Microvascular Research 57, 174 –186 (1999)
Article ID mvre.1998.2127, available online at http://www.idealibrary.com on
Pericardial Fluid Absorption into Lymphatic
Vessels in Sheep
B. Boulanger,* Z. Yuan, M. Flessner,† J. Hay,‡ and M. Johnston
Trauma Research Program and Department of Laboratory Medicine and Pathobiology, Sunnybrook and Women’s
College Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5,
Canada; *Department of Surgery, University of Kentucky, Lexington, Kentucky 40536; †University of
Rochester, Department of Medicine, 601 Elmwood Avenue, Rochester, New York 14642; and ‡Department
of Immunology, University of Toronto, 1 King’s College Circle, Toronto, Ontario M5S 1A8, Canada
Received August 25, 1998
We estimated the volumetric lymphatic clearance rate of
pericardial fluid in sheep. In the first group of studies,
125
I-human serum albumin (HSA) was injected into the
pericardial cavity and after 4 h, various lymph nodes and
tissues were excised and counted for radioactivity. Several lymphatic drainage pathways existed defined by elevated 125I-HSA in the middle and caudal mediastinal, intercostal, and the cardiac nodes located near the root of
the aorta. In a second group of experiments, the plasma
recovery of intrapericardially administered tracer was
compared in sheep with intact lymphatics and in animals
in which thoracic duct lymph was diverted and other
relevant lymphatics ligated. The 4-h plasma recoveries
were reduced significantly from an average of 12.2 6
3.4% injected dose in the lymph-intact group to 3.0 6
1.1% injected dose in the diverted/ligated group (an inhibition of ;75%). In order to estimate the volumetric
clearance of pericardial fluid through lymphatics in conscious sheep, 125I-HSA was administered into the pericardial cavity to serve as the lymph flow marker. 131I-HSA
was injected intravenously to permit calculation of
plasma tracer loss and tracer recirculation into lymphatics. From mass balance equations, total pericardial clearance into lymphatics averaged 1.50 6 0.43 ml/h or ;1.13
ml/h if one was to assume that the average 25% recovered
174
plasma tracer in lymph diverted/ligated animals was due
to nonlymphatic transport. In conclusion, mediastinal
lymphatic pathways remove a volume equivalent to the
pericardial volume (8.1 6 1.1 ml) every 5.4 to
7.2 h. © 1999 Academic Press
Key Words: lymphatics; pericardial cavity; pericardial
fluid; lymphatic clearance; lymph nodes.
INTRODUCTION
The potential importance of lymphatic vessels in
draining the pericardial space has been known for
some time. Indeed, the association between pericardial fluid and lymphatic drainage is evident directly in
chylous pericardium in which thoracic duct obstruction results in reflux of chylous lymph through lymphatics draining the heart and pericardium (Dunn,
1975). To date, the quantitative significance of lymphatic drainage of the pericardial space has remained
elusive and controversial.
In a study on a 14-year-old girl with a massive
pericardial effusion, Stewart et al. (1938) observed that
small molecular weight dyes infused into the pericardial space entered the capillaries easily whereas larger
0026-2862/99 $30.00
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
175
Lymphatic Drainage of Pericardial Fluid
dyes did not. They used the latter observation to conclude that absorption into lymphatics was negligible
or very slow. This conclusion was also reached by
Drinker and Field (1931) in studies on rabbits. Physiological salt solutions were absorbed from the pericardial space rapidly, whereas rabbit or horse serum was
absorbed very slowly. Drinker hypothesized a “low
grade” lymphatic absorption of the pericardial space.
However, other investigators were able to find direct
evidence of lymphatic uptake. Leeds et al. (1977) observed the entry of 131I-albumin into the thoracic and
right lymph duct as early as 15 min after injection into
the pericardial space. Takada et al. (1991) observed
carbon particles in the lumina of lymphatic vessels in
rabbits 60 min after injection into the pericardial cavity. Quantitative studies in the rabbit demonstrated
that 131I-serum albumin injected into the pericardial
cavity entered the plasma rapidly and this entry could
be inhibited significantly with ligation of the right and
thoracic lymph ducts (Gibson and Segal, 1978). Using
the disappearance of a protein tracer from the pericardial cavity as a measure of lymph drainage, Hollenberg and Dougherty (1969) estimated lymph flow
from humans with pericardial effusions. They noted a
range of flows from 46 to 128 ml/day (average 82
ml/day or 3.4 ml/h). However, the use of tracer disappearance from a serous cavity to estimate lymphatic
drainage can be misleading since this method often
results in an overestimation of lymphatic clearance.
Fluid and protein can enter the tissues around the
serous space. While fluid can readily enter capillaries,
the tracer is sieved and is absorbed by the lymphatics
more slowly (Flessner et al., 1983, 1989; Shockley and
Ofsthun, 1992).
The objective of the studies outlined in this report
was to utilize tracer recovery data in plasma and
lymph to make estimates of the volumetric clearance
of pericardial fluid into lymphatics. To achieve this,
mass balances were carried out around the plasma
and the thoracic duct and the resulting series of differential equations was solved simultaneously to provide a method to calculate the rates of fluid transfer
from experimental data. The concentration of an intrapericardially injected protein tracer was measured in
the pericardial fluid, plasma, and thoracic duct lymph
compartments versus time and the thoracic duct flow
rates were recorded. These data were inserted into the
derived equations to provide estimates of the pericardial fluid drainage via lymphatic pathways.
MATERIALS AND METHODS
Surgery
Randomly bred female sheep, ranging in weight
from 22 to 48 kg, were used in this investigation. This
study was approved by the Research Ethics Board of
Sunnybrook Health Science Centre, University of Toronto, and conformed to the guidelines set by the
Canadian Council on Animal Care and the Animals
for Research Act of Ontario. All surgery was performed with fluothane–O 2 anesthesia after induction
with intravenous thiopental sodium (20 mg/kg,
Boehringer Ingelheim, Burlington, Ontario, Canada).
Sheep were intubated endotracheally (HVT cuffed endotracheal tube, 8.0 mm i.d.; Sheridan Catheter Corp.,
Argyle, NY) and ventilated mechanically (Harvard
Apparatus Co., South Natick, MA) during surgery and
during experiments in the anesthetized group.
In all sheep, a catheter (i.d. 1.0 mm, o.d. 1.5 mm;
Critchley, Silverwater, Australia) was inserted into the
pericardial cavity via a left anterior thoracotomy incision and exteriorized for tracer injection and pericardial fluid sampling. The pericardium was sutured
about the catheter to prevent pericatheter fluid leaks.
Also, in all sheep, a left jugular venous line (Cobe 6-in.
pressure monitoring injection line; Lakewood, CO)
was inserted. In sheep that had the thoracic duct
lymph diverted, the thoracic duct was cannulated (i.d.
1.0 mm, o.d. 1.5 mm; Critchley) against the direction of
flow via a left neck incision. The outflow end of the
thoracic duct catheter was exteriorized and positioned
at the level of the right atrium. Thoracic duct lymph
was collected into plastic tubes containing heparin
(;15 U/ml). In those sheep that had lymphatic ligation, all lymph ducts that were visualized in the superior mediastinum were ligated. Sheep in the conscious group were allowed to recover from the
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
176
anesthesia for 3– 4 h and were conscious and standing
for the experiments.
Tracers and Solutions
125
I-Human serum albumin ( 125I-HSA, 0.93 MBq/ml,
10 mg/ml) and 131I-human serum albumin ( 131I-HSA,
37 MBq/ml, 10 mg/ml) were obtained from Frosst
(Kirkland, Quebec, Canada). All tracer solutions were
purified before use by passage through a Centricon
centrifugal concentrator (10,000 MW cutoff) to remove
free 125I or 131I before infusion. To ensure that the
measured radioactivity in any collected sample was
protein associated, a second set of aliquots was assayed after precipitation with 10% trichloroacetic acid.
Free, or nonprotein-associated, 125I or 131I represented
,1% of the total radioactivity in any sample. Lactated
Ringers solution was purchased from Baxter Corporation (Chicago, IL). Radioactivity was determined using a multichannel gamma spectrometer (Compugamma; LKB Wallac, Turku, Finland) with appropriate
window settings and background subtraction.
Anatomical Studies
In nine animals, approximately 25 mCi of 125I-HSA
was injected into the pericardial cavity at the beginning of the experiment and animals were sacrificed 4 h
later. Various lymph nodes and selected other tissues
within the limbs, body cavities, and neck of the animal
were excised, weighed, and counted for radioactivity.
In cases where nodes were bilateral or arranged in
chains, a single point was plotted representing the
mean radioactivity per gram tissue for that group of
nodes. The results of these experiments were calculated as percentages of injected dose per gram of tissue. In some experiments, Evans blue dye was mixed
with sheep protein ex vivo and injected into the pericardial space to outline the relevant lymphatic drainage pathways.
Determination of Lymphatic Contribution to
Pericardial Drainage
In anesthetized animals, we compared the plasma
recovery of intrapericardially injected 125I-HSA under
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
Boulanger et al.
several experimental conditions. For these experiments, the thoracotomy was closed and the animals
were maintained in the right lateral ducubitus position. In one group of sheep (n 5 5), all relevant
lymphatics were left intact. In a second group (n 5
6), the thoracic duct was cannulated and this lymph
diverted from the animal. These animals were also
used to estimate the volumetric clearance of pericardial fluid into lymphatic vessels (method described
later). In a third group (n 5 5), the thoracic duct
lymph was diverted and in addition, the other small
mediastinal lymphatic vessels that had been determined in the anatomical studies to play a role in
pericardial fluid clearance were ligated.
Mathematical Model That Permits Estimation
of Volumetric Pericardial Fluid Clearance
into Lymphatics
The diagram illustrated in Fig. 1 outlines an engineer’s model of the relevant fluid compartments defined by the model. All compartments were assumed
to be well mixed with no spatial variation of tracer
concentration after injection or intercompartmental
transfer. For the mathematical model, it was assumed
that there was no direct transport of tracer into pericardial capillaries and that all tracer was removed
from the pericardial space by lymphatics. All tracer
concentrations (C) were dependent on time (t); however, compartmental volumes were assumed to remain constant. It was assumed that there was no sequestration or destruction of tracer protein in the
thoracic duct lymph. Intercompartmental flow rates
(L, F, and D) were assumed to remain constant during an experiment since the injection of tracer in a
small volume would not be expected to perturb pericardial space pressure.
The pericardial tracer concentration and volume
were designated C per and V per, respectively. From the
pericardial space, there was transfer of the protein via
lymphatics excluding the thoracic duct at a rate D LD
(an unknown). There was also drainage to the thoracic
duct (D TD an unknown). The thoracic duct flow rate
(L TD) and the concentration of tracer in this lymph
(C TD) could be measured directly. The contribution of
177
Lymphatic Drainage of Pericardial Fluid
FIG. 1. Schematic illustrating the essential features of a conceptual model that integrates the pericardial, lymph and plasma compartments.
Abbreviations used: D LD, the rate of transport of pericardial fluid through mediastinal lymphatics (ml/h); D TD, the rate of transport of
pericardial fluid into the thoracic duct (ml/h); C per, concentration of 125I-HSA in pericardial fluid (cpm/ml); C LD, concentration of 125I or 131I-HSA
in mediastinal lymph (cpm/ml); C TD, concentration of 125I or 131I-HSA in thoracic duct lymph (cpm/ml); C P, concentration of 125I or 131I-HSA
in plasma (cpm/ml); L LD, flow rate of mediastinal lymph (ml/h); L TD, observed flow rate of thoracic duct lymph (ml/h); F LD, volumetric transfer
rate of the intrapericardially injected 131I-HSA from the plasma into mediastinal lymph (ml/h); F TD, volumetric transfer rate of the intrapericardially injected 131I-HSA from the plasma into thoracic duct lymph (ml/h); V per, volume of distribution of tracer in pericardial cavity (ml);
V P, volume of distribution of tracer in plasma (ml).
lymphatics other than the thoracic duct (D LD) was
estimated from the 125I-HSA recovered in plasma with
thoracic duct lymph diverted. V P equaled the volume
of distribution of albumin and was estimated by dividing the intravenous dose of 131I-HSA by its C P (t 5
0), which was found by extrapolation of the plasma
concentration to time zero. When 125I-HSA transferred
from the pericardial space to the plasma via the lymphatics, it would simultaneously transfer from the
plasma to other parts of the body. The plasma and
lymph concentrations of the 131I-albumin (injected intravenously) were used to estimate the rate of transfer
of albumin from the plasma to the mediastinal (F LD)
and thoracic duct lymph (F TD). This provided a means
of correcting the 125I-albumin appearance in the lymph
for the transfer of albumin that was carried to the
blood, but recirculated to the tissues which contributed lymph to the right lymph duct and thoracic duct.
Mass Balance Equations
A mass balance around the thoracic duct lymph
compartment assumes that the mass of tracer into the
compartment equals the mass flowing out. All concen-
trations are a function of time (i.e., C (t) ). The mass
balance yields
D TD 5
125
* t0f ~L TDC 125
~t!!dt
TD ~t! 2 F TDC P
* t0f C 125
PER ~t!dt
,
(1)
where C per is the concentration of tracer in pericardial
space. F TD is the volumetric transfer rate of the intrapericardially injected tracer ( 125 I-HSA) from the
plasma. It must be found from a balance around the
thoracic duct for the iv-injected 131I-HSA
F TD 5
* t0f L TDC 131
TD ~t!dt
* t0f C 131
P ~t!dt
.
(2)
If thoracic duct lymph is diverted, then we assume
that any tracer entering the plasma must have transported via the uncannulated mediastinal lymphatics.
We can obtain an estimate of its contribution to pericardial drainage (D LD) by performing a mass balance
around the blood
d~C PV P!
5 L LDC LD 1 L TDC TD
dt
2 K OUT~C PV P! 2 F LDC P 2 F TDC P.
(3)
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
178
Boulanger et al.
From a mass balance around the mediastinal lymphatics
I125
D LDC I125
PER 5 L LDC LD 2 F LDC P.
(4)
I125
@ F LDC P, then
Assuming L LDC LD
D LDC PER 5 L LDC I125
LD
(5)
and substituting into Eq. (3) gives
d~C PV P!
5 D LDC PER 1 L TDC TD
dt
2 K OUT~C PV P! 2 F LDC P 2 F TDC P,
(6)
where all concentrations are functions of time t. Since
thoracic duct lymph flows are diverted, L TD 5 0 in Eq.
(6). Assuming that V P, D LD, F CT, and K OUT are constant,
the above equation takes the form
S
D
dC P~t!
F LD 1 F TD
D LDC PER~t!
C P~t! 5
. (7)
1 K OUT 1
dt
VP
VP
The plasma disappearance curve for 131I-albumin can
be used to determine the overall decay constant for
albumin (K exp) as it filters from the plasma into the
tissues.
S
K exp 5 K out 1
D
F LD 1 F TD
.
VP
(8)
Substituting K exp for the expression in parentheses in
Eq. (7)
dC P~t!
D LDC PER~t!
.
1 K expC P~t! 5
dt
VP
(9)
This can be integrated between t 5 0 and t final (t f):
C P~t f! exp~K exp!t f 2 C P~0!
EF
tf
5
0
G
D LDC PER~t!
exp@~K exp!t#dt.
VP
(10)
Solving this equation for D LD
D LD 5
3
4
C P~t f!exp~K exp!t f 2 C P~0!
.
C PER~t!
tf
*0
exp@~K exp!t#dt
VP
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
The total lymph drainage of the pericardial space is
the sum of D LD and D TD.
Estimates of the volumetric clearance of pericardial
fluid into lymphatics were made in both anesthetized
(n 5 6) and nonanesthetized animals (n 5 5). In the
latter case, the sheep were allowed to recover from the
anesthetic for 3– 4 h and were fully conscious and
standing in the cages.
A 10-ml solution containing 10 mg of 125I-HSA in
0.5% sheep albumin in Ringers was injected into the
pericardial cavity. At the same time, a second tracer
( 131I-HSA) was injected into the venous circulation to
permit (A) calculation of the plasma volume, V P, (B)
determination of a coefficient of elimination from the
plasma, K exp, and (C) estimation of the filtration of
plasma tracer into the thoracic duct, F TD. Radioactivity
was monitored in samples of pericardial space,
plasma, and thoracic duct lymph over 4 h. Pericardial
fluid samples (200 ml) were taken at 2 min for estimating the initial volume in the cavity and then at hourly
intervals. Plasma samples (1.0 ml) were collected every 4 min for 30 min for estimating plasma volume, at
60 min, and then hourly. Lymph from the thoracic
duct was collected continuously. A Macrodex saline
solution (6% Dextran 70, Pharmacia, Quebec) was infused into the jugular venous catheter in volumes
equivalent to those lost from the diverted thoracic
duct.
All data were expressed as the mean 6 SE. In the
graphs, t(0) represents the time that the tracers were
injected into the pericardial space and plasma compartments. The results were analyzed with analysis of
variance. In some cases, the data were assessed with a
paired Student’s t test or a Mann–Whitney rank sum
test as appropriate. We interpreted P , 0.05 as
significant.
RESULTS
Lymphatic Drainage Pathways of the
Pericardial Cavity
(11)
As part of their normal physiological function, lymphatic vessels absorb extravascular protein and, after
Lymphatic Drainage of Pericardial Fluid
179
FIG. 2. Recovery of 125I-HSA in lymph nodes and nonnodal tissues following the injection of the tracer into the pericardial cavity (n 5 5–9).
Results are expressed as percentages of injected dose per gram of tissue. The data were assessed with a Student’s paired t test or a
Mann–Whitney rank sum test as appropriate. * ,# P , 0.05 compared to average cpm per minute in skeletal muscle or prefemoral nodes,
respectively.
passage through various lymph nodes, return it to the
vasculature. Taking advantage of this function, we
injected 125I-HSA into the pericardial cavity to help
elucidate lymphatic pathways that drain pericardial
fluid. Increased radioactivity would indicate the presence of 125I-HSA in transit through the lymphatic channels in the nodes. A sampling of skin, skeletal muscle,
and fat revealed radioactivity between 0.11 and 0.60 3
10 23% injected dose per gram of tissue (Fig. 2). Lymph
nodes that would not be expected to drain the pericardial cavity (superficial nodes, nodes in the abdominal cavity or in the head and neck region) contained
from 0.12 to 0.94 3 10 23% injected dose per gram of
tissue. Of the nodes tested in the thorax, all but the
thymic node had significantly elevated radioactivity
compared to that observed in skeletal muscle which
was used as the control tissue (3.92 to 16.13 3 10 23%
injected dose per gram of tissue). The highest levels of
radioactivity were observed in the cardiac nodes that
were located in the vicinity of the root of the aorta.
One would also expect to see 125I-HSA in the nodal
blood as some of the tracer would have exited the
pericardial cavity by noncannulated lymphatic vessels. However, the observations (1) that most nodes in
the thorax had levels of radioactivity similar to or
greater than plasma and (2) that similarly vascularized
superficial, abdominal, or neck nodes had very low
tracer recovery suggested that the elevated radioactivity measured in the nodes of the thorax indicated
transport of the tracer from the pericardial cavity.
Indeed, statistical comparisons between the radioactivity in the nodes of the thorax (except the thymic
node) and that of the prefemoral node (which is not
expected to drain the pericardial cavity) revealed significant differences (Fig. 2, #).
Studies with Evans blue dye–sheep albumin complex revealed that multiple lymphatic ducts transported pericardial fluid to the plasma. The most intensely stained vessels were small mediastinal vessels
that either emptied into the thoracic duct or appeared
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
180
Boulanger et al.
In one animal, no radioactivity was recovered in
plasma following lymph diversion/ligation.
Estimation of Volumetric Transport of Pericardial
Fluid into Lymphatics
FIG. 3. Cumulative recoveries (% injected dose) of HSA in the
plasma of anesthetized sheep with no lymphatics cannulated (open
circles, n 5 5), in animals with the thoracic duct cannulated and its
lymph diverted from the plasma (closed circles, n 5 6) and in
animals in which the thoracic duct cannulated and its lymph diverted from the plasma and all identifiable mediastinal lymphatics
ligated (open squares, n 5 5). A two-way ANOVA with Greenhouse–Geisser adjusted P values revealed significant differences
between the group with no lymphatics cannulated (open circles)
and the animals in which all lymph was diverted/ligated (open
squares).
to anastomose directly with the veins at the base of the
left neck. The thoracic duct was also stained but less
intensely than the aforementioned vessels. In many of
the animals we were unable to identify the right
lymph duct. However, in the few cases in which a
right lymph duct was identified, we did not observe
any staining.
Effects of Lymph Diversion/Vessel Ligation on the
Transport of a Pericardial Tracer to Plasma
In sheep in which no lymphatics were cannulated or
ligated, the average 4-h plasma recovery of intrapericardially administered tracer was 12.2 6 3.4% injected
dose (n 5 5) (Fig. 3). When thoracic duct lymph was
diverted, plasma recoveries dropped to 6.6 6 1.4%
injected dose at 4 h (n 5 6). When thoracic duct
lymph was diverted and the identifiable mediastinal
ducts were ligated, plasma recoveries were reduced
further to 3.0 6 1.1% injected dose at this time (n 5 5).
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
We were able to estimate the volumetric transport of
pericardial fluid into lymphatics in both an anesthetized (n 5 6) and a conscious group of animals (n 5
5). The calculations were based on tracer recovery
data in pericardial fluid, thoracic duct lymph, and
plasma. Figure 4 illustrates an example of one experiment. Following injection of 125I-HSA into the pericardial cavity, the concentration of this tracer declined in
pericardial fluid and increased over time in the thoracic duct lymph and plasma. The plasma concentration of the intravenously injected 131I-HSA declined
and this tracer was observed in thoracic duct lymph
but not in pericardial fluid. In this example, the directly measured thoracic duct flow rate decreased
slightly. However, on average, the thoracic duct
lymph flow rates did not change significantly over
time in the 11 animals used for the estimates of pericardial lymph drainage rate (Fig. 5).
Tables 1 and 2 illustrate the animal data and estimated flow rates in the anesthetized and conscious
sheep, respectively. The average initial pericardial volume for these two groups of animals was estimated to
be 8.1 6 1.1 ml (n 5 11). In both groups, the pericardial drainage rate into the thoracic duct (0.10 6 0.04
and 0.11 6 0.05 ml/h) was less than that for the
mediastinal vessel transport (0.47 6 0.07 and 1.39 6
0.39 ml/h). Total pericardial drainage into lymphatics
(sum of D LD and D TD) was calculated to be 0.57 6 0.06
ml/h in anesthetized and 1.50 6 0.43 ml/h in conscious sheep.
DISCUSSION
The results from this study argue against the generally held viewpoint that the turnover of pericardial
fluid is low and that pericardial fluid drainage into
lymphatic vessels is of minor significance (Miller,
Lymphatic Drainage of Pericardial Fluid
181
FIG. 4. Example of compartmental HSA concentrations in pericardial fluid (A), thoracic duct (B), and plasma (C) in one of the conscious
animals (sheep No. 2, Table 2). Thoracic duct flow rates measured directly are illustrated in (D). 125I-HSA concentrations are represented by
closed symbols and 131I-HSA concentrations, by open symbols.
1985). In conscious sheep this clearance amounted to
1.5 ml/h or 36 ml/day. Pericardial fluid is produced
as filtrate across the capillaries on the epicardial sur-
face of the heart and possibly from capillaries in the
parietal pericardium. In addition, pericardial fluid
may arise from myocardial interstitial fluid traversing
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
182
Boulanger et al.
FIG. 5. Directly measured thoracic duct flow rates in the anesthetized and nonanesthetized animals (n 5 11). A one-way repeated
measures analysis of variance with Greenhouse–Geisser adjusted P
values revealed no significant changes in flow rates over time.
the epicardium (Stewart et al., 1997). Assuming
steady-state conditions in our experiments, 1.5 ml/h
would also represent the net pericardial fluid formation rate. Considering that the average pericardial volume was 8.1 ml, a volume equivalent to the pericardial
volume was removed by lymphatics every 5.4 h. However, some comment should be made on the assumptions used in the development of our mathematical
approach and on factors that could complicate our
estimates of pericardial fluid transport into lymphatic
vessels.
Potential Clearance of Pericardial Tracer by
Nonlymphatic Mechanisms
Several investigators have suggested that some protein tracers injected into the pericardial space trans-
port directly into capillaries (Szabo and Magyar, 1975;
Leeds et al., 1977). This conclusion is based on studies
in which the lymph from the thoracic and right lymph
duct is diverted from the animals or the vessels ligated. Under these conditions, some tracer still entered the blood. While a very small entry of protein
directly into the capillary circulation can never be
ruled out completely, we believe that it is unlikely that
a significant amount of protein as large as HSA transports directly from the pericardial cavity into blood
without first passing through the lymphatic network.
The junction of the thoracic or right lymph duct with
the venous system can be a complex anatomical structure and it is often difficult to identify and ligate all
ducts. This was certainly the case in the sheep used in
our studies where multiple ducts were observed to
enter the central veins. In addition, other lymph–venous connections may be present (Eliskova et al.,
1995). For example, parasternal lymphatic vessels in
rabbits may enter veins directly without first emptying into the right or thoracic ducts. It is likely that in
cannulating or ligating these vessels, some tracer enters the plasma in unidentified lymphatics in proximity to the dominant duct. For instance, Gibson and
Segal (1978) observed the entry of intrapericardially
injected 131I-serum albumin into the plasma even after
the right lymph duct and thoracic ducts were ligated.
These authors proposed that ligation of the major
lymphatics resulted in a slower drainage through secondary lymphovenous junctions which became patent
as a consequence of increased intralymphatic pres-
TABLE 1
Estimated Lymph Drainage of the Pericardial Cavity in Anesthetized, Thoracic Duct-Cannulated Animals
Sheep
No.
Weight
(kg)
V PER
(ml)
VP
(ml)
K exp
D LD
(ml/h)
D TD
(ml/h)
D LD 1 D TD
(ml/h)
1
2
3
4
5
6
35.0
22.0
37.0
48.0
37.0
39.0
8.8
6.0
11.4
4.9
7.5
4.6
1457
695
1087
1598
1346
1283
0.08
0.06
0.07
0.06
0.04
0.04
0.57
0.45
0.32
0.77
0.42
0.29
0.11
0.10
0.02
0.03
0.09
0.27
0.68
0.55
0.34
0.80
0.51
0.56
Mean
SE
36.3
3.4
7.2
1.1
1244
130
0.06
0.01
0.47
0.07
0.10
0.04
0.57
0.06
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
183
Lymphatic Drainage of Pericardial Fluid
TABLE 2
Estimated Lymph Drainage of the Pericardial Cavity in Conscious, Thoracic Duct-Cannulated Animals
Sheep
(No.)
Weight
(kg)
V PER
(ml)
VP
(ml)
K exp
D LD
(ml/h)
D TD
(ml/h)
D LD 1 D TD
(ml/h)
1
2
3
4
5
40.0
34.0
37.0
35.0
40.0
9.1
12.1
15.1
5.7
4.1
1888
2092
2043
1594
2503
0.05
0.05
0.04
0.01
0.03
0.61
1.81
0.40
2.48
1.63
0.00
0.17
0.01
0.24
0.12
0.61
1.98
0.41
2.72
1.75
Mean
SE
37.2
1.2
9.2
2.0
2024
148
0.04
0.01
1.39
0.39
0.11
0.05
1.50
0.43
sures. Even if other lymph–venous connections existed, this would not be a problem for our model since
these unknown lymphatics would transport the tracer
into the plasma compartment and be accounted for in
the analysis.
A portion of the intrapericardially injected protein
could have initially entered the epicardial tissues of
the heart (Szabo and Magyar, 1975). In support of this,
in three animals we collected cardiac tissues and noted
recoveries between 7.3 and 15.0 3 10 23% injected per
gram of tissue (data not illustrated). Nonetheless, this
tracer would transport ultimately to the cardiac lymphatics. Similarly, protein tracers injected into the
peritoneal cavity enter the tissues surrounding the
serous space including the abdominal wall (Flessner et
al., 1985). By analogy, it seems possible that pericardial
tracers could enter tissues in the chest wall and ultimately be taken up by lymphatics in this location.
It has been suggested that open connections may
exist between the pericardial and pleural spaces in
some species. Following the injection of 131I-serum
albumin into the pericardial space of rabbits, Gibson
and Segal (1978) observed that a significant amount of
tracer passed from the pericardial to pleural cavities.
In rats, golden hamsters, and mice, Nakatani et al.
(1988) demonstrated pores in the pericardial membrane that were up to 50 mm in diameter. There are
three lines of evidence that suggest that significant
pericardial to pleural transport of the tracer did not
occur in our animals. First, in six sheep we lavaged the
pleural cavities and never observed significant radioactivity in the lavaged fluid. Second, in electron microscopic studies performed by our group (data not
illustrated), we could find no evidence that pores existed in the sheep pericardial membrane. The membrane is quite thick (between ;0.3 and 1.0 mm) and
this makes it unlikely that any surface indentation
could extend across the whole membrane and represent a functional pore. Third, in prior experiments, the
infusion of large volumes of isotonic saline into the
pericardial space produced a taut membrane with no
gross evidence of fluid extravasation into the pleural
cavity. Nonetheless, as is the case with tracer entering
the chest wall or epicardial surface of the heart, any
pericardially injected HSA entering the pleural cavity
would be absorbed by lymphatics and returned to the
blood (Courtney Broaddus et al., 1988).
Clearly, the lymphatic drainage pathways of the
pericardial cavity are complex and extremely variable.
Pericardial lymphatic drainage pathways have been
studied systematically in several species including the
dog (Miller et al., 1988), rabbit (Gibson and Segal,
1978), rat (Kluge and Ongre, 1968), and human (Eliskova et al., 1995). Many lymphatic networks appear to
drain pericardial fluid. These include a layer of vessels
in the parietal pericardium and a subpleural network
of ducts. The reflection of the mediastinal pleura onto
the surface of the pericardium contains lymphatics
that connect with pericardial vessels. Connections to
the diaphragmatic ducts are observed on the diaphragmatic surface of the parietal pericardium. Finally, lymphatics from the epicardial surface of the
heart also appear to play a role in pericardial drainage.
Subepicardial lymphatics of the left ventricle are reported to carry tracer to the right lymph duct whereas
vessels of the right ventricle transport tracer to the
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
184
thoracic duct. In dogs, lymphatic drainage of the pericardial space leads ultimately to (A) the principal coronary lymphatic vessel which drains from the left
ventricular muscle and passes to the right upper mediastinum via the cardiac node (to the right lymph
duct), (B) to the lesser coronary lymphatic which
drains the right ventricular muscle and passes to the
left upper mediastinum (to the thoracic duct), and (C)
to bilateral internal mammary (parasternal) ducts—
the right parasternal vessel leads to the right lymph
duct and the left parasternal to the thoracic duct (Miller et al., 1988). In the studies reported here, the majority of pericardial lymph clearance appeared to empty
directly into central veins in the left neck region or into
the thoracic duct. Studies with Evans blue dye–protein
complex demonstrated the occasional small lymphatic
oriented to the right side of the animal to empty
possibly into the right lymph duct. However, in most
animals we were unable to identify the right lymph
duct.
Due to the complexity of the pericardial lymphatic
pathways, it would be very difficult to collect lymph
quantitatively from all of the relevant vessels. For this
reason, we collected only thoracic duct lymph and
assumed that the recovery of the pericardial tracer in
plasma represented lymphatic transport. We demonstrated that the plasma recovery of an intrapericardially administered protein tracer was reduced on average 75% after diversion of thoracic duct lymph and
ligation of all identifiable lymphatics in the neck region. We think it very likely that the remaining 25%
also transported to the plasma via lymphatics. However, if one were to presume that this 25% of injected
tracer was removed by nonlymphatic mechanisms,
then our estimate for pericardial lymph drainage
would be reduced from 1.5 to ;1.13 ml/h and lymphatics would remove a volume equivalent to pericardial fluid volume every 7.2 h.
Mathematical Assumptions
We used tracer recovery data to estimate the volumetric transport of pericardial fluid into lymphatics.
This basic approach has been used previously by the
authors to estimate lymph clearance from the perito-
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
Boulanger et al.
neal cavity (Flessner et al., 1983) and lymph transport
of cerebrospinal fluid from the cranial vault (Boulton
et al., 1998). Essentially, the method calculates the
volume of pericardial fluid (with known tracer concentration) that would have to transport to plasma or
lymph over the appropriate period of time to give the
total mass of tracer measured in the latter two compartments (product of volume and concentration). We
assumed that there was no sieving of tracer at the
interstitial–initial lymphatic interface. This supposition is supported by the ease with which cells enter the
initial lymphatic vessels through the large gaps between the endothelial cells (Gnepp, 1984). Once the
tracer is in the lymphatic vessel, it can be diluted or
concentrated on passage through lymph nodes because of the osmotic gradients present between
plasma and lymph in the nodal tissues (Adair and
Guyton, 1985). This has no impact on our approach
since it is the mass of tracer (product of concentration
times volume of distribution) that is important in the
analysis of the data and this would not change with
the addition or removal of water from the appropriate
compartment.
In deriving the mass balance equations, we assumed
that the volumetric flow rates defined by the letters L,
D, and F (Fig. 1) remained constant for the duration of
the experiment. The observed thoracic duct flow rates
were relatively stable over the 4-h duration of the
experiments (Fig. 5) and we assumed the same would
be true of the uncannulated mediastinal vessels (physiological parameters denoted by L). We wanted a
volume of infusate that would facilitate the sampling
of pericardial fluid over 4 h and yet this volume had to
be small enough to avoid raising pericardial pressure
as this could have an effect on the volume transfer of
pericardial fluid into lymphatics (D). We decided on
an infusate volume of 10 ml which when added to the
estimated residual volume in the cavity (;8 ml)
would give a final pericardial volume of ;18 ml.
Analysis of the pressure–volume relationships in the
sheep pericardium indicated that this volume would
have no significant effect on pericardial fluid pressures (data not illustrated) and we concluded that the
lymph clearance estimates related to resting conditions. The slope of the plasma disappearance curve of
185
Lymphatic Drainage of Pericardial Fluid
intravenously injected 131I-HSA was used to calculate a
coefficient of elimination for labeled HSA and to permit correction for the plasma tracer (and accompanying volume) that refiltered into the lymphatics. The
amount of tracer (and volume) entering the lymphatic
compartment from the plasma would be defined by
the filtration coefficient for each tissue compartment.
No physiological perturbations were attempted in
these experiments and, therefore, the volumetric
transfer of 131I-HSA (F) would not be expected to
change.
There is evidence that venous tracers can enter the
pericardial space (Maurer et al., 1940). Since pericardial fluid is produced as an filtrate across the capillaries on the epicardial surface of the heart and possibly
from capillaries in the parietal pericardium, some
plasma tracer recirculation is to be expected. In the
mathematical approach used in our study, mass balances were performed around the thoracic duct and
plasma but not the pericardial cavity. In the latter case,
we assumed that there would be no significant
amount of tracer recirculation back into the pericardial
space over the 4-h duration of the experiments. This
assumption was supported by the experimental data.
At 4 h, an average of 0.31 6 0.11% injected dose of the
intravenously injected 131I-HSA was recovered in the
pericardial cavity in the 11 anesthetized and conscious
animals from which the data in Tables 1 and 2 were
derived. This result demonstrated that negligible recirculation of the venous tracer into the pericardial
fluid compartment occurred over the course of the
experiment.
Effects of Anesthesia
The estimated pericardial lymph clearance was significantly less in the anesthetized group of animals.
This was to be expected since contractions of lymphatics provide a major source of the energy required to
transport lymph from its collection at the interstitial
level to delivery into the plasma (Johnston, 1995) and
anesthetic agents are known to depress lymphatic contractility (McHale and Thornbury, 1989). In addition,
animal movement and respiratory motion which
could produce passive compression of lymphatic ves-
sel segments would be considerably different in the
two groups.
Summary
The concentration of an intrapericardially injected
HSA in the plasma and thoracic duct lymph compartments was used in conjunction with mass balance
equations to estimate the volumetric pericardial fluid
clearance through lymphatics in sheep. We demonstrated that on average, lymphatics drained pericardial fluid at a rate of 1.5 ml/h in conscious animals
and, therefore, a volume equivalent to the pericardial
volume was cleared by lymphatics every 5.4 h. These
results suggest a relatively rapid turnover of pericardial fluid and point to an important role for lymphatics in regulating pericardial fluid volume.
ACKNOWLEDGMENTS
The authors thank Mr. R. Hancock for technical assistance. This
research was funded by the Heart and Stroke Foundation of Ontario
(NA-3387).
REFERENCES
Adair, T. H., and Guyton, A. C. (1985). Lymph formation and its
modification in the lymphatic system. In “Experimental Biology
of the Lymphatic Circulation” (M. G. Johnston, Ed.), pp. 13– 44.
Elsevier, Amsterdam.
Boulton, M., Flessner, M., Armstrong, D., Hay, J., and Johnston, M.
(1998). Determination of volumetric cerebrospinal fluid absorption into extracranial lymphatics in sheep. Am. J. Physiol. 274,
R88 –R96.
Courtney Broaddus, V., Wiener-Kronish, J. P., Berthiaume, Y., and
Staub, N. C. (1988). Removal of pleural liquid and protein by
lymphatics in awake sheep. J. Appl. Physiol. 64, 384 –390.
Dunn, R. P. (1975). Primary chylopericardium: A review of the
literature and an illustrated case. Am. Heart J. 89, 369 –377.
Drinker, C. K., and Field, M. E. (1931). Absorption from the pericardial cavity. J. Exp. Med. 53, 143–150.
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
186
Eliskova, M., Eliska, O., and Miller, A. J. (1995). The lymphatic
drainage of the parietal pericardium in man. Lymphology 28,
208 –217.
Flessner, M. F., Dedrick, R. L., and Rippe, B. (1989). Letter. ASAIO
Trans. 35, 178 –180.
Flessner, M. F., Fenstermacher, J. D., Blasberg, R. G., and Dedrick,
R. L. (1985). Peritoneal absorption of macromolecules studied
by quantitative autoradiography. Am. J. Physiol. 248, H26 –
H32.
Flessner, M. F., Parker, R. J., and Sieber, S. M. (1983). Peritoneal
lymphatic uptake of fibrinogen and erythrocytes in the rat. Am. J.
Physiol. 244, H89 –H96.
Gibson, A. T., and Segal, M. B. (1978). A study of the routes by
which protein passes from the pericardial cavity to the blood in
rabbits. J. Physiol. 280, 423– 433.
Gnepp, D. R. (1984). Lymphatics. In “Edema” (N. C. Staub and A. E.
Taylor, Eds.), pp. 263–298. Raven Press, New York.
Hollenberg, M., and Dougherty, J. (1969). Lymph flow and 131Ialbumin resorption from pericardial effusions in man. Am. J.
Cardiol. 24, 514 –522.
Johnston, M. G. (1995). Regulation of lymphatic pumping. In “Interstitium, Connective Tissue and Lymphatics” (R. K. Reed, G. A.
Laine, J. L. Bert, P. Winlove, and N. McHale, Eds.), pp. 181–190.
Portland Press.
Kluge, T., and Ongre, A. A. (1968). Pericardial absorption of thorium
dioxide in rats. Acta Pathol. Microbiol. Scand. 72, 87–102.
Leeds, S. E., Uhley, H. N., Meister, R. B., and McCormack, K. R.
(1977). Lymphatic pathways and rate of absorption of 131I-albumin from pericardium of dogs. Lymphology 10, 166 –172.
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
Boulanger et al.
Maurer, F. W., Warren, M. F., and Drinker, C. K. (1940). The composition of mammalian pericardial and peritoneal fluids: Studies
of their protein and chloride contents, and the passage of foreign
substances from the blood stream into these fluids. Am. J. Physiol.
129, 635– 644.
McHale, N. G., and Thornbury, K. D. (1989). The effect of anesthetics
on lymphatic contractility. Microvasc. Res. 37, 70 –76.
Miller, A. J. (1985). The lymph drainage of the heart. In “Experimental Biology of the Lymphatic Circulation” (M. G. Johnston,
Ed.), pp. 231–260. Elsevier, Amsterdam.
Miller, A. J., DeBoer, A., Pick, R., Van Pelt, L., Palmer, A. S., and
Huber, M. P. (1988). The lymphatic drainage of the pericardial
space in the dog. Lymphology 21, 227–233.
Nakatani, T., Shinohara, H., Fukuo, Y., Morisawa, S., and Matsuda,
T. (1988). Pericardium of rodents: Pores connect the pericardial
and pleural cavities. Anat. Rec. 220, 132–137.
Shockley, T. R., and Ofsthun, N. J. (1992). Pathways for fluid loss
from the peritoneal cavity. Blood Purif. 10, 115–121.
Stewart, H. J., Crane, N. F., and Deitrick, J. E. (1938). Absorption
from the pericardial cavity in man. Am. Heart J. 16, 198 –202.
Stewart, R. H., Rohn, D. A., Allen, S. J., and Laine, G. A. (1997). Basic
determinants of epicardial transudation. Am. J. Physiol. 273,
H1408 –H1414.
Szabo, G., and Magyar, Z. (1975). Protein absorption from the pericardial cavity. Res. Exp. Med. 165, 41– 47.
Takada, K., Otsuki, Y., and Magari, S. (1991). Lymphatics and prelymphatics of the rabbit pericardium and epicardium with special emphasis on particulate absorption and milky spot-like structures. Lymphology 24, 116 –124.