LUNG CAPILLARY
system plays a role in the early stages of Goldblatt
hypertension.
In the dogs that received indomethacin, in contrast to
those that did not, the systemic arterial pressure was
elevated within 5 minutes after renal artery constriction and
remained elevated throughout the experiment. Thus the
earlier and more sustained hypertension in indomethacintreated dogs was probably due at least in part to the
inhibition of prostaglandin release. Apparently the opposite
untouched kidney attenuates Goldblatt hypertension, at
least its early stages, not only by its excretory function"
but also by the release of vasodepressor prostaglandins.
Acknowledgments
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
We are deeply grateful to Dr. Francis J. Haddy for help and encouragement during the course of this study and preparation of this manuscript.
During Dr. Ovcrbeck's absence. Dr. J.B Scott provided valuable guidance.
We arc greatly indebted to Dr Michael D Bailey for determinations of
plasma renin concentrations We thank Dr. Sibley Hoobler for supplying us
with antircnin serum and Dr. Erwin Haas, Beaumont Memorial Research
Laboratories, for the renin. We also acknowledge the valuable technical
assistance of Donald L. Anderson, Booker T Swmdall, and Josephine
Johnston, and the excellent secretarial help of Lynn Dietrich.
References
1. Skinner SL, McCubbin JW, Page IH: Angiotensin in blood and lymph
following reduction of renal arterial pcrfusion pressure in dogs Circ Res
13: 336-345. 1963
2. Bianchi G, Baldoli E, Lucca R, Barbin P: Pathogcnesis of arterial
hypertension after the constriction of renal artery leaving the opposite
kidney intact both in the anesthetized and in the conscious dog Clin Sci
42: 651-664, 1972
3 Gutmann FD, Tagawa H, Habcr E, Barger AC: Renal arterial pressure,
renin secretion, and blood pressure control in trained dogs. Am J Physiol
224: 66-72, 1973
PERMEABILITY/Yipintsoi
523
4. Goldblatt H Studies on experimental hypertension and the pathogenesis
of experimental hypertension due to renal ischemia Am J Clin Pathol 10:
40-72, 1940
5. McGiff JC, Crowshaw K, Terragno NA, Lonigro AJ: Release of a
prostaglandin-hkc substance into renal venous blood during renal
ischemia. Circ Res 27: 765-782. 1970
6. Sweet CS, Kadowitz PJ, Forker EL, Brody MJ: Depression of adrcnergic
transmission by a factor in renal venous blood, new evidence for
antihypcrtcnsivc function of the kidney Arch Int Pharmacodyn Trier
198: 229-237, 1972
7. Skeggs LT, Lcntz KE, Hochstrasser H, Kahn JR The purification of
several forms of hog renin substrate. J Exp Med 118: 73-98, 1963
8. Haber E, Kocrner T, Page LB, Kliman B, Purnode A: Application of a
radioimmunoassay for angiotensin I to the physiologic measurements of
plasma renin activity in normal human subjects. J Clin Endocrinol Metab
29: 1349-1355, 1969
9. Haddy FJ, Scott JB, MolnarJI Mechanism of volume replacement and
vascular constriction following hemorrhage Am J Physiol 208: 169-181,
1965
10. Vane JR" Inhibition of prostaglandin synthesis as a mechanism of action
for aspirin like drugs. Nature (New Biol) 231: 232-235, 1971
11. Oparil S, Bailie MD: Mechanism of renal handling of angiotensin II in
the dog. Circ Res 33: 500-507, 1973
12. Leary WP, Ledingham JG: Renal and hepatic inactivation of angiotensin
in rats, influence of sodium balance and renal artery compression Clin
Sci 38: 573, 1970
13. McGiff JC, Crowshaw K, Terragno NA, Lonigro AJ: Release of
prostaglandin-like substance into renal venous blood in response to
angiotensin II Circ Res 26 (suppl I): 121-130, 1970
14. LccJB Cardiovascular-renal effects of prostaglandins, the antihypertcnsivc, natnuretic renal "endocrine" function. Arch Intern Med 133:
56-76, 1974
15. Ishi M, Tobian L- Interstitial cell granules in renal papilla and the solute
composition of renal tissues in rats with Goldblatt hypertension. J Lab
Clin Med 74: 47-52, 1969
16. Ebihara A, Grollman A: Pressor activity of renal venous effluent
following constriction of artery in dogs. Am J Physiol 214: 1-5, 1968
17 Gross F The renin angiotensin system. Ann Intern Med 75: 777-787,
1971
18. Gross F, DandaG, Kazda S, Kyncl J, Mohrmg J, Orth H: Increased fluid
turnover and the activity of the renin-angiotensin system under various
experimental conditions. Circ Res 30/31 (suppl II): 173-181, 1972
Single-Passage Extraction and Permeability Estimation
of Sodium in Normal Dog Lungs
TADA YIPINTSOI, M.D.,
PH.D.
SUMMARY
Bolus injections of "MaCI, '"I-albumin, and
indocyanine green dye were made into the right atria of anesthetized,
ventilated dogs. Blood was sampled from the femoral artery, and
from the dilution curves, instantaneous extractions, E,(t), and area
averaged extractions, E,(t), were calculated for sodium at various
plasma flows ( F p ) . Flow reduction was produced by transient inferior
vena caval obstruction. E,U) and E , m within any dilution curve
generally started off with a high value, then decreased with time. The
E, that occurred at the peak of the albumin-dilution c u n e were about
0.11 for plasma flow of 0.75 liler/min and tended to decrease as flow
increased. Parallel study of the sodium extravascular space at
equilibrium gave values of 0.3-0.4 g of plasma sodium per g of tissue,
suggesting that this volume is not infinitely small. Since the E was low
it is unlikely that the high initial Elt) and the decreasing E(t) were
due to early back flux. Calculation of capillary permeability surface
area product [PS - - F p log. (1 - E)] showed an increasing P S with
plasma flow. The injection of 9-^m microspheres at low and high
total blood flow gave evidence supporting a decrease of capillary
surface area (S) with decreasing total blood flow. Regional pulmonary blood flow also showed marked heterogeneity. Because of the
low average extraction of sodium in the lung, insensitivity of the
method in normal lungs cannot be excluded.
PULMONARY capillary permeability is thought to be
altered in several disease states, independent of whether the
lung is the primary or secondary organ of involvement.1
However, quantification of pulmonary capillary permeabil-
ity in intact animals and man is difficult because (1) there is
an intrinsically high blood flow, (2) regional distribution of
flow in the lung may change when total blood flow increases
or decreases, (3) limitations exist in the techniques currently
used for measurement of capillary permeability.
From the Department of Medicine and Physiology, Montefiore Hospital
and Medical Center (Albert Einstein College of Medicine), Bronx, New
York.
Supported by American Heart Association Grant 74 938 and U.S. Public
Health Service Grant HL 17417, 16476.
Received July 11, 1975: accepted for publication June 3, 1976.
524
CIRCULATION RESEARCH
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Chinard 1 and Yudilevich1 injected sodium and albumin in
the same bolus into the right atrium and sampled blood from
the arterial circulation. Because of the similarity between
the dilution curves of the two isotopically labeled substances, i.e., low value of fractional extraction, it was
thought that either the lung capillary permeability is low or
that the extravascular volume of sodium in the lung
(calculated from the difference in the mean transit time
volumes) is small, or both. Increased capillary permeability
to sodium was implied when this mean transit time volume
was increased or when sodium recovery was decreased.*""
However, in these studies, it was not realized that lung blood
flow per unit of weight is very high in comparison to that in
other muscular organs like the heart. Such high flow would
result in low tissue extraction of the probing molecules.
Subsequently Effros,7 using the osmotic bolus technique,
and Perl et al.,' summarizing their previous indicator-dilution results, reported that the capillary permeability of dog
lungs probably is similar to that of heart or skeletal muscle.
Aside from the latest summary by Perl et al., 8 other
quantification of lung capillary permeability has been
performed in isolated or exteriorized lungs or in situations in
which pulmonary lymphatics could be sampled. Taylor and
Gaar 9 calculated the capillary pore radius to be 40-80 A in
isolated dog lung. Normand et al. 10 calculated pore radius
to be 150 A in exteriorized lungs of fetal lambs.
The purpose of our present study was to use the indicatordilution method and the single-passage extraction technique
described by Crone 11 to evaluate capillary permeability to
sodium in intact canine lungs. The technique demands that
the tissue extraction, E, result from a unidirectional flux
from the intravascular plasma into the interstitium. The
assumptions inherent in the technique are: (1) There is no
back flux of the probing molecule from the interstitial space
into the plasma stream during the period of measurement.
This period of measurement is generally a fraction of a
single transit time through the organ under study. The
degree of back flux depends on the permeability of the
membrane and the extravascular distribution space of the
probing molecules. Greater back flux should occur with the
more permeable substance or with the smaller volume of
distribution or both. Hence the choice of probing molecule is
important and so is the assurance that an adequate volume
of distribution exists. (2) Interlaminar diffusion in the
intravascular compartment is minimal. 12 (3) There is an
adequate extravascular volume of distribution for the hydrophilic probing molecule. (4) There is no marked heterogeneity of perfusion, volume of distribution, or of permeability in
the lung. If marked heterogeneity of these parameters exist,
then we should not allocate a single permeability surface
area product (PS) to the lung (i.e., a lumped model), but
rather a distribution of PS as functions of regional flow or of
regional extravascular volume (i.e., a distributive model).
Since the assumptions listed require knowledge of the
interstitial volume of distribution (v E ) and of heterogeneity
of flow or the available capillary surface area (S), we also
made measurements of the interstitial space and of the
regional blood flow in the lungs.
VOL. 39, No.
4, OCTOBER
1976
Methods
Adult mongrel dogs weighing 18-36 kg were anesthetized
with intravenous pentobarbital sodium (25 mg/kg). All
studies were made with the dogs lying on their backs and
ventilated with a constant volume Harvard respirator. The
tidal volume was not altered during an experiment except for
periodic hyperinflation. The inspiratory airway pressure was
10-15 mm Hg. An external jugular cutdown was performed
and under fluoroscopic control catheters were positioned as
follows: a no. 6 or 7 USCI (Bard) end-hole catheter was
passed into the main pulmonary artery to monitor pressure;
a multiple side-hole catheter (8 pigtail) with an internal volume of 1.2-1.6 ml was placed in the right atrium for the injection of indicators; and a catheter (7 Fr.) with a terminal
inflatable balloon (maximum inflated balloon diameter of 4
cm) was placed so that the inflated balloon was below the
diaphragm in the inferior vena cava (IVC). An 18-gauge, 2inch nylon needle was placed in the femoral artery to monitor pressure and to obtain sequential blood samples. In
some experiments a separate needle was placed in the brachial artery to monitor systemic pressure during collection
of blood from the femoral artery.
An indocyanine green dye (ICG) densitometer (Gilford
103IR) was connected to the femoral artery needle and
blood was withdrawn through this by a rotating pump for
sequential blood collection. The volume of the sampling
system varied between 2 and 3.1 ml, and generally the pump
flow was 60 ml/min when multiple indicators were used.
When only ICG was injected, arterial blood was withdrawn
at 24 ml/min through a Harvard syringe. The calculated
blood flow from ICG was similar with either the syringe or
the rotating pump. When isoproterenol (0.001 mg/ml) was
infused intravenously the rate was adjusted to produce a
stable doubling of the cardiac output.
For the instantaneous extraction study, the bolus injection
(about 0.8 ml) consisted of 3-5 uCi of 131I-albumin (RISA,
Abbott) and 10-30 uCi of 2< NaCl (New England Nuclear),
both freshly diluted in normal saline, and about 1-2 mg of
ICG, freshly diluted in distilled water. The bolus was flushed
into the right atrium with 5-10 ml of normal saline. Blood
was collected sequentially from the femoral artery into
preweighed test tubes. The collection time for each sample
was marked on the photographic record on which we also
monitored the electrocardiogram, the pulmonary and the
systemic arterial pressures, and the ICG-dilution curve. The
rate of blood sampling was as rapid as two samples per
second during the upstroke of the dye curve and was slower,
one sample per 1-2 seconds, prior to the appearance of the
dye and after the peak of the dye curve. The samples then
were covered with wax paper and subsequently were
weighed. In three experiments the free iodide in the injectate
obtained by 10% trichloroacetic acid precipitation represented less than 2% of the total iodide activity.
For each isotope-dilution curve, we analyzed samples of
the injectate suitably diluted in blood. The " N a gamma
activity was counted first (Nuclear Chicago 1085 automatic
gamma system); 4-6 days later (after 6-10 half-lives of
"Na), the remaining " N a and 131I were counted. Aside
LUNG CAPILLARY PERMEABILITY/Yipintsoi
from the first few samples after the appearance time, the
gross counts for " N a and 1S1I alone were in excess of 10,000
counts. The concentration of each blood sample (counts/
min per g of blood weight) then was expressed as a fraction
of the injected dose. The flow was calculated from both the
ICG and the lai I-albumin curve (corrected for density of
blood) using a single exponential extrapolation to the
descending portion of the dilution curve. Generally both
indicators gave flow values within 10% of each other, but in
instances of marked disparity, the albumin curve was
assumed to be correct. The relative concentrations (f) of Na
and albumin at any time were used to calculate the
extraction, E(t).
E,(t) = I - f N .(t)/f Alb (t)
(I)
E,(t) - 1 - o/fN.(X)dX/JfAlb(A)dX
(2)
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
X is the variable for integration up to time t.
The subscripts 2 and 3 refer to the two forms of extraction
described previously,13 where E2 is the instantaneous or
instant-by-instant extraction derived from paired concentration at any time and E , is the area weighted extraction
derived from paired areas of the dilution curves such that the
extraction 4s4nfluenced by the curve up to time t.11 This E, is
1 minus the ratio of the cumulative recovery. From the E, we
calculated the permeability surface area (PS) for the whole
lung as:
PS = - F p log, (1 - E)
(3)
where F p is the plasma flow and E is an arbitrarily chosen
extraction that occurs at the time of the peak (tp) of the
albumin curve.14
The present study does not include the calculation of
recoveries or mean transit time volumes. Perl et al.' have
shown and we have found that the single exponential
extrapolation of the descending slope of the sodium-dilution
curve differs minimally from that of albumin, resulting in a
very small difference in mean transit time volumes.
When we plotted E as a function of flow (see Results) the
values used did not include those obtained after pharmacological interventions.
MEASUREMENT OF ALBUMIN AND SODIUM SPACES
We obtained heparinized arterial blood and tissue samples
at least 30 minutes after the last bolus injection of ' " I albumin and "NaCl. In most dogs this last bolus was the
fourth or fifth, the first having been given several hours previously. In two dogs (7024A and B) only one bolus was given.
A portion of the blood was centrifuged at 2,000-3,000 rpm
for 20 minutes to obtain plasma and packed red cells. The
tissue samples (lung, heart muscle, and skeletal muscle) were
rinsed and blotted dry. No attempt was made to remove
small bronchi from the lung. The tissues, plasma, and
packed red cells were weighed and counted. The other
tissues were weighed and dried at 80-90°C for 2-3 days and
their water content was determined.
The relative isotope content, v, was the concentration of
tissue or red cells divided by that of plasma. The extravascular sodium content, VE_N., was corrected for the albumin
525
content and the equivalent red blood cell (RBC) content:
VE-N. = VN. - vA,b - v Alb -v RBC -Hct/(l - Hct)
(4)
When the RBC sodium space (yRBC) was measured, it
varied from 0.2 to 0.4 g per g of packed red cells. We did not
attempt to correct it to grams of pure red cells because this
correction factor to obtain VE_NI was relatively small.
In two separate experiments, freshly drawn heparinized
blood was kept stirred at 37°C, and labeled albumin and
sodium were added. At varied intervals samples of blood
were removed to estimate the isotopic concentrations of
plasma, red cells, and blood. In these tests it was found that
for constant hematocrit and plasma concentrations of
sodium and albumin, the relative red cell sodium content
was 0.08 g of plasma per g of RBC at 10 minutes and it
increased to 0.3 at 60 or 120 minutes, suggesting that the
time for equilibration of sodium between plasma and RBC
is rather long. Because of this apparent plateau at 0.3 we
used the same value in other dogs in which the vBBc was
not measured.
We reported the lung sodium spaces in two ways. In one,
the lung samples first were considered as a whole prior to
evaluating the extravascular space. In the other, we evaluated VE_N. for each sample, then expressed these as mean (x)
and standard deviation (SD) for n, number of pieces. In this
report simultaneous volumes for heart and skeletal muscle
were obtained so that their values could be compared with
other normal values in the literature.
MEASUREMENT OF THE REGIONAL DISTRIBUTION OF
LUNG BLOOD FLOW AT VARIED FLOWS
We injected labeled 9-^m microspheres (3M Co.) (5
million to 20 million particles) together with ICG solution
into the right atrium. The microspheres were premixed with
2-3 drops of 5% polysorbate 80 (Tween 80) and agitated
with an ultrasonic mixer for 10 minutes prior to injection.
Total flow then was decreased by IVC occlusion or bleeding,
or increased by isoproterenol infusion. When a stable state
was achieved, we injected a differently labeled 9-^m microsphere. Arterial blood was withdrawn at 24 ml/min for the
simultaneous ICG-dilution curve. The activity of microspheres from the collected arterial blood was never significantly above background. The dogs then were killed with
excess pentobarbital, after which the chest was opened by
sternal splitting. The lungs were kept inflated to facilitate
recognition of their hydrostatic position in situ. In some
dogs we analyzed the entire lung; in others, we sampled
regions in which we expected the lowest and highest flows
(e.g., anterior or nondependent and posterior or dependent
portions with the dog lying on its back). The tissues that
were not used for counting were minced and representative
samples were taken to calculate total microsphere content in
both lungs. The wet weight for pieces that were counted
varied from 0.5 to 1.5 g. The relative concentration of
microspheres in each piece was the count/min per g for the
piece divided by the count/min per g for the whole lung.15
This regional distribution of the first microsphere injected at
one total blood flow was plotted against the regional
CIRCULATION RESEARCH
526
Results
ISOTOPIC SPACES
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Table 1 details the 131I-albumin and " N a spaces in the
heart, lung, and skeletal muscle in seven dogs. As shown in
Table 1, some of the studies followed interventions such as
isoproterenol or methoxamine infusion or repeated saline
infusions, and one was after phlebotomy. The values following these interventions were included to show the variation
of these volumes to the water content and the hematocrit.
The present experiments were not designed specifically to
study the isotopic spaces as a function of these other
parameters.
In different lungs there was a wide range of values for vA,b
and vE_Nl but both generally were 2-3 times that of the
heart. The wide range of vA,b should not be used as a true
measure of intravascular space, because no precautions were
made to limit the intravascular volume during tissue sampling. We have no reasonable explanation for the high
albumin space and relatively low sodium spaces in the lung
in experiment 27024 (Table 1). This dog was given an
isoproterenol infusion to increase total blood flow for
regional perfusion study. Except for this dog, the vK NB
generally was 0.33-0.47 g/g.
Since the study on vE was made at an assumed steady
state, we also evaluated sodium uptake at shorter intervals
after injection. In two open-chest dogs (7024A and B),
samples of arterial blood and lung pieces were taken at 2, 5,
10, 20, 30, and 40 minutes after injection of a bolus of
sodium and albumin into the right atrium. The vKBC was
assumed to be zero. The values for vE_N« for the lung in these
quasi-steady state experiments were similar to the values
obtained when the samples were taken an hour later (Table
1, experiments 7024A and B).
In experiment 24094 we evaluated vE_N. following saline
infusion. Here lung and blood samples were taken at three
periods, each of which was 40-50 minutes after a bolus
4, OCTOBER
1976
SODIUM EXTRACTION
Figure 1A and B presents dilution curves and their
extractions [E2(t) and E,(t)] for normal and low blood flows
(produced by IVC balloon) in the same dog. The descending
portions of the dilution curves were quite representative of
our other curves in that there generally was no crossover
between the sodium and albumin, and where the albumin
showed recirculation the sodium concentration continued to
decline, suggesting further extraction of sodium by other
capillary beds.
The initial values of E2(t) and E8(t) were high; these
decreased more gradually for E,(t) than for Ej(t). The
decrement of these extractions often does not allow recognition of a constant value around the peak of the albumin
curve. Neither the initially high sodium extraction values
nor the shape of its subsequent decay was altered by
lengthening the sampling tubing from the femoral artery or
by increasing or decreasing the pump flow rate.
Because of the often decreasing values of E2(t) or Es(t)
and because of lack of distinct crossover between sodium
and albumin, the extractions were evaluated arbitrarily at
the time of peak of the albumin curves and E3(tp) (because of
its area-weighting) was used as the E from which to calculate
PS (Equation 3).
Table 2 gives the relevant hemodynamic data, the E, and
E, (both at t p ), and the calculated PS in eight dogs. We
specifically compared the E during the control period (C) to
that after IVC obstruction (+IVC). Values for other
interventions such as bleeding or saline infusion, or with and
without assisted ventilation (on-and-off respirator) were also
included to show the possible range of values for E, and E,.
The E,(tp) and PS obtained from these latter interventions
Sodium and Albumin Space
Tissue wt (g)
Expt
no.
No.
injection of the tracers into the right atrium; period a was
the control, and periods b and c followed 500-ml infusions of
Ringer's lactate solution. The hematocrit (Table 1)
decreased from 0.44 to 0.39; vA,b decreased from 0.20 to
0.13 g/g; and vHBc increased to 0.3 vE_N. for lung and lung
water remained unchanged.
distribution at another total blood flow. Total blood flow
was calculated from the ICG-dilution curve.
TABLE 1
VOL. 39,
L
12123
7024A
7024 B
27024
5.5
15.7
20.1
64.5
23074
24094*
a
b
c
19124t
122.7
6.8
3.6
11.2
25.2
% water
VAI«(g/g)
V..H.
H
M
L
H
M
Hct
L
H
M
L
x
SD
8.5
3.9
7.1
16.5
4.1
2.9
3.1
—
78
79
78
77
76
76
73
68
71
73
—
—
0.51
0.48
0.38
0.26
0.11
0.08
0.13
0.34
0.03
0.08
0.04
0.07
0.01
0.004
0.44
0.36
0.47
0.25
0.44
0.37
0.46
0.25
0.06
0.05
0.09
0.10
—
79
0.43
0.16
—
—
0.36
0.36
—
—
—
—
81
82
80
78
0.44
0.32
0.29
0.52
0.20
0.22
0.13
0.16
—
—
0.05
0.09
0.44
0.44
0.44
0.33
0.44
0.44
0.45
—
—
—
6.3
9.1
—
—
75
—
—
—
Intervention
before sampling
H
M
4
13
23
59
0.15
0.13
0.16
0.18
0.17
0.04
0.08
—
0.40
—
—
—
0.05
120
—
—
—
Isoproterenol
infusion
Bleeding
0.03
0.01
0.03
—
11
6
12
—
—
—
0.20
0.19
0.20
0.30
0.30
0.19
Saline infusion
Saline infusion
Hypertension
n
VRBC
Tissue wt - the total wet weight used for the analysis; L, H, M - hang, heart, and skeletal muscles; v - the isotopic distribution space; v t N . - the corrected
cxtravascular sodium space where L refers to the value obtained from initially summed up the lung pieces, and x and si) are the mean and standard deviation for the
n lung pieces evaluated individually; vBBC _ the relative sodium content in the red cells. Dash - no measurement made.
• In experiment 24094, a - the control period: b and c - the periods after 5O0-ml infusions of Ringer's lactate solution.
t In experiment 19124, hypertension was induced by methoxamine (2 mg, iv) given in 20 minutes.
LUNG CAPILLARY PERMEABILITY/K/p/n/50/
F
•A
PA
hilt
Hct
W
/
OBUII
2.0 L/«H
11] a i m
16
123 / • *
038
187 o
c
\
* VV
23
17
SECONDS
F - 1,1
A -10t
PA - 14
m • 114
H
\
m
1
/
A
o»o V .
\
M>UI
• - 117
°°\
\
2
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
/
i
\V1
/
A
A
o
20
•
oV
A
B,
24
28
32
36
527
altered by a single intervention (isoproterenol infusion,
transient IVC occlusion, or bleeding). In one dog we first
increased flow by isoproterenol infusion, then after return to
control state we reduced flow by transient IVC occlusion.
Figure 3 shows the regional concentration of microsphere at
normal flow and when flow was increased by isoproterenol
infusion. At initial blood flow of 1,540 ml/min, the pieces of
lung with relative concentration of less than 1 (i.e., less than
average) showed a disproportionate increase when total flow
was elevated to 3,300 ml/min. The reverse was seen in lung
pieces with initially high concentration. In another experiment, the decrease of total blood flow from 2,200 to 1,400
ml/min (by transient IVC occlusion) resulted in a disproportionately greater reduction in regions with a low microsphere content (presumably low flow) (Fig. 4). Of the six
dogs studied, five showed thjs disproportionality in alteration of regional microsphere content when total flow was
changed. The single exception was experiment 23074, in
which total flow was decreased by bleeding. The phlebotomy
resulted in an increase of arterial hematocrit as well as pulmonary artery pressure. In this study, when total flow was
altered the relative microsphere concentration in each region changed proportionally.
SECONDS
FIGURE I A: indicator-dilution curves for **Na (O) and 13 '/albumin ( • ) injected as a bolus into the right atrium and sampled at
the femoral artery. The abscissa is the time from midpoint of the
injection. Extractions, E, ( x ) and E, (A), were calculated as described under Methods. F - blood flow in liters/min: W - the
wet lung weight (excluding large vessels and bronchus) obtained at
the end of the study; BA and PA - mean pressures in brachial and
pulmonary arteries, HR = heart rate. Note the lack of crossing over
of the sodium- and albumin-dilution curves and the decreasing
values of E,(t) and E,(t) with time. B: indicator-dilution curves at a
lower blood flow of I.I liter/min produced by occlusion of the
inferior vena cava during the study. Again, there is a lack of plateau
to E, and E,. The £,(() here were generally higher than E,(t) of
panel A, where the blood flow was greater.
were not used for further illustrations. Figure 2A and B
shows extraction and PS as a function of plasma flow where
the plasma flow was obtained during control or during
transient IVC obstruction. In five of the eight lungs, E
decreased as plasma flow increased (Table 2 and Fig. 2A).
Since the blood flows in these experiments were restricted to
a narrow range by allowing only the estimations made of E
with and without IVC occlusions, and since variations of E
between dogs were large, linear correlation for significance
of the slope between E and plasma flow was not done. The
maximum Ej was 19% and the average Ej at flows of 0.5-1.0
liter/min was 10.5% (n = 16, SD - 3.2%) for an average
plasma flow of 0.75 liter/min.
Figure 2B shows that the calculated PS (ml/min) increased nearly linearly with plasma flow, reaching a maximum PS value of around 150 ml/min. Other interventions
did not produce marked alterations of E or PS except as
they altered flow (Table 2).
REGIONAL DISTRIBUTION OF MICROSPHERES IN THE
LUNGS
The hemodynamic data during the injection of microspheres are given in Table 3. In five of the six dogs flow was
0.2
111
0.1
o
o
,A
1.0
1.S
PLASMA FLOW
(L/mln)
O.S
2.5
1SO
100
a>
• B
OS
to
PLASMA FLOW
(L/mtn)
FIGURE 2 / I : area-averaged extraction (£,) / o r sodium plotted
against plasma flow (liters/min). E, is arbitrarily chosen as the one
that occurs at the time of peak (tp) of the albumin-dilution curve.
There is a trend for E to be less at high flow, but the range of flow
is narrow. B: permeability surface area product (PS), in liters/min,
is plotted against plasma flow (liters/min) and is calculated from
E,(tp) using Equation 3. Note that for plasma flow of about 1 liter/
min, PS ranged from 100 to 150 ml/min. The symbols represent
the experiments shown in Table 2 but only the values for control or
for obstruction of the inferior vena cava are plotted.
CIRCULATION RESEARCH
528
TABLE 2
Fp
A
(liters/
min)
(mm
21084 (A)
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
10094 (•)
24094(0)
16104 ( • )
19124(v)
1976
PA
(mm
Hg)
2.6
0.74
0.74
129
84
80
—
0.70
0.22
0.19
150
100
100
—
+ 1VC
Phleb.
Phleb.
Phlcb.
1.24
0.68
0.39
0.43
0.43
C
+ IVC
+ IVC
Saline
Saline, +IVC
HR
PS
Po,
(mm
Hg)
E,
(*)
E,
(%)
(ml/
min)
WL
(g)
(kg)
WB
H,O
(%)
Hct
0.38
0.38
—
187
180
130
77
75
—
5
9
11
5.0
9.7
10.8
133
76
85
—
36
77
—
0.50
0.50
0.51
200
170
180
82
85
88
10
9
10
12.7
10.4
12.1
96
24
24
—
—
—
113
106
62
70
63
16
14
—
15
14
0.38
0.38
0.44
0.44
0.44
123
114
103
103
103
100
100
80
—
90
9
II
11
12
11
10.8
19.2
16.6
17.3
18.1
142
145
71
82
86
187
22
0.99
0.65
0.74
1.14
0.72
115
105
112
92
90
12
11
14
14
15
0.45
0.45
0.44
0.30
0.32
220
210
210
150
130
51
—
50
64
64
12
13
11
12
14
13.7
14.4
11.7
13.7
17.7
146
101
92
168
140
C
+ IVC
Saline
Saline, +1VC
Saline, +IVC
0.88
0.48
1.12
0.14
0.56
106
106
113
70
92
8
8
10
8
8
0.39
—
0.25
—
0.33
186
190
170
143
135
88
90
70
—
71
10
13
14
22
13
11.4
14.9
17.0
20.7
16.0
107
77
209
32
98
C
+ IVC
Saline 1
Saline,+IVC
Saline 2
Saline, +1 VC
0.83
0.53
0.70
0.49
0.59
0.49
100
100
78
70
70
67
24
24
24
23
23
22
0.43
0.46
0.33
0.29
0.26
0.28
140
130
115
125
110
114
87
80
82
—
92
85
4
11
10
11
14
12
3.9
14.0
10.5
14.0
12.4
15.0
34
80
78
74
78
79
214
22
80
C
+ tubing
Off resp.
On resp.
0.62
0.71
0.71
0.57
146
151
149
142
14
15
15
15
0.43
0.43
0.43
—
120
120
115
140
120
—
96
132
11
11
11
12
13.7
12.7
12.7
16.0
91
96
97
99
198
22
80
C
+ IVC
Hypertension
Hypertension
0.95
0.93
0.39
0.56
118
113
200
207
16
15
15
14
0.51
0.50
0.52
—
160
170
110
120
58
55
60
—
8
10
7
9
12.3
11.2
11.3
14.0
125
111
47
190
27
78
Intervention
C
c
+ IVC
+ IVC
23074 (A)
4, OCTOBER
(beats/
min)
+ IVC
+ IVC
12123 (O)
No.
Extraction and Permeability in Canine Lungs
Expt. no.
and symbol*
11123(»)
VOL. 39,
c
Hg)
—
84
79
81
178
18
81
82
144
21
80
81
82
C - control; +IVC - during IVC obstruction; + tubing - with extra-sampling tube, resp. - respirator, Fp - plasma flow; A and PA - mean systemic and
pulmonary artery pressures. HR - heart rate; Po, - partial pressure of oxygen in arterial blood; WL - lung wet weight; WB - body weight; H,O - percentage
water in the lung obtained by drying; E, and E, are the percent values obtained at tB; PS is defined in the text.
• The symbols denoting experiment are the same as those used in Figure 2. Dash - no measurement made.
Discussion
LUNG EXTRAVASCULAR SPACE
The steady state value for the extravascular space was
0.25-0.47 g/g of wet lung. Previous studies, also at a steady
state, showed an extravascular space of 42-60% for sodium
in human lung,4 30% for sucrose in rabbit lung,4 and 57% for
intra- plus extravascular sodium spaces in dog lung."
Because our technique for estimating capillary permeability
involves only a single transit of the probing molecules, the
apparent extravascular space (sometimes called transient
space) for the exchange may be different from the space
evaluated at steady space.
The previous evidence that suggests a low extravascular
volume for sodium in the lung is derived from indicator-dilution experiments following bolus injections." Mean transit
times are evaluated by using single exponential extrapolation of the tail of the dilution curves. Minimal differences in
the mean transit times of sodium and albumin resulted in
calculated extravascular volumes which have a wide range as
well as low values (from the data of Pearce;* we calculated
these volumes to be 0.001-0.044 ml/g of lung). The
similarity of mean transit times and a consistently positive
extraction of sodium imply (1) dilution curves whose shapes
are similar throughout their major portions, and (2) inability to recover all the injected sodium."- l s We suggest that
extrapolation of the tail of the diffusible indicator-dilution
curve introduces an error because it ignores the late return of
sodium from the tissue, and thus the evidence for a small
VE-N. is derived from an inadequate technique.
It is possible that the v E . N . is not solely interstitial;1"
however, if it is, then the high albumin and the high sodium
LUNG CAPILLARY PERMEABILITY/Yipintsoi
TABLE 3
Parameters Measured for the Regional Lung Blood Flow
HR
(beats/min)
A
(mm Hg)
PA
(mm Hg)
Po,
(mm Hg)
Hct
Blood flow
(liters/min)
"Sc
"Sr
198
170
83
43
—
88
85
0.35
0.29
2.2
0.5
lsoproterenol
IVC balloon
141
Ce
"Sr
175
193
163
140
—
68
64
0.35
0.37
1.8
3.4
lsoproterenol
"Sr
"Sc
125
102
no
68
13
16
100
90
0.33
0.43
1.6
0.8
Phlebotomy
4>
"Sr
Sc
165
160
157
165
50/5
35/3
90
98
0.48
0.46
3.3
2.5
IVC balloon
"Sr
"Sc
135
98
107
83
17
15
111
111
0.44
0.42
2.2
1.4
IVC balloon
"Sc
"Sr
135
130
104
60
12
20
no
0.35
0.33
1.5
3.3
lsoproterenol
Expt no.
Isotope
27024
20034
23074
26114
29114
12124
529
125
Intervention
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Isotopes "Sc, "Sr, and '"Ce are the labels on the 9-/im microspheres. Blood flow is calculated from indocyanine green
dye-dilution curve. The other abbreviations are identified in Table 2. In experiment 26114, the right ventricular pressures
are given instead of the pulmonary artery pressures. The interventions are given between the first and second introduction
of microspheres into the right atrium.
Volumes (2 times that of the heart) fit with the relative
paucity of cellular elements in the structure of the lung.
REGIONAL BLOOD FLOW IN THE LUNG
There are no previous data on regional distribution of
blood flow in discrete areas of the lungs at low and high total
flow. Greenleaf et a l . " measured the distribution of flow in
dogs in the left decubitus position and reported higher
fractional flow per unit of volume in the dependent portion
of the lung. Higher flows also were found at the diaphragmatic portion in comparison to the apex along an isogravitational axis of the thorax. However, in the reports of
Greenleaf et al.1* and Edmunds et al.10 lobar or lobular sizes
were evaluated rather than the 0.5- to 1.5-g pieces reported
here.
The plots of Figures 3 and 4 support the contention that
some lung pieces (e.g., the apex and the nondependent
portion) have so low a flow that if total flow were further
reduced there might be complete cessation of perfusion of
2X1
2.0
CO
Upp«r Mid.
* w t . 14.7 7.S
.7
« Sc.
X Sr. 114 IS
Ant.
Poit.
t
t
\amtl Ace.
rza *7
tu 74
».4 • 1
*
u.
LEFT
RIGHT
Upper Lowtr
ie.i
at
IOO 417
67
4O«
O
•
"
a
•
*:mld-pof tions
O
&
1.0
1.5
2.0
o 1.5
2
O
2.8
LJ 1.0
o
o
o
RELATIVE CONCENTRATION OF Sc
FIGURE 3 Distribution of9-nm microspheres (experiment 12124)
microspheres)
in lung samples at normal flow (Sc - "Sc-labeled
and at high flow induced by intravenous infusion of lsoproterenol
(Sr = "Sr-labeled microspheres). A relative concentration of unity
implies samples that have microsphere concentrations
(counts/min
per g) equal to that of the average for the whole lung. The
percentage distribution for each lung lobe (right upper, middle,
lower, and accessory; left upper and lower) for the microspheres and
for the wet weight are given in the inset. The samples with initially
high microsphere concentration (Sc axis of greater than I _5, where
blood flow during Sc injection was 1J540 ml/min) show disproportionately less increment when total flow increases to 3,300 ml/min
(during Sr injection). The anterior portions of the lungs are denoted
by the open symbols and the posterior or dependent portions of the
lung are denoted by the closed symbols. The asterisks refer to
samples between anterior and posterior.
RIGHT
Upp.Kdj lower
Ant.
Post.
O
.5
1-0
A
A
LEFT
Ace.
Upper Lower
o
•
V
15
o
•
2A
RELATIVE CONCENTRATION OF Sr
FIGURE 4 Distribution of microspheres for experiment 29114. The
"Sr-labeled microspheres were injected at a normal flow of 2J00
ml/min and the "Sc-labeled microspheres were injected at a flow of
1,400 ml I min produced by transient occlusion of the inferior vena
cava. Since not every portion of the lung tissues was sampled, we do
not have the fractional weight or fractional flows of each microsphere. Again, as flow decreases, the samples with initially low flow
(Sr concentration around 0-5) show disproportionately more decrement. For identification of symbols, see inset.
CIRCULATION RESEARCH
530
Ralatlv*
Concentration
of
Ulcroiphim
FIGURE 5 Histogram of tissue weights in an entire lung as a
function of microsphere concentration for experiment 12124 (as of
Fig. 3). At a total flow (F) of IJ40 ml/min (stippled area) 12% of
the lung weight had a relative concenlralion of about O.I (a '/o of
the average for the whole lung). The remainder of the lung had a
relatively even distribution. This evenness of distribution of percentage weights appears more apparent when flow was increased
to 3^300 ml/min by isoproterenol infusion.
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
those areas. If flow is directly related to the surface area
available for blood tissue exchanges, then the S in the PS
term would be less at low total flow than at high total flow,
resulting in some proportionality between PS and flow. In
the present study, neither the pulmonary venous pressure
nor the alveolar pressure was controlled, hence the interaction between these pressures and regional pulmonary blood
flow cannot be determined.
To evaluate the degree of heterogeneity of regional lung
blood flow, a histogram of microsphere distribution vs. the
fractional weights was constructed. In Figure 5, the fractional weights per each unit of microsphere concentration
are distributed uniformly, implying that, unlike the situation in the heart," mean flow is not synonymous with the
mode. This marked heterogeneity of flow becomes important when evaluating the concept of an average PS for the
whole lung.
SODIUM EXTRACTION
The early extractions, defined as E,(t) or E,(t), obtained
within a few samples following the appearance time, were
generally high and declined at various rates independent of
the volume of the sampling system or the speed of sampling.
This could be due to (1) analytical inaccuracy, (2) early back
diffusion of sodium from the extravascular space, (3)
inhomogcneity of flow and permeability, or (4) interlaminar
diffusion in the convective stream.
If analytical inaccuracy affects early E because it involves
taking ratios of small numbers, this should result in E with
random values. Furthermore, the higher subsequent concentration values still result in decremental values for E,(t);
therefore, analytical inaccuracy cannot cause high values of
early E.
Rapid back diffusion of sodium has to be recognized
because it implies a bidirectional and not unidirectional flux.
For back diffusion of sodium to produce high values of early
E, there should be low flow, low vE_N« or high capillary
permeability, either alone or in combination. However, lung
blood flow is high in comparison to the heart, and low vE_N.
is not proven. High permeability is not compatible with peak
VOL. 39, N o . 4, OCTOBER 1976
extraction of less than 20%. For example, in the isolated
heart, where flow is lower than in the lung, a high value of
early E for either sodium or potassium was not a common
feature,"- 2 1 - 2 ! yet the sodium and potassium extractions at
the peak of the albumin curve were 40-60%. In the lung,
tracer water probably shows limitation of flow and therefore
permeability is infinite, but the earliest E for water was
60-70%, decreasing to almost 30% at the time of albumin
peak.1 These values of water extraction exceeded our present
E for Na, which has a lower free diffusion coefficient.
Therefore, rapid back diffusion of sodium could not have
produced high values of early E in the lung.
To implicate heterogeneity of flow or permeability, or
both, as a cause of high values of early E, there must exist
areas in the lung with high PS/F, where F is the plasma flow
into the exchanging sites. However, because of their early
appearance, the probing molecules that traverse the exchanging sites with high P S / F (or low F) also must exit
rapidly from the area; therefore, they also must traverse
nonexchanging vessels with small volumes. In the lung the
anatomical areas that fulfill these implicit criteria would
most likely be those around the large pulmonary vessels.
Interlaminar mixing in the flowing blood was shown by
Lassen and Crone 12 to produce apparent positive extractions
in the brain, which is supposedly impermeable to sodium.
This has been ascribed to large sampling volume or slow
sampling rate, neither of which applies in our study. For
interlaminar diffusion to be detectable, the convective flow
should be relatively slow. It is unlikely that this occurs in the
aorta, since the velocity of blood flow is quite high and, in
the report of Perl et al.,' sampling at the carotid artery did
not appear to alter the appearance of the high initial E.
Hence by exclusion, if interlaminar diffusion is the dominant
cause of high early E, we postulate that it occurs in the
pulmonary venous system where the capacity is high and the
pressure is low.
Because the cause for the high early E is not known, the
use of E(tp) to represent an average E is arbitrary. E, may be
more appropriate because it is less biased toward low
concentration of tracers, although E,(t) often does not have a
plateau value. Ej at the mean transit time of the albumin
curve cannot be used, because by then back diffusion will
have assumed prominence. Since Ea as a function of plasma
flow showed a wide scatter and had values rarely exceeding
15%, we cannot exclude insensitivity of the method as
applied to normal lung. This may negate correct evaluation
of normal PS.
However, if the method is not insensitive, and the choice
of E is appropriate, then the highest value of PS is about 150
ml/min in three dogs. For a lung weight of 180 g, we
calculate the PS to be 0.83 ml/min per g, which is not
markedly different from the average sodium PS of 0.73
ml/min per g in the dog hearts." This similarity of PS
between heart and lung is discussed by Perl et al.'- " In
Weibel's review,24 the pulmonary capillary surface area was
taken to be 72 m2 for a 20- to 24-kg dog, which should have
lung weight of up to 240 g. This would result in capillary
area of 72 x 107240 or 3,000 cm 1 per g, a value 6 times that
of the heart. 14 Hence, in spite of the possible similarity of PS
LUNG CAPILLARY PERMEABILITY/Yipintsoi
between heart and lung, the capillary permeability in the
lung is most likely less than that of the heart.
In conclusion, we have systematically studied the singlepassage extraction technique for sodium in normal lung of
dogs. We show that the sodium distributive volume outside
the exchanging membrane is fairly large. In normal lungs we
suggest that the calculated capillary PS, although giving
results comparable to those of other investigators, may be
incorrect either because of the insensitivity of the method or
because of contamination of the appropriate E by high
values of early E. In evaluating the distribution of regional
lung blood flow with alteration of total flow, we think that as
total flow is reduced, there are lung areas that will not be
perfused from the pulmonary artery. We have also shown
marked heterogeneity of flow such that it may not be
appropriate to demand that the capillaries of the lung be
represented by a single PS value.
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Acknowledgments
Without Ihc technical assistances from the following this work would not
have been possible: Drs. T. Nivatpumin, M. Codini, S Penpargkul and R.
Whitman; J Rosenkrantz, R. Costabile, P. Chandler. M. Torres, and L.
Cioffino. I thank Dr. S Furman for allowing me to use his computer. Marie
Gncco typed Ihe manuscript.
References
1. Robin ED, Cross EC, Zelis R Pulmonary edema N Engl J Med 288:
239-246, 292-304, 1973
2. Chinard FP: The permeability characteristics of the pulmonary blood-gas
barrier. In Advances in Respiratory Physiology, edited by CC Caro.
Baltimore, Williams & Wilkins, 1966, pp 106-147
3. Yudilevich D: Transport of K 4I+ , N a " + and I 1 " " by the pulmonary
circulation. Circ Res 9: 925-935, 1961
4. Bauman A, Rothschild MA, Yallow RS, Berson SA: Pulmonary
circulation and transcapillary exchange of electrolytes. J Appl Physiol
11: 353-361, 1957.
5. Pearce ML: Sodium recovery for normal and edemalous lungs studied by
indicator dilution curves. Circ Res 24: 815-820, 1969
6. Gump FE, Mashima Y, Jorgenscn S, Kinncy JM: Simultaneous use of
three indicators to evaluate pulmonary capillary damage in man. Surgery
70: 262-270, 1971
531
7 Effros RM: Osmotic extraction of hypotonic fluid from the lungs. J Clm
Invest 54: 935-947, 1974
8. Perl W, Silverman F, Delea AC, Chinard FP- Permeability of dog lung
endothelium to sodium, diols amides and water. Am J Physiol 230:
1708-1721. 1976
9. Taylor AE, Gaar KA Estimation of equivalent pore radii of pulmonary
capillary and alveolar membranes. Am J Physiol 218: 1133-1140. 1970
10. Normand 1CS, Olver RE. Reynolds EOR, Strang LB: Permeability of
lung capillaries and alveoli to non-electrolytes in the foetal lamb. J
Physiol (Lond) 219: 303-330, 1971
11. Crone C: Capillary permeability—techniques and problems. In Capillary
Permeability, edited by C Crone, NA Lassen New York. Academic
Press, 1970. pp 15-31
12. Lassen NA, Crone C: The extraction fraction of a capillary bed to
hydropholic molecules; theoretical considerations regarding the single
injection technique with a discussion of the role of diffusion between
laminar streams (Taylor's Effect). In Capillary Permeability edited by C
Crone, NA Lassen New York. Academic Press, 1970, pp 48-59
13. Guller B, Yipintsoi T, Orvis AL, Bassingthwaighte JB Myocardial
sodium extraction at varied coronary flows in the dog; estimation of
capillary permeability by residue and outflow detection Circ Res 37:
359-378, 1975
14. Bassingthwaighte JB: A concurrent flow model for extraction dunng
transcapillary passage Circ Res 35: 483-503, 1974
15 Yipintsoi T, Dobbs WA. Scanlon PD, Knopp TJ. Bassingthwaighte JB:
Regional distribution of diffusible tracers and carbonized microsphcrcs
in the left ventricle of isolated dog hearts. Circ Res 33: 573-587, 1973
16. Fishman AP, Becker EL, Fntts HW, Heinemann HO: Apparent volumes
of distribution of water, electrolytes and hemoglobin within the lung Am
J Physiol 188: 95-98, 1957
17. Chinard FP, Enns T, Nolan MF: Pulmonary extravascular water
volumes from transit time and slope data J Appl Physiol 17: 179-183,
1962
18 Chinard FP Estimation of extravascular lung water by indicator dilution
techniques Circ Res 37: 137-145, 1975
19. Greenleaf JR, Ritman GL, Sass DJ, Wood EH: Spatial distribution of
pulmonary blood flow in dogs in left dccubitus position. Am J Physiol
227: 230-244, 1974
20. Edmunds LG, Gold WM, Heymann MA: Lobar distribution of pulmonary arterial blood flow in awake standing dogs Am J Physiol 219:
1779-1783, 1970
21. Tancrcdi RG, Yipintsoi T, Bassingthwaighte JB: Capillary and cell wall
permeability to potassium in isolated dog hearts. Am J Physiol 229:
537-544, 1975
22. Alvarez DA, Yudilevich DL: Heart capillary permeability to lipid
insoluble molecules J Physiol (Lond) 202: 45-58, 1969
23. Perl W, Chowdhury P, Chinard FP Reflection coefficients of dog lung
endothelium to small hydrophilic solutes. Am J Physiol 228: 797-809,
1975
24. Weibel ER: Morphologic basis of alveolar-capillary gas exchange.
Physiol Rev 53: 419^195, 1973
Single-passage extraction and permeability estimation of sodium in normal dogs lungs.
T Yipintsoi
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Circ Res. 1976;39:523-531
doi: 10.1161/01.RES.39.4.523
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1976 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/39/4/523
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/