Continuous measurement of tissue blood flow
by laser-Doppler spectroscopy
MICHAEL
G. ALLEN
D. STERN, DONALD L. LAPPE, PATRICK D. BOWEN, JOHN E. CHIMOSKY,
HOLLOWAY,
JR., HARRY R. KEISER, AND ROBERT L. BOWMAN
STERN, MICHAEL
D., DONALD L. LAPPE, PATRICK D.
BOWEN, JOHN E. CHIMOSKY, G. ALLEN HOLLOWAY, JR.,
HARRY R. KEISER, AND ROBERT L. BOWMAN. Continuous
measurement of tissue blood flow by Laser-Doppler
spectroscopy.
Am. J. Physiol. 232(4): H441-H448,
1977 or Am. J. Physiol.:
Heart Circ. Physiol. l(4): H441-H448,
1977. -Laser
light scattered from tissue in vivo is broadened in line width as a result
of the Doppler
shift produced
by moving
red cells in the
microcirculation.
A feasibility
study was carried out to demonstrate use of this effect to measure and monitor tissue blood
flow. Light from a helium-neon
laser illuminated
a l-mm area
of tissue (human skin or rat renal cortex), and the backscattered light was detected with a photomultiplier.
The spectrum
of the Doppler beat notes was analyzed directly with a digital
spectrum analyzer, or processed by analog circuitry to yield a
flow parameter
based on the root-mean-square
Doppler line
width. This parameter
was compared with 133Xe washout in
the skin of volunteers
subjected to UV-induced
erythema and
was found to vary in an approximately
linear manner with
skin blood flow. The laser instrument
provided
continuous
monitoring
of blood flow fluctuations,
including
the pulsatile
,component. The instrument
was used to monitor flow in the
outer cortex of the rat kidney during administration
of norepinephrine, angiotensin,
hydralazine,
dextran, dopamine, nitroprusside, and angiotensin
blocked by saralasin. Dynamic and
steady-state
responses were consistent with known pharmacology and renal physiology,
and with the assumption
that
vasoconstrictor
angiotensin
II receptors in the kidney are accessible to blood-borne inhibitors.
The laser-Doppler
method is
a promising
tool for rapid monitoring
of dynamic changes in
tissue perfusion.
microcirculation;
effect; xenon-133;
rine; hydralazine;
capillary
blood flow; coherent light; Doppler
renal blood flow; angiotensin;
norepinephrenal cortex; blood flow velocity
THE NEED TO MEASURE local blood flow in small volumes
of tissue arises in many contexts in physiology, pharmacology, and clinical medicine. Available approaches include, depending on the tissue involved, washout of
radioactive indicators (11, 15; 16), thermal uptake with
heated probes (13), various forms of plethysmography
(22), implanted hydrogen electrodes or radiation detectors (1, lo), and the trapping of labeled microspheres.
An ideal method would provide a continuous reading of
the blood flow in a small, localized region of tissue; it
H441
would not disturb the state of the circulation, and would
be indefinitely
repeatable for the same preparation; it
would also provide absolute values independent of tissue characteristics,
and respond instantaneously
to
changes. At present, no method approaches this ideal.
Laser-Doppler velocimetry (6) has been used to measure the motion of particulates in many applications.
The possibility of using this method to estimate the
velocity of flow in macroscopic vessels (via fiberoptic
catheter) (19) and in retinal arteries and veins (20) has
been demonstrated.
Following up on the observation
that laser light that has passed through a fingertip does
not manifest the speckle phenomenon, we reported previously that coherent light scattered from skin shows
Doppler broadening, which is a sensitive index of flow in
the microvascular
bed. Here we report the results of
feasibility studies showing that this phenomenon can
provide a useful method for measuring the local blood
flow in tissues. Its principal advantages are that it
provides virtually instantaneous and continuous monitoring, good localization, and appears to be essentially
non-invasive (beyond the requirement of exposure of the
tissue). Its drawbacks are that it is limited to the penetration depth of light (unless an invasive fiberoptic
probe is used) and that it requires calibration against
another method of flow measurement, since it is sensitive, in principle ? to changes in the vascular bed geometry and optical properties of the tissues. We have shown
that this method can be calibrated against the 133Xe
uptake technique in h .uman ski .n, and that it can be
used to monitor flow m the outer cortex of the kidney in
the rat during a variety of pharmacologic interventions.
Promising preliminary
results indicate that similar
flow signals can be obtained from brain, liver, and other
tissues.
METHOD
A small area of tissue is illuminated by the beam of a
helium-neon laser (632.8 nm), and one coherence area of
the backscattered radiation is sampled by a pinhole and
allowed to fall on a photomultiplier
tube. The backscattered light consists of Doppler-shifted light which has
interacted with moving red cells and unshifted light
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
Laboratory of Technical Development and Laboratory
of Experimental
Therapeutics,
National Heart, Lung, and Blood Institute, National Institutes of Health,
Bethesda, Maryland 20014; and Division of Bioengineering,
University of
Washington School of Medicine, Seattle, Washington 98195
H442
M.
.=qxzii
- sI)/I
ET
AL.
where I is the mean photocurrent
and s is a constant
which depends on the gain of the photomultiplier.
Equation 2 includes a division by1 which serves to normalize
the flow parameter
so that it is independent
of the
intensity of the laser and the total reflectivity
of the
tissue. In practice, s is most conveniently fixed by zeroing the instrument
on a stationary object; the zero of
flow does not, therefore, depend on a biological calibration. The arguments leading to equation 2 are detailed
in the APPENDIX.
The configuration
of the instrument,
as used for studies of the kidney, is shown in Fig. 1. A 15mW laser with
a beam diameter of about 1 mm (Jodon Engineering
Associates, HN-15) is used. The returning light passes
through a 2-mm aperture at the surface of the tissue,
which limits stray, scattered light, and a .&mm pinhole
1 m away, which selects one coherence area. An interference filter (632.8 nm, 3 nm bandwidth) rejects ambient light; the light is detected by an EM 9658-R photomultiplier with high quantum efficiency in the red. The
use of this wavelength
was dictated by availability
of
the helium-neon
laser. In theory, other wavelengths,
such as the 800-nm isosbestic point of hemoglobin and
oxyhemoglobin,
might be preferable, although the easily available He-Ne line has proved entirely satisfactory
in practice. The mean photocurrent
and the rms detector signal are recorded on a strip chart (Beckman type-R
Dynograph).
The flow parameter may be computed by
an on-line calculator, using equation 2, or by analog
circui .try. Simul taneously, the Doppler spectrum is computed by a real .-time spectrum analyzer (Saicor, 51-B)
and recorded on an X-Y plotter. The spectra shown in
the illustrations
have been corrected for the shot noise
(which is perfectly white) by a computation analogous to
equation 2.
(1)
where P(u) is the power spectrum of the Doppler signal
as a function of Doppler frequency. In reality, changes
in tissue perfusion and red cell content are generally
accompanied by some change in the vascular geometry,
red cell flow is not laminar, and red cells have a spinning motion as they flow. Therefore, the belief that F is
proportional
to flow must be based on empirical evidence. In practice, we find this to be so under a surprisingly wide range of circumstances
(for example, the
skin of different volunteers with or without ultraviolet
(UV)-induced eryth ema). However, there is no a priori
reason to expect that the calibration coefficient will be
the same in different tissues.
The definition of the flow parameter (equation 1) is
somewhat arbitrary; other weighted estimates of bandwidth might do as well. It was chosen because it can be
evaluated in real time by means of relatively simple
analog circuitry, obviating the need for a spectrum analyzer. This is done by electronically differentiating
the
photocurrent
after passing it through a low-pass filter
wide enough to pass the entire Doppler spectrum. The
power of the resulting signal contains two terms, one
due to the Doppler signal, which can be shown to be
equal to F2, and the other due to the shot noise of the
photomultiplier
tube, which is proportional to the mean
photocurrent
and independent of flow. Therefore, ifR is
the output of an rms detector that receives the differentiated signal, the flow parameter F can be computed as:
F = d(R2
STERN
(2)
Measurements
on Human
Skin
Spectra from human skin and the time course of
occlusion and release of the brachial artery by a bloodpressure cuff have been reported previously (18). To
determine
if the flow parameter F could serve as a
quantitative
measure of flow, we compared the method
with washout of 133Xe injected intradermally
in saline.
Studies were conducted on seven healthy Caucasian volunteers of ages 21-32 yr. Erythema was produced on
both forearms 5 cm below the elbow by exposure to a UV
lamp. The methods of producing and grading the erythema, and the 133Xeuptake technique have been previously described (4, 5). On the day following exposure to
UV, blood flow was measured by both techniques in the
erythema region, and in neighboring
normal skin on
either or both forearms. If the subject’s time permitted,
measurements were made again later i n the day when
the erythema had begun to fade. Five points within a
2.2-cm circle (arranged as the 5 spots -on a die) were
measured with the Doppler instrument,
averaging for
20 s at each point. The flow was then measured by
uptake of 13”Xe injected at the center of the circle, after
which the Doppler measurements were repeated. All 10
Doppler measurements
were averaged to minimize the
effects of point-to-point
fluctuations
(mottling) and time
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
which has been scattered by fixed cells and stroma. At
the photomultiplier
cathode the Doppler-shifted
components interfere with the unshifted light and with each
other, producing spectrum of beat notes at audio frequencies related to the velocity distribution
of the red
cells. Since beat frequencies depend on the relative
velocities of the red cells and fixed tissues, overall motion of the entire tissue has little effect on this spectrum .. Because of multiple scattering, the inciden t light
may have its direction partly randomized before i .nteracting with red cells, and frequency multiples can be
produced by scattering from multiple red cells. These
complex effects preclude exact calculation of the shape
of the Doppler spectrum. However, a dimensional argument shows that the bandwidth of the Doppler spectrum
shoul .d scale in proportion to the red cell veloci .ties if the
latter are changed-without
altering the flow geometry.
For viscous flow in a fine network,
this bandwidth
should scale in proportion to red cell flow, given fixed
geometry, and this has in fact been demonstrated experimentally for blood in glass capillary tubes (20). On the
other hand, the magnit ude of the Doppler signa 1 is
expec ted to i ncrease with the nu .mber of red cells in the
flow field (not necessarily linearly because of multiple
scattering). These arguments support the heuristic delinition of the Doppler flow parameter (F) as the unnormalized root-mean .-square (rms) bandwidth of the Doppler signal:
D.
TISSUE
BLOOD
FLOW
BY
LASER
H443
DOPPLER
Analog
,/j
,
)-
Low-Pass
Filter
-
Mean PMT
Current
Differentiator
Ic
Digital
Voltmeter
5
RMS
Detector
- *
FIG. 1. Schematic
setup used in studying
in outer cortex
of rat
Doppler
spectroscopy.
tails of method.
Rate
RMS Detector
Signal
w
Digital
Voltmeter
-
Heart
Current
of experimental
local blood flow
kidney
by laserSee text for de-
Flow Parameter
Calculation
(on-line)
A
v
*
X-Y
Shot Noise
Subtraction+
Corrected
Power
Spectrum
Recorder
Spectrum
Analyzer
*
c
Oscilloscope
(off -line)
-
Analog
Tape
Recorder
PMT
- Photomultiplier
Pressure
variations in local skin blood flow due to sympathetic
tone. Throughout
the measurements
the subject remained supine at an ambient temperature of 22 t 1°C;
30 min equilibration time was allowed prior to starting
the experiment.
Table 1 shows the results of these measurements for
each of the subjects. Plotting the Doppler values (arbitrary units) against the xenon values reveals a positive
biasin the Doppler values. The likely cause of this is the
effect of motion of the forearm, which was controlled
only by an ad hoc arrangement of sandbags. Motion of
the tissue introduces a “motion noise” into the flow
signal which is uncorrelated with the signal due to red
cell flow. It therefore adds in quadrature to the flow
parameter:
F2meas
=
F2true
+ N2
Tube
Transducer
- P23db
TABLE
Statham
1. Results of measurements
Grade of
Erythema*
Skin
Tt;Py
Xenon
Flow,
ml/min
per 100 g
Laser
Flow
arb units
42
3
C
32.1
30.7
42.2
25.3
5.51
1.61
120/65
35.5
2
1
C
2
32.0
30.5
30.7
33.0
41.6
28.4
10.5
50.5
6.55
5.21
1.89
7.00
M
122168
45
3
3
C
32.6
33.0
30.9
51.1
43.7
16.7
7.35
6.50
2.82
22
M
116178
44
3
3
C
33.7
34.9
32.7
61.8
68.3
27.7
7.76
6.31
2.47
G.L.
27
F
120/78
38.5
2
C
2
32.1
31.5
32.1
29.8
5.5
25.5
4.55
2.57
3.15
A.K.
21
F
138184
39
2
32.0
32.3
5.67
G.P.
26
M
132178
43
2
3
34.1
33.2
Age, yr
Sex
Blood
Pressure,
mmHg
P.B.
23
M
120/74
S.W.
32
F
M.L.
24
S.L.
Subject
Hct,
7%
(3)
where N is the flow equivalent of the motion noise. In
order to obtain an unbiased estimate of the calibration
coefficient of the instrument,
a linear regression was
performed between the squares of the xenon and Doppler flow measurements. The data were least-squares fit
to the regression:
(4)
Go, = A2$‘2xenon +B
and the value of A was used as the calibration coefficient to convert the arbitrary Doppler units to milliliters per minute per 100 g. Figure 2 shows a plot of the
calibrated Doppler values against the corresponding xenon values. Since scaling of only the Doppler axis was
done, the shape of this plot and the first-order correlation coefficient of 88% do not depend on the normalizing
procedure just described.
The Doppler
instrument is capable of recording the
-pulsatile components of microvascular
flow; however,
for each subject
-
* C, control;
1 = minimal
redness;
3 = fully developed
mottled
redness.
redness;
54.0
8.63
64.8
8.71
. .
2 = mlnlmal 1 confluent
these are partially obscured by the presence of fluctuations of similar frequency, presumably due to variation
of the tone of the microvessels in the small region samples. To extract an accurate phasic wave form, we averaged the Doppler flow synchronously over a number of
cardiac cycles using the QRS complex of the electrocardiogram (ECG) to gate a digital correlator-probability
analyzer (Saicor, 42-A). In this mode the correlator is
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
-
PMT
M.
15
XENON
30
FLOW
4s
HIL/tlIN/100
60
75
GfW
ls
TIM
RFTER
1
QRS ( SECONOS
1.5
STERN
ET
AL.
2
FIG. 2. Laser-Doppler
flow parameter
plotted
against
corresponding 133Xe washout
flow measurement
for subjects
in Table 1. Laser
axis has been scaled to give absolute
units by procedure
described
in
text. Solid line is line of equality,
not of best fit.
FIG. 3. Pulsatile
component
of skin flow measured
by laser-Doppler method,
obtained
by averaging
many
pulse cycles
synchronously
with ECG. Units
are arbitrary
and only pulsatile
component
is shown.
Steady-state
flow is ca. 10 times larger.
used as a computer of average transients, calculating
the average of the flow parameter from (typically) 150
consecutive cardiac cycles superimposed,
so that the
QRS-complex of each cycle (as detected by a HewlettPackard 780-7A ECG monitor) appears as time zero. The
wave form of the phasic component as obtained from a
fingertip is shown in Fig. 3. It represents about 10% of
the steady flow. The pulse-wave propagation time and
dicrotic notch are visible. The appearance of the dicrotic
notch is of interest, although it is not possible to determine whether or not it is present in true capillaries
since arterioles and venules are also present in the
region of tissue sampled.
by blunt dissection, supported on a glass half-ring, and
bathed in mineral oil at 37°C as for micropuncture.
Arterial pressure was monitored with a strain-gauge
manometer
(Statham, P23db). The electrocardiogram
and heart rate were monitored by a standard ECG
monitor (Hewlett-Packard,
780-7A) connected to a frequency meter (Hewlett-Packard,
500C). Drugs were infused intravenously
with a syringe infusion pump (Harvard Apparatus, 933) at flow rates not exceeding .05 mg/
min. Comparable
infusions of physiologic saline produced no effect. Preparations
in which blood pressure
fell below 90 mmHg or heart rate exceeded 350 beats/
min under control conditions were discarded. These animals were usually found to be bleeding, and were
clearly distinguishable
from the experimental
group.
The following drugs were used in the experiments: norepinephrine
(Levophed,
Winthrop),
hydralazine
HCl
(Apresoline, CIBA), angiotensin II amide (HypertensinCIBA), saralasin acetate (P-113, Norwich), dopamine
HCl (Intropin,
Arnar-Stone),
isoproterenol
(Isuprel,
Winthrop),
Dextran 70 (Macrodex, Pharmacia),
epinephri .ne (Parke-Davis), and sodium nitroprusside
(Nipride, Roche). Steady-state dose-response curves were
obtained for the short-acting
drugs, each drug being
infused at constant rate until all monitored parameters
became steady, at which time a reading was taken. The
animals were allowed to recover to a steady base-line
between points, which were taken in random order.
Doses were chosen to cover a full range from threshold
effects to levels at which intolerable side effects (e.g.,
cardiac arrhythmia,
severe hypotension) occurred.
For long acting hydralazine,
the animal was permitted to achieve a steady base-line (variation
in flow
parameter
of less than 5% over 30 min), after which
single intravenous
doses of .l mg/kg were given at 5min intervals until either a maximum increase in flow
parameter
was obtained or mean blood pressure fell
below 50 mmHg. The mean blood pressure was then
Measurements
on Rat Renal Cortex
The action of pharmacologic
agents on the regional
blood flow of internal organs is a matter of some importance physiologically
and for drug screening. The nonuniform distribution
of blood flow to the anatomic regions of the kidney (12) and the unique sensitivity of the
outer cortex to vasoconstrictors
(2) are of special interest, because the outer cortex of the kidney may be the
target organ for vasoactive interventions,
both deliberate and inadvertent.
We have demonstrated
the use of
the Doppler instrument
for rapid monitoring of the perfusion of the outer cortex of the rat kidney.
Preparation.
Male Sprague-Dawley
rats weighing
350-450 g were anesthetized with sodium pentobarbital
(Veterinary
Laboratories),
6 mg/kg, intramuscularly.
The trachea was cannulated with 3 cm of no. 12 thinwalled polyethylene
tubing.. The carotid artery on one
side was cannulated with .023=inch ID polyethylene tubing and the internal jugular vein was cannulated with
.Oll-inch ID polyethylene tubing. When two drugs were
to be infused simultaneously,
the femoral vein was also
cannulated. The left kidney was exposed by left abdominal approach, freed from underlying
connective tissue
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
0
D.
TISSUE
BLOOD
FLOW
BY
LASER
H445
DOPPLER
restored to 100 mmHg by infusion of 6% Dextran 70
(maximum
3 ml), and a further reading of the flow
parameter was obtained.
RESULTS
100 I@
FLOW
PFlRFlMETER
(% CONTROL)
DOSE
=
10.3X
(MCG/MINl
A
DOSE
q
4.52X
IMCG/MINI
17
DOSE
= 4.44X
tMCG/MIN)
MCG/M IN
2.5
A
MCG/M
1
MCG/M
3.5
MCG/M
4.0
MCG/M
A
0
-5
1
5
FREQUENCY
[KHZ)
LOG
4. A family
of Doppler
spectra
obtained
same rat at varying
dose rates of norepinephrine
nously.
Spectra
were obtained
with a real-time
applied
to output
of photomultiplier
(see text).
has been subtracted.
Units
on vertical
axis are
equal 100 for intercept
of control
spectrum.
FIG.
10
20
SCFlLE
from
kidney
in the
infused
intravespectrum
analyzer
Shot noise constant
arbitrary,
chosen to
0
.5
X = SCQLED
1
NOREPINEPHRINE
1.5
DOSE
2
RFlTE
FIG. 5. Steady-state
dose response
of Doppler
flow parameter
to
intravenous
norepinephrine
in 3 rats. Dose-rate
axis has been scaled
as indicated
for each animal
to fit a standard
exponential
curve,
which
is shown as a solid curve.
Flow parameter
is shown as percent
of control
value for each animal.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
Kidneys in the control state showed values of flow
parameter (arbitrary units) in the range of 1.12-1.4 for
all animals, usually falling within the range 1.25-1.35.
These values were quite stable, varying by about 58%
over the surface of the kidney, and less than 5% over 30
min at a single point. If the renal pedicle was tied, or
when the animal was sacrificed by an overdose of anesthesia, the flow parameter immediately fell to less than
.05 U, and the Doppler spectrum became flat. This uncertainty in the zero value, which also exists for stationary objects, is due to instrument
error (aggravated by
the steep slope of the correction equation (equation 2)
near zero flow), and is not a fundamental
limit on the
ability to measure low flow states (which can be done by
reducing the processing bandwidth).
Figure 4 shows a family of Doppler spectra from the
same kidney during various steady-state dose rates of
intravenous
norepinephrine.
This drug is a potent,
outer cortical vasoconstrictor (14) and produces the expected narrowing of the Doppler spectra. The shape of
the spectra is characteristic
of those that we obtained
from translucent tissues. Its most conspicuous feature, a
logarithmic singularity
at low frequencies, can be accounted for by a heuristic theory of the multiple-scattering process (18).
We have obtained complete dose-response curves for
the flow parameter as a function of norepinephrine
dose
rate in eight animals. The flow parameter always de-
creased monotonically
with increasing dose, and in all
the animals but one, the curves had a convex, generally
exponential shape. The points of three of these animals
are plotted in Fig. 5, with the abscissa scaled for each
animal so as to give a common rate of logarithmic
decrement. The absolute sensitivities of different animals varied as much as threefold. For comparison, over
the range of doses represented in Figs. 4 and 5, the heart
rate varied from 280 to 420 beats/min (falling again at
the highest doses) and the blood pressure from 125 to 250
mmHg. The dynamic response to a bolus of norepinephrine is shown in Fig. 6. With higher bolus doses it is
possible to reduce flow temporarily
to zero; this could
not be done in the steady state because of cardiac arrythmias. Hydralazine,
in contrast, is a known renal
vasodilator when given in doses that do not overly depress blood pressure (21). Cumulative doses of hydralazine produced the expected increase in flow parameter
until blood pressure fell below 80 mmHg, at which point
flow fell or plateaued.
Restoration
of blood pressure
with dextran produced a variable further
increase.
These effects are shown for six rats in Fig. 7. The more
modest effect of vasodilators than of vasoconstrictors is
probably accounted for by the tendency of blood to redistribute away from the outer cortex during states of
vasodilation (17), although a nonlinearity
of the instrument cannot be ruled out at present.
Angio tensin II wa s infused into the j ugular vein at
rates of O-5 pg/min, which ra .ised mean blood pressure
from 100 to -1sO mmHg. A steady-state dose-response
curve was obtained, with points in random order, each
repeated 2 or 3 times (variation less than 3%). A steadystate infus ion of saralasin (a competitive antagonist of
angiotensin II) was begun in the femoral vein at 25 pg/
kg per min, sufficient to block the hypertensive effect of
M.
H446
1 NOREPlNEPHRINE 2.5mcg
I-
TIME
1 minute
D. STERN
ET
AL.
over 5 secdnds
-I
ARTERIAL
PRESSURE
MM HG
0
MICROAMPS
PHOTOCURRENT 03 I
RMS DETECTOR
ARBITRARY
UNITS
’
EKG
HEART
RATE
01
response
of laser
renal
flow parameter
and
FIG. 6. Transient
other physiologic
variables
to a bolus of 2.5 ,ug of norepinephrine
injected
intravenously
over 5 s at time shown
by vertical
arrow.
Flow parameter
is shown
as percent
of its control
value prior to the
(RMS)
2 are
100- ----I---------___
~--w----*---______-------A
n
HYDRAIAZINE
AND DEXTRAN
(+26.1
CONTROL
injection.
Mean photocurrent
and output
of root-mean-square
detector
from which
flow parameter
is computed
by equat&on
shown
as they were recorded
during
experiment.
RENQL
90
CORTICFlL
FLOW PRRRMETER
(% CONTROL1
HYDRALAZINE
a 40
if
2LU +25
2
2
s
0
iii
8
a
80
+20
CONTROL
T=O
+lO
+5
+5
0
0
-5L
A --------A
WITH
-
WITHOUT
P113
70
+15
A
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
0 I
w
-5L
CONTROL
T=30 min.
(+0.64%)
B
FIG. 7. Percent
change
in renal cortical
flow parameter
produced
by hydralazine
given intravenously,
and by hydralazine
followed
by
volume
expansion
with Dextran
70 to restore
mean arterial
pressure
to 100 mmHg
in 6 rats. A: changes
produced
by interventions
compared
to control
state
immediately
prior
to injection.
B: random
variations
occurring
over a comparable
period
(30 min)
without
intervention,
in the same animals,
prior to 1st dose of hydralazine.
60
50
40
30
10
I
I
0
the highest dose of angiotensin, and the entire doseresponse curve was then repeated in the same animal.
The two curves, shown in Fig. 8, reveal total blockade of
renal vasoconstrictor receptors (8). Of interest for drugrequired
screening purposes, this entire experiment
only 5 h (including anesthesia, surgery, and preparation of the illustration).
Saralasin alone produced no
detectable effect.
With the exception of e inephrine, which produced
effects similar to norepinep p1rine, the drugs (isoproterenol, nitroprusside,
dopamine) did not produce major or
consistent effects on the flow parameter in doses at
which their other hemodynamic effects were marked.
Isoproterenol and nitroprusside acted as vasodepressors
P113
0
FINGIOTENSIN
I
1
INFUSION
I
2
RRTE
1
3
(MICROGRRMS
t
4
PER MINUTE)
I
5
FIG. 8. Curves
of steady-state
dose response
to intravenous
angiotensin
II obtained
with and without
simultaneous
infusion
of inhibitor saralasin
(P-113).
Renal
flow parameter
on vertical
axis is
expressed
as percent
of control
value obtained
without
either drug.
All points were obtained
in the same animal,
in random
order, with
return
to base line between
infusions.
in our preparation; the absence of marked changes in
flow parameter presumably indicates either autoregulation or a balance between vasodilation and redistribution of flow away from the outer cortex (3). Dopamine
was the only vasopressor agent that did not produce
marked decrease in outer cortical flow, even at doses
that raised blood pressure to 220 mmHg. This is consist-
TISSUE BLOOD
FLOW
BY
LASER
H447
DOPPLER
ent with the fact that the kidney is believed to possess
vasodilator dopamine receptors (9) in addition to alphaadrenergic vasoconstrictor receptors which respond to
dopamine.
DISCUSSION
Derivation
of Equation
2
The photocurrent
produced
by the backscattered
light consists
of
two components:
the Doppler
signal,
which
is an audio signal
the
spectrum
of which
is the Doppler
shift spectrum
P(O), and the shot
noise,
which
is a perfectly
white
noise of infinite
bandwidth
and
spectral
power
density
proportional
to the mean
photomultiplier
current
(DC). Thus,
we have
i = i, + i,
(Al
1
where
i, is the Doppler
signal
and i,, is the shot noise. The latter
contains
theoretically
infinite
power
(since
it has infinite
bandwidth).
By passing
the signal
through
a low-pass
filter
that
is
slightly
wider than the Doppler
bandwidth,
we limited
the noise to
manageable
levels without
affecting
the signal.
The spectral
power
density
of the combined
signal
is simply
the sum of that for the
Doppler
signal
and the noise,
since the two are statistically
independent.
Therefore,
the total spectrum
is
S(w)
= P(u)
+ al;
W)
w < B
where a is a constant
that depends
on the gain of the phototube,
I is
the mean photocurrent
(DC), andB
is the bandwidth
of the low-pass
filter.
If the entire
signal
is differentiated
at this point,
the power
spectrum
S(W) will simply
be multiplied
by &. The output
of an rms
detector
applied
to this signal
will be the square
root of the total
integrated
power in the spectrum.
R =dF
=~~Bo’p(w)do
The first term under the second square root
parameter
as defined
by equation
1, while
integrated
directly
to give cz1B3/3, or sI,
defined
constant
which
now depends
on the
processing
bandwidth.
The result
is
which
rearranges
Received
is the square
of the flow
the second
term can be
if s is an appropriately
phototube
gain and the
R=dIG-si
(A4
F = dR7.I
(A5)
to
If the flow parameter
is redefined
by dividing
current,
so as to normalize
it to unit intensity,
The
technical
Part
Service
An
American
(A3)
+,[&do
it by the
equation
authors
thank
Don N. Kennedy
and Lee Morin
assistance.
of the work reported
here was conducted
under
Grant
HL15678-02.
Abstract
of this study
was presented
at the
Societies
for Experimental
Biology
Meeting,
for publication
5 May
mean photo2 results.
for invaluable
Public
Health
Federation
April 1976.
of
1976.
REFERENCES
1. AUKLAND,
K., B. V. BOWER, AND R. W. BERLINER.
Measurement
of local blood flow with hydrogen
gas. Circulation
Res. 14: 164187, 1964.
2. BARGER, A. C., AND J. A. HEARD. Renal vascular
anatomy
and
distribution
of blood flow. In: Handbook
of Physiology.
RenaZ
Physiology.
Washington,
D. C. Am. Physiol.
Sot., 1973, sect. 8,
pp. 270-287.
3. BASTRON, R. D., AND G. J. KALOYANIDES.
Effect of sodium
nitroprusside
on function
in the isolated
and intact
dog kidney.
J.
PharmacoZ.
Exptl.
Therap.
181:244-249,
1972.
4. CHIM~~KEY,
J. E. Skin blood flow by 133Xe disappearance
validated by venous
occlusion
plethysmography.
J. AppZ. PhysioZ.
32: 432-435,
1972.
5. CHIMOSKEY,
J. E., W. FLANAGAN,
AND G. A. HOLLOWAY.
Ultraviolet-induced
cutaneous
hyperemia
measured
by xenon
133 in
man. J. Invest.
Dermatol.
63: 367-368,
1974.
6. CUMMINS,
H. Z., N. KNABLE,
AND Y. YEH. Observation
of diffusion broadening
of Reyleight
scattered
light.
[Letter].
PhysioZ.
Rev. 12: 150-153,
1964.
7. DANCE, J. B. PhotoeLectronic
Devices.
London:
Iliffe, 1969, pp. 7273.
8. FREEMAN,
R. H., J. 0. DAVIS, S. J. VITALE, AND J. A. JOHNSON.
Intrarenal
role of angiotensin
II. Homeostatic
regulation
of
renal blood flow in the dog. CircuZution
Res. 32: 692-698,
1973.
9- GOLDBERG,
L. I. Cardiovascular
and renal
action
of dopamine:
potential
clinical
applications.
PharmacoZ.
Rev. 24: l-29,1972.
10 GRANGSJO,
G., H. R. ULFENDALL,
AND M. WOLGEST.
Determination of regional
renal blood flow by means of small semiconductor detectors
and red cells tagged
with phosphorus
32. Nature
211: 1411-2412,
1966.
11 KETY, S. S. Theory
of blood-tissue
exchange
and its application
to the measurement
of blood flow. In: Peripheral
BZood FLOW
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017
The comparison between the Doppler and xenon
methods in human skin indicates that the computed
flow parameter varies in an approximately linear manner with flow. The scatter in this experiment is compounded of a number of factors, the most important of
which are, probably, variation between individuals and
point-to-point and time-to-time variations in skin blood
flow. The latter fluctuations
are quite pronounced in
small areas of skin, and the impossibility of measuring
flow simultaneously
by xenon washout and Doppler
spectroscopy probably introduced excessive apparent
variance between the two methods. In contrast, the
measurements obtained from the kidney of anesthetized
animals showed great stability when the overall physiologic state was stable. The changes produced by pharmacologic agents are consistent with the known physiology of these drugs. In order to calibrate the method for
the kidney, and determine whether or not it is truly
linear there, it is necessary to measure total renal blood
flow and intrarenal flow distribution by the radioactive
microsphere technique, which we are not equipped to
do
The penetration depth of the laser-Doppler method is
difficult to estimate accurately. The theory of multiple
scattering suggests that the average depth of penetration of light before emerging is of the same order as its
lateral spread, giving an estimate of 1 mm for penetration depth. By covering the renal cortex with a slice of
dead rat renal cortex, we have found that the signal is
almost totally attenuated by a l-mm layer. Presumably,
the penetration depth in light-colored tissues (such as
skin) is somewhat greater.
The ease of performing repeated or continuous measurements by the Doppler method suggests that it may
be useful in a number of clinical applications, as well as
in drug screening. Its use in quantitative
physiologic
experiments appears promising; however, more extensive calibration is required to determine if, in fact, this
use is possible.
APPENDIX
H448
12.
13.
14.
15.
16.
17.
Measurement,
edited
by H. D. Bruner.
Chicago:
Year Book,
1960, p. 223-227.
MCKAY,
J. L., AND Y. ABE. Pressure
dependent
heterogeneity
of
renal
cortical
blood flow in dogs. CircuZation
Res. 27: 571-585,
1970.
PERL, W., AND R. L. HIRSCH. Local blood flow in kidney
tissue by
heat clearance
measurement.
J. Theoret.
BioZ. 10: 251-280,1966.
RECTOR, J. B., J. H. STEIN, W. H. BAY, R. W. OSGOOD, AND T. F.
‘FERRIS. Effect
of hemorrhage
and vasopressor
agents
on distribution
of renal blood flow. Am. J. PhysioZ.
222: 1125-1131,
1972.
SERJRSEN,
P. Cutaneous
blood flow in man studied
by freely
diffusible
radioactive
indicators.
Stand.
J. Clin. Lab. Invest.,
Suppl.
93: 52-59, 1967.
STEIN, J. H., S. BOONJARERN,
C. B. WILSON,
AND T. F. FERRIS.
Alterations
in intrarenal
blood flow distribution.
Methods
of
measurement
and relationship
to sodium
balance.
Circukztion
Res. 32: Suppl.
1: 61-72,
1973.
STEIN, J. H., T. F. FERRIS, J. E. HYPRICH,
T. C. SMITH, AND R.
M.
18.
19.
20.
21.
22.
D.
STERN
ET
AL.
W. OSGOOD. Effect
of renal vasodilation
on the distribution
of
cortical
blood flow in the kidney
of the dog. J. CZin. Invest.
50:
1429-1438,
1971.
STERN, M. D. In vivo evaluation
of microcirculation
by coherent
light scattering.
Nature
254: 56-58, 1975.
TANAKA,
T., AND G. B. BENEDEK.
Measurement
of velocity
of
blood flow
(in vivo) using
a fiber optic catheter
and optical
mixing
spectroscopy.
AppZ. Opt. 14: 189-200,
1975.
TANAKA,
T., C. RIVA, AND I. BEN-SIRA.
Velocity
measurements
in human
retinal
vessels.
Science
186: 830-831,
1974.
WILKINSON,
E. L., H. BACKMAN,
AND H. H. HECHT. Cardiovascular and renal adjustments
to a hypotensive
agent (l-hydrazinophthalazine:
Ciba Ba-5968:
apresoline).
J. CZin. Invest.
31: 872879, 1952.
WOUDA,
A. A. Significance
of photoelectric
plethysmography,
thermography
and venous
occlusion
plethysmography
in investigation of peripheral
circulation.
PfZuegers
Arch.
314: 172,197O.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.6 on June 17, 2017