 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The present invention relates generally to the field of blood pressure
monitoring, and more particularly to automatic monitoring of systolic
blood pressure.
The prior art is replete with devices for measuring systolic pressure of a
living subject. An old and simple device is a pressurizeable cuff used in
combination with a mercury manometer which reads pressure in the cuff and
a stethoscope which is used to listen to Korotkof sounds. In another
advanced method of measuring blood pressure, the distance from a blood
pressure cuff to the wall of an artery is accurately determined by
measuring Doppler shifts of sound waves reflected by the artery. In yet
other methods for measuring blood pressure intrusive devices are often
inserted directly into blood vessels.
Oscillometric methods of determining systolic pressure are also well known
in the art. In such methods, the operator observes the representation of
the strength of pulsations of pressure within an artery. This can be done
visually, as by watching the extent of bouncing at the top of a mercury
column in a mercury manometer which is in pressure communication with the
cuff or indirectly as by measuring the occlusion which occurs to a blood
vessel in the pinna of the ear as pressure is exerted thereon. These
oscillometric methods generally define systolic pressure to be the maximum
applied pressure with which threshold oscillations are observed to occur.
With a typical mercury manometer and pressurized cuff, this pressure would
then be the highest pressure which the operator noted bouncing on the top
of the mercury column as the pressure in the cuff was slowly and
relatively uniformly reduced. However there are inaccuracies associated
with this method for determining threshold oscillations, since the inertia
of the mercury column does not allow it to noticeably respond to narrow
width pressure pulses.
Each of the aforementioned techniques or devices for measuring systolic
pressure exhibit some form of shortcoming such as inaccurate response to
narrow width pressure pulses or the requirement for sophisticated and/or
expensive measuring equipment.
There is described in U.S. Patent Application Ser. No. 578,047, filed May
15, 1975 by Link et al for Apparatus and Process for Determing Systolic
Pressure, assigned to the present assignee and incorporated herein by
reference, a method and apparatus for automatically and relatively simply
obtaining accurate systolic blood pressure measurements, thereby
overcoming the shortcomings of the aforementioned devices. That device
determines systolic pressure by applying pressure to a living test subject
by changing pressure in a pressure cuff attached to the subject adjacent a
blood vessel; by measuring at the cuff a quantity proportional to a time
dependent fluctuating component representative of the pulsatile pressure
within the blood vessel, which quantity is proportional to the amplitude
of the pulsatile pressure; by determining the maximum value attained by
the quantity as the applied pressure is changed; by storing a
representation of the maximum value; by determining when the quantity is
substantially equal to about one half of the maximum value for an applied
pressure greater than the pressure applied with the maximum value occurs
or results; and by reading out the applied pressure corresponding to the
quantity being substantially equal to about one half of the maximum value,
the read out pressure corresponding to the systolic pressure of the
subject. The signal from the pressure cuff comprises a fluctuating
quantity proportional to a sum, that sum comprising a time dependent
fluctuating component proportional to the amplitude of the pulsatile
pressure within the blood vessel, which component has a steeply rising
wavefront between end diastole and systole, plus the selectively
changeable pressure applied externally adjacent the blood vessel by the
cuff.
In the apparatus of the aforementioned application U.S. Ser. No. 578,047,
the signal from the pressure cuff is applied to a filtering network to
remove the effects of the cuff pressure ramp. The resulting oscillating
signal is considered to be proportional to the amplitude of the pulsatile
pressure within the blood vessel and a peak-to-peak detector then makes
amplitude measurements utilized to complete the signal processing.
However, random and uncontrollable deviation from the presumed linearity
of the pressure ramp may introduce errors in this amplitude determination.
If a large perturbation in the cuff pressure ramp is encountered, the
filter requires a considerable time to recover and may allow some
variation in the base line from which the fluctuating signal proportional
to the amplitude of the pulsatile pressure within the blood vessel is
measured, thereby resulting in an erroneous output from the peak-to-peak
detector.
The present invention provides a solution to the problems associated with
occasional perturbations in the applied pressure ramp of the cuff in a
systolic blood pressure monitor of the type described in the
aforementioned U.S. patent application Ser. No. 578,047.
SUMMARY OF THE INVENTION
In one sense, the invention comprises an improvement in apparatus for the
determination of systolic pressure of a living test subject, comprising: a
pressure cuff for applying a selectively changeable pressure to the test
subject adjacent a blood vessel; transducer means for measuring a
fluctuating quantity proportional to a sum, the sum comprising a time
dependent fluctuating component proportional to the amplitude of the
pulsatile pressure within the blood vessel plus the selectively changeable
pressure applied by the cuff externally adjacent to blood vessel; means
for determining the maximum value attained by the fluctuating component as
the applied pressure is changed; means for storing the representation of
the maximum value; means for determining when the fluctuating component is
substantially equal to a predetermined fraction of the maximum value for
an applied pressure greater than the pressure applied when the maximum
value results; and means for reading out the applied pressure
corresponding to the fluctuating component being substantially equal to
said predetermined fraction of the maximum value, the readout pressure
corresponding to the systolic pressure of the subject and wherein the
improvement specifically comprises:
means for converting the fluctuating quantity into a representation of a
time derivative of the fluctuating component thereof; means for obtaining
the time integral of the time derivative representation over an interval
of predetermined limits in each of successive blood pressure pulses, the
predetermined intervals of integration occurring between initiation of
systolic rise and the systolic peak in respective blood pressure pulses,
each of the integrals being proportional to the amplitude of the pulsatile
pressure for the respective pulse; and means for extending the time
integral representations to the maximum value determining means.
In another sense, the invention comprises an improved process, or method,
for determining systolic pressure comprising: applying pressure to a
living test subject by changing pressure in a pressure cuff attached to
the subject adjacent a blood vessel; measuring at the cuff a quantity
proportional to a sum, the sum comprising a time dependent fluctuating
component proportional to the amplitude of the pulsatile pressure within
the blood vessel plus the selectively changeable pressure applied by the
cuff externally adjacent the blood vessel; determining the maximum value
attained by the fluctuating component as the applied pressure is changed;
storing a representation of the maximum value; determining when the
fluctuating component is substantially equal to a predetermined fraction
of the maximum value for an applied pressure greater than the pressure
applied when the maximum value results; and reading out the applied
pressure corresponding to the fluctuating component being substantially
equal to said predetermined fraction of the maximum value, the readout
pressure corresponding to the systolic pressure of the subject and wherein
the specific improvement comprises the steps of: converting the quantity
proportional to the sum into a representation of a time derivative of the
fluctuating component thereof:
determining the time integral of the time derivative representation over an
interval of predetermined limits in each of successive blood pressure
pulses, the predetermined interval of integration occurring between
initiation of systolic rise and the systolic peak in respective blood
pressure pulses, the determined integral being proportional to the
amplitude of the pulsatile pressure for the respective pulse; and
using the time integral representations in the determination of the maximum
value attained by the fluctuating component as the applied pressure is
changed.
The present invention recognized that the signal from the pressure cuff,
and more particularly the time dependent fluctuating component thereof
representative of the pulsatile pressure within the blood vessel, may be
differentiated to obtain the first time derivative of said fluctuating
component. Further, it recognizes that this time derivative representation
will cross a zero reference in the positive going direction at the time of
end diastole and will return below the zero reference line when the
systolic peak occurs at the end of the systolic rise. The invention
further recognizes that the "above zero" area under the time derivative
waveform is representative of the peak-to-peak (diastolic to systolic)
amplitude of the respective blood pressure pulse and that such area may be
determined by integrating the time derivative waveform over its "above
zero" extent.
The present invention further recognizes that double differentiation of the
cuff signal in the lower frequency range of the cuff pressure ramp aids in
avoiding off-set of the fluctuating component derivative from the zero
reference.
The method and apparatus of the invention additionally provide for
integrating the time derivative representation when it crosses a reference
magnitude, such as zero in a positive going direction and for terminating
the integration when it crosses the reference in the negative going
direction.
Still further, the invention provides for sampling and holding the integral
value at the end of the period of integration and for clearing the
integration means before an integration is undertaken in a subsequent
blood pressure pulse.
It is thus, then, an object of the present invention to provide an improved
apparatus and process/method for determining systolic pressure. Included
in this object is the provision of an improved apparatus and
process/method which determines systolic pressure with increased accuracy.
These and other objects and advantages of the present invention will be
apparent to those skilled in the art after referral to the detailed
description of the preferred embodiment in conjunction with the appended
drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 illustrates a prior art technique for the determination of the
peak-to-peak magnitude of the pulsatile pressure within a blood vessel;
FIG. 2 illustrates a technique for the determination of peak-to-peak
amplitude of the pulsatile pressure in accordance with the invention;
FIG. 3A illustrates a typical oscillometric envelope of the pulsatile
pressure of a blood vessel;
FIG. 3B illustrates the time derivative of the FIG. 3A waveform;
FIG. 3C illustrates a controlled mode timing diagram in accordance with the
basic method and apparatus of the invention;
FIG. 3D illustrates the time intervals obtained from the waveform of FIG.
3B in accordance with the basic apparatus and process of the invention;
FIG. 4 illustrates, in a block diagram, the apparatus and process of one
embodiment of the invention;
FIG. 5 illustrates a plot of the gain vs. frequency characteristics of a
differentiating network employed in a preferred embodiment of the
invention;
FIG. 6A represents an enlarged portion of the time derivative waveform
illustrated in FIG. 3B showing a validation threshold level and the timing
of various control states associated therewith in accordance with the
embodiment illustrated in FIG. 4;
FIG. 6B illustrates a control state diagram in accordance with FIG. 6A and
the embodiment of FIG. 4;
FIG. 7 illustrates a flow chart or decision tree of the control sequence
employed by the embodiment illustrated in FIG. 4 between successive heart
beats;
FIG. 8 illustrates, in an abbreviated block diagram supplemented by FIG. 4,
the apparatus and process of another embodiment of the invention; and
FIG. 9 illustrates a technique similar to that of FIG. 2 for the
determination of peak-to-peak amplitude of the pulsatile pressure and
further including threshold detection means for identifying particular
"above zero" passages of the waveform derivative as valid systolic rises.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Referring first to FIG. 1, there is illustrated a functional block diagram
of certain portions of the systolic pressure measuring apparatus described
in the aforementioned application U.S. Ser. No. 578,047. More
specifically, the functional blocks of FIG. 1 illustrate a filter network
100 having its output connected through amplifier 102 to the input of a
peak-to-peak detector 104. Filter network 100 receives an input signal
106a on input conductor 106. The input signal 106a comprises a
slowly-increasing ramp indicative of the applied cuff pressure and having
superimposed thereon the time-dependent fluctuating component
representative of pulsatile pressure within the blood vessel of the
subject, which component representative of pulsatile pressure has a
steeply rising wavefront relative to the remaining components during the
rise from end diastole to systole. Filter network 100 was typically
constructed such that its output waveform 100a had the linear effects of
the pressure ramp removed therefrom, however, any random and
uncontrollable deviation from the presumed linearity of the pressure ramp
would introduce errors in the signal 100a. For instance, if a large
perturbation in the cuff pressure ramp was encountered, filter 100
required a considerable time to recover and could allow some variation in
the base line 100b (dotted) from which the fluctuating signal proportional
to the amplitude of the pulsatile pressure within the blood vessel w as
measured. Accordingly, each time peak-to-peak detector 104 is operated in
response to sampling signals 108, the resulting output signal 110a
appearing on conductor 110 included those peak-to-peak errors introduced
by the variation in base line 100b.
In accordance with the present invention as illustrated generally in FIG.
2, an input signal 206a having a waveform identical to that of waveform
106a in FIG. 1 is applied to the input conductor 206 to differentiating
network 200. Differentiating network 200 is constructed such that it
provides single differentiation of signal 206a over that range of
frequencies corresponding with the frequencies of the fluctuating signal
proportional to the amplitude of the pulsatile pressure within the blood
vessel, and doubly differentiates the input signal below that range of
frequencies in order to remove the offset effects of a linear pressure
ramp and any very low frequency perturbations which might have appeared in
the otherwise linear pressure ramp.
A filter or differentiating network having the properties required of
network 200 will possess the Gain v. Frequency characteristics illustrated
in FIG. 5 in which the Gain curve exhibits a -6db per octave slope in the
frequency range f.sub.1 -f.sub.2 and a -12db slope for frequencies below
f.sub.1. The frequency range f.sub.1 -f.sub.2 corresponds with the
bandwidth of the P.sub.ac signal comprising the fluctuating quantity
representative of the pulsatile pressure. That portion of input signal
206a representative of the pulsatile blood pressure is differentiated and
appears at the output of differentiating network 200 as signal 200a,
hereinafter designated P.
This P signal (200a) is applied through amplifier 202 to the input of an
integrator 204 which, by integrating the P signal over a predetermined
interval during each pulse, provides an output value corresponding with
the peak-to-peak pressure of each blood pressure pulse. Sample-and-hold
circuitry 205 associated with integrator 204 serves to sample the value
appearing at the output of integrator 204 at the end of each interval of
integration and to hold that value for an interim period until integration
of the next pressure pulse begins. Control of integrator 204 and
sample-and-hold circuit 205 is provided by the RESET/INTEGRATE/HOLD-signal
208 which controls the period of integration and serves to clear the
integrator prior to each new integration. The output from sample-and-hold
circuit 205 appears on line 210 as waveform 210a having a magnitude which
corresponds with the area under that portion of the waveform P being
integrated.
Referring to FIGS. 3A and 3D for an understanding of the theory underlying
the invention, it will be recalled from the aforementioned U.S. patent
application Ser. No. 578,047 that the systolic pressure is equal to
applied cuff pressure when the fluctuating quantity is about equal to one
half the maximum of value of the fluctuating quantity. The maximum value
of the fluctuating quantity is determined by measuring the diastole and
systole in successive blood pressure pulses. That pulse exhibiting a
maximum P-P amplitude is taken as the maximum value and the applied cuff
pressure is further increased such that the P-P amplitude decreases and
the systolic pressure is determined by noting the applied cuff pressure at
which the P-P pressure becomes one half of the P-P maximum.
FIG. 3A illustrates the time-dependent fluctuating component, P.sub.ac,
representative of pulsatile pressure within a blood vessel. The root ED of
each valley in the P.sub.ac waveform corresponds with the time of
diastole, or more specifically end diastole in a heart beat and the
waveform peak SP corresponds with the time of systole in the heart beat.
As earlier described, the signal from the cuff is differentiated to remove
the applied pressure ramp and low frequency random perturbations, and
results in the derivative P of waveform P.sub.ac, as represented in FIG.
3B. Because waveform P.sub.ac exhibits zero slope at both end diastole
(ED) and the systolic peak (SP), the derivative waveform P will be of zero
magnitude at each of those times. Further, because P.sub.ac exhibits a
positive slope during the systolic rise between ED and SP, the P waveform
lies above the zero reference line during this interval. The zero-crossing
points ED and SP of the P waveform correspond with the points of maximum
amplitude between successive P.sub.ac pulses and thus the area under the P
waveform and above the zero reference between end diastole ED and the
systolic peak SP provides a value which corresponds with the P-P value of
the respective blood pressure pulse. This area is determined by
integrating the "above zero" section of the P waveform. It should be noted
that end diastole (ED) also corresponds essentially with the initiation of
the rise to systolic peak (SP).
FIG. 3C illustrates a control signal generally similar to that of signal
208 in FIG. 2 which clears or resets the integrator prior to the interval
of integration, then integrates the P signal over the interval of
integration, 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 resetting operation
being indicated by R, the integrating operation being represented by S,
and the sample and hold operation being represented by S+H. In fact, the
sampled integral may be held longer than is suggested by the brief
duration of the S+H signal in FIG. 3C.
The results of integrating the P waveform between the limits of ED and SP
are illustrated in FIG. 3D. The magnitude of the integral at the time of
the systolic peak SP corresponds with the P-P value of the respective
blood pressure pulse.
In implementing the concept of integrating the P waveform over the interval
of systolic rise to obtain respective P-P values for the respective blood
pressure pulses or heart beats, standard circuitry may be used to detect
when the P waveform crosses the zero reference in the positive going
direction to begin the integration and to determine when it crosses the
zero reference in the negative going direction to terminate the
integration and/or perform the sample and hold function. The integrator
may be reset immediately after sample and hold and preferably continue
until the next positive going zero-crossing of P. The resulting integral
may then be considered as representing the P-P value of the respective
pulse. However, certain characteristics of the P.sub.ac waveform and/or
the presence of signal artifacts during the diastolic drop may result in P
appearing above the zero reference for a brief time other than between end
diastole and the systolic peak. For instance, as illustrated in FIGS. 3A
and 3B, if random muscular activity introduces a "high frequency" signal
artifact (ART) just prior to end diastole when the slope of the P.sub.ac
waveform is relatively flat, the derivative P waveform may present part of
the artifact as a "greater than zero" value and result in the tentative
values illustrated parenthetically in FIGS. 3C and 3D.
In accordance with an aspect of the invention, illustrated generally in
FIG. 9, a threshold level is established for discriminating between those
P values greater than zero which attend the systolic rise and those
signals, such as artifacts and the like, which do not attend the systolic
rise. The magnitude of P signal associated with the systolic rise is
normally significantly greater than the magnitude of any other (above
zero) portion of the signal (as from artifacts) and accordingly, this
allows discrimination between such signals. The determination that the P
waveform exceeds the threshold level during a particular "above zero"
passage serves to validate the integration of that "above zero" passage
between its respective ED and SP limits.
Referring to FIG. 9, in which those components functionally identical to
corresponding components in FIG. 2 are identically numbered, the input
signal 206a is differentiated by differentiating network 200 to provide
the P waveform which is passed through amplifier 200 to the respective
inputs of integrator 204, a threshold detector 912, and a zero-crossing
detector 907. The zero-crossing detector 907 may correspond with means,
not shown in FIG. 2, which established the interval of integration and
resulted in the control signal 208 therein. The threshold detector 912
establishes a signal magnitude threshold value above which the P waveform
is presumed to be indicative of a valid systolic rise. When the incoming P
waveform exceeds the threshold level of detector 912, a signal is provided
to the input of validating logic 914 indicative of such threshold level
having been exceeded. Similarly, the validating logic 914 receives an
input from the output of the zero-crossing detector 907 to define when the
P waveform crosses a zero reference in the positive going direction and
also in the negative going direction. The output 908' from validating
logic 914 is applied to the RESET input of integrator 204 for resetting
the integrator at least substantially at the beginning of each desired
period of integration beginning with the P waveform crossing the zero
reference in the positive going direction. The output 908 from validating
logic 914 is applied to the "sample" input of the optional sample-and-hold
circuit 205 and serves to store the integral value accumulated by
integrator 204 between the positive going and negative going zero
crossings of the P waveform only if threshold detector 912 has provided an
indication that the P waveform during that interval was in fact a valid
systolic rise. The output 910 of sample-and-hold circuit 205 varies from
the output of 210 of FIG. 2 only where the latter might have included an
invalid output value representative of a systolic rise when in fact only
an artifact was present.
While a threshold of fixed magnitude above the zero reference might be
utilized if an "above zero" portion of the P waveform did not vary in
magnitude in successive pulses, such is not the case, particularly when
using the present oscillometric blood pressure monitoring techniques in
which the ac pressure signal P.sub.ac increases from a small amplitude at
a low applied pressure to a large amplitude at a larger applied pressure
and then to a smaller amplitude at a still larger applied pressure.
Therefore, the threshold level, indicated as TRLD in FIG. 3B, is selected
to be a function of the magnitude of the systolic rise portion of the P
signal over one or more of the immediately preceding blood pressure
pulsations. The increase (and subsequently decrease) in magnitude of
successive systolic rises in the P waveform is sufficiently gradual, and
the relative amplitude of any "above zero" non-systolic rise components of
the P waveform are sufficiently small, that a dynamic threshold which
corresponds with 50% of the maximum "above zero" amplitude of the systolic
rise of the P waveform during the preceding pulse is herein considered
sufficient for recognizing only those "above zero" portions of the P
waveform which, in fact, attend the systolic rise.
It will be appreciated that the dynamic threshold level might be
established by summing and weighting several prior systolic rise portions
of the P waveform in which case threshold TRLD might be at a preselected
level greater or less than 50% of the magnitude of the immediately
preceding systolic rise. An analog example of a dynamic threshold detector
of the type suitable for application herein is described in greater detail
in U.S. Pat. No. 3,590,811 to Harris for Electrocardiographic R-wave
Detector. Digital means for establishing a dynamic threshold level will be
described hereinafter in greater detail.
Reference is now made to FIGS. 4, 6, and 7 for a more detailed description
of the apparatus and process of one aspect and embodiment of the
invention. The apparatus is described with reference to the functional
block diagram of FIG. 4 which provides for the digital processing of the
analog signal received from transducer 23. However, it will be appreciated
that analog implementation is similarly possible. More specifically,
discrete electronic components, discrete digital chips, microprocessor
technology and structure, or digital computer can be employed. FIGS. 6 and
7, respectively, comprise a state diagram and a flow chart, or decision
tree, associated with the processing of the P signal between successive
blood pressure pulses corresponding with successive beats of the heart.
Generally speaking, the signal processing steps of the improved blood
pressure monitoring apparatus and technique of the invention, illustrated
in FIGS. 6 and 7, correspond with that portion of the FIG. 4 apparatus
which integrates the P signal between the limits ED .fwdarw. SP.
The arm 11 of a test subject with artery 13 therein is surrounded by a
typical blood pressure cuff 15. Typically, the brachial artery located in
the upper arm is employed for this type of blood pressure measurement.
Attached to the cuff via conduits 17 and 21 are pump 19 and pressure
transducer 23, respectively. Transducer 23 has a transfer function such
that its electrical output is substantially representative of its pressure
input up to the limit of information contained in the pulse pressure. The
pressure transducer serves to measure the pressure within the cuff, which
pressure is the sum of pressure supplied by the pump and the fraction of
pressure produced by blood pressure fluctuation within the artery, as
represented by waveforms 106a and 206a in FIGS. 1 and 2, respectively. The
fluctuating portion of the output of transducer 23 represents the
amplitude of pulsatile pressure. The output of transducer 23 proceeds, as
represented by line 24, to one input of multiplexing switch 25. The output
of transducer 23 also proceeds, as represented by line 26, through
normally closed switch 27 to the input of differentiating network 28. The
output of differentiating network 28 proceeds, as represented by line 29,
to amplifier 30 and proceeds as represented by line 31, to the other input
of multiplexing switch 25.
The differentiating network 28 differentiates the input signal over the
f.sub.1 -f.sub.2 bandwidth of signal P.sub.ac, as illustrated in FIG. 5,
and additionally provides double differentiation of the frequencies below
f.sub.1. In this manner, substantially the only signal appearing on line
31 is that of the differentiated (P) representation of the P.sub.ac
waveform.
The output of multiplexing switch 25 proceeds, as represented by line 32,
to analog-to-digital (A/D) converter 33. The output of A/D converter 33
proceeds, as represented by line 48, to inputs at gates 40 and 42,
respectively. A clock 34 generates timing pulses which proceed, as
represented by line 35, to a timing-control unit 36 which controls the
switching of multiplexer 25, the conversion of the analog signal to a
digital signal, and the gating of gates 40 and 42. One output of timing
control 36 proceeds, as represented by line 44, to multiplexer 25, A/D
converter 33, and the other input of gate 40 to control the conversion of
the P signal appearing on line 31 to a digital form which is then applied
to gate 40 via line 48. Another output of timing control unit 36 proceeds,
as represented by line 46, to multiplexer 25, A/D converter 33, and the
other input of gate 42 for controlling the conversion of the analog signal
from transducer 23 to a digital form which is applied to the gate 42.
The gating signals appearing on lines 44 and 46, respectively connected to
the inputs of gates 40 and 42, are of sufficient duration that the
digitally-converted data associated therewith and appearing at the other
input to the respective gate is passed through the particular gate. It
will be further appreciated that the control signals represented by lines
44 and 46, as illustrated herein, exist mutually exclusively of one
another such that the data appearing on line 31 or 24 is connected to the
appropriate gate 40 or 42, respectively.
The period between successive blood pressure pulses is normally on the
order of 800-1000 milliseconds with the systolic rise occupying some
10-20% of each period. The P signal may be integrated over the systolic
rise portion by sampling a sufficient number of incremental portions
(P.sub.i) of the P waveform to closely approximate the area under the P
waveform. Typically, 10-20 samples during the 100-200 milliseconds of a
typical systolic rise will be sufficient to provide the requisite number
| | |
