SE438593B - DEVICE FOR MONITORING OF SYSTOLITICAL PRESSURE - Google Patents

Description

15 20 25 30 35 H0 7714589-4 for example by measuring the occlusion that occurs in a blood vessel in the outer ear when pressure is applied to it. These oscillometric methods generally define systolic pressure as the maximum applied pressure at which threshold oscillations are observed to occur. In a typical mercury manometer and pressure cuff, this pressure would then be the highest pressure noted by the operator fluctuating at the top of the mercury column as the pressure in the cuff was slowly and relatively uniformly reduced. However, there are some inaccuracies associated with this method of determining threshold oscillations, since the inertia of the mercury column does not allow it to respond significantly to pulses of small width.

Each of the above-mentioned techniques or devices for measuring systolic pressure has disadvantages of some kind, such as inaccurate response to pressure pulses of small width, or the requirement for sophisticated and/or expensive measurement equipment.

US Patent No. 4,009,709 describes a method and apparatus for automatically and relatively easily obtaining accurate systolic blood pressure measurements, thereby eliminating the disadvantages of the above-mentioned devices.

This device determines systolic pressure by applying pressure to a living test subject by changing the pressure in a pressure cuff attached to said subject adjacent a blood vessel; by measuring on the cuff a quantity proportional to a time-dependent fluctuating component representative of the pulsatile pressure in the blood vessel, which quantity is proportional to the amplitude of the pulsatile pressure; by determining the maximum value attained by said quantity as the applied pressure changes; by storing a representation of the maximum value; by determining when said quantity is substantially equal to about half the maximum value for an applied pressure greater than the pressure applied when the maximum value occurs or results; and by reading out the applied pressure corresponding to the quantity being substantially equal to about half the maximum value, the pressure corresponding to the systolic pressure of the subject. The signal from the pressure cuff comprises a fluctuating quantity proportional to a sum, which sum comprises a time-dependent fluctuating component proportional to the amplitude of the pulsatile pressure in the blood vessel, which component has a steeply rising wavefront between end-diastole and systole plus the selectively variable pressure applied externally adjacent the blood vessel through the cuff.

In the apparatus of the above-mentioned U.S. Patent 4,009,709, the oscillating signal is considered to be proportional to the amplitude of the pulsating pressure in the blood vessel, and a peak-to-peak detector then makes amplitude measurements which are utilized to complete the signal processing. However, random and uncontrollable deviation from the assumed linearity of the pressure ramp can introduce errors into this amplitude determination. If a large disturbance in the cuff pressure ramp occurs, this filter requires a considerable time for recovery and may allow some variation in the baseline from which the fluctuating signal, which is proportional to the amplitude of the pulsating pressure in the blood vessel, is measured, resulting in an erroneous output signal from the peak-to-peak detector.

The present invention provides a solution to the problems associated with temporary disturbances in the applied pressure ramp of the cuff in a systolic blood pressure monitoring device of the type described in the aforementioned U.S. Patent 4,009,709.

The apparatus of the invention is particularly useful in blood pressure monitoring equipment where it is desired to know the peak-to-peak value of each blood pressure pulse over the systolic rise (steeply rising wave front) from end-diastole to systole. Such knowledge of the peak-to-peak value of each blood pressure pulse is utilized in a preferred embodiment to determine the systolic pressure of a living test subject. Input signals may be obtained from a pressure cuff or other device.

Means are provided for determining the maximum value reached by successive identified (i.e. valid) integral values, which values correspond to the peak-to-peak value of the respective pulses.

The maximum determined integral value is stored and means are provided for determining when such integral value is substantially equal to about half the maximum integral for an applied pressure greater than the pressure applied by the cuff when the maximum integral value is obtained, the applied pressure at which the particular half maximum integral value occurs being read out as the systolic pressure.

It is thus an object of the invention to provide an improved apparatus and method for determining systolic pressure. Included in this object of the invention is the provision of an improved apparatus and method which determines systolic pressure with increased accuracy.

Another object of the invention is to provide an improved apparatus and method for identifying and quantizing a substantially periodic, steeply rising wavefront in the presence, or possible presence, of low-amplitude interference, and possibly also low-frequency, high-amplitude interference.

These and other objects and advantages of the invention will become apparent from the following detailed description of the preferred embodiment taken in conjunction with the drawing.

Fig. 1 of the drawing illustrates prior art technology for measuring the peak-to-peak value of the pulsatile pressure in a blood vessel.

Fig. 2 shows a technique for determining the peak-to-peak amplitude of the pulsatile pressure in accordance with the invention.

Fig. 5A shows a typical oscillometric envelope of the pulsatile pressure in a blood vessel.

Fig. SB shows the time derivative of the waveform in Fig. SA.

Fig. 5C shows a controlled timing diagram in accordance with the basic apparatus of the invention.

Fig. 3D shows the time intervals obtained from the path shape in Fig. 3B in accordance with the basic apparatus of the invention.

Fig. 4 shows in a block diagram the apparatus according to an embodiment of the invention.

Fig. 5 shows a graph of the gain versus frequency of a differentiating network used in a preferred embodiment of the invention.

FIG. 6A shows an enlarged portion of the time derivative waveform of FIG. 6B, illustrating a validation threshold level and the timing of various control states associated therewith in accordance with the embodiment of FIG. 4.

Fig. 6B shows a control state diagram in accordance with Fig. 6A and the embodiment according to Fig. 4.

Fig. 7 shows a flow chart or decision tree of the control frequency used in the embodiment of Fig. 4 between successive heartbeats.

Fig. 8 shows, in an abbreviated block diagram supplemented by Fig. 4, the apparatus and method of another embodiment of the invention. 10 15 20 25 30 35 40 7714589-4 Fig. 9 illustrates a technique similar to that of Fig. 2 for determining the peak-to-peak amplitude of the pulsatile pressure, and also includes threshold detection means for identifying particular "above zero" passages of the waveform derivative as valid systolic rises.

In Fig. 1, a functional block diagram of certain parts of the apparatus for measuring systolic pressure described in the above-mentioned U.S. Patent 4,009,709 is shown. In particular, the functional blocks in Fig. 1 show a filter network 100 whose output is connected through an amplifier 102 to the input of a peak-to-peak detector 104. The filter network 100 receives an input signal 106a on input conductor 106. The input signal 106a includes a slowly increasing ramp signal indicative of the applied pressure and has superimposed thereon a time-dependent fluctuating component representing pulse pressure in the subject's blood vessels, which component representative of pulse pressure has a steeply rising wavefront relative to the other components during the rise from end-diastole to systole. The filter network 100 was typically constructed in such a way that its output waveform 100a had the linear effects of the pressure ramp eliminated from it, but any temporary and uncontrollable deviation from the assumed linearity of the pressure ramp would introduce errors into the signal 100a.

For example, if a large disturbance occurred in the cuff pressure ramp, a considerable time was required for the filter 100 to reset, which could allow for some variation in the baseline 100b (dashed) from which the fluctuating signal, proportional to the amplitude of the pulse pressure in the blood vessel, was measured. Thus, each time the peak-to-peak detector 104 was actuated in response to sampling signals 108, the resulting output signal 100a appearing on conductor 110 included the peak-to-peak errors introduced by the variation in baseline 100b.

In accordance with the invention, as generally shown in FIG. 2, an input signal Z06a having a waveform identical to waveform 106a in FIG. 1 is applied to input conductor 206 of differentiator network 200. Differentiator network 200 is constructed to provide single differentiation of signal 206a within the frequency range corresponding to the frequencies of the fluctuating signal proportional to the amplitude of the pulse pressure in the blood vessel and double differentiation of the input signal below this frequency range in order to eliminate the offset effects of a linear pressure ramp and any very low frequency disturbances that may have occurred in the otherwise linear pressure ramp.

A filter or differentiating network with dc characteristics required of the network 200 has the gain/frequency characteristics shown in Fig. 5, where the gain curve exhibits a slope of -0dB per octave in the frequency range f - fz and a slope of ~12db for frequencies below f1. The frequency range fï - fz corresponds to the bandwidth of the Pac signal which includes the fluctuating quantity representative of the pulsatile pressure. This portion of the input signal Z06a representative of the pulsatile blood pressure is differentiated and appears at the output of the differentiating network 200 as signal 200a, hereinafter designated P.

This signal P (200a) is applied through the amplifier 202 to the input of an integrator 204 which, by integrating the signal P over a predetermined interval during each pulse, provides an output value corresponding to the peak-to-peak pressure of each blood pressure pulse. The sample-and-hold circuit 205 associated with the integrator 204 serves to sample the value appearing at the output of the integrator 204 at the end of each integration interval and to hold this value for a transition period until the integration of the next pressure pulse begins. Control of the integrator 204 and the sample-and-hold circuit 205 is provided by the reset-integration-hold signal 208 which controls the integration period and serves to reset the integrator before each new integration. The output signal from the sample-and-hold circuit 205 appears on line 210 as waveform 210a with a magnitude corresponding to the area under the portion of the waveform 0 being integrated.

With regard to Figs. 3A and 3D which are intended to explain the theory of the invention, it should be noted that according to the above-mentioned US Patent 4,009,709, the systolic pressure is equal to the applied cuff pressure when the fluctuating quantity is approximately equal to half of its maximum value. The maximum value of the fluctuating quantity is determined by measuring the diastole and systole in successive blood pressure pulses.

The pulse that exhibits a maximum P-P amplitude is taken as the maximum value and the applied cuff pressure is further increased so that the P-P amplitude is reduced, and the system determines the EGUOW notation of the applied cuff pressure at which the P-P pressure becomes half the P-P maximum.

Fig. 3A shows the time-dependent fluctuating component, P dc', which is representative of the pulsatile pressure in a blood vessel.

The "root" ED of each "valley" in the Pac waveform corresponds to the time of diastole, or more specifically end-diastole of a heartbeat, and the waveform peak SP corresponds to the systole time of the heartbeat. As described above, the signal from the cuff is differentiated to remove the applied pressure ramp and low-frequency random noise, resulting in the derivative P of the Pac waveform, as shown in Fig. 3B.

Since the waveform Pag exhibits a zero slope at both end-diastole (ED) and peak systole (SP), the derivative waveform P takes the value zero at each of these times. Furthermore, since Pac has a positive slope during the systolic rise between ED and SP, the waveform D lies above the zero reference line during this interval. The zero crossing points ED and SP of the waveform D correspond to the points of maximum amplitude between successive Pag pulses and thus the area under the waveform P and above the zero reference line between end-diastole ED and peak systole SP will give a value corresponding to the P-P value of the respective blood pressure pulse. This area is determined by integrating the section "above zero" of the waveform D. It should be noted that end-diastole (ED) also essentially corresponds to the beginning of the rise to peak systole (SP).

Fig. 3C shows a control signal substantially analogous to signal 208 in Fig. 2, which resets the integrator prior to the integration interval, then integrates signal D over the integration interval, and finally samples and holds the value of the integration as a representation of the P-P value of the respective blood pressure pulse. This sequence of control events is repeated with the reset operation indicated by R, the integration operation represented by S, and the sample and hold operation represented by S+H. The sampled integral may be held for longer than is indicated by the short duration of signal S+H in Fig. 3C.

The results of integrating the waveform D at the boundaries of ED and SP are shown in Fig. 3D. The magnitude of the integral at the time of the systolic peak SP corresponds to the P-P value of the respective blood pressure pulse.

To implement the idea of integrating waveform D over the interval of systolic rise to obtain the respective values P-P for the respective blood pressure pulses or heartbeats, standard circuits can be used for detecting when waveform D crosses the zero reference line in the positive direction to begin integration and determining when it crosses the zero reference line in the negative direction to terminate integration and/or perform the sample and hold function. The integrator may be reset immediately after the sample and hold operation and preferably continue to the next positive-going zero-line crossing of É. The resulting integral may then be considered as representing the P-P value of the respective pulse. However, certain characteristics of the waveform Pac and/or the presence of signal artifacts during the diastolic pressure drop may result in P briefly appearing above the zero reference line during times other than between end-diastole and the systolic peak. For example, if, as shown in Figs. 3A and 3B, random muscle activity introduces a "high frequency" signal artifact (ART)sxrax prior to end diastole when the slope of the Pac waveform is relatively flat, then the derivative-É waveform may exhibit some of the artifact as a "greater than zero" value and result in the experimental values shown parenthetically in Figs. 3C and 3D.

According to a feature of the invention, as generally shown in Fig. 9, a threshold level is provided for discriminating between the values P which are greater than zero and which belong to the systolic rise, and those signals, such as artifacts and the like, which do not belong to the systolic rise. The magnitude of the signal É associated with the systolic rise is generally significantly greater than the magnitude of any other (greater than zero) portion of the signal (such as from artifacts), allowing discrimination between such signals. This determination that the waveform É exceeds the threshold level during a particular passage "above zero" serves to validate the integration of that passage (greater than zero) between its respective ED and SP limits.

In FIG. 9, in which the components functionally identical to the corresponding components in FIG. 2 have the same designations, the input signal 206a is differentiated by the differentiation network 200 to produce the waveform É which is sent through the amplifier 200 to the respective inputs of the integrator 204, a threshold detector 912 and a zero-crossing detector 907. The zero-crossing detector 907 may correspond to means, not shown in FIG. 2, which determined the integration interval and resulted in the control signal 208. The threshold detector 912 determines a signal magnitude threshold above which the waveform É is assumed to indicate a valid systolic rise. When the incoming waveform É exceeds the threshold level of the detector 912, a signal is produced to the input of the validating logic circuit 91k, indicating that the threshold level has been exceeded. Likewise, the validating logic circuit 91k receives an input signal from the output of the zero crossing detector 907 to define when the waveform É crosses a zero reference line in the positive direction and also in the negative direction. The output 9081 of the validating logic circuit 91U is applied to the reset input of the integrator 20H to reset it at least substantially at the beginning of each desired integration period, starting with the waveform P crossing the zero reference line in the positive direction. The output signal 908 from the validating logic circuit 914 is applied to the "sample input" of the optional sample and hold circuit 205 and serves to store the integral value accumulated by the integrator 26% between the positive-going and negative-going zero-crossings of the waveform f only if the threshold detector 912 has given an indication that the waveform f during that interval was indeed a valid systolic rise. The output 910 of the sample and hold circuit 205 varies from the output of 210 in Fig. 2 only when the latter includes an output value without validity, representative of a systolic rise when in fact there was only an artifact.

While a fixed magnitude threshold above the zero reference line could be utilized if the "above zero" portion of the waveform É did not vary in magnitude during successive pulses, this is not the case, particularly when using the present oscillometric blood pressure monitoring technique, in which the ac pressure signal Pac is increased from a small amplitude at low applied pressure to a large amplitude at a greater applied pressure and then to a smaller amplitude at an even greater applied pressure.

Therefore, the threshold level, shown as 'HKEK in Fig. 35, is chosen to be a function of the systolic rise portion of the signal É over one or more of the immediately preceding blood pressure pulsations. The increase (and then decrease) in the magnitude of successive systolic rises in the waveform Û occurs sufficiently gradually, and the relative amplitude of possible non-systolic rise components, "above zero", of the waveform Û are sufficiently small that a dynamic threshold corresponding to 50% of the maximum amplitude "above zero" of the systolic rise of the waveform É during the preceding pulse shall be considered here as sufficient for identifying only those "above zero" portions of the waveform Ü which actually belong to the systolic rise.

It should be noted that the dynamic threshold level could be determined by summing and weighting several previous systolic rise portions of the waveform U, in which case the threshold IRUSK could be at a preselected level higher or lower than 50 2 of 10 15 20 25 30 35 H0 '7714589-4 10 the magnitude of the immediately preceding systolic rise. An analogous example of a dynamic threshold detector of the type suitable for use herein is described in more detail in U.S. Patent 3,590,811. Digital means for determining a dynamic threshold level will be described in more detail below.

Reference is now made to Figs. 1, 6 and 7 for a more detailed description of the apparatus and method with respect to a feature and an embodiment of the invention. The apparatus is described with reference to the functional block diagram of Fig. 4, which shows the digital processing of the analog signal received from the transformer 23. However, analog implementation is also possible. In particular, discrete electronic components, discrete digital chips, microprocessor technology and structure or digital computer may be used.

Fig. 6 and 7 show a state diagram and a flowchart, or decision tree, associated with the processing of the signal É between successive blood pressure pulses, corresponding to successive heartbeats. Generally speaking, the signal processing steps of the improved blood pressure monitoring apparatus and technique of the invention, shown in Figs. 6 and 7, correspond to that portion of the apparatus of Fig. 1 which integrates the signal É between the boundaries ED-¿>SP.

The arm 11 of the test subject with the artery 13 is surrounded by a typical blood pressure cuff 15. In the usual way, the brachial artery in the upper arm is used for this type of blood pressure measurement. Via lines 17 and 21, the pump 19 and the pressure transducer 23 are connected to the cuff. The transducer 23 has such a transfer function that its electrical output signal is essentially representative of its pressure input up to the limit of the information contained in the pulse pressure.

The pressure transducer serves to measure the pressure in the cuff, which pressure is the sum of the pressure supplied by the pump and the pressure portion or fraction generated by the blood pressure fluctuation in the artery, as shown by waveforms 106a and 206a in Figs. 1 and 2, respectively. The fluctuating portion of the output of the transducer 23 represents the amplitude of the pulsatile pressure. The output signal of the converter 23 goes, as shown by line 24, to an input of the multiplexing switch 25. The output signal of the converter 23 also goes, as shown by line 26, through the normally closed switch 27 to the input of the differentiating network 28. The output signal from the differentiating network 28 goes, as shown by line 29, to the gain amplifier 30 and goes, as shown by line 31, to the other input of the multiplexing switch 2. 10 15 20 25 30 H0 7714589-4 11 The differentiating network 28 differentiates the input signal over the bandwidth f1-f2 of the signal Pac, as shown in Fig. 5, and also provides double differentiation of the frequencies below f1.

In this way, essentially the only signal appearing on line 31 is the differentiated (Û) representation of the waveform Pac.

The output signal from the multiplexing switch 25 goes, as shown by line 32, to the analog-to-digital (A/D) converter 33. The output signal from the A/D converter 33 goes, as shown by line HB, to inputs of gates H0 and H2, respectively. A clock 3H generates timing pulses which, as shown by line 35, go to a timing controller 36, which controls the switching of the multiplexer 25, the conversion of the analog signal to a digital signal, and the control of the gates H0 and 42. An output signal from the timing controller 36 goes, as shown by line HH, to the multiplexer 25, the A/D converter 33, and the second input of the gate 40 to control the conversion of the signal É appearing on line 31 to a digital form, which is then supplied to the gate H0 via line H8. Another output from the timing controller 36 is, as shown by line H6, to the multiplexer 25, the A/D converter 33, and the second input of gate H2 to control the conversion of the analog signal from converter 23 to a digital form which is applied to gate H2. The opening signals on lines HH and M6, which are connected to the inputs of gates H0 and H2, are of sufficient duration to allow the digitally converted data associated therewith appearing on the second input of the respective gate to pass through the gate in question. It should also be noted that the control signals represented by lines MH and Hb, as shown here, exist only one or the other so that the data appearing on line 31 or 2% is coupled to the associated gate H0 and M2, respectively.

The period between successive blood pressure pulses is normally of the order of 800-1000 ms, with the systolic rise taking up about 10-20 s of each period. The signal P can be integrated over the systolic rise portion by sampling a sufficient number of incremental portions (Éí) of the waveform Û to closely approximate the area under the waveform É. Typically, 10-20 samples during the 100-200 ms of a typical systolic rise are sufficient to provide the required number of Éi increments, and the speed and control signals on line HH are selected accordingly. At least once during each blood pressure pulse, a control signal acts on line 46 to convert the signal on line ZH to digital form and to input the digitized signal into the calculation and averaging unit 83 as a measure of the applied pressure in the cuff 15 as will be described hereinafter.

Each time a conversion and opening signal appears on line 44 from timing controller 36, a correspondingly timed incremental sample Di of waveform P appears at the output of gate 40. This incremental sampling of waveform D preferably occurs repetitively during the blood pressure pulse and at least during the interval in which waveform D exceeds the zero reference level. The timing of the system also provides for a sampling of the waveform on line 2U at least once during each pressure pulse at a time (S) that does not conflict with the sampling of waveform D on line 31.

Each sample Éí is supplied from the gate H0, which is represented by line 50, to the respective inputs of the gate 52, the comparator SH, the comparator 56 and the comparator 58. The input É¿ to the comparator SH is compared with a so-called zero reference value ~(P,) which is represented by line 60. The output of the comparator 54, which is represented by line 62, is set to a level representative of 1, if the value Éi appearing at the input of the comparator is greater than (>) the zero reference value PZ, and to a level representative of zero, if it is equal to or less than PZ. The reference value P2 appearing on line 60 is intended to correspond to the zero reference value in Fig. 3B. The actual voltage appearing on the line 60 is adjustable and appears, for example, as the voltage on the movable contact of a potentiometer 20 whose end terminals are connected to voltages higher and lower than a voltage that could correspond to the voltage Pz in question. In order to accurately determine the correct voltage setting Pz, the switch 27 can be temporarily actuated in such a way that a "zero value" appears as the output signal Di from the gate HU on the line 50, and the value Pz on the line 60 is adjusted to match that on the line 50. It should be noted that the voltage level P2 on the line 60 appears in digital form (via the A/D conversion circuit, not shown) for comparison with the digitized value fi appearing on the line 50.

As soon as the value Éi exceeds P7, the unity gain on line 62 opens gate 52 to pass the particular value Pi to the gate output, which is represented by line GH. Thus, the data appearing on line SH is representative of a value Pi greater than PZ.

Each such value Pi appearing on line 64 is applied to and accumulated in an accumulator 66 which, by summing successive increments Pi, integrates the waveform Pi over the relevant interval from which the incremental samples are taken. The accumulated sum of the increments Pi is available at the output of accumulator 66 which is represented by line 68, after addition of each successive increment Pi to the sum. This output signal 68 from accumulator 66 is applied as an input to gate 78 for controlled application to the "beat-to-beat" circuit to be described hereinafter.

The interval over which the accumulator 66 accumulates the increment ON is at its later end bounded by the increment Pi being equal to or less than (g) the zero reference value PZ. To identify this boundary, the value Pi on line 50 and the value Pz on line 60 are applied as the two input signals to the comparator 56, the output of which is represented by line 70 and is set from a zero level to a one level when Pi is equal to or less than PZ.

A bistable element 72 receives the output of comparator SH on line 62 which goes to a "set" input of the same for setting its Q output when Pi first exceeds Pz and the output of comparator SH goes from zero to one. Similarly, the output of comparator 56 on line 70 is applied through delay element 7H to the "reset" input of bistable element 72 so that its output Ö is switched from zero to one substantially at the time when Pi is no longer greater than zero. In this manner, bistable element 72 serves to identify the interval during which the Pi increment of waveform P exceeds Pz.

The output signal Q of the bistable element 72 passes through the line 76 to the "reset" input of the accumulator 66 for emptying or resetting the accumulator to zero when the output signal Ö is one. This resetting of the accumulator 66 is intended to include not only the time when the waveform Pi first goes to or below the reference level P2, but also during the time when it remains at or below Pz. In this way, the accumulator 66 is prevented from integrating any portion of the waveform P other than that which exceeds the reference level Pz. The delay element 7H, which may comprise a monostable flip-flop or the like, serves to momentarily delay the reset of the bistable element 72 by an interval which is less than the interval between successive samples Pi but which is large enough to delay the reset of the accumulator 66 until the final integral value on the line 58 is transferred through the gate 78, as will be explained hereinafter.

An opening pulse on line 80 is applied to gate 78 at the end of each systolic rise interval such that the time integral of waveform D between boundaries ED and SP appears at the gate output and goes, as shown by line 82, to the beat-to-beat signal processing circuit. In order to ensure that the opening pulses on line 80 are generated only at the end of systolic rise and not at the end of any other "above zero" portion of waveform P which may be associated with the dicrotic notch or the like, the Pi increments on line 50 are compared in comparator 58 with a stored threshold value on line SH for setting a bistable element 86 if the threshold value is exceeded.

The stored threshold value on line 84 is determined to be one-half the value of the maximum magnitude Pi increment during the systolic rise of the preceding blood pressure pulse, as will be described hereinafter. The output of comparator 58, as shown by line 88, is fed to the "one-ring" input of flip-flop 86, and goes from zero to one when the first Di increment exceeding the stored threshold occurs.

This sets the output Q of the flip-flop 86 to the one level, at which it remains until reset by a signal on line 70, which goes to the "reset" input of the flip-flop. The zero-to-one signal transition on line 70 occurs when Pi first becomes ¿¿Pz. At this later instant, the output Ö of the flip-flop 86 goes from zero to one, which output signal goes as shown through line 90 to the inputs of the pulse generators 92 and SH, respectively.

The pulse generator 92 is responsive to the zero-to-one input transition to generate the output opening pulse represented by line 80 to pass the integral value appearing on line 68 through gate 78 to its output 82. It should be noted that the delay provided by delay element 7H is sufficient to permit transmission of the integral value on line 68 through gate 78 prior to the reset of accumulator 68 by the "reset" signal on line 76.

In order to provide a dynamic threshold against which the waveform É is compared for validation or rejection of the "above zero" passage of the waveform as a systolic rise, the signal on line 64 representative of the Éí increment being > PZ is applied to one input of comparator 98 and one input of gate 96. Comparator 95, represented by line 97, controls gate 96. Gate 96 serves to allow selected increments Pi to be loaded into memory unit 98 via line 99. The Éi value stored in storage unit 98 is applied, via line 120, to the other input of comparator 95. The output of comparator 95, represented by line 97, goes from a zero level to a gate-opening one level when the incoming éi increment is of greater magnitude than the Pi increment retained in storage unit 98. In this way, the Pi increment retained in storage unit 98 will represent the maximum size Pi increment at that time within the particular "above zero" portion of the P-waveform.

At the time the É waveform again crosses the P2 reference level downwards, the value stored in unit 98 is representative of the maximum êi (Éí:max.) occurring during the immediately preceding "above zero" P waveform. The retained Pi maximum value is supplied to the "divide-by-two" circuit 12? via line 12H for development of the threshold value to be entered into storage unit 126. However, the actual division of the Pi maximum value and its storage in unit 126 will not occur until receipt of a trigger or opening signal obtained on line 128 from the output of pulse generator 9M. As with the opening signal 80 from pulse generator 192, the opening signal 128 is obtained only if the current "above zero" passage of the E waveform has been identified as a valid systolic rise after comparison with a threshold established by the preceding systolic rise, and occurs at the time the E waveform passes below the PZ reference level. At this time, the new value Eimax , 2 is entered via line 130 into threshold storage circuit 126 to become the new threshold value with which the next "above zero" passage of the P waveform is compared. The "reset" signal represented by line 76 also goes to the "reset" or empty input of the storage unit 98 shortly after the new threshold value has been stored in the unit 126 for emptying the storage unit 98. It should be noted that the reset of the fi-max storage unit 98 occurs not only when an "above zero" passage of the P-waveform has been recognized as a systolic rise, but also after every other "above zero" passage of the a-waveform, which latter operation is required to avoid the introduction of an artifact fi value in the determination of fi-max and the development of a new threshold value.

Now considering the value represented by line 82 which comprises the time integral of the E waveform between the ED-SP boundaries, this value is available at the end of each respective systolic rise for application to the beat-to-beat circuit described in the above-mentioned U.S. Patent 4,009,709. Briefly, on line 82 through an optional averaging unit 39 (as stated, the P time integral quantity appearing is represented by dashed lines) which can average several successive blood pressure pulses, for example four pulses, and then through line H9 to inputs of comparator H3, gate H7 and systolic comparator 63 respectively. Comparator H3, represented by line H5, controls gate H7. Gate H7 serves to allow averaging unit 39, as indicated by LCD-nfifim 39, to load the selected value of the (possibly averaged) ε-systolic rise integral representative of the P-P value of the pulse, represented by line 51, into storage unit 53.

The value of the quantity stored in the storage unit 53 is supplied to the comparator H3, which is represented by the line 55. In the comparator 43, stored previous representations of the preliminary maximum value of the systolic rise integral are compared with the current values of said quantity entered into the comparator H3, which is represented by the line H9. When the value of said quantity supplied to the comparator H3 through the line H9 is greater than the quantity preliminarily stored in the storage unit 53 and supplied to the comparator H3 through the line 55, the gate H7 is activated by the comparator 43 via the line H5 and the greater value of said quantity replaces the preliminary maximum value in the storage unit 23.

The preliminary maximum value of said P-P quantity is fed, through line 57, into a halving unit (division by two) 59 in which it is divided by two. The halved value is fed, through line 61, to the systolic comparator §3. The (current) current value of the (possibly averaged) P-systolic rise interval is fed to the systolic comparator 63 through line H9. When the systolic comparator 63 determines that the magnitude supplied to it through line H9 is less than or equal to half of the preliminary maximum value supplied to it through line 61, the systolic comparator 63, through line 67, commands the switching device 69 to stop the pump 19 and to drain the cuff 15 through solenoid control valve line 20, the stop and drain command being represented by line 71.

The coupling device 69 and the systolic comparators 63 also command, via lines 73 and 93, the interpolation unit 75 to interpolate between the values of the applied pressure, i.e. pressure applied to the cuff 15 by the pump 19, in order to determine the exact applied pressure corresponding to said quantity (P-P) being equal to about half of said maximum value.

Applied pressure values are supplied to the interpolation unit 75 via lines 77 and 79. Line 77 represents the input of the applied pressure value for measurement just before said quantity becomes less than half of the maximum value, and line 79 represents applied pressure when the quantity was equal to or slightly less than half of the maximum value. In other words, line 79 represents the more recent value of applied pressure.

These applied pressure values are obtained by inputting the digitized waveform of line ZH from converter 23 through gate H2 to a "last" (most recent) storage unit 83 which averages the signal over the last "nb" (e.g., four) most recent blood pressure pulses. As each new blood pressure pulse occurs, the average obtained in averaging and storage unit 83 is updated with the immediately preceding average and passes through line 89 to storage unit 81 which then stores the "nearest preceding" average.

It is generally sufficient that the signal on line 2H is converted to its digitized value substantially at the time of the systolic peak SP, for example by using the signal on line 90 from the flip-flop 86 as a timing input to the timing control unit 35. Essentially, the analog signal on line 2H, finally in digitized form on line H8, consists essentially of the slowly increasing pressure ramp with only a very small fluctuating magnitude representing the pulsatile pressure superimposed thereon. Thus, the value appearing on line 2H is substantially proportional to the pressure applied through the cuff 15 and is represented as P; however, low-pass filters may be introduced on line 24 if it is desirable to eliminate the pulsating component therefrom.

Certain obvious features and control functions of the embodiment shown in Fig. 11 have not been shown therein and include, for example, an initialization signal which is supplied when the operation of the pump 19 is initiated to empty and/or restore substantially all of the storage elements to a suitable initial state.

Normally this initial state represents a "zero" state. However, it may be desirable to initially store some non-zero minimum threshold value in the threshold storage unit 126 such that the systolic rise validation procedure is effective at the very first blood pressure pulse.

Furthermore, the systolic rise portion of the V waveform could be integrated over slightly larger or smaller limits than the preferred ED-SP limits to obtain a value that is generally representative of the P-P value of the respective blood pressure pulse.

In such a case, the so-called P2 reference appearing on line 80 and utilized by the various comparators could assume a value slightly above or below zero.

Although Pi max could be used with a suitable divisor of nearly the same value, if man were divided by two, it is clear that some other cuff and/or the patient's arm has such a structure that the cuff applies maximum pressure to one side or the other of the normal position center length of the cuff.

Furthermore, the illustration in FIG. 1, particularly with respect to the digitization of the P-waveform, the determination of the limits of the systolic rise, and the determination of a valid systolic rise, is intended to depict only one of a number of possible alternative circuit configurations for implementing the principles of the invention. For example, the determination of when the P-waveform goes above or below Pz and whether a valid systolic rise has occurred could be made on the analog P-waveform prior to its digitization, and such determinations could then be used to control which portions of the P-waveform appear as increments for input to accumulator 66. Furthermore, comparator 56 and the "state-indicating" flip-flops 72 and 86 could be omitted if comparators 54 and 58 operate only on analog waveforms or otherwise have digital signal information continuously input to them.

It will be appreciated by those skilled in the art that the implementation of the various functions shown in FIG. 1 is accomplished by means of commercially available component parts. Except for the pneumatics, pneumatic controls, and initial analog circuitry associated with the transducer 23, the remaining functional blocks are constructed substantially from commercially available microprocessors and other digital circuitry. A brief description of a state diagram shown in FIG. 6B and associated with the waveform shown in FIG. 6A is provided below, as well as a flow chart or decision tree shown in FIG. 7 to clearly indicate the decision and operation sequence used between successive blood pressure pulses or heartbeats.

Fig. A shows a portion of the É waveform and divides the waveform with respect to time into five different states associated with the signal processing state diagram shown in Fig. 6B. The state diagram and the more detailed decision tree in Fig. 7 derived therefrom provide sufficient information for one skilled in the art to implement the ideas of the invention in a digital manner, such as with microprocessors.

Referring to Fig. 7 and the sequence of instructions therein, between the systolic peak (SP) and the end diastole (ED) of the preceding blood pressure pulse, and thus prior to the initiation of the systolic rise of the current pressure pulse, the storage unit (98) containing Éi max is emptied, and the storage unit (accumulator 6§1_eom) containing the Pi sum SP . ' (i.e. 1 P) I LD is then also emptied. Thereafter, a short-term time increment of the P waveform is converted to a digital value represented by Pi indicating the magnitude of the particular increment. For the moment, it is assumed that the system is in state "one" or possibly "two", of the waveform P, and then Pi is compared with PZ. If f i is less than or equal to PZ, then the system remains in state "one" and the sequence returns to the step "empty Pi sum" and repeats.

If, on the other hand, f is greater than Pz, then the "tvd" state is entered and the magnitude of Pi over reference 13 is added to the sum of the Pi increments (in accumulator 66) thereby increasing the value of Pi sum. 10 15 20 25 30 35 H0 77145 89-4 20 A determination is then made as to whether a valid systolic rise has been detected by setting "V. Threshold Level".

Assuming that the "V. flag" has not yet been set, as is the case until state "three" in Figs. BA and SB is entered, the sequence then determines whether Pi is less than the threshold. Since the system is still in state "two", the program branches and the sequence beginning with the conversion of the next increment of the P waveform u to Pi is repeated. The sequence of state "two" continues until Pi is 2¿threshold level at which time the "threshold flag" is set and the transition from state "two" to state "three" takes place.

The sequence then determines whether the particular Pi is less than a stored value Pmax. Since Pi max is determined by the maximum previous Pi increment, a new Pi will not normally be less than Pi during the entire "three" state. max Thus, the sequence causes the current value of Pi to be loaded into the storage location Pi max to yield a new Pmax. which shall The sequence then determines whether Pi is greater than PZ.

Assuming that Pi is either in the "three" or "four" state, it is greater than P2 and thus the sequence returns to the step in which the P waveform is converted to a new Pi increment and the sequence is repeated. In this way, the Pi increment will be continuously added to the stored Pi sum during the "three" and "four" states as well as the "two" state.

When the P waveform peak is reached, at the transition from state 'three' to state 'four', a maximum Pi occurs and any remaining maximum is stored at the peak. Thus the sequence branches around the subsequent Pi normally taking on a smaller value than the Pi step, which would otherwise introduce the value of the now smaller Pi into the Pi maximum storage.

Upon transition from state "four" to state "five" a new Pi is no longer greater than P2 and the sequence recognizes that integration has been completed and Pi can then be scaled as required. The scaled Pi sum is stored at a position representative of the peak-to-peak (P-P) magnitude of the respective blood pressure pulse for use in interbeat therapy.

Finally, the value previously stored in Pi max is halved and this value is stored in a position representative of the threshold value for determining the threshold for the subsequent pulse.

It should be observed that the threshold value is increased when the size of the last P.

I max in each pulse is increased, and will likewise decrease as this last k i max value decreases. The "V.threshold flag", which was previously set, is reset either now at the end of the current sequence or alternatively at the initiation of the sequence in the next blood pressure pulse.

It should be noted that if a particular state "two" never reaches the threshold level (state "three") before falling below éz (as in the case of an artifact), then the "TV. threshold flag" will remain in its reset state. Although the Pi increments are continuously added to the Pi sum during this state "two" interval, they are not subsequently treated as a valid value since the "V»tPöSkC flag" is not yet set, and the sequence branches back to the "ready éi sum" step while the waveform is below éZ to continuously empty the Pi sum storage element.

In Fig. 8, an alternative embodiment of the invention is shown in which it may be desirable to store a larger portion of the P waveform than is represented by the single Éi increment. In such a system, the integration of the relevant intervals of the P waveform may be accomplished by reading the above-zero Pi data stored in memory backward out of memory and into integrating means, such as accumulator 66 in Fig. R, from which it is then processed in the same manner as described with reference to Fig. H. Those elements of the embodiment of Fig. H which are common to the elements of the embodiment of Fig. 8 and which do not contribute to the former are not shown in Fig. 8. Likewise, those elements and functions of the embodiment of Fig. 4 which are identical in the embodiment of Fig. 8 are numbered in the same manner as in Fig. 8.

The pi increments 50 go to the forward input of a last-in-first-out (LIFO) storage element or memory 559. The memory 569 has a word length of "m" units where "m" corresponds to the number of ri increments that would be obtained during a systolic rise of maximum anticipated duration (i.e., slowest anticipated pulse rate). This may include about 20-25 ri increments. The zero-crossing and threshold detection circuit of FIG. 11 operates to drive the Q output of the flip-flop 86 to one at the instant that the ri waveform is equal to or less than zero, assuming that the threshold level has been exceeded. This positive transition of the O output signal from the flip-flop 86 occurs on line 90 and is input to the pulse generator SGU, which in turn produces an output pulse on line 565 which is fed to the opening capture nos LIFO driver circuit 566. The LIFO driver circuit 586 receives an input signal from the timing control circuit 36' via line 567 and is operative, when activated, to produce inverted drive zeros to the LIFO memory 569. The inverted drive pulses cause the Pi data stored in the LIFO memory 569 to be read out therefrom on line 6%' in the reverse order in which they were entered. The timing controller 36' can provide drive pulses at a sufficient rate to allow the entire contents of memory 569 to be read out in reverse between successive inputs of new Εi increments. However, if the rate of the drive pulses from timing controller 36' is slower, it may be necessary to inhibit the introduction of new Εi increments into memory 589 during LIPO readout. The Εi increment which resulted in the positive-going transition of the Q output from flip-flop 86 is not itself stored in memory 569, and thus the last Εi increment entered will have a small positive value. Thus, when the contents of memory 569 are read in reverse order, the data on line EH' will first have a small positive value and then be incremented in the positive direction and then decremented towards the zero reference, in the opposite sequence to that in which it was originally stored. Line BH' includes an input to a comparator 560, the other input of which is the P2 reference 60. Comparator 560 goes from zero to one when the data appearing on line SH' is equal to or less than the P2 reference. The output of comparator 560 goes via line 561 to the input of a pulse generator 562 for generating an output pulse when the comparator makes a transition from zero to one. The output of pulse generator 562 goes via line 563' to the disable input of LIFO driver 566 for inhibiting further supply of reversed drive pulses to the LIFO memory 569.

In this way, the Pi data on lead GU' becomes the time-reversed, above-zero data that is representative of a valid systolic rise for a blood pressure pulse.

The data on line GU' goes to accumulator 56 in which it is summed as described above. The output from accumulator 65 goes via line 68' to the input of gate 78. The control pulse 80 applied to gate 78 serves to pass the resulting sum accumulated in accumulator 66 through gate 75 to its output, represented by line 82, and then to the processing circuit (not shown) described above with reference to FIG. U.

The invention may be embodied in other specific forms without departing from the scope of the invention. The embodiments described are therefore to be considered as illustrative only and not restrictive in any way, the invention being defined by the appended claims, and all modifications falling within the scope of equivalence of the claims are thus intended to be covered by the same.

Claims (18)

7714589-4 of Patent Claims 1. Apparatus for determining systolic pressure of a living test subject, the apparatus comprising: means (19) for applying a preselected variable pressure to the test subject adjacent a blood vessel (131); means (23) for measuring a fluctuating quantity proportional to a sum, which sum comprises a time-varying fluctuating component proportional to the amplitude of the pulsatile pressure in the blood vessel plus the variable pressure applied externally adjacent the blood vessel; means (H3) for determining the maximum value that said fluctuating component attains when the applied pressure is changed; means (53) for storing a representation of said maximum value; means (63) for determining when the fluctuating component is substantially equal to a predetermined fraction of the maximum value of an applied pressure greater than the pressure applied when said maximum value is obtained; and means (75) for reading out said applied pressure corresponding to the fluctuating component being substantially equal to a predetermined fraction of said maximum value, said read-out pressure corresponding to the systolic pressure of the object, characterized in that the apparatus comprises: means (28) for converting said quantity into a representation of a time derivative of said fluctuating component thereof; means (56, 72, 52, 86, 92) for forming the time integral of said time derivative representation over an interval of predetermined limits (ED, SP) in each of successive blood pressure pulses, said integration interval occurring between initiation of systolic rise and the systolic peak in the respective blood pressure pulses, each such integral being proportional to the amplitude of said pulsating pressure for the respective pulse; and means (39, 49) for supplying the time integral representations to said maximum value determining body (U3). 2. Apparatus according to claim 1, characterized in that the predetermined integration intervals extend substantially from the initiation of systolic rise to the systolic peak in the respective blood pressure pulses. 3. Apparatus according to claim 1, characterised in that said means for forming the time integral comprises: resettable integrating means (66); means (SU, 72) responsive to said time derivative representation for resetting said integrating means to an initial state; means (52, 62) responsive to the time derivative representation for initiating integration of the time derivative representation at a selected initialization time; and means (85, 92) responsive to the time derivative representation for sampling the integral value of said integrating means (66) at a selected sampling time, for supplying this value to said maximum value determining means (H3). H. 4. Apparatus according to claim 3, characterized in that said time derivative representation crosses a reference value in one direction at said initiation of systolic rise and crosses the reference value in the opposite direction at said systolic peak, said integration initiating means (52, 62) and said sampling means (86, 92) being responsive to said time derivative representation crossing the reference value in said first and said opposite directions, respectively, for initiating integration and sampling of the integrated value, respectively. 5. Apparatus according to claim U, characterized in that said resetting means (SH, 72) are arranged to respond to the time derivative representation for resetting said integrating means (66) when the time derivative representation passes the reference value in said opposite direction or falls below this value. 6. Apparatus according to claim 5, characterized in that said resetting means (Su, 72) are arranged to interrupt the resetting of said integrating means substantially at the point at which the integration is initiated. 7. Apparatus according to claim 6, characterized in that the reference value is substantially zero and that the magnitude of the time derivative representation is greater than zero during systolic rise. 8. Apparatus according to claim 3, characterised in that it comprises means (HO) for converting the time derivative representation into consecutively timed increments, which increments are supplied to said integrating means (56) for integration therein. 9. Apparatus according to claim 1, characterized in that said quantity conversion means comprises a differentiating network (28) for converting said quantity into a representation of the first time derivative thereof within the frequency band of said fluctuating component, and for converting said quantity into a representation of the second time derivative thereof at the lower frequency corresponding to the selectively variable pressure. 10. Apparatus for determining systolic pressure of a living test subject comprising: means (19) for applying a selectively variable pressure to the test subject adjacent a blood vessel (13); means (23) for measuring a fluctuating quantity proportional to a sum, which sum comprises a time-dependent fluctuating component proportional to the amplitude of the pulsatile pressure in the blood vessel plus the selectively variable pressure applied externally adjacent the blood vessel; characterised in that the apparatus comprises: means (28) for converting said quantity into a representation of a time derivative of its fluctuating component; means (66) for forming the time integral of said time derivative over an interval of predetermined limits (ED, SP) in each of successive blood pressure pulses, which interval corresponds to the time during which said time derivative representation exceeds a predetermined reference level; means (H3) for determining the maximum value attained by successive said integrals as the applied pressure changes; means (53) for storing a representation of said maximum integral value; means (63) for determining when said integral is substantially equal to a predetermined fraction of said maximum integral value for an applied pressure greater than the pressure applied when the maximum integral value is obtained; means (75) for reading out said applied pressure corresponding to said integral being substantially equal to said predetermined fraction of the maximum value, which read-out pressure corresponds to the systolic pressure of said subject; means (85, 98, 126) for providing a threshold level signal representative of the magnitude of substantially only the time derivative representing systolic rise; means (58) for comparing said time derivative representation of said magnitude fluctuating component with said threshold value signal when the time derivative representation exceeds the reference level and for providing a control signal indicating whether said time derivative representation exceeding the reference level is representative of a valid systolic rise; and means (86) responsive only to a validating indication that a particular said time derivative representation exceeding the reference level is a systolic rise, for inputting said particular representation to said maximum integral value determining means (H3) and said fraction of maximum integral value determining means (63). 11. Apparatus according to claim 10, wherein said threshold level signal providing means L«n:3mm~r said threshold level as a function of the amount by which at least the immediately preceding time derivative representation exceeding the reference level exceeds a second reference level. 12. Apparatus according to claim 11, characterized in that said reference level and said second reference level are the same. 13. Apparatus according to claim 12, characterized in that said threshold level is proportional only to the amount by which said immediately preceding time derivative representation exceeds said reference level. 1H. 14. Apparatus according to claim 13, characterized in that said threshold level is substantially 50% of the reference level-exceeding magnitude of said immediately preceding time derivative representation. 15. Apparatus according to any one of claims 10-1H, characterised in that said means for forming the time integral of the time derivative representation comprises accumulation means (66), wherein the time derivative representation is supplied to and accumulated in said accumulation means during each interval in which it exceeds said reference level so that the accumulated value of all said reference level-exceeding intervals forms a respective preliminary said integral, and said systolic rise validating means (58) is operative to forward only the integrals belonging to respective valid systolic rises to said maximum integral value determining means (H3). 16. Apparatus according to claim 15, characterised in that it comprises means (52) for converting at least said reference level-exceeding portions of the time derivative representation into consecutively timed discrete increments thereof, and in that said accumulation means is arranged to sum the discrete increments exceeding the reference level. 17. Apparatus according to any one of claims 10-1%, characterised in that said means for providing the time integral of said time derivative representation comprises: -..........__....-.___..__.,._._....._.......___, _ ”_ _, _ __ 7714589-4 gg accumulation means (6 6 ); means (569) for temporarily storing that portion of said time derivative representation which extends over an immediately preceding interval, which immediately preceding interval is at least as long as the maximum anticipated duration of the systolic rise portion of the blood pressure pulses, 'means (86) responsive to said time derivative representation going from a level exceeding said reference level to a level not exceeding the reference level and responsive to an indication of valid systolic rise from said validating means (58) for directing a last-in-first-out readout of said portion of the time derivative representation stored in said storage means (559); means (SN) for forwarding said readout from said storage means to said accumulation means (66) and for accumulating only those values of said readout which exceed the reference level, the value of said time derivative representation readout accumulated in said accumulation means being forwarded to said maximum integral value determining means (H3). (Fig. 8). 18. Apparatus according to claim 15 or 17, characterised in that it comprises means (72) for emptying said accumulation means (66) before each subsequent integration accumulation.