 |
Description  |
|
|
BACKGROUND OF THE INVENTION
(a) Field of the Invention
This invention relates to arterial blood pressure monitoring and more
particularly to a method and apparatus for continuous and nonn-contactive
blood pressure measurement.
(b) Description of the Prior Art
Continuous blood pressure monitoring techniques suffer from the problems of
either being invasive or occluding the flow of blood. Invasive
measurement, by way of arterial catheterization has been commonly used in
intensive care units and operating rooms for a number of years. However,
the risks of infection, thrombus formation, hemorrhage, etc., have given
rise to a search for non-invasive approaches which would provide the
desired continuous and accurate measurement.
The typical non-invasive approaches taken to date include devices using the
so-called tracking cuff principle, e.g., see U.S. Pat. No. 4,524,777. The
device disclosed in this patent utilizes a hydraulic servocontrol system
to maintain a finger arterial volume constant (in an unloaded state) so
that the counter cuff pressure follows the intraarterial blood pressure
thus giving an instantaneous arterial blood pressure measurement from the
counter cuff pressure. Although this method does provide continuous and
non-invasive measurement of arterial blood pressure, long-term maintenance
of cuff pressure restrains microcirculation in the finger capillaries and
this causes pain. Also, a downward drift occurs for some period of time
before stability is reached.
One other non-invasive approach, but which is not continuous, uses a
sphygmomanometric technique based on Riva-Rocci's principle in which
auscultatory (sound) measurements of blood flow are made to determine the
arterial blood pressure (systolic and diastolic only), again using cuff
pressure. These measurements are less desirable not only because they are
not continuous but also because they are not as accurate.
Another non-invasive approach, known as the oscillometric method, utilizes
volume measurements, rather than auscultatory measurements, and cuff
pressure to more accurately determine systolic blood pressure. This method
also measures mean blood pressure, but not diastolic pressure.
An accurate, non-invasive, continuous and non-occlusive (does not occlude
the flow of blood) apparatus and method would be valuable for use in
intensive care units and operating rooms to avoid complications which can
arise with the above-described prior art devices and methods.
SUMMARY OF THE INVENTION
It is an object of the invention to provide a method and apparatus for
measuring arterial blood pressure in a continuous, non-occlusive fashion.
The above and other objects of the invention are realized in a specific
illustrative embodiment thereof which includes a continuous, indirect and
non-occlusive blood pressure monitor operable under control of a
microprocessor. The monitor icludes an annular-inflatable cuff for
placement about a patient's finger. Positioned in the cuff are a pressure
transducer for producing a signal indicating the cuff pressure, a light
emitting diode positioned on one interior side of the cuff, and a
photoelectric transducer positioned on the opposite interior side of the
cuff. The light emitting diode produces light which is partially
transmitted through a patient's finger to the photoelectric transducer
which detects the light level or intensity and produces a signal
indicating arterial volumetric changes in the finger. The
volume-indicating and pressure-indicating signals are amplified and then
supplied to anolog-to-digital converters, with the resultant digital
output being supplied to the microprocessor. The microprocessor controls a
ramp pressure generator which is coupled to the cuff to alternately
inflate the cuff (causing the pressure to increase linearly) and deflate
the cuff.
The above-described monitor is first used in a calibration cycle in a
temporarily occlusive manner to determine certain parameters which will
then be used by the monitor to continuously and non-occlusively measure
the arterial pressure waveform which includes the mean arterial blood
pressure P.sub.m, systolic blood pressure P.sub.s and diastolic blood
pressure P.sub.d, which are the measurements of interest. At first, the
microprocessor determines the mean blood pressure P.sub.m and systolic
blood pressure P.sub.s in a conventional way (using the well-known
oscillometric method) from the measured volume and pressure signals. With
the relative volume signal collected during cuff pressure application, the
cuff pressure is then relieved. The microprocessor then makes a first
estimate of diastolic blood pressure P.sub.d (using a formula for
estimating P.sub.d from P.sub.m and P.sub.s), and uses that estimated
value, along with the measured mean blood pressure and systolic blood
pressure, in a recursive procedure based on a pressure-volume relationship
known as the Hardy model, to derive a calculated mean blood pressure
designated as P.sub.M. The calculated mean blood pressure P.sub.M is
compared with the measured mean blood pressure P.sub.m and if the
difference between the two is greater than some predetermined standard,
then a new estimated diastolic blood pressure P.sub.d is used in the
recursive procedure to obtain a new mean calculated blood pressure
P.sub.M. This recursive procedure continues until the calculated mean
blood pressure P.sub.M is within a certain range of the measured mean
blood pressure P.sub.m.
In the course of determining the calculated mean blood pressure P.sub.M,
three parameters are also developed to define a Hardy model compliance
curve for the particular patient in question. These parameters include k
which represents the compliance index for the blood vessels of the patient
being rreated and is unique to that patient, V.sub.m which represents the
maximum volume of the vessels in the patient's finger (being examined),
and V.sub.0 which represents the volume of the patient's finger vessels at
zero pressure. With these parameters determined for the patient, the Hardy
model compliance curve may then be used to relate the relative blood
vessel volume the patient's arterial blood pressure and vice versa. With
the Hardy model parameters and compliance curve for the patient in
question being determined, the blood pressure monitor may now be utilized
to continuously measure the relative volume V as a function of time from
which the continuous arterial pressure waveform as a function of time can
be determined using the Hardy model.
Periodically, recalibration will be carried out, i.e., new Hardy model
parameters will be developed for use in making the continuous blood
pressure measurements.
As can be seen, except for the calibration cycle, continous and
non-occlusive blood pressure measurement may be carried out. This avoids
the pain and trauma of both the invasive blood pressure measuring
techniques and the occluding blood pressure measuring techniques.
BRIEF DESCRIPTION OF THE DRAWINGS
The above and other objects, features and advantages of the invention will
become apparent from a consideration of the following detailed description
presented in connection with the accompanying drawings in which:
FIG. 1 is a circuit diagram of an automatic, continuous and non-occlusive
blood pressure monitor made in accordance with the principles of the
present invention; and
FIGS. 2A through 2C show typical waveforms of the pressure-volume
relationship of a blood vessel, of a photoelectric plethysmograph, and of
arterial blood pressure, respectively.
DETAILED DESCRIPTION
The present invention, except for occasional calibration requirements,
provides for continuous, noncontactive or nonpressure imposed arterial
blood pressure monitoring. The blood pressure measurements, in the form of
a pressure waveform, are obtained by calculation from a measured arterial
volume signal. The arterial volume signal, which is a relative measurement
of the arterial volume over time, is developed using optical sensing
techniques which are generally known. The ultimate determination of the
pressure waveform is made using the so-called compliance model of Hardy
and Collins (hereinafter referred to as the Hardy model) discussed in
Hardy, H. H. and Collins, R. E., "On the Pressure-Volume Relationship in
Circulatory Elements", Med. and Biol., Eng. and Comput., September, 1982,
pages 565-570.
The Hardy model shows that under static conditions, the pressure-volume
(P-V) relationship of blood vessels may be described by
k(V.sub.m -V)=dV/dP, P.gtoreq.0 (1)
where P represents the transmural pressure (the difference in pressure
inside and outside the artery), V.sub.m is the limiting or maximum volume
of the blood vessel in question and k is a physiological constant which
characterizes the elasticity of the vascular wall. The constant k is
sometimes referred to as the vascular compliance index. Solving the
differential equation (1) yields the following pressure/volume
relationship
V=V.sub.m -(V.sub.m -V.sub.0)e.sup.-kP, P.gtoreq.0 (2)
where V.sub.0 is the vessel volume under zero transmural pressure. Equation
2 is referred to as the transformation of the P-V relationship. Combining
equations 1 and 2 gives the following equation
dV/dP=k(V.sub.m -V.sub.0)e.sup.-kP, P.gtoreq.0 (3)
which defines the vascular compliance of a vessel (unique to each
individual).
With the above described Hardy model, the relationship between pressure P
and absolute volume V is established. However, it would be advantageous to
define the pressure waveform P(t) in terms of relative volume V(t) which
is a parameter measureable by a photoelectric plethysmogram. The
relationship between absolute volume V(t) and relative volume V(t) is
given by
V(t)=V.sub.x +B V(t) (4)
where B is the mapping coefficient from the analog signal V(t) to the
absolute arterial volume as illustrated in FIG. 2A, and V.sub.x is a
parameter defining the contribution of tissue absorption of transmitted
light emitted by the photoelectric plethysmogram. In other words, there is
a linear relationship between the relative volume V(t) and absolute volume
V(t), in which V.sub.x is the intercept, and B is the slope.
Now define systolic volume V.sub.s and diastolic volume V.sub.d in
accordance with equations 2 and 4 to get
##EQU1##
where V.sub.s and V.sub.d are analogue systolic and diastolic volumes
respectively. Substracting equation 6 from equation 5 gives the mapping
coefficient
##EQU2##
The addition of equation 6 and equation 5 gives the mapping intercept
V.sub.x.
##EQU3##
Taking the log of both sides of equation 2, substituting equation 4, and
rearranging terms, gives the pressure waveform
##EQU4##
which may be called the inverse transformation of the P-V relationship
with respect to equation 2. Combining equations 7, 8, and 9, the
calibrated blood pressure as a function of the analogue arterial volume
signal is obtained, i.e., the inverse transformation of P(t)-V(t)
relationship, given as
##EQU5##
where V(t) is the continuous relative volume as a function of time,
P.sub.s is the systolic pressure, P.sub.d is the diastolic pressure,
V.sub.s the systolic relative volume, and V.sub.d is the diastolic
relative volume all of which can be noninvasively measured as will be
discussed hereafter. Thus, if these last mentioned parameters and the
arterial compliance index k are noninvasively determined in advance, the
arterial blood pressure (P(t) when cuff pressure is zero) can be developed
by measuring the relative volume V(t) and using equation (10).
The relative systolic volume V.sub.s and relative diastolic volume V.sub.d
may be non-invasively determined in a conventional fashion as shown in
FIG. 2A. The systolic pressure P.sub.s may be non-invasively determined
using conventional oscillometric methods, as mentioned earlier.
Specifically, the systolic pressure may be determined by the cuff pressure
at which the pulsatile plethysmograph developed using the oscillometric
method disappears. This is well known in the art. The diastolic pressure
P.sub.d cannot be determined directly from pulsatile information of a
plethysmograph, but can be determined using an iterative procedure to be
descibed momentarily. Calculation of the compliance index k is dependent
on parameters which include the diastolic pressure P.sub.d. The iterative
procedure to be described will yield both P.sub.d and k to thus provide
the five parameters V.sub.s, V.sub.d,P.sub.s,P.sub.d and k needed in
equation 10.
A least square approach may be used to determine the compliance index k
which will now be described. First, refer to FIG. 2B which shows the
photoelectric plethysmograph volume .DELTA.V under the corresponding
transmural pressure P=P.sub.b -P.sub.c, where P.sub.b and P.sub.c are
intra-arterial and cuff (measured) pressures respectively. Note the
established fact that the amplitude of the pulsatile volume .DELTA.V is
maximum when P.sub.c equals the mean blood pressure P.sub.m (see FIG. 2B).
Equation 3 can be used to relate mean blood pressure P.sub.m cuff
pressure P.sub.c, the amplitude of pulsatile photoelectric plethysmograph
.DELTA.V, and pulsatile blood pressure .DELTA.P as follows:
.DELTA.V/.DELTA.P.perspectiveto.k(V.sub.m -V.sub.0)e.sup.-k (P.sub.m
-P.sub.c), (11)
where .DELTA.P=(P.sub.s -P.sub.c)-(P.sub.d -P.sub.c)=P.sub.s -P.sub.d,
P.sub.s and P.sub.d being systolic and diastolic blood pressure
respectively. Assume that intra-arterial blood pressure remains the same
in the time of calibration, i.e., ramp cuff pressure application. A set of
n equations representing the same relationship of equation 11, under
different ramp cuff pressures can then be acquired as follows
##EQU6##
Equation set 12 is linearized by taking the natural logarithm with respect
to equation 11, giving
##EQU7##
which has the linear form of
y=ax+b (14)
where
##EQU8##
which transforms the set of n nonlinear equations 12 into the set of n
linear equations, given as
##EQU9##
where y.sub.i =ln(.DELTA.V.sub.i /.DELTA.P),x.sub.i =P.sub.ci,(i=1,2, . .
. ,n) The coefficients and their uncertainties are obtained from equation
set 16 applying standard least mean square error analysis, thus providing
the estimate of the arterial compliance index k of the Hardy model, in
terms of x and y, i.e.,
##EQU10##
in which
##EQU11##
Had an accurate estimate of P.sub.d been used in equations 15-17, the Hardy
model parameters would be correctly determined. However, to obtain an
accurate value for P.sub.d, the following iterative procedure is used:
1. Select an initial value of P.sub.d to be P.sub.dj =3/2 P.sub.m -1/2
P.sub.s where P.sub.m and P.sub.s are measured using known oscillometric
methods.
2. Calculate the compliance index k.sub.j based on P.sub.dj using equation
17.
3. Calculate P(t).sub.j in accordance with equation 10 using k.sub.j and
the measured relative volume V(t).
4. Calculate a mean pressure in accordance with the formula
##EQU12##
5. Compare the calculated mean pressure P.sub.Mj with the measured mean
presure P.sub.m and if the difference is less than one mmHg the procedure
is stopped, otherwise another "estimated" diastolic pressure P.sub.d(j+1)
is determined by the gradient method from the formula
P.sub.d(j+1) =P.sub.dj -(P.sub.m -P.sub.mj).differential.(P.sub.m
-P.sub.mj)/.differential.P.sub.dj (20)
The procedure then returns to step 2 above.
When the procedure yields a calculated mean pressure which is within the
predetermined range of the measured mean pressure, needed parameters for
the Hardy model will have been obtained, namely k, V.sub.m and V.sub.0. In
effect, the parameters necessary to develop a Hardy model compliance curve
for a patient are determined by the above iterative procedure. With this
information, the blood pressure waveform P(t) can be produced from the
measured relative volume V(t) on a continuous, nonocclusive basis.
Apparatus for carrying out the desired measurements, both for obtaining the
parameters for the Hardy model and for producing the waveform P(t), is
shown in FIG. 1. The apparatus includes an inflatable annular finger cuff
4 having a rigid outer wall 8 and a resilient inflatable annular bag 12
held in place by inwardly extending end walls 16 and 20. The bag 12 is
filled with air or other fluid for producing a pressure on a finger 24
when the finger is inserted into the cuff. Disposed on the interior wall
of the bag 12 between the bag and the finger 24 and on one side of the
cuff 4 is a light emitting diode 28. Disposed on the interior wall of the
bag 12 on the other side of the cuff 4 is a photoelectric transducer 32
for detecting light transmitted from the light emitting diode 28 through
the finger 24 to the photoelectric transducer. The amount of light
reaching the transducer 32 is proportional to the volume of the blood
vessel or vessels positioned between the diode 28 and transducer 32. The
output signal of the transducer therefore represents the relative volum
V(t) of the blood vessel or vessels in question.
Disposed inside the bag 12 is a pressure transducer 36 for producing a
signal representing the cuff pressure P.sub.c being applied to the finger
24 by the bag 12 when it is inflated. A ramp pressure generator 40, of
conventional design, supplies air to the cuff 4 in response to analog
signals from a digital-to-analog converter 44 which receives the digital
counterparts of the analog signals form a microprocessor or microcomputer
48. The microcomputer produces signals for causing the ramp pressure
generator 40 to alternately inflate the bag 12 with a linearly increasing
ramp pressure, and then deflate the bag.
The relative volume signal V(t) produced by the photoelectric transducer 32
is supplied to an A.C. volume amplifier 52 which amplifies the A.C.
component of the relative volume signal, and to a D.C. volume amplifier 56
which amplifies the total relative volume signal. The outputs of the two
amplifiers 52 and 56 shown graphically at 60 and 64 respectively are
supplied to analog-to-digital converters 68 where the analog signals are
converted to digital form for application to the microcomputer 48. The
output P.sub.c of the pressure transducer 36 is also supplied to an
amplifier 68 which amplifies the signal shown at 72 for application to the
analog-to-digital converters 68 for conversion to digital form and
ultimate transfer to the microcomputer 48.
The signals received from the analog-to-digital converters 68 are used by
the microcomputer to both calculate the parameters needed for the Hardy
model (calibration for a particular patient), and produce the arterial
blood pressure waveform P(t), shown at 76, from which the systolic,
diastolic and mean blood pressures can be obtained. This information is
displayed in real time on a display unit 80 which might illustratively
include a CRT and digital displays. A power supply unit 84 provides power
for operation of the monitor.
Use of the blood pressure monitor shown in FIG. 1 will now be described. A
patient whose blood pressure is to be determined inserts his finger into
the annular cuff 4, and the bag 12 is inflated to apply pressure to the
finger. As this is being done, relative volume signals V(t) and cuff
pressure signals P.sub.c are developed by the photoelectric transducer 32
and pressure transducer 36 respectively and supplied to corresponding
amplifiers. The signals are amplified and supplied to the
analog-to-digital converters 68, with the digital versions being supplied
to the microcomputer 48. From these signals, the microcomputer 48 develops
the mean pressure P.sub.m and systolic pressure P.sub.s using conventional
oscillometric methods. The cuff pressure is then released so that there is
no occlusion of the finger blood vessels. The microcomputer 48 then begins
the iterative procedure described earlier by calculating a first estimated
diastolic pressure P.sub.d which, along with the measured mean pressure
and systolic pressure are used to obtain Hardy model parameters for
defining a Hardy model compliance curve. From this compliance curve, a
blood pressure waveform P(t) is generated and then from that a mean
| | |
