width with the amount of N=f.sub.S/n yields the most accurate values for the central systolic blood pressure at a denominator n value of 4 when the heart rate is 70 bpm. A window width of 125 is obtained with f.sub.S=500 Hz. With the preferred application of a symmetrical time window, the value 62 is obtained for m, because N=2m+1. Thus, in the present exemplary embodiment, a time window of 125 measured values, the average of each being formed, is plotted successively on the time axis of x(k), the absolute peripheral blood pressure curve of the patient. The resulting averages are multiplied times the coefficient of the moving average filter, also referred to as the average coefficient, and provided with the all-pass filter component. The evaluation unit determines the maximum as the central systolic blood pressure from this filtered blood pressure curve y(k), calculated by the evaluation unit 3, and stores it temporarily.

[0065] After the average curve has been determined over the respective time window of the window width N, the measured values that are no longer needed and intermediate calculations can be deleted or overwritten. This reduces the storage capacity burden on the microcontroller of the evaluation unit 3.

[0066] In this exemplary embodiment, the blood treatment machine optionally has a device 7 for determining the blood vessel stiffness; this device can be combined with the device for determining the central systolic blood pressure. The device for determining the blood vessel stiffness measures and evaluates the pulse wave velocity by pulse curve analysis as an indicator for vascular stiffness. This is also done in the evaluation unit 3 in the present case. Together with the vascular stiffness, the value of the central systolic blood pressure allows even more accurate information about cardiovascular diseases, if any. By combining the devices, an easy-to-operate overall unit that is highly compact is achieved with which these characteristic data are determined and detected with little effort during a blood treatment.

[0067] By approximation of the filter transfer function according to equation (2) to a calibration transfer function, for example, it is possible to determine A, B and N.

[00007]$\begin{array}{cc}H\left(f\right)=A+\frac{B}{N}.Math.\frac{\sin \left(.Math..Math.\mathrm{Nf}/f_s\right)}{\sin \left(.Math..Math.f/f_s\right)}& \left(2\right)\end{array}$

[0068] For example, a generalized transfer function determined on a patient population as reported by Y. T. Shih et al. (Comparison of two generalized transfer functions for measuring central systolic blood pressure by an oscillometric blood pressure monitor in: Journal of Human Hypertension 3, 204-210 (2013)) may be used as the calibration transfer function. FIG. 2 shows the inverted amplitude curve of this generalized transfer function as the curve labeled with reference numeral 10.

[0069] This yields, for example, a transfer function 20, such as that shown in FIG. 2, in addition to the calibration transfer function GTF at values of A=B=0.5, F.sub.S=500 Hz, n=4, N=125. This shows the very good correspondence in the low frequency range with f4 Hz in particular.

[0070] The approximation described here need be performed only once for the patient. Therefore, it may be done in a separate unit of evaluation unit 3, which is not shown here but has a larger computation and storage capacity than the evaluation unit 3 itself. The approximation may also be performed on a completely separate computation unit, such as a personal computer. In the following sessions, the device 4 may be operated with stored values A, B, N, n, which are used in the filter equation (1), used for determining the central systolic pressure alone. The stored values may be retrieved by inserting a patient card into the evaluation unit 3, for example, which makes it possible to call up patient-specific data from the separate unit.

[0071] In the present exemplary embodiment, the patient population from which the calibration transfer function originates, had an average pulse of approximately 70 bpm, which corresponds to a heart rate of 1.17 Hz.

[0072] The patient in the exemplary embodiment also coincidentally has a pulse of 70 bpm in the measurement of the peripheral blood pressure curve that was performed. Therefore, a pulse-dependent correction of the described determination of the central systolic blood pressure is necessary.

[0073] However, in a second measurement of the peripheral blood pressure curve in another session for blood treatment using the blood treatment machine according to the invention, the patient has an elevated pulse of 90 bpm, which is thus outside of the predefined core range with a lower limit HR.sub.S of 50 bpm and an upper limit HR.sub.H of 85 bpm, so it is advantageous according to equation (3) to determine a pulse-corrected window width N.sub.1, which is derived from the window width N:

[00008]$\begin{array}{cc}{N}_1=\frac{f_s}{\mathrm{nk}}.Math.\frac{1}{{\mathrm{HR}}_{\mathrm{pat}}/{\mathrm{HR}}_{\mathrm{Ref}}}& \left(3\right)\end{array}$

[0074] The pulse-corrected window width N.sub.1 of the quotient f.sub.S/n.sub.1 is therefore calculated, where n.sub.1 is a pulse-corrected denominator. The pulse-corrected denominator is itself calculated by multiplying the denominator n from the determination of the window width N for the patient population that has a reference heart rate HR.sub.Ref with a quotient HR.sub.Pat/HR.sub.Ref, where HR.sub.Pat is the current heart rate of the patient in determination of the peripheral blood pressure, i.e. 90 bpm in the current case. The reference heart rate HR.sub.Ref is the same heart rate as in the determination of the calibration pressure curve, i.e., 70 bpm here. This was calculated as the average heart rate of the patient population. Thus n.sub.1 is calculated as 490/70=5.1. The corrected window width N.sub.1 is obtained in turn at a scanning frequency of 500 Hz, rounded to the number of 98 measuring points in the time window.

[0075] According to one variant, the pulse-dependent denominator n.sub.1 is again multiplied times a correction factor K, which permits a further adjustment of the filtered blood pressure curve to take into account additional factors such as medication, disease symptoms, etc. In the present case, K is assumed to be 1, so as not to perform a corresponding correction.

[0076] FIG. 3 shows as an example in one flowchart the sequence of the process according to the invention, where the values of A, B, N are already stored, in this case available in the memory. In this example, a heart rate of 70 bpm is selected as the reference heart rate HR.sub.Ref. The lower limit HR.sub.S of the core region is 50 bpm and the upper limit HR.sub.H is 85 bpm.

[0077] All the steps just described may be performed by the blood treatment machine 1 as fully automatic or semiautomatic processes. They may also be performed with the device 4 for determining the central systolic blood pressure in a fully automatic or semiautomatic process, even without the blood treatment machine 1. In addition, they may also be performed with one or two of these devices by at least partial manual operation. However, the process described for this exemplary embodiment may also be carried out in principle with measurement equipment, controllers and/or regulators and other devices in addition to those explained here as examples.