Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
APPARATUS AND METHOD FOR MEAS[~RING BLOOD PRESSllRE
Field of the Invention: This invention relates in
general to blood pressure measuring apparatus and, in
particular, to blood pressure measuring devices which
measure arterial blood pressure by oscillometry - the
monitoring of pressure oscillations produced by
arterial blood pulsations within a pressurized air
cuf~
Background of the Invention: The variations of blood
pressure occurring during various physiological
states of a patient is of great interest in modern
medical diagnostic procedures. The traditional
method of characterizing blood pressure is a determ-
ination of the systolic and diastolic pressure
values. Another measurement variable, the mean
arterial pressure (MAP), has also been determined to
be useful as an indication of blood pressure. The
mean arterial blood pressure is defined as the time
average of the instantaneous blood pressure or as a
weightea average o~ the systolic and diastolic
pressures. In particular, if blooa pressure is
plotte~ relative to time, the MAP is a level chosen
so that the area between the systolic section of the
curve and the MAP level equals the area between the
~AP level and the diastolic section of the curveO
The MAP level can be roughly estimated from the
z~
systolic and diastolic values according to the
following formula:
MAP = Diastolic + 1/3 tSystolic - Diastolic)
The value determined by this equation may be in-
accurate in shock cases, in an operating room
environment, or where certain diseases are involved
due to changes in the blood pulse waveform.
There are presently several methods of measuring
the various values of arterial blood pressure which
are in common use. The most accurate method is
direct measurement of arterial pressure by using an
arterial cannula. However, invasive techniques are
often inconvenient and may give rise to considerable
patient ~iscomfort.
Accordingly, several noninvasive techniques have
been developed. One of the earliest techniques which
is in common use involves occluding the bloo~ vessels
in a patient's limb by means of in~latable cuff which
encircles the limb. The pressure (typically, air) in
the cuff is then slowly decreased. When the decreas-
ing pressure equals the arterial systolic press~re,
characteristic sounds commonly known as KoLotkoff
sounds can be heard by auditory monitoring of the
blood flow. When the decreasing pressure in the cuff
reaches the arterial diastolic pressure, the
Korotkoff sounds also change in a characteristic
manner. These phenomena can be easily used to
measure the systolic and diastolic blood pressure by
observing the cuff pressure by means of conventional
mercury or aneroid sphygmomanometer while manually
listening to the blood flow in the arteries. The
technique has also been automated by detecting the
Korotkoff sounds using microphones or ultrasound
transducers in the inflatable cuff. One problem with
this method is that it cannot be used to directly
measure the mean arterial pressure which must be
estimated from the systolic and diastolic values us-
ing the formula reterre~ to above. This formula may
be inaccurate due to a variety o~ factors including
disease or shock.
A more recently discovered technique is the
oscillometriG method of detecting and quantifying
blood pressure values. This technique utilizes a
blood vessel-occluding air cuff as in the ~orotkoff
technique, but senses blood pressure values by a
different means. 5pecifically, as the air pressure
in the inflatable air cufE is decreased below the
systolic blood pressure, small pressure oscillations
can be observed above the baseline cuff pressure.
These small pressure oscillations are reflected in
the air pressure of the surrounding cuff as a result
of expansion and contraction of the arteries produced
by the pulsatile blood flow. The pressure
72:~
oscillations increase in amplitude and reach a rnaxi-
mum as the cuff pressure becomes equal to the mean
arterial blood pressure. The oscillations then
decrease in amplitude until they entirely disappear
below a threshhold value of the applied cuff-
pressure. The mean arterial pressure is then easily
measured by detecting the air cuff pressure at which
the maximum amplitude of the pressure oscillations in
the air cu~f occurs. This measurement technique is
easily automated and is especially useful in blood
pressure and measuring devices that are controlled by
microprocessors.
However, one problem with prior art blood
pressure measuring devices using the oscillometric
method is that although the mean arterial pressure
can easily be measured, no simple, accurate method
for measuring either systolic or diastolic pressures
has been developed.
Consequently, most prior art devices rely on an
extrapolation of the systolic and diastolic pressures
from the measured mean arterial pressure. For
example, it has been observed that the systolic and
diastolic pressures occur at points where the
pressure oscillations in the air cuff reach a magni-
tude which is approximately one half the magnitude of
the oscillations at the mean arterial pressure~ This
method provi~es an easy way of calculating the syst-
olic and diastolic pressures from the mean arterial
~7'7Z~L
pressure. However, it is subject to several
additional problems. First, artifacts introduced by
patient movement or outside interference may produce
erroneous results if they occur at cuff pressure
measurements in the vicinity of the diastolic or
systolic pressures. Second~y, the one-halt magnitude
relation of the oscillation amplitudes at mean
pressure and systlic/diastolic pressures is not
exactly correct. Therefore the systolic and dias-
tolic pressures calculated by this technique are only
approximations as to the true systolic and diastolic
pressures.
Accordingly, it is an object of the invention to
provide a more accurate method for determining syst-
olic and diastolic blood pressure values in an
oscillometric-mode blood-pressure measuring system.
It is another object of the invention to obtain
accurate systolic àn~ diastolic blood pressure read-
ings in the presence of noise and other external
~isturbances and in the case of shock, operating room
environments and disease situations.
It is a further object of the invention to obtain
increased artifact rejection in obtaining systolic
and diastolic blood pressure readings.
Summary of the Invention: The above problems are
solved and the objects accomplished in an illustra-
tive embodiment of the invention in which a more
'7~
accurate determination of systolic and diastolic
pressure values is produced and erroneous results
which might be produced by artifacts are avoided by
calculating the systolic and diastolic pressures from
a series of measurements taken at cuff pressures in
the region of the systolic and/or the diastolic
pressure rather than from just a single measurement
of the mean arteria] pressure or from a single
measurement taken in the vicinity of the systolic and
diastolic pressure.
Specifically, it has been determined that as the
cuf~ air pressure in the blooo vessel-occluaing (air)
cuff is decreased from a value above the systolic
blood pressurel the oscillations which occur in the
cuff air pressure slowly increase in amplitude at a
gradual and a first appro~imately constant rate.
However, when the cuff air pressure reaches the vic-
inity of the systolic pressure the rate of increase
of the oscillation magnitudes sharply increases. The
oscillation magnitudes then continue to grow at
approximately the second constant increased rate as
the cufi air pressure is decreased~ until the mean
arterial pressure is reached and the maximum ampli-
tude of oscillation occurs. The oscillation magni-
tuces then decrease at approximately the same rate as
the secono increased rate until the aiastolic
pressure is reached. At this point the rate of
decrease o~ oscillation magnitudes changes to a
secon~ more gradual rate until a cuff pressure is
reached at which the oscillations disappear. The
present invention determines the systolic pressure by
determinlng the cuff pressure at which there is a
change in the rate of increase o t the oscillation
magnitudes as the cuff pressure is passing through
the pressure corresponding to the systolic pressure.
The diastolic pressure is then determined by measur-
ing the cuff air pressure at an oscillation magnitude
corresponding to the oscillation magnitude at the
measured systolic pressureO
Specifically, in an illustrative embodiment of
the invention, a series of "readings" are taken as
the cuff pressure is decreased. Each reading con-
sists of the peak oscillation magnitude and the
corresponding baseline cuff air pressure~ A plural-
ity of magnitude readings are selected which occur at
respecti~e cuff pressures above the expected systolic
pressure. In addition, a plurality of magnitude
readings are selected which occur at respective cuff
pressures below the expected systolic pressure. Two
relationships representing, respectively, the change
of the peak amplitudes of eawch set of readings with
the change of the cuff pressure, are derived by
well-known methods from each set of readinys. The
relationships, which, for example, may be straight
line equations, are manipulated to determine values
of cuff pressure and corresponding oscillation
'7~
magnitudes which satisfy both relationships; in a
mathematical sense, they are set equal and solved for
the systolic pressure. In a graphical sense, the two
functions represent curves (illustratively straight
lines, connecting the peak values of each set); such
curves extended, intersect at a cuff pressure value
equal to the systolic pressure.
The diastolic pressure is subsequently determined
by detecting a cuff pressure amplitude which produced
an oscillation magnitude equal to the oscillation
magnitude at the calculated systolic pressure. It
will be appreciated that this proce~ure could be
reversed; the diastolic pressure can be oetermined
first, and the systolic pressure can then be determ-
ined from it.
Brief Description of the_Drawing:
Figure 1 shows a block schematic diagram of an
illustrative blood pressure measuring device suitable
for use with the invention.
Figure 2 shows a typical set of readings takerl by
means of the apparatus in Figure 1.
Figure 3 is a flow diagram of the operation of
the circuitry shown in Figure 1 used to take the
readings shown in Figure 2.
Figure 4 is a flow diagram of the routine used to
'7~
calculate the systolic and diastolic pressures.
Detailed Description: Referring to Figure 1, an air
cuff 101 is placed around the limb of a patient,
preferably on the upper arm or upper leg in order to
controllably occlude the blood vessels in preparation
for the measurement of the arterial blood pressure.
Cuff 101 is a well-known device which typically con-
tains an air bladder that has two flexible tubing
connections 102 and 103. The air bladder in cuff 101
may be inflated or deflated by means of flexible tube
103 which is connected to valve 104. Valve 104 is
controlled, via lead 105, by control circuit 150 as
will be hereinafter described and operates to inflate
cuff 101 by means of pressuri7ed air which is provid-
ed from a compressed air source (not shown) through
tube 106. Valve 104 may also operate under control
of control circuit 150 to release air pressure from
cuff 101 in graduated steps. The air bladder in cuff
101 is also provided with a second flexible connec~
tion 102 which is attached to a transducer unit 107.
The second connection allows measurements to be made
on the air pressureo in the cuff without interference
from the air flow on the tube 103. Transducer unit
107 responds to the air pressure in the cuff relative
to atmospheric pressure and produces an electronic
signal. The magnitude of the electronic signal is
proportional to the pressure in cuff 101. The output
--10--
7'~
of transducer 107 is provided via leads 108 and 109
to amplifiers 111 and 112. Amplifier 112 is a D.C.
amplifier which produces at its output 114 a signal
that is representative of the average baseline
pressure in cuff 101. Amplifier 111 is an A.C.
coupled amplifier (shown schematically as amplifier
111 in a series connection with capacitor 110) which
produces a signal at its output 113 in response to
the oscillations in pressure that result from the
pulsation of the blood vessels within the patient's
limb. Either output 113 of amplifier 111 oe output
114 of amplifier 112 can be selected by multiplexer
115 under control of control circuit 150, via lead
120, and applied, via lead 125, to analog-to-digital
converter 130. Converter 130 converts the analog
signals produced by amplifiers 111 and 112 to digital
signals which are used by the processin circuitry in
order to compute the mean arterial, systolic and
diastolic blood pressures.
Therefore, in summary, control circuit 150 may
operate multiplexer 115 and analog-to ditigal con-
verter 130 to produce a digital signal which is rep-
resentatlve of the baseline pressure in air cuff 101,
via amplifier 112, or the amplitude of the pressure
oscillation in air cuff 101 produced by the blood
flow in the patient's blood vessels, via amplifier
111.
As will be hereinafter described in further
detail control circuit 150 operates memory 140 and
computation circuit 160 to selectively make a plural-
ity of readings, each reading consisting if a base-
line pressure an~ corresponding oscillation magni-
tude. Af~er completing a plurality of such readings,
control circuit 150 operates memory 140 to transfer
selected readings to computation circuit 160, via
channel 141. Control circuit 150 then operates in
conjunction with computation circuit 160 to compute
the mean, systolic and diastolic blood pressures
which are producèd on output 161.
In the illustrative embodiment described herein,
analog-to-digital converter 130, memory 140, control
circuit 150 and computation circuit 160 (all as shown
in the enclosed dotted box 170) may be implemented as
a microprocessor. Alternatively, conventional cir-
cuitry may be used to implement the functions which
will be hereinafter described.
Figure 2 shows a typical set of readings which
are taken by means of the circuitry shown in Figure
1. The Figure consists of a graph in which the hor-
izontal axis represents baseline cuff pressure
increasing towards the right and the vertical axis
represents systolic and diastolic magnitude increas-
ing in the upward direction. The series of dots or
"points" on the graph each represent a single reading
which has corresponding oscillation magnitude and
'7~
baseline cuff pressure as read on the horizontal and
vertical axes. Figure 2 will be used in connection
with Figures 3 and 4 to describe the operation cycle
of the circuitry disclosed in Figure 1.
Specifically, in Figure 3, the flow diagram ill-
ustrates the operation of the circuitry in Figure 1
during the taking of data readings on the pressure
signals produced by air cuff 101. In step 301, con-
trol circuit 150 operates valve 104 to inflate air
cuff 101 to a pressure which is higher than the
expected systolic pressure. In the illustrative
embodiment cuff 101 is inflated to 170 Torr.
In step 302, control circuit 150 causes the out-
put signal of analog-to-digital convertor 130 repre
sentative of the baseline cuff pressure to be stored
in memory 140. In step 303, control circuit 150
causes a sample of the oscillation amplitude (in dig-
ital form) produced by a convertor to be stored in
memory 140. Typically, oscillations occur in the air
pressure in cuff 101 at the frequency of the
pa~ient's pulse rate which is typically 60 to 80
Hertz. The sampling operation on the amplitude, how-
ever, is conducted at a much higher frequency so that
the variation in the oscillation amplitude during the
sample period is minimal. In step 304 the sampled
amplitude from the output of convertor 130 is stored
in memory 140.
7~'~
In step 305, the oscillation amplitude is sampled
again. This sampling operation takes place at prede-
termined in~ervals. In step 306, the two samples
previously obtained are compared. If the second
sample is less than the ~irst, step 307 is executed.
If the second sample is greater than the first, step
310 is executed. Assuming that the second sample i5
less, in step 307, control circuit 150 takes an
additional sample of the oscillation amplitude and
compares it (in step 313) to the sample previously
stored in memor~ 140. If the present sample is less
than the stored sample (indicating that the oscilla-
tion magnitude is still decreasing) the most recent
sample is stored in place of the previous stored
sample (step 311) and a new sample is taken (step
307~. This operation is continued until a present
sample is greater than the stored sample indicating
that the minimum of the oscillation magnitude has
been reached. In this case control circuit 150
executes step 315 and aesignates the stored sample as
a minimumO
Control circuit lS0 then determines whether a
maximum value of the oscillation amplitude has been
determined at step 321.- If not, steps 310s 312, 314,
and 316 are executed in which additional samples are
taken and compared to a previously stored sample
until a present sample is less than the stored sample
indicating a maximum has been found. When this
14-
7~
occurs, the maximum is designated in step 316. ~ince
a minimum has already been found in step 320, control
circuit 150 prog~esses to step 325 in which the mini-
mum value is subtracted from the maximum value to
generate the peak-to-peak magnitude of the oscilla~
tion. This value is stored in step 330 and control
circuit 150 then compares the present oscillation
magnitude stored in step 330 to those previously
stored to determine whether oscillation magnitude is
increasing or decreasing.
Assuming that the oscillation magnitude is
increasing (indicating that the mean arterial
pzessure has not been reached yet) step 345 is
executed in which the air pressure in cuff 101 is
deflated by a predetermined amount. This amount may
illustratively be in the range of 10 to 20 Torr.
new set of readings is taken and operation in this
manner continues until, in step 335, it is determined
that the present oscillation magnitude is less than
the previous magnitude. In this case control circuit
150 executes step 340 and determines whether at least
five out of the six previous stored magnitudes are
greater than the present magnitude. If this condi-
tion is satisfied, it indicates that the baseline
cuff pressure has passed through the diastolic
pressure point and that measurements may be discon-
tinued. If this condition is not satisfied, opera-
tion is continued until the baseline pressure does
-15-
~'7'~
pass through the diastolic point.
When the end of the operation shown in flow chart
3 has been reached, a series of readings consisting
of pairs of measured points will have been stored in
memory 140. If plotted on graph paper these points
would appear as shown in Figure 20 Before proceeding
to a specific description of the calculation of the
systolic and diastolic blood pressures the specific
characteristics of the readings shown in Figure 2
will be discussed. When plotted as in Figure 2, the
readings assume a "bell-shaped" curve~ As is well-
kno~n in the art, the baseline cuff pressure corres-
ponding to the maximum of the curve (point B) is
equivalent to the mean arterial blood pressure. The
systolic and diastolic pressure points occur where
the oscillation magnitude decreases to approximately
half its value at the maximum. At this point, or
example, in the vicinity of points A and C the slope
of the curve exhibits a marked change or "break-
point", (which is exaggerated in the figure to clar-
ify the description).
In accordance with the present inven~ion~ a cal-
culation of the breakpoint 203 in the slope of the
curve is made using additional data points in the
vicinity of the expected systolic pressure in order
to achieve an accurate calculation of the systolic
pressure. Specifically, a reading is chosen as a
starting point in which the oscillation magnitude is
-16-
~L8~'7~:~
approximately one half of the peak amplitude. This
would correspond to point 203 in Figure 2. Two sets
of three readings each are taken around point 203
corresponding to three readings with the baseline
cuff pressure greater than the pressure at point 203
and three readings with the baseline cuff pressure
less than the cuff pressure at point 20320 In the
illustrative embodiments these groups are chosen as
201 and 202 respectively. Using the three points
contained in group 201, a straight line approximation
(shown as 205) to the curve is made using standard
mathematical procedures. Similarly, a straight line
approximation (206) is made using the three points in
group 202 The pressure corresponding to the inter-
section of the two straight lines ~point D) corre~
sponding to cuf pressure E on Figure 2, is the cal-
culated systolic pressure. In order to calculate the
diastolic pressure, a reading is chosen on the dias-
tolic side of the curve in which the magnitude of the
oscillations shown by line 210 is equivalent to that
calculated for the systolic pressure. The corre-
sponding baseline cuff pressure (shown at point A) is
the calculated diastolic pressure.
As will be appreciated, these measurements can be
performed by hand, using a graph plotted from
measurements taken by the apparatus, so as to repro-
duce Figure 2, and constructing the straight lines
205 and 206 on the graph.
-17-
z~
In Figure 4, the operations performed by the
circuitry shown in Figure 1 in order to calculate the
mean, systolic and diastolic pressures are shown.
After obtaining and storing readings correspond-
ing to baseline cuff pressures in a range including
the expected systolic and diastolic pressures as
~escribed above, control circuit 150 selects two
readings corresponding to the highest baseline cuff
pressure and the next to highest baseline cuff
pressure (readinss 215 and 220 respectively in Figure
2). The magnitude of the oscillations in the two
readings are compared in step 403. If the magnitude
read last is greater than the magnitude read immedi-
ately before, step 402 is repeated and an additional
magnitude is read and compared to the one ~hich was
just previously read. This process is repeated until
the present magnitude reading is less than the prev-
ious magnitude reading. This occurs at point B in
Figure 2 corresponding to the mean arterial pressure.
In order to select a starting point for the syst-
olic pressure calculation, the value of the oscilla-
tion magnitude at the mean arterial pressure point (B
in figure two) is divided by two in step 404. In
step 405, the corresponding cuff pressure is determ-
ined. This would correspond to point 203 and
pressure C in Figure 2. Having determined the start-
ing point for the systolic pressure determination,
control circuit lS0 then reads from memory 140 the
-18-
7;2~
oscillation magnitudes and baseline pressure values
o~ a plurality of readings which were made previous
to the rea~ing 203. In the illustrative embodiment~
three readings are selectedO These would correspond
to the readings at points-202 in figure two. The
values are read into computation circuit 160. Under
control of control circuit 150, computation circuit
160 determines a mathematical relationship which best
"fits" the plurality of points read from memory 140.
Illustratively, the relationship may be a straight
line equation.
The determination of such a straight line equa-
tion can be accomplished in any number of well-known
mathematical techniques. Illustratively~ as a simple
approximation, one of the three readings may be
assumed to coincide with the mathematical averages of
the three readings. The sum of the errors is then
minimized, resulting in a line with a slope that is
the average of the slopes of two lines passing
through each of the remaining readings and a point
corresponding to the averages of the three readings.
Using this method, an equation is derived from the
readings which has the form.
0 = CM + B.
--lg--
7t~2~
Where O is the oscillation magnitude, C is the base-
line cuff pressure and M and B are constants equal to
the slope of the line and the Y-axis intercept.
Assuming, for the purposes of illustration, that the
points in an illustrative group have the coordinates
l' Cl' 2' C2; O3~ C3~ in accordance
wi~h the above approximation method the constants
and B are given by the following equations:
= 1/2 ~(U~ ) + (03 0
~Cl - C-) (C3 - C )~
and
B = O - C~
Where O and C are simple averages of the point coord-
inates given by the following equations:
= l + Q2 + 3
C = Cl + C2 ~ C3
The well-known technique of "least squares" approxim-
ation may also be used~ In accordance with the least
squares method of straight line approximation~ the
slope of the derived equation is as follows:
-20-
(C C-)(0 -0) + (C2-C )~2 ~ ) (C3 )~ 3
-
=
(Cl-C )2 + (C2-C)2 + (C3-C )2
~here O, C, and B are given by the previous equa-
tions. Utilizing either of these equations and the
coordinates of the readings for the points in group
203, a straight line approximation of the form:
0 = C ~l + B
P PP P
is obtained. The subscript p indicates the coeffic-
ients ~p and Bp are for readings taken previous to
the expected systolic point 203.
After determining the value of the ~ and B
coefficients for the first set of points, in step
408, control circuit 150 then reads from memory 140
the oscillation magnitude and corresponding baseline
pressure values of the three readings taken sub-
sequently to point 203 (corresponding to set 201 in
Figure 2). Using the readings thus obtained, in step
410, control circuit 150 determines a second equation
o~ the best fit to the readings. As previously
mentioned, an equation is obtained of the form:
s = CSMs + Bs
-21-
'7~
Where the subscript is indicates that the coeffic-
ients are determined for the set o~ readings taken
subsequently to point 203.
In accordance with the teaching of the invention,
the calculated systolic pressure appears at point D
in Figure 2 where the two lines determined by the
best-fit equations intersect. This point, o~ course
occurs where the calculated oscillation magnitudes
Op ana Os are equal. (In addition, the two base-
line cuff pressures will be equivalent at that point
Cp = Cs = Csys~olic ). To determine this point,
the two calculated eq-lations are set equal and solved
for the common baseline cuff pressure~ The calculat-
ed systolic blood pressure is given by the ~ollowing
equation:
Csystolic
~ ~ Ms
The calculated systolic pressure in Figure 2 corre-
sponds to point E. In step 415, the calculated syst-
olic pressure is stored.
The determination of the systolic pressure in
-22-
72~
accor~ance with the method of the invention results
in an accurate determination of the systolic value.
Even if one of the readings used in the approximation
should be erroneous due to patient movement or other
external noise, a reasonable approximation can still
be obtained. This operation is in contrast to
prior-art methods, such as differentiationl which are
quite sensitive to artifacts in the region of the
systolic pressure. In extremely noisy conditions
better noise and artifact rejection may be obtalned
by increasing the number of readings which are used
to make the approxima~ionO
In accordance with another aspect of the inven-
tion, the diastolic pressure is determined in accord-
ance with step 414 by first determining the oscilla-
tion magnitude corresponding to the calculated
systolic pressure. This can be easily determined by
substituting the calculated systolic pressure value
into either of the derived equations giving the
corresponding oscillation magnitude ~ystolic as
8ystolic Csystolic ~ ~
Control circuit 150, in accordance with step 414,
then searches through memory 140 for readings on the
diastolic portion of the oscillation magnitude ~-
cuff pressure curve to find the cuff pressure which
corresponds to the oscillation magnitude calculated
2~
immediately above (point A in Figure 2). In step 416
the calculated diastolic pressure is stored.
The storea systolic an~ ~iastolic pressures may
be displayed in any suitable manner by means of dig-
ital or analog devices which are well-known to those
skilled in the art.
Variations of the technique and apparatus dis-
closed herein with the spirit of the invention will
be obvious to those skilled in the art. For example,
the diastolic pressure may be cal~ulated first by
making two linear approximations to sets of points in
the vicinity of the expected diastolic pressure and
setting the approximations equal exactly in the
manner disclosed herein for calculating the systolic
pressure. The systolic pressure may then be derived
by determining the cuff pressure at which the oscill-
ation amplitude is equivalent in magnitude.
In addition, mathematical relationships other than
straight line equations may be derived from the sets
of readings according to well-known approximation
techniques.
-24-
