Respiration Physiology (1979) 37, 45-59 © Elsevier/North-Holland Biomedical Press
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS OF BULLFROG
MASASHI HORIMOTO, TOMIYASU KOYAMA, HIROMICHI MISHINA, TOSHIMITSU ASAKURA and MAKOTO MURAO ~ Research Institute of Applied ElectriciO,, Hokkaido University (! 1st Department of Internal Medicine, School of Medicine), Sapporo 060, Japan
Abstract. Flow velocity in the pulmonary microvessels of the exposed lung of bullfrogs was measured by means of a laser Doppler microscope of an oblique backward mode, together with a signalanalyzing system having a time sharing circuit triggered by the R-wave of the ECG. By these means, measurements of the changes of flow velocity contour in the cardiac cycle were made. Flow velocity was clearly pulsatile in response to cardiac cycles in all microvessels including capillaries. Flow velocities in the arteriole and venule consistently decreased for a short period after the R-wave (84 :t: 33 msec (mean + SD) in the arteriole and 130 + 31 msec in the venule, respectively) and rapidly increased up to a maximum value. The mean flow velocities in arterioles (diameter 50 + 17 I~m) and venules (39 + 9 ~tm) were 2.29 + 0,32 and 2.30 + 0.27 ram/see. The amplitudes of puisatile flow in these vessels were 0.83 + 0.31 and 0.63 + 0.16 mm/sec, respectively. In the capillary the times from the R-wave to the minimum and maximum values were variable. In some cases the velocity gradually increased without first decreasing and the increase sharply accelerated a certain time after the R-wave. The tnean velocity in the pulmonary capillary and the amplitude of the pulsatile flow were 1.78 :t: 0.31 and 0.37 + 0.12 mm/sec, respectively. The ratios of the pulsatile amplitude to the mean velocity in the pulmonary capillary, venule and arteriole averaged 0.21, 0.28 and 0.36, respectively. Blood flow velocity Pulmonary capillaries
Pulmonary circulation Pulsatile blood flow
Measurements of the blood flow in the pulmonary capillary are indispensable for understanding gas exchange and hemodynamics in pulmonary circulation. Hales (1733) reported pulsatile flow in the pulmo: ary capillary of frogs by a microscopic observation. Roeder and Roeder (1932) observed a pulsatile inflow into the frog's pulmonary precapillary but did not mention pulsatile flow in the capillary. Wearn et al. (1934) saw no pulsatile flow in the capillary of a cat's lung but did Acceptedfor publicaqon 20 January 1979
45
46
M. HORIMOTO et aL
observe an intermittent flow, probably due to an irregular repetition of opening and closing of pulmonary arterioles without any direct relation to cardiac cycles. According to Hall (1925), a pulsatile flow was observed only in the larger arterioles but neither intermittent nor pulsatile blood flow were found in the capillaries and smaller arterioles of a rabbit's lung. Vogel (1947) and Schlosser et al. (1965) working with a high speed cinematographic technique combined with a microscope did not mention pulsatile flow in the capillary. Dynamic pressure in the arterioles of the rabbit's lung was measured with a manometer by Rappaport et al. (1959) and found to be pulsatile, but the pressure in the pulmonary capillary was not mentioned. Wagner and Filley (1965)observed that the blood flow in the arterioles, capillaries and venules of a dog's lung was pulsatile but gave no quantitative values. Concerning the strong pulsatile blood flow detected in the human pulmonary capillary by means of N_,O body plethysmography (Lee and DuBois, 1955), the possibility was suggested that the N,O method only measured the rate of blood flow in the arteriolar end of the capillary. Maloney et al. (1968) presented questions on the strongly pulsatile capillary flow and called attention to the diminution of pulsatile amplitude caused by the pulmonary circulatory system. Pinkerson (1967) observed a marked attenuation of pressure and flow pulse after their transmission through the entire pulmonary circulatory system in dogs. Moreover, our own miirroscopic inspection revealed no evident pulsatile flow in the capillary of frog's lung. Thus, the lack of the quantitative measurements of pulmonary blood flow and the interest in the attenuation of the pulsatile flow pattern through the pulmonary microvessels motivated us to measure blood flow velocity in the microvessels of frog lungs by means of a modern technique, laser Doppler microscope combined with a time-sharing circuit and computer system.
Methods MATERIALS
The bullfrog was anesthetized by immersion in the 0.08% solution of MS 222® and placed on a tray which could be moved in steps of 5 ~tm over a wide range. After a careful unilateral incision of chest wall and pleura, without injuring the lung surface, a balloon catheter was inserted into the trachea. The ipsilateral lung was then inflated by sending 3 ml air through the catheter so that the microvessels on the lung surface could be clearly observed through the microscope. The ballistic oscillation due to cardiac ejection was in some animals strong enough to disturb the measurements, because it makes the intersecting area of the two laser beams move off from the vessel under observation. A lead frame of a U-shape was placed over the exposed lung so as to press it slightly down and minimize the oscillation. Although the breathing of the frog was not artificially maintained, the frog could endure the experiments for more than 3 hours. During the measurements, saline
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
47
was gently poured onto the exposed lung surface when necessary in order to avoid drying. The diameter of the arterioles, capillaries and venules chosen for the present study were 25-90, 10-20 and 30-60 l~m, respectively. As the diameter of the laser beams was 40 l~m, the probing area contained two or occasionally three neighbouring capillaries running almost parallel in the case of measurements of capillary flow velocity. When the flow velocity was measured in the arterioles and venules, the probing area contained only one vessel. The central portion of the probing area was always adjusted to be on the center line of the vessel by manipulating the position of the frog lung. LASER DOPPLER MICROSCOPE
Figure 1A shows the principle of the measuring system employed. A 2 mw He-Ne laser beam (2 = 0.63 x l0 -3 mm) is split into dual beams through a beam splitter. The dual beams are crossed on a microvessel visualized on the lung surface through a microscope (magnification is x 200). The two beams intersect at an angle of 18.77 °. The direction of the beam splitter holder is carefully regulated in such a way that the plane formed by the dual beams can contain the microvessel to be studied. vessel
beam
splitter
//~,~ objective
•
w evire .
.
° ' ° ' ° ° -
.
.
.
.
.
.
~ ~.
ph.ot.om . .u.ltip.lie.
r
Capillary
Red bl DO d ce[[
1Dual laser beams
I n ter feren ce fringe
Scat t e red tight
Fig. 1. Schematic illustration of the optical system of laser Doppler microscope arranged in an oblique backward scatter mode (A) and the interference fringes formed in the intersecting area (B).
48
M. HORIMOTO et al.
By these two beams, interference fringes are formed in the intersecting area perpendicularly to the flow in the vessel as schematically shown in fig. lB. When red blood cells are present in the flow and pass across the interference fringes, the dual incident beams are scattered by the flowing cells and independently undergo a Doppler shift related to the flow velocity. The scattered light is collected by a microscope objective and detected by a photomultiplier as a beat signal lasting for the few milliseconds required for the passage of red blood cells through the fringes. An example of the beat signals is shown in fig. 2A. Red blood cells flow through
Fig. 2. Low- (A) and high-speed (B) view ofoscillogram of bursted beat signals (upper traces) and clipped beat signals for measuring the wave period (lower traces) obtained in a slowly flowing capillary.
BLOOD FLOW VELOCITY IN P U L M O N A R Y MICROVESSELS
49
a capillary one after another, causing sequential beat signals lasting only for a few milliseconds. A high speed synchroscopic recording of one beat signal is shown in fig. 2B. The wave period Wp of the beat signal relates to the flow velocity v of a red blood cell by the following equation; ,
1/Wp = fd = v. 2 sin } / z
(1)
where fd, ( and 2 represent the frequency of a beat signal, the intersecting angle and the wavelength of the incident laser beams. The output signals from the photomultiplier are transmitted to a '~nd pass filter. This filter cuts off from the signals both the high frequency noises, the so-called white noises inherent in a photomultiplier, and the low frequency pedestal voltages which are commonly attributed to the intensity variation of laser beams, so as to yield noiseless beat signals. Then the beat signals are transmitted to the beat analyzer. The position of the frog lung was precisely adjusted by the micromanipulator so that the direction of the blood flow in the vessel under observation became perpendicular to the center line of the intersecting angle made by two incident beams. Parts of the incident laser beams are reflected at the lung surface. If the reflected light strikes the photomultiplier, its output signals contain strong noises making the detection of the Doppler shift signals impossible. In order to eliminate such noises optically, the lung surface was covered with a water-containing plastic disc (Koyama et ai.. 1979). This simple procedure enabled the microscope connected with the photomultiplier to be set on the same side as the laser tube against the lung surface. This setting of an optical system may be called an oblique backward scatter mode according to the technical terminology in the field of optics. Figures 3A and 3B show a schematic diagram of the experimental arrangement and an actual example of a measurement, respectively.
DATA PROCESSING SYSTEM
Each of the Doppler beat signals consisting of a small number of waves as seen in fig. 2B was analyzed by a on-line wave-period measuring circuit constructed by Mishina et al. (1974, 1975, 1978) (manufactured by Nihon-kagaku-kogyo Co. Japan). Since the wave period of the beat signals proportional to the reciprocal of the flow velocity changes with the velocity of the red blood cells during one cardiac cycle, the measured wave period was successively and separately pooled in 16 channels of a 4-k words memory by a time-sharing circuit trigge~:ed by the ECG of the frog. The time interval which each channel covered could be selected manually as either 60 msec or 200 msec. When the 60-msec interval was selected for one channel, the 16 channels covered 960 msec after the R-wave of ECG. When one cardiac cycle was longer than 960 msec, the flow velocity during the earlier phase of the
M. HORIMOTO et aL
-,-1 I
HiStograms -multipl.
velocity signals
Fig. 3. Schematic illustration of the experimental arrangement (A) and photograph of the actual measurements (B). Two microscope objective lenses (A for the incident laser dual beams and B for the collection of the scattered light) and the exposed lung covered at the upper portion with a plastic disc can be seen in (B).
cardiac cycle was measured with the 60-msec interval and then the velocity contour over the whole cardiac cycle was determined with the 200-msec interval. Since one red blood cell yields only one beat signal which falls in a certain time period during the cardiac cycle, measurements of the wave period of beat signals must be continued over several cardiac cycles. Velocity signals of red blood cells flowing through the probing area at the same time-section of several cardiac cycles were summarized in one channel defined for the time-section. After measurements
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
51
for several cardiac cycles, each channel constructed a velocity appearance histogram on a synchroscope. In each histogram, the abscissa indicates the flow velocity corresponding to the wave period of beat signals and the ordinate indicates the appearance frequency which is nearly proportional to the number of red blood cells flowing at the given flow velocity through the probing area during the time interval. Although each channel covered successive time periods of 60 msec after the R-wave, the data of flow velocities were recorded only during the first 30 msec of each time interval in the present study. Thus, the data on flow velocity are summarized as 16 sequential but intermittent histograms with the lapse of time after the R-wave of ECG. By a manual selection of computer programs, these 16 histograms were displayed on a X-Y plotter and the mean value and standard deviation of the wave period for each histogram calculated. Then, the flow velocity corresponding to the mean wave period is calculated from Eq. (1) and plotted against the time elapsed after the R-wave. The flow velocity contour is obtained by smoothly connecting the velocities at the 16 time-sections. The mean flow velocity is calculated by timeaveraging the flow velocity contour. Two problems concerning the accuracy of the present method may be mentioned here. They have been discussed in detail and in a more general form in the field of laser Doppler anemometers by Durrani and Greated (1977) and Cochrane and Earnshaw (1978). Since the laser dual beams have a finite diameter, the probing area at their intersection covers a finite volume within the blood vessel where a velocity distribution exists. This is the case when there is one blood cell in the probing volume at a time. The red blood cells flowing successively into the different portions of the probing area cause different Doppler shifts according to the velocity distribution. If this phenomenon occurs at a certain time-section during sequential cardiac cycles, the values obtained from different Doppler shifts are stored in a definite memory channel corresponding to the time-section, whose accumulated data constitute the histogram. The first problem is the fact that the number of red blood cells passing through a portion of the probing area depends on the velocity distribution and that the histogram constructed is the appearance frequency curve weighted by the number of passing red blood cells. Consequently the mean value of the histogram deviates slightly from the true blood flow velocity. Such a weighted mean effect is discussed by Baker and Wayland (1974) with respect to the double slits method and by Owen and Rogers (1975) with respect to the laser Doppler anemometer; they call this the +particle bias effect'. Both sets of authors postulated a principally similar equation, root mean square velocity, for describing the measured value of steady fluid flow. The true flow velocity is approximately 1.3 times the measured blood flow velocity, which is subject to tlae particle bias effect at the center line of a steady flow (estimated from the data saown by Baker and Wayland, 1974). However, no study on this effect is available for ::lightly pulsatile flows, although a similar equation would seem possible. Thz particle bias effect is not taken into consideration in the calculation of the blood flow velocity in the
52
M. HORIMOTO et al.
present study, and the present data for microvessels, except for the capillaries where thi~ effect does not obtain, must be considered with a certain reserve. In any event, it can be said that the flow velocity measured in the present study represents the average velocity of red blood cells flowing through the f'mite volume of the probing area in the vessels.s Another problem may be the spatial velocity distribution. If more than one blood cell exists in the probing area at one time and if they flow at different velocities at different positions, the scattered light receiving frequency shifts of different degrees interfere with one another, yielding beat signals slightly different from those of the true flow velocities. Fortunately tile red blood cells of the frog are large (13 x 21 ~tm) and the hematocrit of frog blood is low (5%). The particle density is so low that more than one blood cell is not expected in the probing area at a time, yet it is sufficiently high for a long-time accumulation and averaging.
Results
An example of 16 histograms obtained in an arteriole having a diameter of 30 pm is shown in fig. 4A, where the wave period of the Doppler shift, i.e. the reciprocal of the flow velocity, is shown on the abscissa versus the appearance frequency on the ordinate. The lowest curve indicates the histogram obtained during 30 msec after the R-wave (lst channel). The second and the third lowest curves show the histograms obtained during 60-90-msec period (2nd channel) and 120-150-msec period (3rd channel) after the R-wave, respectively "The other histogram follow the same relation for time lapses of 30 msec. The gradual shifts of the histograms can be clearly seen. The histogram contour begins to deviate toward the left at the 5th channel and continues to deviate further until the 11th channel (600-630 msec after the R-wave). Thereafter it returns gradually toward the right. The mean value of each histogram was calculated and plotted to construct a velocity contour shown in fig. 4B. The contour shows a minimum value of 1.91 mm/sec at 60-90 msec and a maximum of 2.48 mm/sec at 600-630 msec after the R-wave. The flow velocity of red blood cells flowing through the probing area during a certain time-section is not constant over a number of cardiac cycles. The fluctuations of the velocity cause the spread of the histograms. The standard deviation of the velocity at each time section is given by the width of the peak of each histogram. The difference between the minimum and the maximum velocity respectively estimated from the 2nd and l lth histogram was significant (P < 0.01) by the Student's non paired t-test (n = 1250). Five typical flow velocity contours of the alveolar capillary ob~ined in four frogs are shown in fig. 5. Three curves gradually decrease for a small time period after the R-wave and reach a minimum value, while the other two slowly increase after the R-wave, forming a small concavity, and reach a maximum value. The flow patterns were similar in the neighbouring capillaries but a variation was
PULM. ARTERIOLE (D=30#)
53
900-930 ms¢c
J
m 10.93
0.16
?
NAVE
0-30 ms~c
IB8
#¢
2
........ i 21 .'/0
,
,,,,,,
I
32.464
PERIOD ............. I'
FLOW VELOCITY ram/see
MM/SEC
3.0
2.5
2,0"
1.5"
300 TIRE AFTER R WAVE
600
900 MSEC
Fig. 4. (A) The 16 histograms of the flow velocity signals, obtained in an arteriole (diameter, 30 pm and heart rate, 32/min), sequentially time-shared after the R-wave of electrocardiogram. (B) A velocity contour constructed by interpolating the average values of the velocity signals contained in each histogram shown in fig. 4A.
54
M. HORIMOTO eta',. Im/SEC 2.5
BLOODFLOHVELOCITYIN PULII)HARYCAPILLARIES
>. 2'0 l
u_
~, 1.5 l
1'01
.....
i "' 3O0
, 60~
TIHE AFTER R WAVE
9O0
MSEC
Fig. 5. Five typical flow velocity contours obtained in pulmonary capillaries.
observed according to the different situations of capillaries; the velocity pulsatility in the capillaries near the arteriolar end was observed to be larger than that in capillaries far fro~ this end, although the difference was not statistically significant. Such a local difference in velocity pulsatility was more clearly observed in the frogs whose blood flow velocity in the pulmonary capillaries diminished substantially, presumably because of the prolonged duration of the experiment. In any event, a pulsatile blood flow was evident in all capillaries. When the intersecting area of two laser beams was moved from the capillary to the venule along the direction of blood flow, the mean flow velocity and the amplitude of pulsation increased with the gradual increase of vascular diameter. As an example, fig. 6 shows a set of velocity contours obtained serially in a capillary and the venule connected with the capillary. The mean velocity and the pulsatile amplitude in the capillary were 1.31 and 0.26 mm/sec, respectively. The smaller venule (diameter, 30 pro) situated near the capillary and the larger one (diameter, 45 pin) downstream had a mean flow velocity and pulsatile amplitude of 2.06 and 0.5 mm/sec, and 2.65 and 0.74 mm/sec, respectively. Typical velocity contours of various kinds of microvessels in one frog (heart rate 48/rain) are shown in fig. 7. The value of the blood flow velocity in an artery (150 pm in diameter) in the interstitium between alveoli attained a rriinimum at 60 msec and then rapidly increased up to a maximum 360 msec after the R-wave. Thereafter, it gradually decreased .and reached the initial value, showing some oscillations on the downhill slope.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
55
Nfl[SEC q
~ J ~
,
~
~
VENULE (1)=30).1)
3~o
60o
g6o
MSEC
TIME AFTER R WAVE
Fig. 6. Flow velocity contours obtained in a pulmonary capillary and at two portions in the venule with which the capillary was connected.
,~dsec
ARTERY(~lSOP)
3
)2 §
Y
%
""'L
o
3~o
~
TIME AFTER R NAVE
Fig. 7. A variety of flow velocities in the pulmonary microvessels of ot~e frog whose heart rate remained at 48/min during the repetition of measurements.
56
M. HORIMOTO et aL
The features were similar in all microvessels, although the phases and amplitudes of blood flow were different. The time after the R-wave at which minimum velocity was attained increased from 60 msec in the arteriole to 90 msec in the venule and 150 msec in the capillary, while that for the maximum increased from 360 msec in the arteriole to 420 msec in the capillary and venule. The mean flow velocities and pulsatile amplitudes in the pulmonary microvessels obtained in fifteen frogs are summarized in fig. 8. These values are listed in table 1 together with the ratio of the pulsatile amplitude to the mean flow velocity, and standard deviations. The ratio in the capillary was 20.9 + 7.59/0 which is significantly smaller than those in the arteriole and venule. MEAN BLOOD FLOW VELOCITY AND PULSATILE AMPLITUDE IN PULM ARTERIOLE,CAPILLARYAND VENULE
•
PILLS, AMP
rnmls£c 3 |~ t
rr
5 IZ tD
Fig. 8. A diagram showing the relation between the mean flow velocities and puisatile amplitudes on an average.
TABLE 1 The diameters of the microvessels studied, mean velocities, pulsatile amplitudes and their ratios against mean velocities are listed together with standard deviations. Statistical significances were checked by Student's non-paired t-test N
Capillary Venule Arteriole
37 14 29
Diameter
Mean velocity
Pulsatile amplitude
(~m)
(mm/see)
(mm/sec)
Pulsatile amplitude / Mean velocity °/ bo)
10-20 30-60 25-90
1.78+0.31 2.30 +_0.27 2.29 + 0.32
0.37+0.12 0.63 + 0.16 0.83 + 0.31
20.91 + 7.50 27.70 + 7.16 36.49 + 13.76
P
< 0.01 < 0.05
P-value for the statistical significance was smaller than 0.001 for the difference between capillary and arteriole in the ratio of pulsatile amplitude against the mean velocity.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
57
Five measurements successively carried out at the same portion of an arteriole yielded the mean value and standard deviation of 2.35 + 0.12 mm/sec, showing a sufficient reproducibility of the present method.
Discussion
The finding that the puisatility in the arteriole is greater than that in the capillary and venule agrees with the observation in dogs by Wagner and Filley (1965). The values of the flow velocity in the pulmonary capillary coincides with the ones reported by Hales (1733) and Vogel (1947) in the frog, and slightly exceeds the flow velocity observed by Schlosser et al. (1965) in a rabbit's lung. "Ihe ratios of the pulsatile amplitude to the mean velocity in the artery, arteriole and capillary were 56.0, 36.5 and 20.9%, respectively. This result presumably supports the marked attenuation of pressure and flow pulses after the transmission through the pulmonary circulatory system of dog lungs (Pinkerson, 1967) and the diminution of pulsatile flow amplitude through the pulmonary microvessels of isolate6 lungs of dogs (Maloney et al., 1968). In past studies, different conclusions on the nature of blood flow in the pulmonary capillary were obtained by microscopic inspection as was mentioned in the introduction. Although the flow velocity was so great that direct microscopic observation revealed no pulsatility in the microvessels, the present experimental measurements by means of the laser Doppler microscope consistently showed there to be a pulsatile blood flow in all microvessels. There are so few branchings from the arteriole that the pulsatile flow in the artery can easily be transmitted to capillaries. The arteriolar end is barely tapered in the pulmonary microvessels (Roeder and Roeder, 1932) and blunt termination of arterioles is frequently observed (Irwin et al., 1954). In addition, pulmonary vessels have a relatively low resistance to blood flow. These factors contribute to the transmission of the pulsatile flow to the pulmonary capillaries. However, there is a quantitative discrepancy between the present results and those obtained by the N20 method (Lee and DuBois, 1955; Linderholm et al., 1962), namely that the pulsatile amplitude in the capillary flow was here only 20.9Y/o of the mean flow velocity. Since the local capillary flow velocity seems to have different time delays in phase changes specific to each location, the overall pulmonary flow rate given by the summation of the local flow velocities is smoothed out and the pulsatile amplitude of the overall pulmonary capillary flow decreases probably from that of the local capillary flow. The pulmonary flow rate of the whole lung determined by the N20 method in human subject at rest showed a flow rate contour having a minimum rate of 40-70 ml/sec and a maximum of 203-426 ml/sec depending on the point in the cardiac cycle. This flow rate corresponds to a pulsatile amplitude ratio of 141% if simple algebraic averaging is permitted. A direct comparison of the present results with those by the N:O method may have no
58
M. HORIMOTO et al.
significance because of the difference in species of the experimental animals. However, since the amphibian lung is analogous to the mammalian lung, an interpretation of the frog's pulmonary circulation may be worthwhile for the understanding of the human pulmonary circulation, and the above difference seems to be noteworthy. One possible cause for the difference may be the diffusion of inert gases into the relatively large vessels during body plethysmography. Oxygen and hydrogen are taken up by pulmonary artery branches having a diameter of 2 mm (Jameson, 1964) and up to 3 mm (Sobol et al., 1963). Ether dissolved in alcohol diffuses out of the arterioles when injected into the pulmonary artery (Sackner et al., 1964). These results suggest that some N~O may be taken up by the pulmonary blood flowing through the arteries whose flow rate is strongly pulsatile. The pulsatile amplitude seems to increase with the diameter of arterial vessels even in the frog. For instance, the pulsatile amplitude ratio in the artery shown in fig. 7 was 56%. In a larger artery, the pulsatile amplitude is probably much higher. Another possible cause for the above difference is that the lung surface was exposed in the present study. If measurements could be made in the intact lung protected in the thoracic cage as in the case of the N:O method, the pulmonary capillary flow might eventually skow a stronger pulsatility. This point, however, remains for future stud~. In any event, pulsatiie flow velocity exists in the pulmonary capillary of the exposed lung of the frog.
References Baker, M. and H. Wayland (1974). On-line volume flow rate and velocity profile measurement for blood in microvessels. Microvasc. Res. 7: 131-143. Cochrane, T. and J.C. Earnshaw (1978). Laser Doppler measurements in spatially restricted flows. J. Phys. D (Appl. Phys.) 11: 1509-1517. Durrani, T. S. and C.A. Greated (1977). Laser Systems in Flow Measurements. New York and London, Plenum Press. Hales, S. (1733). Statical essays: Haemastaticks, (3rd ed). London, Wilson and Nicol, 1969, Vol. 2, pp. 66-67. Cited from A. P. Fishman (1963). Dynamics of the pulmonary circulation. In: Handbook of Physiology. Section 2. Circulation, Vol. 2, edited by W.F. Hamilton, American Physiological Society, Washington D.C., pp. 166%1743. Hall, H. L. (1925). A study of the pulmonary circulation by the transillumination method. Am. J. Physiol. 72: 446-457. Irwin, J.W., W. S. BtLrrage, C.E. Aimar and R.W. Chesnut, Jr. (1954). Microscopical observations of the pulmonary arterioles, capillaries and venules of living guinea pigs and rabbits. Anat. Rec. 119: 391-407. Jameson, A.G. (1964). Gaseous diffusion from alveoli into pulmonary arteries. J. Appl. Physiol. 19: 448-456. Koyama, T., M. Horimoto, H. Mishina, T. Asakura, M. Horimoto and M. Murao (1979). Laser Doppler microscope in an oblique-backward mode and pulsatile blood flow velocity in pulmonary arteriole. Experientia (in press). Lee, G. de J. and A.B. DuBois (1955). Pulmonary capillary blood flow in man. d. Clin. Invest. 34: 1380-1390.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
59
Linderholm, H., P. Kimbel, D.H. Lewis and A.B. DuBois (1962). Pulmonary capillary blood flow during cardiac catheterization. J. AppL Physiol. 17: 135-141. Maloney, J.E., D.H. Bergel, J.B. Glazier, J.M.B. Hughes and J.B. West (1968). Transmission of pulsatile blood pressure and flow through the isolated lung. Circ. Res. 23:11-24. Mishina, H., T. Ushizaka, S. Tokui and T. Asakura (1974). Beat signal analyzing systems in a laser Doppler microscope. Bull. Res. Inst. Appl. Electr. 26:51--66. Mishina, H. and T. Asakura (1975). Measurement of velocity fluctuations in laser Doppler microscope by the system employing the time-to pulse height converter. Appl. Phys. 5:351-359. Mishina, H., K. Hironaga, T. Ushizaka, T. Koyama and M. Horimoto (1978). Newly developed signal analyzing system for Doppler beat by using period measurement. Trans. Soc. Instrum. Contr. Eng. (in Japanese) 14: 324-331. Owen, J. M. and R.H. Rogers (1975). Velocity biasing in laser Doppler anemometers. Proceedings of the LDA-Symposium Copenhagen, pp. 89-113. Pinkerson, A.L. (1967). Pulse-wave propagation through the pulmonary vascular bed of dogs. Am. J. Physiol. 213 : 450-454. Rappaport, M.B., E.H. Bloch and J.W. Irwin (1959). A manometer for measuring dynamic pressures in the microvascular system. J. Appl. Physiol. 14: 651--655. Roeder, T. and F. Roeder (1932). Beobachtungen an Lungenkapillaren. 1. Teil. Z. Gesamte Exp. Med. 80: 540-561. Sackner, M.A., K.A. Feisal and D. N. Karsch (1964). Size of gas exchange vessels in the lung. J. Clin. Invest. 43: 1847-1855. Sehlosser, D., E. Heyse and H. Bartels (1965). Flow rate of erythrocytes in the capillaries of the lung. J. AppL Physiol. 20:110-112. Sobol, B.J., G. Bottex, C. Emirgil and H. Gissen (1963). Gaseous diffusion from alveoli to pulmonary vessels of considerable size. Circ. Res. 13: 71-79. Vogel, H. (1947). Die Geschwindigkeit des Blutes in den Lungenkapillaren. Heir. Pllvsiol. Acta 5:105-121. Wagner, W.W., Jr. and G. F. Filley (1965). Microscopic observation of the lung in vivo. Vasc. Dis. 2: 229-241. Wearn, J.T., A.C. Ernstene, A.W. Bromer, J.S. Barr, W.J. German and L.J. Zschiesche (1934). The normal behavior of the pulmonary blood vessels with observations on the intermittence of the flow of blood in the arterioles and capillaries. Am. J. Physiol. 109: 236-256.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS OF BULLFROG
MASASHI HORIMOTO, TOMIYASU KOYAMA, HIROMICHI MISHINA, TOSHIMITSU ASAKURA and MAKOTO MURAO ~ Research Institute of Applied ElectriciO,, Hokkaido University (! 1st Department of Internal Medicine, School of Medicine), Sapporo 060, Japan
Abstract. Flow velocity in the pulmonary microvessels of the exposed lung of bullfrogs was measured by means of a laser Doppler microscope of an oblique backward mode, together with a signalanalyzing system having a time sharing circuit triggered by the R-wave of the ECG. By these means, measurements of the changes of flow velocity contour in the cardiac cycle were made. Flow velocity was clearly pulsatile in response to cardiac cycles in all microvessels including capillaries. Flow velocities in the arteriole and venule consistently decreased for a short period after the R-wave (84 :t: 33 msec (mean + SD) in the arteriole and 130 + 31 msec in the venule, respectively) and rapidly increased up to a maximum value. The mean flow velocities in arterioles (diameter 50 + 17 I~m) and venules (39 + 9 ~tm) were 2.29 + 0,32 and 2.30 + 0.27 ram/see. The amplitudes of puisatile flow in these vessels were 0.83 + 0.31 and 0.63 + 0.16 mm/sec, respectively. In the capillary the times from the R-wave to the minimum and maximum values were variable. In some cases the velocity gradually increased without first decreasing and the increase sharply accelerated a certain time after the R-wave. The tnean velocity in the pulmonary capillary and the amplitude of the pulsatile flow were 1.78 :t: 0.31 and 0.37 + 0.12 mm/sec, respectively. The ratios of the pulsatile amplitude to the mean velocity in the pulmonary capillary, venule and arteriole averaged 0.21, 0.28 and 0.36, respectively. Blood flow velocity Pulmonary capillaries
Pulmonary circulation Pulsatile blood flow
Measurements of the blood flow in the pulmonary capillary are indispensable for understanding gas exchange and hemodynamics in pulmonary circulation. Hales (1733) reported pulsatile flow in the pulmo: ary capillary of frogs by a microscopic observation. Roeder and Roeder (1932) observed a pulsatile inflow into the frog's pulmonary precapillary but did not mention pulsatile flow in the capillary. Wearn et al. (1934) saw no pulsatile flow in the capillary of a cat's lung but did Acceptedfor publicaqon 20 January 1979
45
46
M. HORIMOTO et aL
observe an intermittent flow, probably due to an irregular repetition of opening and closing of pulmonary arterioles without any direct relation to cardiac cycles. According to Hall (1925), a pulsatile flow was observed only in the larger arterioles but neither intermittent nor pulsatile blood flow were found in the capillaries and smaller arterioles of a rabbit's lung. Vogel (1947) and Schlosser et al. (1965) working with a high speed cinematographic technique combined with a microscope did not mention pulsatile flow in the capillary. Dynamic pressure in the arterioles of the rabbit's lung was measured with a manometer by Rappaport et al. (1959) and found to be pulsatile, but the pressure in the pulmonary capillary was not mentioned. Wagner and Filley (1965)observed that the blood flow in the arterioles, capillaries and venules of a dog's lung was pulsatile but gave no quantitative values. Concerning the strong pulsatile blood flow detected in the human pulmonary capillary by means of N_,O body plethysmography (Lee and DuBois, 1955), the possibility was suggested that the N,O method only measured the rate of blood flow in the arteriolar end of the capillary. Maloney et al. (1968) presented questions on the strongly pulsatile capillary flow and called attention to the diminution of pulsatile amplitude caused by the pulmonary circulatory system. Pinkerson (1967) observed a marked attenuation of pressure and flow pulse after their transmission through the entire pulmonary circulatory system in dogs. Moreover, our own miirroscopic inspection revealed no evident pulsatile flow in the capillary of frog's lung. Thus, the lack of the quantitative measurements of pulmonary blood flow and the interest in the attenuation of the pulsatile flow pattern through the pulmonary microvessels motivated us to measure blood flow velocity in the microvessels of frog lungs by means of a modern technique, laser Doppler microscope combined with a time-sharing circuit and computer system.
Methods MATERIALS
The bullfrog was anesthetized by immersion in the 0.08% solution of MS 222® and placed on a tray which could be moved in steps of 5 ~tm over a wide range. After a careful unilateral incision of chest wall and pleura, without injuring the lung surface, a balloon catheter was inserted into the trachea. The ipsilateral lung was then inflated by sending 3 ml air through the catheter so that the microvessels on the lung surface could be clearly observed through the microscope. The ballistic oscillation due to cardiac ejection was in some animals strong enough to disturb the measurements, because it makes the intersecting area of the two laser beams move off from the vessel under observation. A lead frame of a U-shape was placed over the exposed lung so as to press it slightly down and minimize the oscillation. Although the breathing of the frog was not artificially maintained, the frog could endure the experiments for more than 3 hours. During the measurements, saline
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
47
was gently poured onto the exposed lung surface when necessary in order to avoid drying. The diameter of the arterioles, capillaries and venules chosen for the present study were 25-90, 10-20 and 30-60 l~m, respectively. As the diameter of the laser beams was 40 l~m, the probing area contained two or occasionally three neighbouring capillaries running almost parallel in the case of measurements of capillary flow velocity. When the flow velocity was measured in the arterioles and venules, the probing area contained only one vessel. The central portion of the probing area was always adjusted to be on the center line of the vessel by manipulating the position of the frog lung. LASER DOPPLER MICROSCOPE
Figure 1A shows the principle of the measuring system employed. A 2 mw He-Ne laser beam (2 = 0.63 x l0 -3 mm) is split into dual beams through a beam splitter. The dual beams are crossed on a microvessel visualized on the lung surface through a microscope (magnification is x 200). The two beams intersect at an angle of 18.77 °. The direction of the beam splitter holder is carefully regulated in such a way that the plane formed by the dual beams can contain the microvessel to be studied. vessel
beam
splitter
//~,~ objective
•
w evire .
.
° ' ° ' ° ° -
.
.
.
.
.
.
~ ~.
ph.ot.om . .u.ltip.lie.
r
Capillary
Red bl DO d ce[[
1Dual laser beams
I n ter feren ce fringe
Scat t e red tight
Fig. 1. Schematic illustration of the optical system of laser Doppler microscope arranged in an oblique backward scatter mode (A) and the interference fringes formed in the intersecting area (B).
48
M. HORIMOTO et al.
By these two beams, interference fringes are formed in the intersecting area perpendicularly to the flow in the vessel as schematically shown in fig. lB. When red blood cells are present in the flow and pass across the interference fringes, the dual incident beams are scattered by the flowing cells and independently undergo a Doppler shift related to the flow velocity. The scattered light is collected by a microscope objective and detected by a photomultiplier as a beat signal lasting for the few milliseconds required for the passage of red blood cells through the fringes. An example of the beat signals is shown in fig. 2A. Red blood cells flow through
Fig. 2. Low- (A) and high-speed (B) view ofoscillogram of bursted beat signals (upper traces) and clipped beat signals for measuring the wave period (lower traces) obtained in a slowly flowing capillary.
BLOOD FLOW VELOCITY IN P U L M O N A R Y MICROVESSELS
49
a capillary one after another, causing sequential beat signals lasting only for a few milliseconds. A high speed synchroscopic recording of one beat signal is shown in fig. 2B. The wave period Wp of the beat signal relates to the flow velocity v of a red blood cell by the following equation; ,
1/Wp = fd = v. 2 sin } / z
(1)
where fd, ( and 2 represent the frequency of a beat signal, the intersecting angle and the wavelength of the incident laser beams. The output signals from the photomultiplier are transmitted to a '~nd pass filter. This filter cuts off from the signals both the high frequency noises, the so-called white noises inherent in a photomultiplier, and the low frequency pedestal voltages which are commonly attributed to the intensity variation of laser beams, so as to yield noiseless beat signals. Then the beat signals are transmitted to the beat analyzer. The position of the frog lung was precisely adjusted by the micromanipulator so that the direction of the blood flow in the vessel under observation became perpendicular to the center line of the intersecting angle made by two incident beams. Parts of the incident laser beams are reflected at the lung surface. If the reflected light strikes the photomultiplier, its output signals contain strong noises making the detection of the Doppler shift signals impossible. In order to eliminate such noises optically, the lung surface was covered with a water-containing plastic disc (Koyama et ai.. 1979). This simple procedure enabled the microscope connected with the photomultiplier to be set on the same side as the laser tube against the lung surface. This setting of an optical system may be called an oblique backward scatter mode according to the technical terminology in the field of optics. Figures 3A and 3B show a schematic diagram of the experimental arrangement and an actual example of a measurement, respectively.
DATA PROCESSING SYSTEM
Each of the Doppler beat signals consisting of a small number of waves as seen in fig. 2B was analyzed by a on-line wave-period measuring circuit constructed by Mishina et al. (1974, 1975, 1978) (manufactured by Nihon-kagaku-kogyo Co. Japan). Since the wave period of the beat signals proportional to the reciprocal of the flow velocity changes with the velocity of the red blood cells during one cardiac cycle, the measured wave period was successively and separately pooled in 16 channels of a 4-k words memory by a time-sharing circuit trigge~:ed by the ECG of the frog. The time interval which each channel covered could be selected manually as either 60 msec or 200 msec. When the 60-msec interval was selected for one channel, the 16 channels covered 960 msec after the R-wave of ECG. When one cardiac cycle was longer than 960 msec, the flow velocity during the earlier phase of the
M. HORIMOTO et aL
-,-1 I
HiStograms -multipl.
velocity signals
Fig. 3. Schematic illustration of the experimental arrangement (A) and photograph of the actual measurements (B). Two microscope objective lenses (A for the incident laser dual beams and B for the collection of the scattered light) and the exposed lung covered at the upper portion with a plastic disc can be seen in (B).
cardiac cycle was measured with the 60-msec interval and then the velocity contour over the whole cardiac cycle was determined with the 200-msec interval. Since one red blood cell yields only one beat signal which falls in a certain time period during the cardiac cycle, measurements of the wave period of beat signals must be continued over several cardiac cycles. Velocity signals of red blood cells flowing through the probing area at the same time-section of several cardiac cycles were summarized in one channel defined for the time-section. After measurements
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
51
for several cardiac cycles, each channel constructed a velocity appearance histogram on a synchroscope. In each histogram, the abscissa indicates the flow velocity corresponding to the wave period of beat signals and the ordinate indicates the appearance frequency which is nearly proportional to the number of red blood cells flowing at the given flow velocity through the probing area during the time interval. Although each channel covered successive time periods of 60 msec after the R-wave, the data of flow velocities were recorded only during the first 30 msec of each time interval in the present study. Thus, the data on flow velocity are summarized as 16 sequential but intermittent histograms with the lapse of time after the R-wave of ECG. By a manual selection of computer programs, these 16 histograms were displayed on a X-Y plotter and the mean value and standard deviation of the wave period for each histogram calculated. Then, the flow velocity corresponding to the mean wave period is calculated from Eq. (1) and plotted against the time elapsed after the R-wave. The flow velocity contour is obtained by smoothly connecting the velocities at the 16 time-sections. The mean flow velocity is calculated by timeaveraging the flow velocity contour. Two problems concerning the accuracy of the present method may be mentioned here. They have been discussed in detail and in a more general form in the field of laser Doppler anemometers by Durrani and Greated (1977) and Cochrane and Earnshaw (1978). Since the laser dual beams have a finite diameter, the probing area at their intersection covers a finite volume within the blood vessel where a velocity distribution exists. This is the case when there is one blood cell in the probing volume at a time. The red blood cells flowing successively into the different portions of the probing area cause different Doppler shifts according to the velocity distribution. If this phenomenon occurs at a certain time-section during sequential cardiac cycles, the values obtained from different Doppler shifts are stored in a definite memory channel corresponding to the time-section, whose accumulated data constitute the histogram. The first problem is the fact that the number of red blood cells passing through a portion of the probing area depends on the velocity distribution and that the histogram constructed is the appearance frequency curve weighted by the number of passing red blood cells. Consequently the mean value of the histogram deviates slightly from the true blood flow velocity. Such a weighted mean effect is discussed by Baker and Wayland (1974) with respect to the double slits method and by Owen and Rogers (1975) with respect to the laser Doppler anemometer; they call this the +particle bias effect'. Both sets of authors postulated a principally similar equation, root mean square velocity, for describing the measured value of steady fluid flow. The true flow velocity is approximately 1.3 times the measured blood flow velocity, which is subject to tlae particle bias effect at the center line of a steady flow (estimated from the data saown by Baker and Wayland, 1974). However, no study on this effect is available for ::lightly pulsatile flows, although a similar equation would seem possible. Thz particle bias effect is not taken into consideration in the calculation of the blood flow velocity in the
52
M. HORIMOTO et al.
present study, and the present data for microvessels, except for the capillaries where thi~ effect does not obtain, must be considered with a certain reserve. In any event, it can be said that the flow velocity measured in the present study represents the average velocity of red blood cells flowing through the f'mite volume of the probing area in the vessels.s Another problem may be the spatial velocity distribution. If more than one blood cell exists in the probing area at one time and if they flow at different velocities at different positions, the scattered light receiving frequency shifts of different degrees interfere with one another, yielding beat signals slightly different from those of the true flow velocities. Fortunately tile red blood cells of the frog are large (13 x 21 ~tm) and the hematocrit of frog blood is low (5%). The particle density is so low that more than one blood cell is not expected in the probing area at a time, yet it is sufficiently high for a long-time accumulation and averaging.
Results
An example of 16 histograms obtained in an arteriole having a diameter of 30 pm is shown in fig. 4A, where the wave period of the Doppler shift, i.e. the reciprocal of the flow velocity, is shown on the abscissa versus the appearance frequency on the ordinate. The lowest curve indicates the histogram obtained during 30 msec after the R-wave (lst channel). The second and the third lowest curves show the histograms obtained during 60-90-msec period (2nd channel) and 120-150-msec period (3rd channel) after the R-wave, respectively "The other histogram follow the same relation for time lapses of 30 msec. The gradual shifts of the histograms can be clearly seen. The histogram contour begins to deviate toward the left at the 5th channel and continues to deviate further until the 11th channel (600-630 msec after the R-wave). Thereafter it returns gradually toward the right. The mean value of each histogram was calculated and plotted to construct a velocity contour shown in fig. 4B. The contour shows a minimum value of 1.91 mm/sec at 60-90 msec and a maximum of 2.48 mm/sec at 600-630 msec after the R-wave. The flow velocity of red blood cells flowing through the probing area during a certain time-section is not constant over a number of cardiac cycles. The fluctuations of the velocity cause the spread of the histograms. The standard deviation of the velocity at each time section is given by the width of the peak of each histogram. The difference between the minimum and the maximum velocity respectively estimated from the 2nd and l lth histogram was significant (P < 0.01) by the Student's non paired t-test (n = 1250). Five typical flow velocity contours of the alveolar capillary ob~ined in four frogs are shown in fig. 5. Three curves gradually decrease for a small time period after the R-wave and reach a minimum value, while the other two slowly increase after the R-wave, forming a small concavity, and reach a maximum value. The flow patterns were similar in the neighbouring capillaries but a variation was
PULM. ARTERIOLE (D=30#)
53
900-930 ms¢c
J
m 10.93
0.16
?
NAVE
0-30 ms~c
IB8
#¢
2
........ i 21 .'/0
,
,,,,,,
I
32.464
PERIOD ............. I'
FLOW VELOCITY ram/see
MM/SEC
3.0
2.5
2,0"
1.5"
300 TIRE AFTER R WAVE
600
900 MSEC
Fig. 4. (A) The 16 histograms of the flow velocity signals, obtained in an arteriole (diameter, 30 pm and heart rate, 32/min), sequentially time-shared after the R-wave of electrocardiogram. (B) A velocity contour constructed by interpolating the average values of the velocity signals contained in each histogram shown in fig. 4A.
54
M. HORIMOTO eta',. Im/SEC 2.5
BLOODFLOHVELOCITYIN PULII)HARYCAPILLARIES
>. 2'0 l
u_
~, 1.5 l
1'01
.....
i "' 3O0
, 60~
TIHE AFTER R WAVE
9O0
MSEC
Fig. 5. Five typical flow velocity contours obtained in pulmonary capillaries.
observed according to the different situations of capillaries; the velocity pulsatility in the capillaries near the arteriolar end was observed to be larger than that in capillaries far fro~ this end, although the difference was not statistically significant. Such a local difference in velocity pulsatility was more clearly observed in the frogs whose blood flow velocity in the pulmonary capillaries diminished substantially, presumably because of the prolonged duration of the experiment. In any event, a pulsatile blood flow was evident in all capillaries. When the intersecting area of two laser beams was moved from the capillary to the venule along the direction of blood flow, the mean flow velocity and the amplitude of pulsation increased with the gradual increase of vascular diameter. As an example, fig. 6 shows a set of velocity contours obtained serially in a capillary and the venule connected with the capillary. The mean velocity and the pulsatile amplitude in the capillary were 1.31 and 0.26 mm/sec, respectively. The smaller venule (diameter, 30 pro) situated near the capillary and the larger one (diameter, 45 pin) downstream had a mean flow velocity and pulsatile amplitude of 2.06 and 0.5 mm/sec, and 2.65 and 0.74 mm/sec, respectively. Typical velocity contours of various kinds of microvessels in one frog (heart rate 48/rain) are shown in fig. 7. The value of the blood flow velocity in an artery (150 pm in diameter) in the interstitium between alveoli attained a rriinimum at 60 msec and then rapidly increased up to a maximum 360 msec after the R-wave. Thereafter, it gradually decreased .and reached the initial value, showing some oscillations on the downhill slope.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
55
Nfl[SEC q
~ J ~
,
~
~
VENULE (1)=30).1)
3~o
60o
g6o
MSEC
TIME AFTER R WAVE
Fig. 6. Flow velocity contours obtained in a pulmonary capillary and at two portions in the venule with which the capillary was connected.
,~dsec
ARTERY(~lSOP)
3
)2 §
Y
%
""'L
o
3~o
~
TIME AFTER R NAVE
Fig. 7. A variety of flow velocities in the pulmonary microvessels of ot~e frog whose heart rate remained at 48/min during the repetition of measurements.
56
M. HORIMOTO et aL
The features were similar in all microvessels, although the phases and amplitudes of blood flow were different. The time after the R-wave at which minimum velocity was attained increased from 60 msec in the arteriole to 90 msec in the venule and 150 msec in the capillary, while that for the maximum increased from 360 msec in the arteriole to 420 msec in the capillary and venule. The mean flow velocities and pulsatile amplitudes in the pulmonary microvessels obtained in fifteen frogs are summarized in fig. 8. These values are listed in table 1 together with the ratio of the pulsatile amplitude to the mean flow velocity, and standard deviations. The ratio in the capillary was 20.9 + 7.59/0 which is significantly smaller than those in the arteriole and venule. MEAN BLOOD FLOW VELOCITY AND PULSATILE AMPLITUDE IN PULM ARTERIOLE,CAPILLARYAND VENULE
•
PILLS, AMP
rnmls£c 3 |~ t
rr
5 IZ tD
Fig. 8. A diagram showing the relation between the mean flow velocities and puisatile amplitudes on an average.
TABLE 1 The diameters of the microvessels studied, mean velocities, pulsatile amplitudes and their ratios against mean velocities are listed together with standard deviations. Statistical significances were checked by Student's non-paired t-test N
Capillary Venule Arteriole
37 14 29
Diameter
Mean velocity
Pulsatile amplitude
(~m)
(mm/see)
(mm/sec)
Pulsatile amplitude / Mean velocity °/ bo)
10-20 30-60 25-90
1.78+0.31 2.30 +_0.27 2.29 + 0.32
0.37+0.12 0.63 + 0.16 0.83 + 0.31
20.91 + 7.50 27.70 + 7.16 36.49 + 13.76
P
< 0.01 < 0.05
P-value for the statistical significance was smaller than 0.001 for the difference between capillary and arteriole in the ratio of pulsatile amplitude against the mean velocity.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
57
Five measurements successively carried out at the same portion of an arteriole yielded the mean value and standard deviation of 2.35 + 0.12 mm/sec, showing a sufficient reproducibility of the present method.
Discussion
The finding that the puisatility in the arteriole is greater than that in the capillary and venule agrees with the observation in dogs by Wagner and Filley (1965). The values of the flow velocity in the pulmonary capillary coincides with the ones reported by Hales (1733) and Vogel (1947) in the frog, and slightly exceeds the flow velocity observed by Schlosser et al. (1965) in a rabbit's lung. "Ihe ratios of the pulsatile amplitude to the mean velocity in the artery, arteriole and capillary were 56.0, 36.5 and 20.9%, respectively. This result presumably supports the marked attenuation of pressure and flow pulses after the transmission through the pulmonary circulatory system of dog lungs (Pinkerson, 1967) and the diminution of pulsatile flow amplitude through the pulmonary microvessels of isolate6 lungs of dogs (Maloney et al., 1968). In past studies, different conclusions on the nature of blood flow in the pulmonary capillary were obtained by microscopic inspection as was mentioned in the introduction. Although the flow velocity was so great that direct microscopic observation revealed no pulsatility in the microvessels, the present experimental measurements by means of the laser Doppler microscope consistently showed there to be a pulsatile blood flow in all microvessels. There are so few branchings from the arteriole that the pulsatile flow in the artery can easily be transmitted to capillaries. The arteriolar end is barely tapered in the pulmonary microvessels (Roeder and Roeder, 1932) and blunt termination of arterioles is frequently observed (Irwin et al., 1954). In addition, pulmonary vessels have a relatively low resistance to blood flow. These factors contribute to the transmission of the pulsatile flow to the pulmonary capillaries. However, there is a quantitative discrepancy between the present results and those obtained by the N20 method (Lee and DuBois, 1955; Linderholm et al., 1962), namely that the pulsatile amplitude in the capillary flow was here only 20.9Y/o of the mean flow velocity. Since the local capillary flow velocity seems to have different time delays in phase changes specific to each location, the overall pulmonary flow rate given by the summation of the local flow velocities is smoothed out and the pulsatile amplitude of the overall pulmonary capillary flow decreases probably from that of the local capillary flow. The pulmonary flow rate of the whole lung determined by the N20 method in human subject at rest showed a flow rate contour having a minimum rate of 40-70 ml/sec and a maximum of 203-426 ml/sec depending on the point in the cardiac cycle. This flow rate corresponds to a pulsatile amplitude ratio of 141% if simple algebraic averaging is permitted. A direct comparison of the present results with those by the N:O method may have no
58
M. HORIMOTO et al.
significance because of the difference in species of the experimental animals. However, since the amphibian lung is analogous to the mammalian lung, an interpretation of the frog's pulmonary circulation may be worthwhile for the understanding of the human pulmonary circulation, and the above difference seems to be noteworthy. One possible cause for the difference may be the diffusion of inert gases into the relatively large vessels during body plethysmography. Oxygen and hydrogen are taken up by pulmonary artery branches having a diameter of 2 mm (Jameson, 1964) and up to 3 mm (Sobol et al., 1963). Ether dissolved in alcohol diffuses out of the arterioles when injected into the pulmonary artery (Sackner et al., 1964). These results suggest that some N~O may be taken up by the pulmonary blood flowing through the arteries whose flow rate is strongly pulsatile. The pulsatile amplitude seems to increase with the diameter of arterial vessels even in the frog. For instance, the pulsatile amplitude ratio in the artery shown in fig. 7 was 56%. In a larger artery, the pulsatile amplitude is probably much higher. Another possible cause for the above difference is that the lung surface was exposed in the present study. If measurements could be made in the intact lung protected in the thoracic cage as in the case of the N:O method, the pulmonary capillary flow might eventually skow a stronger pulsatility. This point, however, remains for future stud~. In any event, pulsatiie flow velocity exists in the pulmonary capillary of the exposed lung of the frog.
References Baker, M. and H. Wayland (1974). On-line volume flow rate and velocity profile measurement for blood in microvessels. Microvasc. Res. 7: 131-143. Cochrane, T. and J.C. Earnshaw (1978). Laser Doppler measurements in spatially restricted flows. J. Phys. D (Appl. Phys.) 11: 1509-1517. Durrani, T. S. and C.A. Greated (1977). Laser Systems in Flow Measurements. New York and London, Plenum Press. Hales, S. (1733). Statical essays: Haemastaticks, (3rd ed). London, Wilson and Nicol, 1969, Vol. 2, pp. 66-67. Cited from A. P. Fishman (1963). Dynamics of the pulmonary circulation. In: Handbook of Physiology. Section 2. Circulation, Vol. 2, edited by W.F. Hamilton, American Physiological Society, Washington D.C., pp. 166%1743. Hall, H. L. (1925). A study of the pulmonary circulation by the transillumination method. Am. J. Physiol. 72: 446-457. Irwin, J.W., W. S. BtLrrage, C.E. Aimar and R.W. Chesnut, Jr. (1954). Microscopical observations of the pulmonary arterioles, capillaries and venules of living guinea pigs and rabbits. Anat. Rec. 119: 391-407. Jameson, A.G. (1964). Gaseous diffusion from alveoli into pulmonary arteries. J. Appl. Physiol. 19: 448-456. Koyama, T., M. Horimoto, H. Mishina, T. Asakura, M. Horimoto and M. Murao (1979). Laser Doppler microscope in an oblique-backward mode and pulsatile blood flow velocity in pulmonary arteriole. Experientia (in press). Lee, G. de J. and A.B. DuBois (1955). Pulmonary capillary blood flow in man. d. Clin. Invest. 34: 1380-1390.
BLOOD FLOW VELOCITY IN PULMONARY MICROVESSELS
59
Linderholm, H., P. Kimbel, D.H. Lewis and A.B. DuBois (1962). Pulmonary capillary blood flow during cardiac catheterization. J. AppL Physiol. 17: 135-141. Maloney, J.E., D.H. Bergel, J.B. Glazier, J.M.B. Hughes and J.B. West (1968). Transmission of pulsatile blood pressure and flow through the isolated lung. Circ. Res. 23:11-24. Mishina, H., T. Ushizaka, S. Tokui and T. Asakura (1974). Beat signal analyzing systems in a laser Doppler microscope. Bull. Res. Inst. Appl. Electr. 26:51--66. Mishina, H. and T. Asakura (1975). Measurement of velocity fluctuations in laser Doppler microscope by the system employing the time-to pulse height converter. Appl. Phys. 5:351-359. Mishina, H., K. Hironaga, T. Ushizaka, T. Koyama and M. Horimoto (1978). Newly developed signal analyzing system for Doppler beat by using period measurement. Trans. Soc. Instrum. Contr. Eng. (in Japanese) 14: 324-331. Owen, J. M. and R.H. Rogers (1975). Velocity biasing in laser Doppler anemometers. Proceedings of the LDA-Symposium Copenhagen, pp. 89-113. Pinkerson, A.L. (1967). Pulse-wave propagation through the pulmonary vascular bed of dogs. Am. J. Physiol. 213 : 450-454. Rappaport, M.B., E.H. Bloch and J.W. Irwin (1959). A manometer for measuring dynamic pressures in the microvascular system. J. Appl. Physiol. 14: 651--655. Roeder, T. and F. Roeder (1932). Beobachtungen an Lungenkapillaren. 1. Teil. Z. Gesamte Exp. Med. 80: 540-561. Sackner, M.A., K.A. Feisal and D. N. Karsch (1964). Size of gas exchange vessels in the lung. J. Clin. Invest. 43: 1847-1855. Sehlosser, D., E. Heyse and H. Bartels (1965). Flow rate of erythrocytes in the capillaries of the lung. J. AppL Physiol. 20:110-112. Sobol, B.J., G. Bottex, C. Emirgil and H. Gissen (1963). Gaseous diffusion from alveoli to pulmonary vessels of considerable size. Circ. Res. 13: 71-79. Vogel, H. (1947). Die Geschwindigkeit des Blutes in den Lungenkapillaren. Heir. Pllvsiol. Acta 5:105-121. Wagner, W.W., Jr. and G. F. Filley (1965). Microscopic observation of the lung in vivo. Vasc. Dis. 2: 229-241. Wearn, J.T., A.C. Ernstene, A.W. Bromer, J.S. Barr, W.J. German and L.J. Zschiesche (1934). The normal behavior of the pulmonary blood vessels with observations on the intermittence of the flow of blood in the arterioles and capillaries. Am. J. Physiol. 109: 236-256.
