Cardiovascular Research, 1975, 9, 715-721.
An evaluation of the pericardial sac as a safety factor during tamponade' BARBARA L. PEGRAM
and
VERNON
s.
BISHOP
AUTHORS' S Y N O P S I S Studies were conducted in iitro to determine the permeability characteristics
of the rabbit and dog pericardium. The hydraulic conductance (Lp),which was significantly higher g.cm-2.min-'.cm H,O-'* (mean? than other tissue, was 1.6f0.19~ and 1.78f0.18~ SEM), respectively, for the rabbit and dog pericardium. For various solutes, the permeability coefficient (P) of the pericardium for the rabbit was: 1.40fO.lOx cm-s-' for water, 0.50f0.03x cm.s-' for glucose, and 0.73i-0.005~ cm.s-' for albumin. The reflection coefficient, u, for various solutes was: glucose 8.89f0.86~ sucrose 15.7f 1.5 x dextran (molecular weight 40 OOO) 0.39f 0.03,albumin 0.42i- 0.05, and for haemoglobin 0.58 i- 0.06.These results, which indicate that the pericardium offers little resistance to the bulk transfer of liquids and to the passage of large molecules, are discussed as possible safety factors during pericardial effusion. Obstruction of cardiac venous drainage and lymph flow results in a marked increase in the rate of pericardial effusion (Miller et a/, 1971). Thus, pericardial effusion is apparently similar to fluid accumulation in peripheral tissue since both are dependent upon the net forces across the capillaries and the flow of lymph (Taylor e t a / , 1970;Guyton et a,, 1971). However, in the case of pericardial effusion, the fluid is confined to a limited space by the pericardial sac. Thus, the distensibility of the pericardial sac and its ability to contain fluids and various size molecules may be an important determinant of the consequences of pericardial effusion. Currently, there is no information concerning the sieving characteristics of the pericardial sac to various This study was supported in part by the National Institutes of Health Grant No. 2 ROI HL12415. Air Force Contract No. AFOSR-712074, and the Texas Heart Association (San Antonio Chapter). Reprint requests to V.S.E., Department of Pharmacology. The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas 78284, USA. * The SI units of pressure are kPa. To convert g.cm-2. min-'.cm H20-l to 9 . cm-2. min-'. kPa values should be divided by 0.098.
size molecules or to the influence of pressure on the transfer of fluid across the sac. The present study was conducted to establish the permeability characteristics of the pericardial sac and the alterations during pressure fluctuations. To ascertain the permeability properties of a tissue, one must measure three phenomenological coefficients (Kedem and Katchalsky, 1958). These coefficients, the hydraulic conductance (volume flow per unit of hydrostatic pressure), reflection coefficient (the ratio of the osmotic flow produced by a concentration gradient of a test molecule to that produced by a known pressure head), and the permeability coefficient (a velocity term measured at zero volume flow, indicating the ability of substances to diffuse across the membrane) were calculated in the isolated pericardial preparation.
Methods The pericardia used in this study were obtained from white albino rabbits weighing 1-1.4 kg or from mongrel dogs weighing 11-18 kg. The rabbits
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
From the Department of Pharmacology, The Uniiiersity of Texas Health Science Center at San Antonio, San Antonio, Texas 78284, USA
716 Pegram and Bishop were stunned by a blow on the head, exsanguinated, and the chest opened. The pericardium was removed and placed in physiological Tyrode’s solution. The dogs were anaesthetized with 30 mg/kg sodium pentobarbital, the chest opened, and the pericardium removed and placed in physiological Tyrode’s solution.
L -
Awt
g
- (P) (A) (At) cmHa0.cm2.min1
where AWt=change in weight in g; P=mean hydrostatic pressure in cmHaO; A = surface area of pericardium in cm2; At = change in time in minutes. Reflection coefficient The reflection coefficient is defined as the ratio of the osmotic flow (Lp0) produced by a concentration gradient of a test molecule to that produced by a hydrostatic pressure gradient, ie, hydraulic conductance (L,) (Kedem and Katchalsky, 1958). The osmotic filtration coefficient was determined using a procedure similar to that used for determining the hydraulic filtration coefficient ( Lp). The pericardial sac was tied to the end of the glass capillary tube and filled with physiological Tyrode’s solution. After weighing, the preparation was transferred to a beaker containing a hypertonic solution of the test molecule (glucose, sucrose, dextrose [molecular weight 40 0001 albumin, or haemoglobin). Following a predetermined time, the preparation was removed and weighed to the nearest 0.2 mg. This procedure was repeated four times for each test solute. The volume of the pericardial sac was determined and the surface area calculated as above. The osmotic coefficient (L,,o) was then calculated as follows: AWt/At L,,D = AAn where AWt, At, and A have the same units as above and AT is Van’t Hoff osmotic pressure difference (cmH,O) of the solute across the pericardial sac. The L,,o reflects the ability of the test solute to exert an osmotic pressure on the contents of the pericardial sac. In the case of a semipermeable membrane, molecules which are impermeable would exert a greater osmotic pressure resulting in a loss of fluid from the sac, ie, a large Lpn. If the membrane is freely permeable to the test molecule, LPDwould approach zero since little or no osmotic pressure would be exerted across the sac. The more permeable the membrane, the less L,” will be. From the values obtained for the hydraulic conductance and the osmotic coefficient, u was calculated from their ratios (LpD/Lp).The reflection coefficient is equal to 1 ( o = l ) when the membrane is impermeable to the test molecules and is equal to zero when the test solute and solvent cross the membrane at identical rates. Thus, the hydraulic conductance provides important information con-
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
Hydraulic conductance of the pericardium The filtration coefficients, L,, of rabbit and dog pericardia were determined using a modification of the method described by Diamond (1962). The pericardial sac was tied on the end of a capillary tube 145 mm long (ID, 1.85 mm). One end of the capillary tube was slightly enlarged by a sleeve of plastic tubing so that the pericardial sac could be tied in place and held without fear of slipping when a hydrostatic pressure was applied to its interior. Using a syringe, the pericardium was then filled with physiological Tyrode’s solution previously equilibrated with 95% O2and 5% C02 buffered at pH 7.4. Care was taken to insure that no visible air bubbles were introduced into the system. The preparation was then suspended by a wire hook from a precision balance (Federal Pacific Electric Company) and weighed to the nearest 0.2 mg. The preparation was then transferred to a beaker of Tyrode’s solution located in a water bath which was maintained at 37°C. The hydrostatic pressure was measured to the nearest 0.1 cm referenced to the midline of the pericardial sac. At the end of a predetermined period of time (usually 5-10 min), the hydrostatic pressure was again measured. The preparation was removed from the beaker and reweighed. The amount of surface water remaining on the outside of the pericardium could introduce a major error at this point. A consistent procedure of draining the pericardium against the side of the beaker was adopted (Diamond, 1962). Repeated trials on the same preparation reproduced the weight to within 5 mg. The weighing procedure required approximately 30 s. This entire procedure was conducted a minimum of four times for each preparation. At the end of the experiment, all fluid in the pericardium and capillary tube was removed. The volume of this fluid was then determined to the nearest 0.01 ml using microsyringes. Correcting for the amount of fluid contained in the capillary tube, the volume of the pericardial sac was obtained. The pericardial sac, when filled with fluid, always assumed a spherical shape. Thus, knowing the volume, we could calculate the radius (from the formula: V = + nr3, where V=volume and r = radius of the sphere). The surface area of the preparation was determined from the following formula: A = 4 ma,where A = surface area (Diamond,
1962). The hydraulic conductivity for the pericardium was then calculated as follows:
717 The pericardial sac as a safety factor cerning the passage of fluid across the pericardial sac when pressure is altered while the reflection coefficient provides an index to the sieving properties of the pericardium to various molecules.
where P = permeability coefficient of the pericardium to test solute in cm/s; V1=volume in compartment one in ml; V2=volume in compartment two in ml ; A = surface area of pericardium in cm2; t+= half-time in s of approach to equilibrium. or
p=-
cz vz
Permeability coefficient of the pericardium to various solutes The permeability of the pericardium t o various solutes obtained in these studies is indicated in
A A t C1
where P = permeability coefficient of the pericardium to test solute in cm/s; C1=concentration of test solute in compartment one in CPM/ml (to); Ca = concentration of test solute in compartment two in CPM/ml (tn); Vz=volume in compartment two in ml; A = surface area of pericardium in cm2; t=time interval in s when C2 is determined. The first method is particularly useful for solutes which have a high permeability coefficient, while the second method is more convenient for those solutes with a low permeability coefficient. In a few experiments, a magnetic stirrer was used to insure mixing on each side of the tissue. The permeability coefficients determined with or without the stirrer were similar, thus negating the influences of an unstirred layer near the membrane. Statistical analysis For statistical evaluation of the data, Student's t test was used.
TABLE I
The permeability coefficient P of various solutes for the rabbit pericardium Solure
P x 10-4,
Molecular weight (g .mol- l )
( c m .s-l)
Ha0
Urea Creatine Glucose Albumin:
1.40 2 0.10 I .48 0.09t 89.40 & 8.70 41.802 3.00
18
*
60
0.500+0.003
0.730k0.005
131 I80 60 000
n
12 I2 8 6 8 6
~~
* The
solute was placed on the cardiac surface of the pericardium. N o significant difference was observed in the permeability coefficient when the solute was placed on the pleural surface of the pericardium. t The permeability coefficient determined after the pericardium was subjected to an average hydrostatic pressure of 1.5 or 6.6 cmHzO for one hour. 2 Human serum albumin.
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
Permeability coefficient of the pericardium Using a modification of an Ussing flux chamber with a 95% oxygen, 5% carbon dioxide pump system to circulate the fluid, the permeability coefficient, P, of various size labelled solutes was measured across the pericardium. Permeability coefficients for Ha30, 14C-labelled glucose, and 1251-labelled albumin were determined under equilibrium conditions (ie, no pressure difference) and after the membrane had been subjected to hydrostatic pressure differences of 1.0 to 6.6 cmHzO. Samples (0.1 ml) were obtained periodically from both sides of the preparation and counted using standard liquid scintillation techniques and a Nuclear-Chicago Mark I1 scintillation counter. In all cases, the 95% confidence level was obtained. Permeability coefficient of the test solutes were calculated as follows (Taylor et al, 1965):
Results Hydraulic conductance of the pericardium The hydraulic conductance coefficient, L,, of the rabbit pericardium was 1 . 6 6 k 0 . 1 9 ~ g - ~ m - ~ . m i n - ~ . c m H , (mean? O - ~ SEM; n=7). For dog pericardium, the filtration coefficient was 1.78?0.18x lo-* g ~ ~ r n - ~ ~ r n i n - ~ - c m H ~ O - ' (n = 15). There was no significant difference in the filtration coefficient of the pericardia from these two different species. The filtration coefficient was independent of pressure over the range of 3 to 12 cmH,O. In comparison to other membranes, the L, for the pericardium is 100 t o lo00 times greater (Stein, 1967). The filtration coefficient for many tissues range from to g cm - min - 1. cmH,O- l. For instance, the filtration coefficients for the luminal surface of the rat intestinal mucosa, human erythrocytes, and frog skeletal muscle are 3.5, 5.4, and 5.6 x lo-' g ~ c m - 2 ~ m i n - ' ~ c m H a O - 1respec, tively (Stein, 1967). On the other hand, capillaries (Landis and Pappenheimer, 1963) and connective tissue (Granger and Taylor, 1975) have similar hydraulic filtration coefficients.
Pegram and Bishop TABLE 2
Osmotic coefficient ( L p ~and ) reflection coefficient(a)for rabbit pericardium for various solutes L p px 10-7
Solute
-
(g.ma- 2.miti-
Glucose Sucrose Dextran Albumin' Haemoglobint
l)
1.52 k 0.14 2.60k0.25 641 k52 103 f 90 964f 99
U
Molecular weight
(L p D / L p )
(g.mol-')
8.950.9~1 0 - 4 1.6k0.2~ 3.9k0.3x 10-l 4.2k0.5~lo-' 5.8 k0.6x 10-1
180 342 40 000 45 OOO 68 000
n
8 8
8 6 6
Egg albumin. Bovine haemoglobin.
Hovig, 1967a), rabbit, or dog (Pegram and Bishop, 1969), have shown that the pericardium is predominantly dense connective tissue lined on either side with mesothelial cells. Tight junctions between the cells are common, but direct syncytial continuity has not been observed (Kluge and Hovig, 1967b). Numerous microvilli project from the epicardium as well as the pleural and cardiac aspects of the parietal pericardium. Providing an extensive surface area for the mesothelium, it has been suggested that the pericardium acts as a lubricating surface for the heart (Luisada, 1962). The microvilli may then help to hold a film of lubricating pericardial fluid allowing the heart to glide freely during the cardiac cycle. Cells with an abundant roughsurfaced endoplasmic reticulum have been observed in the epicardium. Kluge and Hovig (1967a) have speculated that these cells may be involved in the production of pericardial fluid components. Additional functions which have Reflection coefficient of the pericardium to been proposed for the pericardium include the facilitation of atrial filling (Holt, 1970), the various solutes Listed in Table 2 are the osmotic coefficients, maintenance of a compensating hydrostatic presL,D, and reflection coefficients, u, for glucose, sure on the outside of the heart (Kenner and sucrose, dextran (molecular weight 40 000) Wood, 1966), and the prevention of cardiac albumin, and haemoglobin obtained in these dilatation (Holt, 1970). The results of our study are concerned with the studies. It is particularly significant that a u equal to one was not obtained. The largest importance of the pericardial sac during abreflection coefficient observed was 0.58 (haemo- normal pericardial effusion. Normally, a small globin), indicating that the pericardium is amount of interstitial fluid leaves the myoreadily permeable to even large molecular cardium and is collected between the parietal and visceral pericardium, the volume regulation weight compounds. depending on normal capillary and lymphatic function in the myocardium and pericardium. Discussion However, excess fluid can collect within this Electron micrographs of the parietal peri- space during experimental cardiac venous and cardium, whether it be human, rat (Kluge and lymphatic obstruction and pathological states Table 1. The permeability of the pericardium to water, 1.40f0.10~ cm.s-', compared favourably with the permeability rates for water of other tissues. Exposure of the tissue to a hydrostatic pressure of either 1.5 cmH,O or 6.6 cmH,O for I h did not damage the preparation, since the permeability to water (1.48 f 0 . 0 9 ~lo-* cm.s-l), when returned to equilibrium, was not significantly different from the permeability to water determined before exposure to a hydrostatic pressure (1.40 f0.10 x cm.s-'). However, during the exposure to a hydrostatic pressure of 6.6 cmH,O, the permeability coefficient was increased two-fold. The permeability coefficients were similar when determined from either the cardiac to pleural surface or from the pleural to cardiac surface. Larger molecular weight compounds such as glucose and albumin had lower permeability rates (Table 1).
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
t
,r m H 2 0 -
719 The pericardial sac as a safety factor large molecular weight molecules such as albumin. Thus, the reflection coefficient (u)was small which clearly indicates that the pericardium poses only a slight barrier for the passage of large molecules (Kedem and Katchalsky, 1958). The ratio of the free diffusion coefficient for albumin (5.9 x cm2*s-') to the measured diffusion coefficient (P=7.29x cmas-') is 0.54 (assuming the thickness of the pericardial membrane to be 150 p). This is additional evidence that there is little restriction by the pericardium to the passage of large molecules. Capillaries, on the other hand, have a high reflection coefficient and effectively transmit osmotic pressure. Previous studies in rabbit (Drinker and Yoffey, 1941) and man (Stewart et al, 1938) have failed to demonstrate significant removal of large molecular weight substances from the pericardial cavity. Recent studies in patients with chronic pericardial effusion have shown that the loss of albumin from the pericardial sac, presumably via the lymphatics, is considerable (Hollenberg and Dougherty, 1969). The above studies were not designed to evaluate the transport of fluids or solutes across the pericardial membrane and, thus, it is difficult to compare their results with the sensitive measurements employed in this study. When considering the passage of molecules across the pericardium, it is important to remember that the structure is composed of several different tissue types. Each structural component may influence the transfer according to its own physical and chemical properties. An important limiting factor for the passage of molecules through this tissue may be the mesothelial cells. Diffusion through connective tissue elements is known to be rapid (Guyton, 1963; Granger and Taylor, 1975) and since the pericardium is composed predominantly of connective tissue elements, little restriction to the passage of molecules would be expected. Therefore, the main determinant for the passage of any substance would be the mesothelial cells with a small contribution from the thickness of the membrane. The vascularity of the pericardium is slight; therefore, the transfer of substances by vascular reabsorption can be ignored. In view of our observations, it seems reasonable to postulate that the pericardium, which
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
such as congestive heart failure or myopericarditis (Miller et al, 1971). This excess formation of pericardial fluid is analogous to oedemaformation in the peripheral tissue (Guyton et al, 1971). However, in this case, the fluid is confined to a finite area; the pericardial sac. With continued excessive pericardial effusion, cardiac performance is, of course, reduced. However, the total volume accumulation may be limited by the hydraulic conductance (Lp), reflection coefficient (u) and permeability coefficient (P). When compared to other tissues (Stein, 1967) (with the exception of capillaries and connective tissue), the pericardium has a large hydraulic conductance, suggesting that with increasing hydrostatic pressure during fluid accumulation, a bulk transfer of fluid would occur across the parietal pericardium and into the pleural space. Although under normal conditions the pressure gradient from the pericardium to pleura may vary with position (Kenner and Wood, 1966), an average pericardial to pleural pressure gradient near 1 mm Hg (0.133 kPa) exists during inspiration and expiration (Morgan et al, 1965; Kenner and Wood, 1966). Using the hydraulic conductance for the dog pericardium and a pericardial surface area of 100 cm2, one can calculate that 1.1 ml.h-'.cmH20-1 of fluid cancross the pericardium. This is approximately one-third the normal cardiac lymph flow. Thus, during acute pericardial eflusion when the pericardial to pleural pressure gradient is high, a large amount of fluid can be lost from the pericardial sac; the order of magnitude being similar to the lymph drainage (Miller et al, 1964). The permeability of the pericardium to water ems-') compares favourably to that (1.4 x reported for other tissue. An earlier report (Takashina et nl, 1962) on the permeability of water ( 3 . 0 6 ~ cmms-l) could not be substantiated in the present study. Only damage to the membrane or a very high artificially-induced pressure gradient could account for this value. In the present study, we could increase the permeability coefficient two-fold by maintaining a pressure gradient of 6.6 cmH,O. However, reducing the gradient to zero returned the permeability coefficient to the normal value (1.4~ cm-s-'). For the pericardium, the volume flow induced by an osmotic pressure gradient was low even for
720 Pegram and Bishop
Jv,
=
J v p + Jvi.
The relationship for the volume transported across the pericardium is:
J,, = L, A(dP,
+ up
n,)
where L, is the hydraulic conductance of the pericardium, A is the area of the pericardium, ,is the reflection coefficient of the pericardium, and up, the colloid osmotic pressure difference across the pericardium and corresponds to about one-half of the colloidal osmotic pressure of plasma (Hollenberg and Dougherty, 1969). Because the reflection coefficient and the effective osmotic term is small, the transport of fluid across the pericardium is primarily dependent upon the hydraulic coefficient and the pressure gradient between the pericardium and pleural, ie, L, A d P p L , A updvp. Thus, the original equation can be approximated by:
J,,
=
L, A P+ Jvl.
Normally, the lymph flow (J,J is two to three times the flow due to hydraulic conductance. However, during pericardial effusion, the volume transported (J,) across the pericardium as a result of hydraulic conductance may approach or exceed the lymph flow (Jvl) since lymph flow, as is the case in peripheral oedema, may reach a maximum before significant collection of fluid is evident (Taylor et a/, 1970, Guyton et a/, 1971). Furthermore, certain types of pericardial effusions may be related to poor lymph drainage. During pericardial effusion associated with increasing volume of fluid within the pericardial sac, the volume of fluid accumulating (Jvs)
exceeds the volume transported across the pericardium (J,,) and by lymph flow (Jvl), ie, Jvs
>
Jvp
+
JVI.
The magnitude of this imbalance will determine the severity of the tamponade and may result from : ( 1 ) large increases in capillary filtration (Jv,); (2) a low lymph flow (Jvl); (3) a decrease in transport of fluid from the pericardium (Jvp); (4) a combination of any of the above. In the above analysis, we have assumed that the pericardium does not secrete or actively transport significant quantities of fluid. Based upon existing data, there is no reason to assume the contrary. The important concept is that regardless of the alterations responsible for pericardial effusion fluid transport across the pericardium can be an important safety factor in the prevention of the accumulation of dangerously large fluid volumes between the heart and pericardium. The authors wish to express their appreciation to Miss Linda Fox and Mr Ben Wiggins for their technical support on this project.
References Diamond, J . M. (1962). The mechanism o f water transport by the gall-bladder. Journal of Physiology, 161, 503-527. Drinker, C. K., and Yoffey, J . M. (1941). Lymphatics, Lymph, and Lymphoid Tissue, p. 105. Harvard University Press, Cambridge, Mass. Granger, H. J., and Taylor, A. E. (1975). Permeability o f connective tissue linings isolated from implanted capsules: Implications for interstitial pressure measurements. Circulation Research, 36, 222-228. Guyton, A. C. (1963). Concept o f negative interstitial pressure based on pressures in implanted perforated capsules. Circulation Research, 12, 399404. Guyton, A . C., Granger, H. J., and Taylor, A. E. (1971). Interstitial fluid pressure. Physiological Reoiews, 51, 527-563, Hollenberg, M., and Dougherty, J. (1969). Lymph flow and ':"I-Albumin resorption from pericardial effusions in man. American Journal of Cardiology, 24, 514-52 I . Holt, J . P. (1970). The normal pericardiurn. American Journal of Cardiology. 26, 4 5 5 4 6 5 . Isaacs, J. P., Berglund, E., and Sarnoff, S. J . (1954). Ventricular function. 111. The pathologic physiology o f acute cardiac tamponade studied by means by ventricular function curves. American Heart Journal, 48, 66-76. Kedern, O., and Katchalsky, A. (1958). A thermodynamic analysis of the permeability o f biological membranes to non-electrolytes. Biochimica el Biophysica Acta, 27, 229246. Kenner, H. M., and Wood, E. H. (1966). lntrapericardial, intrapleural, and intracardiac pressures during acute heart failure in dogs studied without thoracotomy. Circulation Research, 19, 1071-1079.
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
offers little resistance to fluid flow, may provide an additional safety factor during excessive pericardial effusion. This additional pathway for the removal of fluid is important in light of the stiffness of the pericardium (Holt, 1970). Slight increases in pericardial volume can significantly limit diastolic expansion of the heart (Isaacs et a/, 1954). Under equilibrium conditions, the volume of fluid accumulating between the pericardial sac and the myocardium (Jvs), as a result of capillary filtration, must be equal to the sum of the volume of fluid transported by the cardiac and pericardial lymphatics (J,]) and the volume of fluid transported across the pericardial sac (Jvp). This relationship is:
721
The pericardial sac as a safety factor drainage from the heart muscle in the dog. American Journal of Cardiology, 28, 463466.
Morgan, B. C., Guntheroth, W. G., and Dillard, D. H. (1965). Relationship of pericardial to pleural pressure during quiet respiration and cardiac tamponadc. Circulation Research, 16, 493498. Pegram, B. L., a n d Bishop, V. S. (1969). Physical properties of the pericardium. Physiologist, 12, 325. Stein, W. D. (1967). The Movement of Molecules Across the Cell Membranes, pp. 109-1 12. Academic Press, New York. Stewart, H. J., Crane, N. F., and Deitrick, J. E. (1938). Studies of the circulation in pericardial effusion. American Heart Journal, 16, 189-197. Takashina, T., Lazzara, R., Cronvich, J. A., and Burch, G. E. (1962). Transfer of Cd116" across the pericardium of dogs. American Heart Journal, 64, 431432. Taylor, A. E., Gaar, K. A., Jr, and Gibson, H. (1970). Forces affecting lymph flow (Abstr.) Biophysical Journal, 10, 4Sa. Taylor. A. E., Guyton, A. C., and Bishop, V. S. (1965). Permeability o f the aveolar membrane t o solutes. Circulution Research, 16, 353-362.
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
Kluge, T., and Hovig, T. (1967a). The ultrastructure of human and rat pericardium. 1. Parietal and visceral mesothelium. Acta Parhologica et Microbiologica Scandiiiavica, 71, 529-546. Kluge, T., and Hovig, T. (1967b). The ultrastructure of human and rat pericardium. 11. lntercellular spaces and junctions. Acta Pathologicu et Microbiologira Srandinavica, 71, 547-563. Landis, E. M..and Pappenheimer, J. R. (1963). Exchange of substances through the capillary walls. In Handbook of Ph.vsiology, Sect. 2. vol. 2, Circulation, edited by W. F. Hamilton and P. Dow. pp. 961-1034. American Physiological Society, Washington, D.C. Luisada. A. A. (1962). Cardiovascular Functions. McGrawHill, New York. Miller, A. J., Ellis, A., and Katz, L. N. (1964). Cardiac lymph: flow rates and composition in dogs. American Joiirnul of Ph.vsiology, 206, 63-66. Miller, A. J.. Pick, R., and Johnson, P. J. (1971). The production of acute pericardial effusion: the effects of various degrees of interference with venous blood and lymph
An evaluation of the pericardial sac as a safety factor during tamponade' BARBARA L. PEGRAM
and
VERNON
s.
BISHOP
AUTHORS' S Y N O P S I S Studies were conducted in iitro to determine the permeability characteristics
of the rabbit and dog pericardium. The hydraulic conductance (Lp),which was significantly higher g.cm-2.min-'.cm H,O-'* (mean? than other tissue, was 1.6f0.19~ and 1.78f0.18~ SEM), respectively, for the rabbit and dog pericardium. For various solutes, the permeability coefficient (P) of the pericardium for the rabbit was: 1.40fO.lOx cm-s-' for water, 0.50f0.03x cm.s-' for glucose, and 0.73i-0.005~ cm.s-' for albumin. The reflection coefficient, u, for various solutes was: glucose 8.89f0.86~ sucrose 15.7f 1.5 x dextran (molecular weight 40 OOO) 0.39f 0.03,albumin 0.42i- 0.05, and for haemoglobin 0.58 i- 0.06.These results, which indicate that the pericardium offers little resistance to the bulk transfer of liquids and to the passage of large molecules, are discussed as possible safety factors during pericardial effusion. Obstruction of cardiac venous drainage and lymph flow results in a marked increase in the rate of pericardial effusion (Miller et a/, 1971). Thus, pericardial effusion is apparently similar to fluid accumulation in peripheral tissue since both are dependent upon the net forces across the capillaries and the flow of lymph (Taylor e t a / , 1970;Guyton et a,, 1971). However, in the case of pericardial effusion, the fluid is confined to a limited space by the pericardial sac. Thus, the distensibility of the pericardial sac and its ability to contain fluids and various size molecules may be an important determinant of the consequences of pericardial effusion. Currently, there is no information concerning the sieving characteristics of the pericardial sac to various This study was supported in part by the National Institutes of Health Grant No. 2 ROI HL12415. Air Force Contract No. AFOSR-712074, and the Texas Heart Association (San Antonio Chapter). Reprint requests to V.S.E., Department of Pharmacology. The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas 78284, USA. * The SI units of pressure are kPa. To convert g.cm-2. min-'.cm H20-l to 9 . cm-2. min-'. kPa values should be divided by 0.098.
size molecules or to the influence of pressure on the transfer of fluid across the sac. The present study was conducted to establish the permeability characteristics of the pericardial sac and the alterations during pressure fluctuations. To ascertain the permeability properties of a tissue, one must measure three phenomenological coefficients (Kedem and Katchalsky, 1958). These coefficients, the hydraulic conductance (volume flow per unit of hydrostatic pressure), reflection coefficient (the ratio of the osmotic flow produced by a concentration gradient of a test molecule to that produced by a known pressure head), and the permeability coefficient (a velocity term measured at zero volume flow, indicating the ability of substances to diffuse across the membrane) were calculated in the isolated pericardial preparation.
Methods The pericardia used in this study were obtained from white albino rabbits weighing 1-1.4 kg or from mongrel dogs weighing 11-18 kg. The rabbits
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
From the Department of Pharmacology, The Uniiiersity of Texas Health Science Center at San Antonio, San Antonio, Texas 78284, USA
716 Pegram and Bishop were stunned by a blow on the head, exsanguinated, and the chest opened. The pericardium was removed and placed in physiological Tyrode’s solution. The dogs were anaesthetized with 30 mg/kg sodium pentobarbital, the chest opened, and the pericardium removed and placed in physiological Tyrode’s solution.
L -
Awt
g
- (P) (A) (At) cmHa0.cm2.min1
where AWt=change in weight in g; P=mean hydrostatic pressure in cmHaO; A = surface area of pericardium in cm2; At = change in time in minutes. Reflection coefficient The reflection coefficient is defined as the ratio of the osmotic flow (Lp0) produced by a concentration gradient of a test molecule to that produced by a hydrostatic pressure gradient, ie, hydraulic conductance (L,) (Kedem and Katchalsky, 1958). The osmotic filtration coefficient was determined using a procedure similar to that used for determining the hydraulic filtration coefficient ( Lp). The pericardial sac was tied to the end of the glass capillary tube and filled with physiological Tyrode’s solution. After weighing, the preparation was transferred to a beaker containing a hypertonic solution of the test molecule (glucose, sucrose, dextrose [molecular weight 40 0001 albumin, or haemoglobin). Following a predetermined time, the preparation was removed and weighed to the nearest 0.2 mg. This procedure was repeated four times for each test solute. The volume of the pericardial sac was determined and the surface area calculated as above. The osmotic coefficient (L,,o) was then calculated as follows: AWt/At L,,D = AAn where AWt, At, and A have the same units as above and AT is Van’t Hoff osmotic pressure difference (cmH,O) of the solute across the pericardial sac. The L,,o reflects the ability of the test solute to exert an osmotic pressure on the contents of the pericardial sac. In the case of a semipermeable membrane, molecules which are impermeable would exert a greater osmotic pressure resulting in a loss of fluid from the sac, ie, a large Lpn. If the membrane is freely permeable to the test molecule, LPDwould approach zero since little or no osmotic pressure would be exerted across the sac. The more permeable the membrane, the less L,” will be. From the values obtained for the hydraulic conductance and the osmotic coefficient, u was calculated from their ratios (LpD/Lp).The reflection coefficient is equal to 1 ( o = l ) when the membrane is impermeable to the test molecules and is equal to zero when the test solute and solvent cross the membrane at identical rates. Thus, the hydraulic conductance provides important information con-
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
Hydraulic conductance of the pericardium The filtration coefficients, L,, of rabbit and dog pericardia were determined using a modification of the method described by Diamond (1962). The pericardial sac was tied on the end of a capillary tube 145 mm long (ID, 1.85 mm). One end of the capillary tube was slightly enlarged by a sleeve of plastic tubing so that the pericardial sac could be tied in place and held without fear of slipping when a hydrostatic pressure was applied to its interior. Using a syringe, the pericardium was then filled with physiological Tyrode’s solution previously equilibrated with 95% O2and 5% C02 buffered at pH 7.4. Care was taken to insure that no visible air bubbles were introduced into the system. The preparation was then suspended by a wire hook from a precision balance (Federal Pacific Electric Company) and weighed to the nearest 0.2 mg. The preparation was then transferred to a beaker of Tyrode’s solution located in a water bath which was maintained at 37°C. The hydrostatic pressure was measured to the nearest 0.1 cm referenced to the midline of the pericardial sac. At the end of a predetermined period of time (usually 5-10 min), the hydrostatic pressure was again measured. The preparation was removed from the beaker and reweighed. The amount of surface water remaining on the outside of the pericardium could introduce a major error at this point. A consistent procedure of draining the pericardium against the side of the beaker was adopted (Diamond, 1962). Repeated trials on the same preparation reproduced the weight to within 5 mg. The weighing procedure required approximately 30 s. This entire procedure was conducted a minimum of four times for each preparation. At the end of the experiment, all fluid in the pericardium and capillary tube was removed. The volume of this fluid was then determined to the nearest 0.01 ml using microsyringes. Correcting for the amount of fluid contained in the capillary tube, the volume of the pericardial sac was obtained. The pericardial sac, when filled with fluid, always assumed a spherical shape. Thus, knowing the volume, we could calculate the radius (from the formula: V = + nr3, where V=volume and r = radius of the sphere). The surface area of the preparation was determined from the following formula: A = 4 ma,where A = surface area (Diamond,
1962). The hydraulic conductivity for the pericardium was then calculated as follows:
717 The pericardial sac as a safety factor cerning the passage of fluid across the pericardial sac when pressure is altered while the reflection coefficient provides an index to the sieving properties of the pericardium to various molecules.
where P = permeability coefficient of the pericardium to test solute in cm/s; V1=volume in compartment one in ml; V2=volume in compartment two in ml ; A = surface area of pericardium in cm2; t+= half-time in s of approach to equilibrium. or
p=-
cz vz
Permeability coefficient of the pericardium to various solutes The permeability of the pericardium t o various solutes obtained in these studies is indicated in
A A t C1
where P = permeability coefficient of the pericardium to test solute in cm/s; C1=concentration of test solute in compartment one in CPM/ml (to); Ca = concentration of test solute in compartment two in CPM/ml (tn); Vz=volume in compartment two in ml; A = surface area of pericardium in cm2; t=time interval in s when C2 is determined. The first method is particularly useful for solutes which have a high permeability coefficient, while the second method is more convenient for those solutes with a low permeability coefficient. In a few experiments, a magnetic stirrer was used to insure mixing on each side of the tissue. The permeability coefficients determined with or without the stirrer were similar, thus negating the influences of an unstirred layer near the membrane. Statistical analysis For statistical evaluation of the data, Student's t test was used.
TABLE I
The permeability coefficient P of various solutes for the rabbit pericardium Solure
P x 10-4,
Molecular weight (g .mol- l )
( c m .s-l)
Ha0
Urea Creatine Glucose Albumin:
1.40 2 0.10 I .48 0.09t 89.40 & 8.70 41.802 3.00
18
*
60
0.500+0.003
0.730k0.005
131 I80 60 000
n
12 I2 8 6 8 6
~~
* The
solute was placed on the cardiac surface of the pericardium. N o significant difference was observed in the permeability coefficient when the solute was placed on the pleural surface of the pericardium. t The permeability coefficient determined after the pericardium was subjected to an average hydrostatic pressure of 1.5 or 6.6 cmHzO for one hour. 2 Human serum albumin.
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
Permeability coefficient of the pericardium Using a modification of an Ussing flux chamber with a 95% oxygen, 5% carbon dioxide pump system to circulate the fluid, the permeability coefficient, P, of various size labelled solutes was measured across the pericardium. Permeability coefficients for Ha30, 14C-labelled glucose, and 1251-labelled albumin were determined under equilibrium conditions (ie, no pressure difference) and after the membrane had been subjected to hydrostatic pressure differences of 1.0 to 6.6 cmHzO. Samples (0.1 ml) were obtained periodically from both sides of the preparation and counted using standard liquid scintillation techniques and a Nuclear-Chicago Mark I1 scintillation counter. In all cases, the 95% confidence level was obtained. Permeability coefficient of the test solutes were calculated as follows (Taylor et al, 1965):
Results Hydraulic conductance of the pericardium The hydraulic conductance coefficient, L,, of the rabbit pericardium was 1 . 6 6 k 0 . 1 9 ~ g - ~ m - ~ . m i n - ~ . c m H , (mean? O - ~ SEM; n=7). For dog pericardium, the filtration coefficient was 1.78?0.18x lo-* g ~ ~ r n - ~ ~ r n i n - ~ - c m H ~ O - ' (n = 15). There was no significant difference in the filtration coefficient of the pericardia from these two different species. The filtration coefficient was independent of pressure over the range of 3 to 12 cmH,O. In comparison to other membranes, the L, for the pericardium is 100 t o lo00 times greater (Stein, 1967). The filtration coefficient for many tissues range from to g cm - min - 1. cmH,O- l. For instance, the filtration coefficients for the luminal surface of the rat intestinal mucosa, human erythrocytes, and frog skeletal muscle are 3.5, 5.4, and 5.6 x lo-' g ~ c m - 2 ~ m i n - ' ~ c m H a O - 1respec, tively (Stein, 1967). On the other hand, capillaries (Landis and Pappenheimer, 1963) and connective tissue (Granger and Taylor, 1975) have similar hydraulic filtration coefficients.
Pegram and Bishop TABLE 2
Osmotic coefficient ( L p ~and ) reflection coefficient(a)for rabbit pericardium for various solutes L p px 10-7
Solute
-
(g.ma- 2.miti-
Glucose Sucrose Dextran Albumin' Haemoglobint
l)
1.52 k 0.14 2.60k0.25 641 k52 103 f 90 964f 99
U
Molecular weight
(L p D / L p )
(g.mol-')
8.950.9~1 0 - 4 1.6k0.2~ 3.9k0.3x 10-l 4.2k0.5~lo-' 5.8 k0.6x 10-1
180 342 40 000 45 OOO 68 000
n
8 8
8 6 6
Egg albumin. Bovine haemoglobin.
Hovig, 1967a), rabbit, or dog (Pegram and Bishop, 1969), have shown that the pericardium is predominantly dense connective tissue lined on either side with mesothelial cells. Tight junctions between the cells are common, but direct syncytial continuity has not been observed (Kluge and Hovig, 1967b). Numerous microvilli project from the epicardium as well as the pleural and cardiac aspects of the parietal pericardium. Providing an extensive surface area for the mesothelium, it has been suggested that the pericardium acts as a lubricating surface for the heart (Luisada, 1962). The microvilli may then help to hold a film of lubricating pericardial fluid allowing the heart to glide freely during the cardiac cycle. Cells with an abundant roughsurfaced endoplasmic reticulum have been observed in the epicardium. Kluge and Hovig (1967a) have speculated that these cells may be involved in the production of pericardial fluid components. Additional functions which have Reflection coefficient of the pericardium to been proposed for the pericardium include the facilitation of atrial filling (Holt, 1970), the various solutes Listed in Table 2 are the osmotic coefficients, maintenance of a compensating hydrostatic presL,D, and reflection coefficients, u, for glucose, sure on the outside of the heart (Kenner and sucrose, dextran (molecular weight 40 000) Wood, 1966), and the prevention of cardiac albumin, and haemoglobin obtained in these dilatation (Holt, 1970). The results of our study are concerned with the studies. It is particularly significant that a u equal to one was not obtained. The largest importance of the pericardial sac during abreflection coefficient observed was 0.58 (haemo- normal pericardial effusion. Normally, a small globin), indicating that the pericardium is amount of interstitial fluid leaves the myoreadily permeable to even large molecular cardium and is collected between the parietal and visceral pericardium, the volume regulation weight compounds. depending on normal capillary and lymphatic function in the myocardium and pericardium. Discussion However, excess fluid can collect within this Electron micrographs of the parietal peri- space during experimental cardiac venous and cardium, whether it be human, rat (Kluge and lymphatic obstruction and pathological states Table 1. The permeability of the pericardium to water, 1.40f0.10~ cm.s-', compared favourably with the permeability rates for water of other tissues. Exposure of the tissue to a hydrostatic pressure of either 1.5 cmH,O or 6.6 cmH,O for I h did not damage the preparation, since the permeability to water (1.48 f 0 . 0 9 ~lo-* cm.s-l), when returned to equilibrium, was not significantly different from the permeability to water determined before exposure to a hydrostatic pressure (1.40 f0.10 x cm.s-'). However, during the exposure to a hydrostatic pressure of 6.6 cmH,O, the permeability coefficient was increased two-fold. The permeability coefficients were similar when determined from either the cardiac to pleural surface or from the pleural to cardiac surface. Larger molecular weight compounds such as glucose and albumin had lower permeability rates (Table 1).
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
t
,r m H 2 0 -
719 The pericardial sac as a safety factor large molecular weight molecules such as albumin. Thus, the reflection coefficient (u)was small which clearly indicates that the pericardium poses only a slight barrier for the passage of large molecules (Kedem and Katchalsky, 1958). The ratio of the free diffusion coefficient for albumin (5.9 x cm2*s-') to the measured diffusion coefficient (P=7.29x cmas-') is 0.54 (assuming the thickness of the pericardial membrane to be 150 p). This is additional evidence that there is little restriction by the pericardium to the passage of large molecules. Capillaries, on the other hand, have a high reflection coefficient and effectively transmit osmotic pressure. Previous studies in rabbit (Drinker and Yoffey, 1941) and man (Stewart et al, 1938) have failed to demonstrate significant removal of large molecular weight substances from the pericardial cavity. Recent studies in patients with chronic pericardial effusion have shown that the loss of albumin from the pericardial sac, presumably via the lymphatics, is considerable (Hollenberg and Dougherty, 1969). The above studies were not designed to evaluate the transport of fluids or solutes across the pericardial membrane and, thus, it is difficult to compare their results with the sensitive measurements employed in this study. When considering the passage of molecules across the pericardium, it is important to remember that the structure is composed of several different tissue types. Each structural component may influence the transfer according to its own physical and chemical properties. An important limiting factor for the passage of molecules through this tissue may be the mesothelial cells. Diffusion through connective tissue elements is known to be rapid (Guyton, 1963; Granger and Taylor, 1975) and since the pericardium is composed predominantly of connective tissue elements, little restriction to the passage of molecules would be expected. Therefore, the main determinant for the passage of any substance would be the mesothelial cells with a small contribution from the thickness of the membrane. The vascularity of the pericardium is slight; therefore, the transfer of substances by vascular reabsorption can be ignored. In view of our observations, it seems reasonable to postulate that the pericardium, which
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
such as congestive heart failure or myopericarditis (Miller et al, 1971). This excess formation of pericardial fluid is analogous to oedemaformation in the peripheral tissue (Guyton et al, 1971). However, in this case, the fluid is confined to a finite area; the pericardial sac. With continued excessive pericardial effusion, cardiac performance is, of course, reduced. However, the total volume accumulation may be limited by the hydraulic conductance (Lp), reflection coefficient (u) and permeability coefficient (P). When compared to other tissues (Stein, 1967) (with the exception of capillaries and connective tissue), the pericardium has a large hydraulic conductance, suggesting that with increasing hydrostatic pressure during fluid accumulation, a bulk transfer of fluid would occur across the parietal pericardium and into the pleural space. Although under normal conditions the pressure gradient from the pericardium to pleura may vary with position (Kenner and Wood, 1966), an average pericardial to pleural pressure gradient near 1 mm Hg (0.133 kPa) exists during inspiration and expiration (Morgan et al, 1965; Kenner and Wood, 1966). Using the hydraulic conductance for the dog pericardium and a pericardial surface area of 100 cm2, one can calculate that 1.1 ml.h-'.cmH20-1 of fluid cancross the pericardium. This is approximately one-third the normal cardiac lymph flow. Thus, during acute pericardial eflusion when the pericardial to pleural pressure gradient is high, a large amount of fluid can be lost from the pericardial sac; the order of magnitude being similar to the lymph drainage (Miller et al, 1964). The permeability of the pericardium to water ems-') compares favourably to that (1.4 x reported for other tissue. An earlier report (Takashina et nl, 1962) on the permeability of water ( 3 . 0 6 ~ cmms-l) could not be substantiated in the present study. Only damage to the membrane or a very high artificially-induced pressure gradient could account for this value. In the present study, we could increase the permeability coefficient two-fold by maintaining a pressure gradient of 6.6 cmH,O. However, reducing the gradient to zero returned the permeability coefficient to the normal value (1.4~ cm-s-'). For the pericardium, the volume flow induced by an osmotic pressure gradient was low even for
720 Pegram and Bishop
Jv,
=
J v p + Jvi.
The relationship for the volume transported across the pericardium is:
J,, = L, A(dP,
+ up
n,)
where L, is the hydraulic conductance of the pericardium, A is the area of the pericardium, ,is the reflection coefficient of the pericardium, and up, the colloid osmotic pressure difference across the pericardium and corresponds to about one-half of the colloidal osmotic pressure of plasma (Hollenberg and Dougherty, 1969). Because the reflection coefficient and the effective osmotic term is small, the transport of fluid across the pericardium is primarily dependent upon the hydraulic coefficient and the pressure gradient between the pericardium and pleural, ie, L, A d P p L , A updvp. Thus, the original equation can be approximated by:
J,,
=
L, A P+ Jvl.
Normally, the lymph flow (J,J is two to three times the flow due to hydraulic conductance. However, during pericardial effusion, the volume transported (J,) across the pericardium as a result of hydraulic conductance may approach or exceed the lymph flow (Jvl) since lymph flow, as is the case in peripheral oedema, may reach a maximum before significant collection of fluid is evident (Taylor et a/, 1970, Guyton et a/, 1971). Furthermore, certain types of pericardial effusions may be related to poor lymph drainage. During pericardial effusion associated with increasing volume of fluid within the pericardial sac, the volume of fluid accumulating (Jvs)
exceeds the volume transported across the pericardium (J,,) and by lymph flow (Jvl), ie, Jvs
>
Jvp
+
JVI.
The magnitude of this imbalance will determine the severity of the tamponade and may result from : ( 1 ) large increases in capillary filtration (Jv,); (2) a low lymph flow (Jvl); (3) a decrease in transport of fluid from the pericardium (Jvp); (4) a combination of any of the above. In the above analysis, we have assumed that the pericardium does not secrete or actively transport significant quantities of fluid. Based upon existing data, there is no reason to assume the contrary. The important concept is that regardless of the alterations responsible for pericardial effusion fluid transport across the pericardium can be an important safety factor in the prevention of the accumulation of dangerously large fluid volumes between the heart and pericardium. The authors wish to express their appreciation to Miss Linda Fox and Mr Ben Wiggins for their technical support on this project.
References Diamond, J . M. (1962). The mechanism o f water transport by the gall-bladder. Journal of Physiology, 161, 503-527. Drinker, C. K., and Yoffey, J . M. (1941). Lymphatics, Lymph, and Lymphoid Tissue, p. 105. Harvard University Press, Cambridge, Mass. Granger, H. J., and Taylor, A. E. (1975). Permeability o f connective tissue linings isolated from implanted capsules: Implications for interstitial pressure measurements. Circulation Research, 36, 222-228. Guyton, A. C. (1963). Concept o f negative interstitial pressure based on pressures in implanted perforated capsules. Circulation Research, 12, 399404. Guyton, A . C., Granger, H. J., and Taylor, A. E. (1971). Interstitial fluid pressure. Physiological Reoiews, 51, 527-563, Hollenberg, M., and Dougherty, J. (1969). Lymph flow and ':"I-Albumin resorption from pericardial effusions in man. American Journal of Cardiology, 24, 514-52 I . Holt, J . P. (1970). The normal pericardiurn. American Journal of Cardiology. 26, 4 5 5 4 6 5 . Isaacs, J. P., Berglund, E., and Sarnoff, S. J . (1954). Ventricular function. 111. The pathologic physiology o f acute cardiac tamponade studied by means by ventricular function curves. American Heart Journal, 48, 66-76. Kedern, O., and Katchalsky, A. (1958). A thermodynamic analysis of the permeability o f biological membranes to non-electrolytes. Biochimica el Biophysica Acta, 27, 229246. Kenner, H. M., and Wood, E. H. (1966). lntrapericardial, intrapleural, and intracardiac pressures during acute heart failure in dogs studied without thoracotomy. Circulation Research, 19, 1071-1079.
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
offers little resistance to fluid flow, may provide an additional safety factor during excessive pericardial effusion. This additional pathway for the removal of fluid is important in light of the stiffness of the pericardium (Holt, 1970). Slight increases in pericardial volume can significantly limit diastolic expansion of the heart (Isaacs et a/, 1954). Under equilibrium conditions, the volume of fluid accumulating between the pericardial sac and the myocardium (Jvs), as a result of capillary filtration, must be equal to the sum of the volume of fluid transported by the cardiac and pericardial lymphatics (J,]) and the volume of fluid transported across the pericardial sac (Jvp). This relationship is:
721
The pericardial sac as a safety factor drainage from the heart muscle in the dog. American Journal of Cardiology, 28, 463466.
Morgan, B. C., Guntheroth, W. G., and Dillard, D. H. (1965). Relationship of pericardial to pleural pressure during quiet respiration and cardiac tamponadc. Circulation Research, 16, 493498. Pegram, B. L., a n d Bishop, V. S. (1969). Physical properties of the pericardium. Physiologist, 12, 325. Stein, W. D. (1967). The Movement of Molecules Across the Cell Membranes, pp. 109-1 12. Academic Press, New York. Stewart, H. J., Crane, N. F., and Deitrick, J. E. (1938). Studies of the circulation in pericardial effusion. American Heart Journal, 16, 189-197. Takashina, T., Lazzara, R., Cronvich, J. A., and Burch, G. E. (1962). Transfer of Cd116" across the pericardium of dogs. American Heart Journal, 64, 431432. Taylor, A. E., Gaar, K. A., Jr, and Gibson, H. (1970). Forces affecting lymph flow (Abstr.) Biophysical Journal, 10, 4Sa. Taylor. A. E., Guyton, A. C., and Bishop, V. S. (1965). Permeability o f the aveolar membrane t o solutes. Circulution Research, 16, 353-362.
Downloaded from http://cardiovascres.oxfordjournals.org/ at Emory University on April 10, 2016
Kluge, T., and Hovig, T. (1967a). The ultrastructure of human and rat pericardium. 1. Parietal and visceral mesothelium. Acta Parhologica et Microbiologica Scandiiiavica, 71, 529-546. Kluge, T., and Hovig, T. (1967b). The ultrastructure of human and rat pericardium. 11. lntercellular spaces and junctions. Acta Pathologicu et Microbiologira Srandinavica, 71, 547-563. Landis, E. M..and Pappenheimer, J. R. (1963). Exchange of substances through the capillary walls. In Handbook of Ph.vsiology, Sect. 2. vol. 2, Circulation, edited by W. F. Hamilton and P. Dow. pp. 961-1034. American Physiological Society, Washington, D.C. Luisada. A. A. (1962). Cardiovascular Functions. McGrawHill, New York. Miller, A. J., Ellis, A., and Katz, L. N. (1964). Cardiac lymph: flow rates and composition in dogs. American Joiirnul of Ph.vsiology, 206, 63-66. Miller, A. J.. Pick, R., and Johnson, P. J. (1971). The production of acute pericardial effusion: the effects of various degrees of interference with venous blood and lymph
