Download Exercise-induced intrapulmonary arteriovenous shunting in healthy

J Appl Physiol 97: 797– 805, 2004.
First published April 23, 2004; 10.1152/japplphysiol.00137.2004.
Exercise-induced intrapulmonary arteriovenous shunting in healthy humans
Marlowe W. Eldridge,1,2 Jerome A. Dempsey,1 Hans C. Haverkamp,1
Andrew T. Lovering,1 and John S. Hokanson2
1
John Rankin Laboratory of Pulmonary Medicine, Population Health Sciences and
Department of Pediatrics, University of Wisconsin, Madison, Wisconsin 53705
2
Submitted 5 February 2004; accepted in final form 20 April 2004
Eldridge, Marlowe W., Jerome A. Dempsey, Hans C.
Haverkamp, Andrew T. Lovering, and John S. Hokanson. Exercise-induced intrapulmonary arteriovenous shunting in healthy humans. J Appl Physiol 97: 797– 805, 2004. First published April 23,
2004; 10.1152/japplphysiol.00137.2004.—We hypothesized that increasing exercise intensity recruits dormant arteriovenous intrapulmonary shunts, which may contribute to the widened alveolar-arterial
oxygen difference seen with exercise. Twenty-three healthy volunteers (13 men and 10 women, aged 23– 48 yr) with normal lung
function and a wide range of fitness (mean maximal oxygen uptake ⫽
126% predicted; range ⫽ 78 –200% predicted) were studied by
agitated saline contrast echocardiography (4-chamber apical view).
All 23 subjects had normal resting contrast echocardiograms without
evidence of intracardiac or intrapulmonary shunting. However, with
cycle ergometer exercise, 21 of 23 (91%) of the subjects showed a
delayed (⬎3 cardiac cycles) appearance of contrast bubbles in the left
heart. This pattern is consistent with passage of contrast bubbles
through the pulmonary circulation. Because the contrast bubbles are
known to be significantly larger than pulmonary capillaries, we
propose that they are traveling through direct arteriovenous intrapulmonary shunts. In all cases, the intrapulmonary shunting developed at
submaximal oxygen uptakes [%maximal oxygen uptake ⫽ 59 ⫾ 20
(SD)] and once evident persisted at all subsequent work rates. Within
3 min of exercise termination, the contrast echocardiograms with
bubble injection showed no evidence of intrapulmonary shunting.
These dynamic shunts will contribute significantly to the widened
alveolar-arterial oxygen difference seen with exercise. They may also
act as a protective parallel vascular network limiting the rise in
regional pulmonary vascular pressure while preserving cardiac output
during exercise.
pulmonary circulation; pulmonary gas exchange; exercise-induced
hypoxemia
WITH EXERCISE, GAS-EXCHANGE
efficiency, as quantified by the
difference between the alveolar and the arterial blood oxygen
tensions (A-aDO2), progressively worsens in an intensitydependent manner (1, 10, 53, 60). At maximal exercise, the
A-aDO2 reaches values of 20 –30 Torr in normal, healthy,
untrained subjects and can be as high as 35–50 Torr in some
elite athletes (9, 20). In contrast, fixed-workload high-intensity
endurance exercise does not result in a time-dependent increasing of gas-exchange efficiency (59), indicating that the magnitude of the A-aDO2 is determined primarily by metabolic
rate, rather than exercise duration.
The A-aDO2 is a complex physiological variable and as such
is determined by a variety of mechanisms during rest and
exercise. Imperfect matching of the distributions of alveolar
Address for reprint requests and other correspondence: M. W. Eldridge,
John Rankin Laboratory of Pulmonary Medicine, Univ. of Wisconsin, Medical
School, H4/422 Clinical Sciences Center, 600 Highland Ave., Madison, WI
53792-4108 (E-mail: [email protected]).
http://www. jap.org
ventilation (V̇A) and pulmonary blood flow (Q̇), otherwise
known as the V̇A/Q̇ ratio, contributes to the A-aDO2 during
both rest and exercise (16, 55). With exercise, overall V̇A/Q̇
nonuniformity increases slightly as estimated by the multiple
inert-gas elimination technique (MIGET) (16, 55). It should be
stressed that the V̇A/Q̇ nonuniformity during exercise fails to
explain all of the widening of the A-aDO2. Indeed, not all
subjects increase V̇A/Q̇ nonuniformity during exercise (45),
whereas a widened A-aDO2 is universally observed. It is argued
that any difference between the actual A-aDO2 and that predicted from the measured amount of V̇A/Q̇ nonuniformity
given by MIGET can be attributed to diffusion limitation. With
the use of this technique, significant amounts of diffusion
limitation, up to two-thirds of the total A-aDO2, have been
predicted at metabolic rates as low as 2.0 l/min (19, 55).
However, it is unlikely that a diffusion limitation would occur
at these moderate metabolic rates when pulmonary capillary
blood volume and pulmonary blood flow are submaximal and
mean transit time is still ⬎450 ms (56).
Another potential contributor to the widened A-aDO2 with
exercise is venous admixture from intrapulmonary arteriovenous shunts. Wagner and colleagues (55) used 100% oxygen
breathing and MIGET to test for shunt during exercise. By
using the 100% oxygen test, the calculated shunt fractions were
found to be ⬃2%. The authors suggested that this represented
postpulmonary shunt because the MIGET did not detect intrapulmonary shunting. However, the validity of these standard
tests for detecting intrapulmonary shunts may be questionable
because significant prepulmonary capillary gas exchange may
occur and its magnitude is critically dependent on the gas
concentration gradient (6, 11, 24, 49). Therefore, intrapulmonary arteriovenous connections that act as shunts at low or
normal oxygen tensions may participate in gas exchange at
high inspired oxygen fraction, leading to an underestimation of
shunted blood. Furthermore, when a high inspired oxygen
fraction is used, measurements of blood oxygen tension are not
sufficiently accurate to distinguish shunts ⬍10% of the cardiac
output. Similarly, MIGET may underestimate intrapulmonary
shunting because the detection of shunting is dependent on
retention of sulfur hexafluoride, an inert gas with a very low
solubility in blood. The very low solubility and the large
concentration gradient may allow for precapillary elimination
of sulfur hexafluoride, like that which occurs with the 100%
oxygen technique, thus underestimating intrapulmonary shunting. The MIGET is also prone to spurious and even impossible
results as evidenced by findings of a wider predicted (from
V̇A/Q̇ estimates) than actually measured A-aDO2 (19).
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
8750-7587/04 $5.00 Copyright © 2004 the American Physiological Society
797
798
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
Morphological studies demonstrate the existence of direct
vascular conduits between pulmonary arteries and veins in both
dog (4, 5, 34, 36, 37) and human lungs (51, 52, 62). Furthermore, increases in the fraction of intrapulmonary shunting in
these animal models appear to occur with increasing pulmonary arterial pressure and flow (5). Thus it is reasonable to
postulate that even moderate-intensity exercise may augment
this shunt fraction. Because the shunted blood is deoxygenated
and becomes more so as oxygen extraction by the contracting
muscles increases, only a small fraction of the cardiac output as
shunt is necessary to widen the A-aDO2.
We hypothesized that intrapulmonary arteriovenous shunts
are recruited in healthy humans during exercise. To test our
hypothesis, we performed agitated saline contrast bubble echocardiography at rest and during progressive exercise in healthy
individuals with a wide continuum of maximal oxygen uptakes.
METHODS
This study received approval from the University of WisconsinMadison Human Subjects Committee, and each subject gave his or her
written, informed consent before participation. All studies were performed according to the Declaration of Helsinki.
Subjects. Twenty-six healthy, nonsmoking volunteers (15 men and
11 women), aged 18 – 49 yr were recruited and, after written, informed
consent was given, agreed to further study. A screening cardiopulmonary history and physical examination were performed, and all subjects appeared to be free of cardiopulmonary disease. The resting
contrast echocardiograms performed just before the exercise protocol
revealed a previously unrecognized patent foramen ovale in two
subjects, and a third subject had a contrast bubble echocardiogram
consistent with a pulmonary arteriovenous malformation. These three
subjects were excluded from the exercise studies.
Pulmonary function and lung diffusion capacity for carbon monoxide testing. Baseline pulmonary function (Pulmonizer model PFT
3000, Med Science, St. Louis, MO) including forced vital capacity,
forced expiratory volume in 1 s, forced mid-expiratory flows, and
peak expiratory flow were determined as described previously (50).
Lung diffusion capacity for carbon monoxide (DLCO) was determined
by a single-breath breath-holding method (35).
Exercise protocol. Twenty-three subjects (13 men and 10 women)
completed a progressive incremental exercise test to exhaustion on a
magnetically braked cycle ergometer. After a 2 min warm-up with the
initial work rate set at 65 W, the work rate was increased by 30 W
every 2 min until volitional fatigue. During the exercise protocol,
subjects breathed through a low-resistance two-way nonrebreathing
valve (model 2400, Hans Rudolph) with expired gases sampled at the
mouth and after a mixing chamber (8.64 liters) via a mass spectrometer (Perkin-Elmer model 110). Inspiratory and expiratory flow rates
were measured separately by two pneumotachographs. All signals
were displayed on a chart recorder, sent through an analog-to-digital
board, and sampled on a computer at 75 Hz. Heart rate was measured
with a three-lead ECG and recorded continuously.
Contrast echocardiography during progressive exercise. An apical
four-chamber contrast echocardiogram with harmonic imaging was
performed (Cypress Ultrasound Systems, Acuson/Semins, Mountain
View, CA) at rest, during the last minute of each 2-min exercise stage,
and 3 min after the final exercise load. In each subject, a 20-gauge
intravenous catheter with a saline solution lock was placed in the
median basilic vein. A three-way stopcock was attached, and two
10-ml syringes were attached to the other two ports. One syringe
contained 1 ml of air, and the other contained 5 ml of sterile saline and
1 ml of the subject’s blood. The contrast bubbles were created by
flushing the saline solution from one syringe to another. A forceful
hand injection of the agitated saline solution was performed while
J Appl Physiol • VOL
images were obtained simultaneously in the apical four-chamber
view. Manual agitation of the saline creates contrast bubbles, which
are highly echogenic and are readily visualized in the right heart after
venous injection (17, 33).
Without right-to-left shunting, peripherally infused contrast bubbles are visualized as a cloud of echoes in the right heart and then
gradually disappear as the bubbles become trapped and eliminated in
the pulmonary microcirculation (4, 31, 32). Timing of the appearance
of contrast bubbles in the left heart after right heart filling yields
important anatomic information. If shunting exists at the atrial or
ventricular level, contrast bubbles will rapidly fill the left heart (44).
If the contrast bubbles pass through the lungs, they will appear in the
left heart after a delay of at least three cardiac cycles. The delayed
appearance of bubbles in the left heart indicates transpulmonary
passage of contrast bubbles either through abnormally dilated capillaries (43) or through intrapulmonary arteriovenous shunts (2, 18, 21,
27, 48). Harmonic imaging enhances detection of the nonlinear
backscatter from the contrast bubbles, thus improving signal-to-noise
and greatly improving visualization of bubble contrast in the cardiac
chambers. All of the contrast echocardiograms were performed with
the subject seated on the cycle ergometer with the mouthpiece and
nose clip in place.
Data analysis. Descriptive and physiological data are presented as
means ⫾ SD (Sigma Stat 2.03, Aspire Software International, Leesburg, VA). All of the echocardiograms were digitally recorded and
analyzed offline (Camtronics Medical System, Hartland, WI). This
system allows for analysis of the echocardiograms at 30 frames/s.
RESULTS
Lung function and maximal oxygen uptake data. Anthropometric, pulmonary function, DLCO, and exercise data for the 23
subjects that completed the exercise protocol are shown in
Table 1. All of these subjects had resting pulmonary function
and DLCO that were within normal values. There was a wide
continuum in fitness level with predicted maximal oxygen
uptake (V̇O2 max) ranging from 78 to 200%.
Contrast echocardiography during progressive exercise.
Shown in Figs. 1, 2, and 3 are contrast echocardiograms
obtained from a 48-yr-old male subject (%predicted V̇O2 max ⫽
136%) during his progressive exercise study. The images
presented are typical of those obtained from all 23 subjects
studied. In all cases the quality of the images was good. With
digital image recording and a frame rate of 30 images/s, we
were often able to see bubbles emerging from the pulmonary
veins and then track these bubbles as they passed through the
left heart chambers.
Table 1. Subject characteristics and resting lung function
Men (n ⫽ 13)
Women (n ⫽ 10)
Age, yr
32.2⫾8.0
27.2⫾7.6
Height, cm
177.6⫾8.0
165.1⫾6.6
Weight, kg
73.2⫾11.9
58.8⫾3.0
FVC, liters
5.3⫾0.6 (100.9⫾11.7)
4.1⫾0.4 (112.1⫾10.2)
4.3⫾0.5 (98.5⫾11.7)
3.6⫾0.4 (115.0⫾9.9)
FEV1, liters
FEV1/FVC
0.81⫾0.08
0.88⫾0.06
FEF25–75, l/s
4.4⫾1.6 (93.7⫾31.1)
4.2⫾0.8 (115.0⫾21.7)
⫺1
⫺1
DLCO, ml 䡠 min 䡠 Torr
38.8⫾4.3 (89.9⫾10.4)
26.7⫾4.7 (86.0⫾12.8)
V̇O2max, ml 䡠 kg⫺1 䡠 min⫺1 51.9⫾10.1 (120.0⫾24.6) 41.7⫾8.2 (132.5⫾33.1)
Values are means ⫾ SD. Values in parentheses are percent predicted (3, 25,
26). FVC, forced vital capacity; FEV1, forced expiratory volume in 1 s; FEF1
forced expiratory flow of midexpiratory volume; DLCO, diffusion capacity for
carbon monoxide; V̇O2max, relative maximal oxygen uptake.
97 • SEPTEMBER 2004 •
www.jap.org
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
Figure 1 shows the apical four-chamber contrast echocardiograms obtained at rest. The subject is seated on the cycle
ergometer with the mouthpiece and nose clip in place. Immediately after injection of contrast bubbles, both the right atrium
(RA) and right ventricle (RV) are densely opacified with
contrast bubbles. The left heart is free of contrast consistent
with the absence of intracardiac shunts. After several cardiac
cycles the left heart remains free of contrast, indicating that the
pulmonary circulation has effectively trapped and eliminated
the contrast bubbles.
Figure 2 shows contrast echocardiograms obtained at exercise
intensities of 100, 230, and 260 W. At 100 W (%V̇O2 max ⫽ 40%),
immediately after contrast bubble injection, the RA and RV are
densely opacified, whereas the left heart is free of contrast indicating no intracardiac shunting. Five cardiac cycles later the left
heart remains clear, with no evidence of transpulmonary passage
of contrast bubbles. In this subject, transpulmonary passage of
contrast bubbles occurred at 230 W (%V̇O2 max ⫽ 84%). Again,
immediately after contrast injection the RA and RV are densely
opacified, whereas the left-heart remains clear. Five seconds (or
10 cardiac cycles) later, contrast bubbles are clearly seen in the left
heart. The delayed arrival of contrast in the left ventricle indicates
passage of bubbles through the pulmonary circulation (also see
Fig. 4). A similar pattern with a delayed appearance of contrast is
seen at 260 W (%V̇O2 max ⫽ 94%). Note that the density of
contrast bubbles appearing in the left ventricle is qualitatively
greater at 260 W than at 230 W.
Figure 3 shows the contrast echocardiograms obtained 3 min
after termination of the exercise bout. The subject remained on
the cycle ergometer in the riding position. After contrast
injection, there was no evidence of intracardiac shunting or
transpulmonary passage of contrast bubbles.
Detection of intracardiac vs. intrapulmonary right-to-left
shunts with contrast echocardiography. Time of the appearance of contrast bubbles in the left heart after right heart filling
is used to distinguish intracardiac from intrapulmonary shunts.
If intracardiac right-to-left shunting exists, contrast bubbles
will rapidly fill the left heart (44). If the contrast bubbles pass
through the lungs, they will appear in the left heart after a delay
of at least three cardiac cycles. Figure 4 shows six sequential
apical four-chamber contrast echocardiograms obtained from
a 28-yr-old female subject during submaximal exercise
(%V̇O2 max ⫽ 40). The first image shows the peripheral contrast
injection with contrast bubbles filling the right heart. Each
subsequent image is separated in time by 1 s. There is no
evidence of contrast bubbles in the left heart until the fifth
image, which is eight cardiac cycles after the contrast injection.
The delayed appearance of bubbles in the left heart indicates
transpulmonary passage of contrast bubbles either through
abnormally dilated capillaries (43) or through intrapulmonary
arteriovenous shunts (2, 18, 21, 27, 48). We found that with
submaximal exercise 91% (21 of 23) of the subjects showed a
delayed (⬎3 cardiac cycles) appearance of contrast bubbles in
the left heart. However, no shunting was seen after termination
of exercise.
Intersubject variability. Table 2 shows the cardiopulmonary
measures at the onset of intrapulmonary arteriovenous shunting. In all cases, the intrapulmonary shunting developed at
submaximal exercise levels (13– 84%V̇O2 max) and, once
present, persisted with each subsequent work rate. Shown in
Fig. 5 is the frequency distribution, among the 23 subjects, of
the exercise intensity at the onset of shunting. The %V̇O2 max at
the onset of shunting was broadly distributed, with the distribution skewed toward the higher exercise intensities. Two
male subjects did not develop intrapulmonary arteriovenous
shunting at any exercise intensity. Otherwise, these subjects
were not different from the subjects that demonstrated exercise-induced shunting.
DISCUSSION
The goal of this study was to determine whether intrapulmonary arteriovenous shunts develop in healthy humans during
exercise. We used agitated saline contrast echocardiography
during exercise and found that contrast bubbles traversed the
pulmonary circulation and appeared in the left heart in 91% (21
of 23) of subjects tested. The transpulmonary passage of
contrast bubbles was not evident at rest but developed at
submaximal oxygen consumptions and persisted with each
subsequent work rate. However, the shunting was not seen 3
min after the termination of maximal exercise. We believe our
Fig. 1. Agitated saline contrast echocardiograms from a 48-yr-old male subject at rest. After several cardiac cycles (panel at right),
the left heart remains free of contrast, indicating that the pulmonary circulation has trapped and eliminated the contrast bubbles.
All images are apical 4-chamber views obtained immediately before the exercise. RA, right atrium; RV, right ventricle; LA, left
atrium; LV, left ventricle.
J Appl Physiol • VOL
799
97 • SEPTEMBER 2004 •
www.jap.org
800
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
Fig. 2. Contrast echocardiograms during exercise. At 100 W, there is no evidence of intracardiac or intrapulmonary shunting, as
the left heart remains free of contrast bubbles. In this subject, the first evidence of intrapulmonary shunting is seen at 230 W [percent
maximal oxygen uptake (%V̇O2 max) ⫽ 84%]. Note the delayed appearance of contrast bubbles in the left heart. The same pattern
is seen at 260 W. Again all images are apical 4-chamber views.
Fig. 3. Contrast echocardiograms 3 min after the termination of maximal exercise. Note that there is no evidence of intracardiac
or intrapulmonary shunting, as the left heart remains free of contrast bubbles. All images are apical 4-chamber views.
J Appl Physiol • VOL
97 • SEPTEMBER 2004 •
www.jap.org
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
801
Fig. 4. Contrast echocardiograms from a 28-yr-old female subject during exercise (%V̇O2 max ⫽ 40%). The sequential images show
the delayed appearance of contrast bubbles in the left heart. Each sequential image is separated in time by 1 s. The first evidence
of contrast bubbles in the left heart is found in the 5th image, which is 8 cardiac cycles after contrast appears in the right atrium.
All images are apical 4-chamber views.
findings provide good evidence for the recruitment of arteriovenous intrapulmonary shunts with exercise.
Is contrast echocardiography valid for detecting intrapulmonary arteriovenous shunting during exercise? Contrast
echocardiography is a standard clinical method for detecting
anatomic intrapulmonary shunts at rest (2, 18, 21, 48). Recently, Lee and colleagues (27) showed that contrast echocardiography was 100% sensitive compared with pulmonary angiography for detecting pulmonary arteriovenous malformations in patients with hereditary hemorrhagic telangiectasia.
However, the specificity of contrast echocardiography for detecting intrapulmonary arteriovenous shunts has not been carefully examined. Confidence in contrast echocardiography for
detecting anatomic intrapulmonary shunts during exercise is
dependent on adequately addressing several concerns involving contrast bubbles and pulmonary capillary morphology.
What are the effects of exercise on bubble survival during
passage through the pulmonary circulation? A possible explanation for our findings is that, with the shortened pulmonary
circulation time during exercise, small contrast bubbles (⬍10
␮m) may survive long enough to pass through pulmonary
capillaries and appear in the left heart chambers. The injected
saline contrast bubbles have a wide spectrum of diameters.
However, the small-diameter (⬍10 ␮m) contrast bubbles that
could pass through normal pulmonary capillaries collapse rapidly (29, 31, 32, 38). The remaining larger bubbles are filtered
and eliminated by the pulmonary microcirculation (4, 31, 32,
44). Meltzer and colleagues (32) calculated survival times of
contrast bubbles in degassed stationary whole blood by applying the theoretical principles of bubble behavior in fluids (14,
Table 2. Physiological data at the onset of transpulmonary
passage of contrast
Men (n ⫽ 11)
Women (n ⫽ 10)
All Subjects (n ⫽ 21)
Work rate, W 154⫾72 (65–230) 107⫾62 (16–196) 132⫾70 (16–230)
HR, beats/min 124⫾19 (102–154) 133⫾23 (111–178) 128⫾21 (102–178)
V̇E, l/min
56.4⫾20.1 (33–96) 35.5⫾15.3 (11–62) 46.4⫾20.6 (11–96)
2.3⫾0.7 (1.0–3.4) 1.4⫾0.7 (0.7–2.1) 1.9⫾0.8 (0.7–3.4)
V̇O2, l/min
%V̇O2max, %
61⫾18 (30–84)
56⫾23 (13–85)
59⫾20 (13–85)
Values are means ⫾ SD. Values in parentheses are ranges. HR, heart rate;
V̇E, expiratory ventilation; V̇O2, oxygen uptake; %V̇O2max, percent of relative
maximum oxygen uptake.
J Appl Physiol • VOL
Fig. 5. Histograms showing the distribution of exercise intensity (%V̇O2 max)
at the onset of shunting. Numbers above the bars indicate number of subjects.
Note that onset of intrapulmonary shunting is skewed toward the higher
exercise intensities.
97 • SEPTEMBER 2004 •
www.jap.org
802
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
63) to experimental data (64). These calculations estimated that
in static blood 8-␮m-diameter contrast bubbles would have a
life span of ⬍200 ms. Furthermore, theoretical and experimental work demonstrates accelerated bubble dissolution with
increased fluid pressure (54) and flow velocity (65). Thus,
during exercise, with increased vascular pressures and blood
flow velocity, contrast bubble survival times would likely be
even shorter than at rest. The mean pulmonary capillary transit
time is ⬃750 ms at rest and decreases during exercise. However, even in well-trained athletes, with cardiac outputs as high
as 30 l/min, mean pulmonary transit time does not fall below
450 ms at maximal exercise (56). Thus it is highly unlikely that
survival of contrast bubbles able to pass through normal
pulmonary capillaries during exercise explains our findings.
The contrast bubble diameters in vivo are not known precisely.
However, given the inherent instability of small bubbles in the
circulation (see above), it is estimated that the size distribution
of peripherally injected contrast bubbles entering the pulmonary microcirculation is in the range of 60 –90 ␮m (43).
What is the relationship between pulmonary capillary
distention and bubble size during exercise? Contrast echocardiography cannot distinguish between distinct anatomic
intrapulmonary arteriovenous shunts and dilated pulmonary
capillaries. Structural derangements of the pulmonary microcirculation, including intrapulmonary arteriovenous shunts and
capillary distention with capillary diameters of 60 – 80 ␮m, are
reported in hepatopulmonary syndrome (8, 43). These patients
have a widened A-aDO2 and a positive contrast echocardiogram. However, there are no data to support this magnitude of
capillary distention in the normal lung at rest or during exercise.
With progressive exercise, pulmonary vascular pressures
and flow increase. These forces act to recruit and distend the
pulmonary microcirculation (41). The true extent to which the
pulmonary capillaries distend in humans during exercise is not
known. Morphological studies in isolated perfused greyhound
lungs suggest that capillary distensibility is limited and is
unlikely to exceed 15 ␮m despite distending pressures as high
as 73 Torr (15). Reeves and Taylor (41) applied a distensibility
model for the pulmonary microcirculation (28) to pulmonary
hemodynamic data obtained in healthy humans during both
supine and upright exercise. They estimated the distensibility
coefficient (defined as the percent change in capillary diameter
per unit change in pressure) for human pulmonary microcirculation to be 1.35% per Torr. In humans, during peak exercise,
the mean pulmonary capillary distending pressure in zone III
has been estimated to be 36 Torr (58). Thus these calculations
suggest that pulmonary capillary distention would not exceed
20 ␮m, even at peak exercise intensity. Pulmonary capillary
disruption, not excessive distention, has been suggested as the
more likely result with such high pulmonary vascular pressures
(57). Thus our findings are not explained by passage of contrast
bubbles through distended pulmonary capillaries.
During exercise with increased pulmonary vascular driving
pressures (pulmonary artery pressure minus left atrial pressure), it is conceivable that contrast bubbles larger than pulmonary capillaries could be forced through the pulmonary
microcirculation. In humans, rapid, forceful injection of agitated saline contrast bubbles through a tightly wedged pulmonary artery catheter forced contrast bubbles through the normal
pulmonary microcirculation (31). This required a firm occluJ Appl Physiol • VOL
sive wedge position and an injection pressure of 300 Torr (44).
In humans at maximal exercise, pulmonary vascular driving
pressure remains relatively low, never exceeding 20 Torr (40).
With these low driving pressures it is unlikely that trapped
contrast bubbles are being forced through the pulmonary microcirculation even at the highest exercise levels.
Spontaneous echo formation has been reported in the literature but tends to occur in structures prone to blood stasis, such
as the inferior vena cava (30) and left atrium (7). Red blood
cell clumping and rouleaux are believed to be the most likely
source of the spontaneous echoes (39). Because increased
blood flow and shear stress abrogate these spontaneous echoes,
they are extremely unlikely during exercise and should not
create ambiguity in our study.
In summary, our findings in combination with published
theoretical and experimental data suggest that the contrast
bubbles are passing through distinct intrapulmonary arteriovenous shunts. However, our findings using saline contrast
echocardiography are qualitative. Furthermore, contrast echocardiography is limited to detecting anatomic intrapulmonary
shunts and thus is blind to intrapulmonary shunting due to
atelectasis and alveolar flooding. In addition, contrast echocardiography will not detect postpulmonary shunting. Further
studies are needed to define shunt vessel diameters and to
quantify the shunt fraction.
What is the anatomic basis for intrapulmonary arteriovenous shunts? Intrapulmonary arteriovenous shunts have been
demonstrated in human lungs (51, 52, 62). Wilkinson and
Fagan (62) examined the lungs of 49 infants (⬍44 wk old) who
died suddenly. They showed that very low injection pressures
⬍7.5 Torr were required to drive a 2% gelatin solution across
the pulmonary vascular bed in 61% (30 of 49) of the lungs.
Furthermore, they found that 50-␮m polymethylmethacrylate
beads also passed through the pulmonary circulation. The
authors suggested that the gelatin and beads were flowing
through a distinct low-resistance vascular network and that
these large-diameter conduits were remnant intrapulmonary
arteriovenous shunts. Tobin (51) examined plastic casts of
normal human lungs prepared specifically to preserve the
microcirculation. As expected, he found that the alveoli within
a primary lobule are supplied by a single branch of the arteriole
accompanying the bronchiole into the lobule. However, in 47%
of the lobules examined he found secondary glomuslike vessels
that branch from the parent arterioles at right angles. Some of
these vessels bypassed the capillary network, to terminate in a
pulmonary venule. Furthermore, infusion of glass or resin
beads, 50 –200 ␮m in diameter, into the pulmonary artery or
inferior vena cava bypassed the pulmonary microcirculation
and subsequently appeared in the pulmonary veins (51, 52).
The authors suggested that these vessels may function as
intrapulmonary arteriovenous shunts.
Elliot and Reid (13) described, in human lungs, a network of
small muscular arteries that branch from the conventional
pulmonary arteries at right angles and, because they have no
accompanying airways, are referred to as supernumerary arteries. Conventional pulmonary arteries accompany the airway to
supply the alveolar capillary from within the lobule. The
supernumerary arteries rapidly divide and enter the lobule at its
edge (42). Interestingly, Elliott and Reid (13) could not identify
supernumerary arteries in pulmonary angiograms performed in
humans at rest. At their origin, supernumerary arteries have a
97 • SEPTEMBER 2004 •
www.jap.org
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
sphincter (13) or muscular baffle valve (47) that appears to
regulate blood entry. The baffle valve is situated such that
parent artery dynamics determine whether the valve is opened
or closed. Under low-flow conditions or active vasoconstriction of the parent artery, the baffle valve is pulled closed.
However, as pulmonary blood flow increases, the parent artery
is distended and the baffle valve opens to allow blood to enter
the supernumerary artery (47). We believe that the supernumerary arteries may be the same vessels described by Tobin
and colleague (51, 52) (see above). Furthermore, a supernumerary-like vessel with a baffle valve could explain our findings that recruitment of intrapulmonary arteriovenous shunts
during exercise appear to be regulated, in part, by pulmonary
vascular pressures and flow. Clearly, further investigation is
needed to better define the structural characteristics and functional regulation of these dynamic intrapulmonary arteriovenous conduits.
What is the physiological relevance of intrapulmonary arteriovenous shunting during exercise? Intrapulmonary shunting through arteriovenous conduits will contribute to venous
admixture and widen the A-aDO2. Because shunted blood is
deoxygenated and becomes more so with exercise as oxygen
extraction increases, only a small fraction of the cardiac output
as shunt is necessary to widen the A-aDO2 significantly. Indeed, during moderate-intensity exercise an assumed 2% shunt
of mixed venous blood can account for one-half of the widened
A-aDO2, with increased V̇A/Q̇ nonuniformity as measured, with
MIGET accounting for the remainder (16). During maximal
exercise, Wagner and colleagues (55) also used MIGET to
determine that one-third of the measured A-aDO2 of 25 Torr
was due to V̇A/Q̇ nonuniformities. Using data from Wagner et
al. and the measured mixed venous oxygen content of 5.2 ml
O2/100 ml, we calculated that the remaining A-aDO2 could be
accounted for by a shunt fraction of 2.2% of the cardiac output.
In well-trained endurance athletes with high V̇O2 max, some
develop excessive A-aDO2 widening during both moderateand high-intensity exercise, whereas others do not (9, 20, 22).
Perhaps a significant source of this intersubject variability lies
in small interindividual differences in exercise effects on intrapulmonary shunts or variability in shunt vessels between
subjects.
The pulmonary vascular bed is unique. Being in series with
the left heart and the systemic circulation, the pulmonary
vascular bed must accommodate the entire cardiac output,
which will increase four- to sixfold from rest to exercise in
healthy humans. Local recruitment of arteriovenous intrapulmonary shunts, as exercise intensity increases, may help to
maintain low regional pulmonary vascular resistance. This may
allow for small increases in cardiac output after maximal
pulmonary capillary recruitment and distention has occurred.
With shunting, the arterial blood oxygen content will fall.
However, an increased cardiac output would at least partially
compensate for the fall in oxygen content and potentially
preserve systemic oxygen transport. There is a paucity of
studies examining the impact of intrapulmonary right-to-left
shunting on exercise. An extreme example is the work of
Whyte and colleagues (61), who investigated the effects of
intrapulmonary right-to-left shunting on pulmonary hemodynamics and gas exchange during exercise in patients with
substantial pulmonary arteriovenous malformations. These patients had surprisingly good maximal exercise capacity (⬎70%
J Appl Physiol • VOL
803
predicted) despite profound arterial oxygen desaturation. During progressive exercise, these individuals generate higher than
normal cardiac outputs, which serve to prevent large decrements in systemic oxygen delivery.
Local recruitment of intrapulmonary shunts may also act as
a parallel vascular network that opens to divert potentially
damaging hydrodynamic energy away from fragile pulmonary
capillaries. Without these shunts, some regions of the lung may
see high pressures and flow with resultant vascular and capillary injury. Indeed, some elite athletes develop exercise-induced pulmonary hemorrhage during very intense exercise
(23). In addition, some individuals develop capillary-leak pulmonary edema with rapid ascent to altitude (46). A postulated
mechanism for high-altitude pulmonary edema is capillary
stress failure (57). With increasing exercise intensity, individuals prone to high-altitude pulmonary edema have higher
pulmonary vascular pressures and resistance than matched
controls (12). Are these individuals more susceptible to pulmonary vascular injury because they do not have or fail to open
arteriovenous intrapulmonary shunts under these hyperdynamic conditions?
In summary, we have discovered what we believe to be
intrapulmonary arteriovenous shunts that are recruited during
exercise in healthy humans. The intrapulmonary shunts appear
to be distinct vascular conduits that open as pulmonary blood
flow and vascular pressures climb with increasing exercise
intensity. These intrapulmonary shunts will contribute to the
widened A-aDO2 seen with exercise. To a much lesser extent,
they may act as a protective parallel circuit limiting the rise in
pulmonary vascular pressure while preserving cardiac output
during exercise. In our study, only 2 of the 23 subjects did not
demonstrate intrapulmonary shunting. Are these individuals at
higher risk for pulmonary capillary injury during heavy exercise at high cardiac output but less prone to excessive widening
of the A-aDO2? Clearly, the existence of these dynamically
controlled intrapulmonary shunts has far-reaching impact on
how we view the pulmonary circulation in both health and
disease. Further investigations are needed to quantify these
exercise-induced shunts and to define their anatomical characteristics.
ACKNOWLEDGMENTS
We thank our subjects for enthusiastic participation in this study. The
technical support of Jamie Beebe, Sofia Spanoudaki, and David Pegelow is
gratefully acknowledged, as is the valuable insight provided by Micheal
Stickland.
GRANTS
Support for this project was provided by the Department of Pediatrics,
University of Wisconsin-Madison and National Heart, Lung, and Blood
Institute (NHLBI) RO1 Grant HL-15469-33. H. C. Haverkamp and A. T.
Lovering were supported by a NHLBI Training Grant-T32 HL-07654-16.
REFERENCES
1. Assmussen E and Nielsen M. Alveolar-arterial gas exchange at rest and
during work at different O2 tensions. Acta Physiol Scand 50: 153–166,
1960.
2. Barzilai B, Waggoner AD, Spessert D, Picus D, and Goodenberger D.
Two-dimensional contrast echocardiography in the detection and follow-up of congenital pulmonary arteriovenous malformations. Am J Cardiol 68: 1507–1510, 1991.
3. Bruce RA, Kusumi F, and Hosmer D. Maximal oxygen intake and
normographic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 85: 546 –562, 1973.
97 • SEPTEMBER 2004 •
www.jap.org
804
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
4. Butler BA and Hills BA. The lung as a filter for microbubbles. J Appl
Physiol 47: 537–543, 1979.
5. Cheney FW, Pavlin J, Ferens J, and Allen D. Effect of pulmonary
microembolism on arteriovenous shunt flow. J Thorac Cardiovasc Surg
76: 473– 478, 1978.
6. Conhaim RL and Staub NC. Reflection spectrophotometric measurements of O2 uptake in pulmonary arterioles of cats. J Appl Physiol 48:
848 – 856, 1980.
7. Daniel WG, Nellessen U, and Schroeder E. Left atrial spontaneous echo
contrast in mitral valve disease: an indictor for increased thromboembolic
risk. J Am Coll Cardiol 11: 1204 –1211, 1985.
8. Davis HH, Schwartz DJ, Lefrak SS, Susman N, and Schainker BA.
Alveolar-capillary oxygen disequilibrium in hepatic cirrhosis. Chest 73:
507–511, 1978.
9. Dempsey JA, Hanson PG, and Henderson KS. Exercise-induced arterial
hypoxaemia in healthy human subjects at sea level. J Appl Physiol 355:
161–175, 1984.
10. Dempsey JA, Reddan W, Rankin J, and Balke B. Alveolar-arterial gas
exchange during muscular work in obesity. J Appl Physiol 21: 1807–1814,
1966.
11. Duling B and Berne R. Longitudinal gradients in periarteriolar oxygen
tension. Circ Res 27: 669 – 678, 1970.
12. Eldridge MW, Podolsky A, Richardson RS, Johnson DH, Knight DR,
Johnson EC, Hopkins SR, Michimata H, Grassi B, Feiner J, Kurdak
SS, Bickler PE, Wagner PD, and Severinghaus JW. Pulmonary hemodynamic response to exercise in subjects with prior high-altitude pulmonary edema. J Appl Physiol 81: 911–921, 1996.
13. Elliott FM and Reid L. Some new facts about the pulmonary artery and
its branching pattern. Clin Radiol 16: 193–198, 1965.
14. Epstein PS and Plesset MS. On the stability of gas bubbles in liquid-gas
solutions. J Chem Phys 18: 1505–1509, 1950.
15. Glazier JB, Hughes JM, Maloney JE, and West JB. Measurements of
capillary dimensions and blood volume in rapidly frozen lungs. J Appl
Physiol 26: 65–76, 1969.
16. Gledhill N, Froese AB, and Dempsey JA. Ventilation to perfusion
distribution during exercise in health. In: Muscular Exercise and the Lung,
edited by Dempsey JA and Reed CE. Madison, WI: University of Wisconsin Press, 1977, p. 325–344.
17. Gramiak R, Shah PM, and Kramer DH. Ultrasound cardiology. Contrast studies in anatomy and function. Radiology 92: 939 –948, 1968.
18. Gudavalli A, Kalaria VG, Chen X, and Schwarz KQ. Intrapulmonary
arteriovenous shunt: diagnosis by saline contrast bubbles in the pulmonary
veins. J Am Soc Echocardiogr 15: 1012–1014, 2002.
19. Hammond MD, Gale GE, Kapitan KS, Ries A, and Wagner PD.
Pulmonary gas exchange in humans during exercise at sea level. J Appl
Physiol 60: 1590 –1598, 1986.
20. Harms CA, McClaran SR, Nickele GA, Pegelow DF, Nelson WB, and
Dempsey JA. Exercise-induced arterial hypoxaemia in healthy young
women. J Physiol 507: 619 – 628, 1998.
21. Hernandez A, Strauss AW, McKnight R, and Hartman AFJ. Diagnosis of pulmonary arteriovenous fistula by contrast echocardiography.
J Pediatr 93: 258 –261, 1978.
22. Hopkins SR, Barker RC, Brutsaert TD, Gavin TP, Entin P, Olfert IM,
Veisel S, and Wagner PD. Pulmonary gas exchange during exercise in
women: effects of exercise type and work increment. J Appl Physiol 89:
721–730, 2000.
23. Hopkins SR, Schoene RB, Henderson WR, Spragg RG, Martin TR,
and West JB. Intense exercise impairs the integrity of the pulmonary
blood-gas barrier in elite athletes. Am J Respir Crit Care Med 155:
1090 –1094, 1997.
24. Jameson A. Diffusion of gases from alveolus to precapillary arteries.
Science 139: 826 – 828, 1963.
25. Knudson RJ, Kaltenborn WT, Knudson DE, and Burrows B. The
single-breath carbon monoxide diffusing capacity. Reference equations
derived from a healthy nonsmoking population and effects of hematocrit.
Am Rev Respir Dis 135: 805– 811, 1987.
26. Knudson RJ, Lebowitz MD, Holber CJ, and Burrows B. Changes in
normal maximal expiratory flow-volume curve with growth and aging. Am
Rev Respir Dis 127: 725–734, 1983.
27. Lee WL, Graham AF, Pugash RA, Hutchison SJ, Grande P, Hyland
RH, and Faughnan ME. Contrast echocardiography remains positive
after treatment of pulmonary arteriovenous malformations. Chest 123:
351–358, 2003.
J Appl Physiol • VOL
28. Linehan JH, Haworth ST, Nelin LD, Krenz GS, and Dawson CA. A
simple distensible vessel model for interpreting pulmonary vascular pressure-flow curves. J Appl Physiol 73: 987–994, 1992.
29. Meerbaum S. Principles of echo contrast. In: Advances in Echo Imaging
Using Contrast Enhancement, edited by Nanda N and Schlief R. Boston,
MA: Kluwer Academic Publishers, 1993, p. 9 – 42.
30. Meltzer RS, Klig V, Visser CA, and Teichholz LE. Spontaneous
echocardiographic contrast in the inferior vena cava. Am Heart J 110:
826 – 830, 1985.
31. Meltzer RS, Sartorius OEH, Lancee GT, Srerruys PW, Verdouw PD,
Essed CE, and Roelandt J. Transmission of echocardiographic contrast
through the lungs. Ultrasound Med Biol 7: 377–384, 1981.
32. Meltzer RS, Tickner EG, and Popp RL. Why do the lungs clear
ultrasonic contrast? Ultrasound Med Biol 6: 263–269, 1980.
33. Meltzer RS, Tickner EG, Salines TP, and Popp RL. The source of
ultrasound contrast effect. J Clin Ultrasound 8: 121–127, 1980.
34. Niden AN and Aviado DM. Effects of pulmonary embolism on the
pulmonary circulation with reference to arterio-venous shunts in the lung.
Circ Res 4: 67–73, 1956.
35. Ogilvie C, Forster R, Blakemore W, and Morton J. A standardized
breath holding technique for the clinical measurement of the diffusing
capacity of the lung for carbon monoxide. J Clin Invest 36: 1–17, 1957.
36. Prinzmetal M, Ornitz ME, Simkin B, and Bergman HC. Arteriovenous anastomoses in liver, spleen and lungs. Am J Physiol 152:
48 –52, 1948.
37. Rahn H, Stroud RC, and Tobin CE. Visualization of arteriovenous
shunts by cineflurography in the lungs of normal dogs. Proc Soc Exp Biol
Med 80: 239 –241, 1952.
38. Raisinghani A and DeMaria AN. Physical principles of microbubble
ultrasound contrast agents. Am J Cardiol 90, Suppl: 3J–7J, 2002.
39. Reeder GS, Charlesworth JE, and Moore SB. Cause of spontaneous
echocardiographic contrast as assessed by scanning electron microscopy.
J Am Soc Echocardiogr 7: 169 –173, 1994.
40. Reeves JT, Dempsey JA, and Grover RF. Pulmonary circulation during
exercise. In: Pulmonary Vascular Physiology and Pathophysiology, edited
by Weir EK and Reeves JT. New York: Dekker, 1988, p. 107–133.
41. Reeves JT and Taylor AE. Pulmonary hemodynamics and fluid exchange
in the lung during exercise. In: Handbook of Physiology. Exercise:
Regulation and Integration of Multiple Systems. Bethesda, MD: Am.
Physiol. Soc., 1996, sect. 12, chapt. 13, p. 585– 613.
42. Reid LM. Structural remodelling of the pulmonary vasculature by environmental change and disease. In: The Pulmonary Circulation and Gas
Exchange, edited by Wagner WW Jr and Weir EK. New York: Futura,
1994, p. 77–110.
43. Rodriguez-Roisin R, Agusti AGN, and Roca J. The hepatopulmonary
syndrome: new name, old complexities. Thorax 47: 897–902, 1992.
44. Roelandt J. Contrast echocardiography. Ultrasound Med Biol 8: 471–
492, 1982.
45. Schaffartzik W, Poole DC, Derion T, Tsukimoto K, Hogan MC, Arcos
JP, Bebout DE, and Wagner PD. V̇A/Q̇ distribution during heavy
exercise and recovery in humans: implications for pulmonary edema.
J Appl Physiol 72: 1657–1667, 1992.
46. Schoene RB, Swenson ER, Pizzo CJ, Hackett PH, Roach RC, Mills
RM Jr, Henderson WR Jr, and Martin TR. The lung at altitude:
bronchoalveolar lavage in acute mountain sickness and pulmonary edema.
J Appl Physiol 64: 2605–2613, 1983.
47. Shaw AM, Bunton DC, Fisher A, McGrath JC, Montgomery I, and
MacDonald A. V-shaped cushion at the origin of bovine pulmonary
supernumerary arteries: structure and putative function. J Appl Physiol 87:
2348 –2356, 1999.
48. Shub C, Tajik AJ, Seward JB, and Dines DE. Detecting intrapulmonary
right-to-left shunt with contrast echocardiography. Observations in a
patient with diffuse pulmonary arteriovenous fistulas. Mayo Clin Proc 51:
81– 84, 1976.
49. Sobol B, Bottex G, Emirgil C, and Gissen H. Gaseous diffusion from
alveoli to pulmonary vessels of considerable size. Circ Res 13: 71–79,
1963.
50. Sonetti D, Wetter T, Pegelow D, and Dempsey J. Effects of respiratory
muscle training versus placebo on endurance exercise performance. Respir
Physiol 127: 185–199, 2001.
51. Tobin CE. Arteriovenous shunts in the peripheral pulmonary circulation
in the human lung. Thorax 21: 197–204, 1966.
97 • SEPTEMBER 2004 •
www.jap.org
INTRAPULMONARY ARTERIOVENOUS SHUNTING DURING EXERCISE
52. Tobin CE and Zariquiey MO. Arteriovenous shunts in the lung. Proc
Soc Exp Biol Med 75: 827– 819, 1950.
53. Torre-Bueno JR, Wagner PD, Saltzman HA, Gale GE, and Moon RE.
Diffusion limitation in normal humans during exercise at sea level and
simulated altitude. J Appl Physiol 58: 989 –995, 1985.
54. Tsujino T and Shima A. The behavior of gas bubbles in blood subjected
to an oscillating pressure. J Biomech 13: 408 – 416, 1980.
55. Wagner PD, Gale GE, Moon RE, Torre-Bueno JR, Stolp BW, and
Saltzman HA. Pulmonary gas exchange in humans exercising at sea level
and simulated altitude. J Appl Physiol 61: 260 –270, 1986.
56. Warren GL, Cureton KJ, Middendorf WF, Ray CA, and Warren JA.
Red blood cell pulmonary capillary transit time during exercise in athletes.
Med Sci Sports Exerc 23: 1353–1361, 1991.
57. West JB. Invited review: Pulmonary capillary stress failure. J Appl
Physiol 89: 2483–2489, 2000.
58. West JB. Left ventricular filling pressures during exercise. A cardiologist
blind spot? Chest 113: 1695–1697, 1998.
59. Wetter TJ, St. Croix CM, Pegelow DF, Sonetti DA, and Dempsey
JA. Effects of exhaustive endurance exercise on pulmonary gas ex-
J Appl Physiol • VOL
60.
61.
62.
63.
64.
65.
805
change and airway function in women. J Appl Physiol 91: 847– 858,
2001.
Whipp BJ and Wasserman K. Alveolar-arterial gas tension differences
during graded exercise. J Appl Physiol 27: 361–365, 1969.
Whyte MK, Hughes JM, Jackson JE, Peters AM, Hempleman SC,
Moore DP, and Jones HA. Cardiopulmonary response to exercise in
patients with intrapulmonary vascular shunts. J Appl Physiol 75: 321–328,
1993.
Wilkinson MJ and Fagan DG. Postmortem demonstration of intrapulmonary arteriovenous shunting. Arch Dis Child 65: 435– 437,
1990.
Yang WJ. Dynamics of gas bubbles in whole blood and plasma. J Biomech 4: 119 –125, 1971.
Yang WJ, Echigo R, Wotten DR, and Hwang JB. Experimental studies
of the dissolution of gas bubbles in whole blood and plasma. I. Stationary
bubbles. J Biomech 4: 275–281, 1971.
Yang WJ, Echigo R, Wotten DR, and Hwang JB. Experimental studies
of the dissolution of gas bubbles in whole blood and plasma. II. Moving
bubbles or liquids. J Biomech 4: 283–288, 1971.
97 • SEPTEMBER 2004 •
www.jap.org