Transvascular Fluid Shifts Research Articles

The role of the glycocalyx in transvascular fluid shifts

SUMMARY Perioperative fluid management and fluid shifting have been main topics of debate in the recent years. Liberal fluid application is still a common practice and fluid overload believed to be a minor problem. Recent studies, however, have shown that liberal fluid management facilitates fluid shifts into the interstitial space especially in the context of major abdominal surgery. Consequently, the integrity of the vascular barrier is a rediscovered target of research. The endothelial glycocalyx, which covers the endothelial cells on the luminal side of every healthy vessel, plays a key role concerning the barrier properties of healthy blood vessels. Together with bound plasma proteins the glycocalyx forms the endothelial surface layer with a thickness of about 1 µm. With these packed proteins, this layer has its own colloid osmotic force. Damage of the endothelial glycocalyx can lead to protein extravasation and tissue edema. The endothelial glycocalyx is also important concerning leukocyte and platelet adhesion and shear-stressed nitric oxide-release and therefore, very probably, with regard to the pathophysiology of diseases like sepsis and ischemia/reperfusion injury. The purpose of this review is to elucidate the physiological and functional impact of the endothelial glycocalyx in transvascular fluid shifts.

Micropuncture measurements of interstitial fluid pressure in rat nasal mucosa during early inflammatory reactions.

Interstitial fluid pressure (Pif) has been studied in rat nasal mucosa during early inflammatory reactions induced by dextran anaphylaxis and local application of histamine. Pif was measured by using sharpened micropipettes connected to a servo-controlled counterpressure system. Access to the nasal mucosa was obtained from the facial side of the head through a small cavity drilled in the nasal bone. During dextran anaphylaxis, Pif increased significantly from control values of 2.2 +/- 0.4 to 3.8 +/- 0.21 mmHg (P < 0.05) within 1 h. Corresponding Pif values for histamine were 1.6 +/- 0.9 and 2.9 +/- 0.9 mmHg (P < 0.05), respectively. These measurements support the hypothesis that a major driving force for the rapid exudation across inflamed respiratory mucosa is a hydrostatic pressure gradient created by increased mucosa Pif. When the transvascular fluid shifts accompanying the inflammatory reactions are prevented by circulatory arrest, Pif decreased significantly to subatmospheric values, -0.8 +/- 0.8 and -3.3 +/- 1.2 mmHg in the dextran and histamine group, respectively (P < 0.05). The decrease in Pif in the nasal mucosa after inflammatory stimuli, during circulatory arrest, provides further evidence for "active" modulation of Pif through changes in mechanical properties of the interstitial matrix. The decrease in Pif seen under these circumstances reveals a possible mechanism participating in the rapid and initial edema formation after inflammatory provocations.

Increased negativity of interstitial fluid pressure contributes to development of oedema in rat skin following application of xylene

Intradermal interstitial fluid pressure (Pi) has been studied in rat skin during formation of inflammatory oedema caused by application of xylene. Pi was measured with sharpened micropipettes connected to a servocontrolled counter-pressure system. Control Pi averaged -1.3 +/- 0.6 (SD) mmHg. Following xylene application Pi decreased to -5.0 mmHg after 5 min and then increased to stabilize at about 0 mmHg at 45-60 min and later. When the transvascular fluid shifts accompanying the inflammatory reaction were prevented by inducing circulatory arrest prior to xylene application, Pi fell to -7.5 mmHg within 5 min and remained at this level throughout the observation period of 90 min. Aprotinin in large doses (80,000 KIE kg-1) before xylene application reduced the fall in Pi, whereas indomethacin had no effect. The increased negativity in Pi will add directly to a normal transcapillary net filtration pressure of about 0.5 mmHg, resulting in a 10- to 20-fold increase in this pressure. The present experiments therefore suggest that the interstitium plays an active role in oedema formation in the initial phase of xylene-induced inflammation in rat skin through the development of an increased negativity of Pi.

Early fluid and protein shifts in men during water immersion.

High precision blood and plasma densitometry was used to measure transvascular fluid shifts during water immersion to the neck. Six men (28-49 years) undertook 30 min of standing immersion in water at 35.0 +/- 0.2 degrees C; immersion was preceded by 30 min control standing in air at 28 +/- 1 degrees C. Blood was sampled from an antecubital catheter for determination of blood density (BD), plasma density (PD), haematocrit (Ht), total plasma protein concentration (PPC), and plasma albumin concentration (PAC). Compared to control, significant decreases (p less than 0.01) in all these measures were observed after 20 min immersion. At 30 min, plasma volume had increased by 11.0 +/- 2.8%; the average density of the fluid shifted from extravascular fluid into the vascular compartment was 1006.3 g.l-1; albumin moved with the fluid and its albumin concentration was about one-third of the plasma protein concentration during early immersion. These calculations are based on the assumption that the F-cell ratio remained unchanged. No changes in erythrocyte water content during immersion were found. Thus, immersion-induced haemodilution is probably accompanied by protein (mainly albumin) augmentation which accompanies the intravascular fluid shift.

Continuous Blood Densitometry

To evaluate rapid fluid shifts after graded hemorrhage in splenectomized animals, four pigs and two dogs were bled 15-23 ml/kg body wt in steps of 2.2-6.0 ml/kg. Arterial blood density (BD), mean arterial pressure (MAP), and central venous pressure (CVP) were recorded continuously, and arterial plasma density (PD) and hematocrit (Hct) were determined from blood samples. Erythrocyte density was computed from PD, BD, and Hct. Starting with stable control conditions, MAP, CVP, and BD fell from the beginning of hemorrhage. Each blood withdrawal was followed by an immediate and rapid decrease in BD, even at the lowest (less than 3 ml/kg) initial blood losses. The time course of BD change mirrored that of the volume replacement, with time constants of 3.0-9.6 min and amplitudes depending on the magnitude of the relative volume loss. The PD decrease was significant (P less than 0.01) after 5.4 +/- 0.7 ml/kg hemorrhage. At 15 ml/kg blood loss the mean PD and BD had dropped by 0.99 +/- 0.15 and 2.42 +/- 0.26 g/l, respectively, and Hct had dropped by 2.40 +/- 0.47 units. Calculations suggest that either the inward-shifted fluid has a higher density than normal ultrafiltrate and/or there is a rise of the whole-body-to-large vessel Hct (F cell ratio). The rapid fluid replacement ranged from 5.8 +/- 0.8 to 10.6 +/- 2.0% of the initial plasma volume, or one-fifth to one-third of the lost volume with a 20% hemorrhage. Transvascular fluid shifts can be monitored with continuous high-precision blood densitometry.

Plasma volume changes during head-out water immersion in conscious dogs.

Blood volume (51Cr-erythrocyte dilution method), hematocrit, and arterial (Pa), central venous (Pv), plasma colloid osmotic (IIpl), and interstitial fluid hydrostatic (Pcps, Guyton's capsule method) pressures were measured continuously to determine the kinetics of the transvascular fluid shift during 100 min of water immersion (WI) at 37 degrees C in six splenectomized dogs. Urine flow increased by 180% above control levels (P less than 0.05) by 30 min of WI. Plasma volume (PV) started to increase at 5 min of WI and rose by 7.2% (P less than 0.05) above control levels by 35 min of WI, and then it decreased gradually. PV returned to control levels immediately after WI. Plasma protein concentration and IIpl decreased significantly by 0.2 g/100 ml and 1.2 mmHg, respectively, at 35 min of WI, while plasma osmolality and Na+ concentration were constant. Pa and Pv increased (P less than 0.05) by 25 and 12 mmHg, respectively. Mean capillary pressure, which was calculated from Pa, Pv, and an estimated pre-to-postcapillary resistance ratio of 5-12, increased by 13-14 mmHg while Pcps increased (P less than 0.05) by 17 and 26 mmHg at upper hindlimb and lower forelimb, respectively. The changes in mean capillary pressure and IIpl tend to promote capillary filtration in WI; however, the greater elevation of Pcps more than offsets these forces and leads to a net transvascular shift into the plasma compartment.

Continuous blood densitometry: fluid shifts after graded hemorrhage in animals.

To evaluate rapid fluid shifts after graded hemorrhage in splenectomized animals, four pigs and two dogs were bled 15-23 ml/kg body wt in steps of 2.2-6.0 ml/kg. Arterial blood density (BD), mean arterial pressure (MAP), and central venous pressure (CVP) were recorded continuously, and arterial plasma density (PD) and hematocrit (Hct) were determined from blood samples. Erythrocyte density was computed from PD, BD, and Hct. Starting with stable control conditions, MAP, CVP, and BD fell from the beginning of hemorrhage. Each blood withdrawal was followed by an immediate and rapid decrease in BD, even at the lowest (less than 3 ml/kg) initial blood losses. The time course of BD change mirrored that of the volume replacement, with time constants of 3.0-9.6 min and amplitudes depending on the magnitude of the relative volume loss. The PD decrease was significant (P less than 0.01) after 5.4 +/- 0.7 ml/kg hemorrhage. At 15 ml/kg blood loss the mean PD and BD had dropped by 0.99 +/- 0.15 and 2.42 +/- 0.26 g/l, respectively, and Hct had dropped by 2.40 +/- 0.47 units. Calculations suggest that either the inward-shifted fluid has a higher density than normal ultrafiltrate and/or there is a rise of the whole-body-to-large vessel Hct (F cell ratio). The rapid fluid replacement ranged from 5.8 +/- 0.8 to 10.6 +/- 2.0% of the initial plasma volume, or one-fifth to one-third of the lost volume with a 20% hemorrhage. Transvascular fluid shifts can be monitored with continuous high-precision blood densitometry.

Transvascular fluid shift and thoracic duct lymph: analysis of lymph formation in the rat.

To analyze the effect of changes in interstitial fluid on lymph production, intravascular infusions of saline were given to splenectomized rats under pentobarbital anesthesia at 3 different rates (2, 3, and 4 ml/100 g of body weight) over 10 min. Change of blood volume was continuously monitored and, simultaneously, thoracic duct lymph was collected during and after the infusion. Equilibrium was attained approximately 40 min after the infusion; regardless of the infusion rate, 10% of the infused volume was incorporated into the vascular space and 90% was filtered into the extravascular space. Thus, the amount of transvascular fluid shift showed a linear relationship with the infused volume. However, the drainage from thoracic duct lymph amounted to 5.9%, 11.4%, and 17.8% of the infused saline volume when given at the rate of 2, 3, and 4 ml/100 g, respectively. The relation of lymph flow and infused volume could be regarded as a nonlinear system. By means of a simulation study, this relation was found to be attributed to the nonlinearity of conductance for fluid movement from tissue to lymph duct, which was only one sixth of that determined for the capillary membrane. The drainage of lymph protein following the infusion was only about 50% of that expected from the interstitial protein concentration in even distribution of the infused saline. These characteristics of interstitial fluid space play an important role in absorbing water and, hence, buffering changes in circulating blood volume after volume loading.

Transcapillary protein flux following blood volume modification in dog.

The net fluid and protein movements between intravascular and interstitial space following blood withdrawal and retransfusion of 15% of circulating blood volume were measured in dogs using a continuous monitoring method of circulating blood volume. A significant (p less than 0.01) increase in transvascular fluid shift was observed after the start of blood withdrawal and a new equilibrium state was reached within 15 to 20 min. Associated with the fluid shift, a significant increase in plasma protein of about 1 g was observed. On the other hand, blood retransfusion caused significant (p less than 0.01) increases in the shift of fluid and protein from intravascular space to interstitial space. The magnitudes of the shift of fluid and protein were almost identical in both blood withdrawal and retransfusion. The Kedem-Katchalsky transport equation was applied to the results obtained in the present study in order to analyze the relative role of diffusion and convection and to estimate the reflection coefficient for protein. A significant (p less than 0.01) linear relationship was observed between fluid and protein movement. These results suggest that the convective process is involved in the shift of protein between intravascular and interstitial space observed after blood volume modification.

Tissue pressure and plasma oncotic pressure during exercise.

Six healthy male subjects exercised on a cycle ergometer for 3 min for assessment of forces involved in transvascular fluid shift during intense exercise. The work load was at 105% of peak O2 uptake of the subjects. This caused a 17.2 +/- 1.2% reduction in plasma volume. The plasma volume loss was associated with an increase in plasma sodium, from 142.6 +/- 0.5 to 148.1 +/- 1.0 meq X 1(-1) (P less than 0.005); chloride, from 101.8 +/- 0.6 to 104.6 +/- 0.9 meq X 1(-1) (P less than 0.005); lactate, from 1.4 +/- 0.2 to 14.0 +/- 1.5 meq X 1(-1) (P less than 0.005); and osmolality, from 283 +/- 2 to 299 +/- 3 mosmol X kg-1 H2O (P less than 0.005) within 2 min after cessation of exercise. Plasma protein increased from 7.0 +/- 0.2 to 8.1 +/- 0.3 g X dl-1 (P less than 0.005), and plasma colloid osmotic pressure from 25.1 +/- 0.6 to 30.6 +/- 1.4 mmHg (P less than 0.005) after exercise. Interstitial fluid pressure in the exercising vastus lateralis muscle increased from a base-line value (SE) of -1.0 +/- 0.9 to + 1.5 +/- 1.1 cmH2O, 14 min after the end of exercise (P less than 0.05). Interstitial fluid pressure of the triceps brachii (inactive) did not change significantly after exercise. Our data suggest that increased transvascular colloid osmotic pressure and elevation of interstitial fluid pressure become increasingly important in preventing loss of plasma volume during maximal exercise.

Venous pressures in the alert rat: effects of activity and single injections of catecholamines.

Short bursts of activity are accompanied by a sharp rise, then a sharp drop, followed facultatively by a tonic rise of pressure in the portal and mesenteric vein. Caval vein and abdominal pressure show only a monophasic rise during activity. The drop of portal pressure immediately after the activity peak can be shown to be passive, due to a fall in portal flow rate. The tonic rise after activity can be reproduced by single injections of catecholamines, noradrenalin being more effective than adrenalin. Systemic injections as against intramesenteric injections show, on the other hand, that adrenalin more effectively produces the postactivity pressure drop in the V. portae than noradrenalin. Raising venous pressures in the abdominal cavity by 15-40 mm H2O through acute introduction of air into the abdomen produces no change of venous hematocrit values. The results are discussed in connection with transvascular fluid shifts caused by activity and catecholamines. Portal vein pressure fluctuates with the same frequency, but with a phase displacement of 180 degrees, as Mayer waves.

Canine blood volume and cardiovascular function during hyperthermia.

The effect of acute hyperthermia on hemodynamic functions and blood volume regulation was examined on eight splenectomized dogs. Elevation of core body temperature by 2 degrees C over 90 min caused significant increase in cardiac output (11.2 +/- 12.5 ml X min-1 X kg-1 or about 10%) and significant decrease in total peripheral resistance (TPR; -1.3 +/- 1.0 mmHg X s X ml-1 or about 20%), whereas blood volume (BV), plasma oncotic pressure, and intravascular protein mass remained unchanged. Thus the raised core temperature caused peripheral vasodilation with decreased TPR and compensatory increase in cardiac output. Because BV remained unchanged during warming, mobilization of extravascular fluid did not occur; only the redistribution of blood to the vasodilated cutaneous circulation took place. To assess the effects of heat stress on transvascular fluid equilibrium, Ringer solution (10.7 ml X kg-1 X 10 min-1) was infused under normothermic and hyperthermic conditions. The volume of fluid retained within the intravascular space under equilibrium state was 33.5% in hyperthermia and 9.4% in normothermia. In hyperthermia, the transvascular fluid shift and urinary output were decreased both during and after infusion. The role of preferential fluid retention within the intravascular space observed during hyperthermia was discussed in relation to the mechanism to maintain cardiovascular function and BV under heat stress.

Transvascular fluid shifts accompanying short bodily activity in the alert rat and their relation to arterial pressure.

Hematocrit and arterial blood pressure were measured continuously in alert rats with chronic vascular catheters implanted in the aorta and in various veins. Changes of arterial pressure and hematocrit in the various veins occurred either spontaneously or were induced by mild stimulation of the animal. Results: 1. Slow, spontaneous fluctuations of hematocrit without or with very little activity of the animal are directly related to blood pressure changes, such that a definite hematocrit corresponds to a definite blood pressure. 2. Fast Hct changes instantaneous to 2–20 sec of activity are negative in blood from the liver, the kidney and the gut and positive in blood from the lower extremities. Immediately after this first phase of Hct change, there is a rise of Hct in all veins and the aorta (second phase). This is related to the preceding arterial pressure rise. Intense continued activity is followed in a third phase by an ultimate drop of Hct. 3. Heart rate may either drop or rise on stimulation of the animal, but is invariably elevated during the second phase of Hct response. 4. Evidence is presented that the Hct changes (δHct), being the results of vascular reflexes, are due to appropriate fluid shifts (Δv) in the tissues, which can be calculated by considering time-integrated Hct and flow rates. Thus Δv is, for instance, 0.15 ml from the liver, 0.08 ml from the kidneys, 0.04 ml from the gut, and 0.05 ml into tissues of the lower extremities in a few seconds of activity. The slow spontaneous Hct fluctuations indicate a Δv of only 56μl per mm Hg of arterial pressure change. Blood-volume increase after 3–10 min of activity is by about 5%.

Fluid Exchange in the isolated perfused lung.

The advantages of an isolated, perfused lung preparation in studies of transvascular fluid shifts are described. We find the permeability characteristics of the exchange vessels in the rabbit lung preparation to be similar to those obtained in lungs in situ. The filtration coefficient is similar during plasma and during Krebs-Ringer dextran T 70 perfusion. Therefore proteins in the perfusate do not modify the permeability of the exchange vessels in these lungs as had been reported for other vascular beds. Interstitial fluid pressure in both the alveolar walls and the alveolar corner regions increases with a rise in alveolar pressure. The result is a reduction in the fluid filtration rate under both zone-III and zone-I conditions. The resistance of the extravascular pathway and the fluid pressure in the periarterial/perivenous spaces both appear to be reduced when lung volume is increased by a reduction in pleural pressure. The net effect of positive airway pressure on the transvascular fluid exchange will depend on the relative contribution of the alveolar pressure effects and the lung volume effects. Fluid reabsorption can be observed to occur in the isolated, perfused rabbit lung. This reabsorption does not take place at the level of the loose perivascular tissue spaces since the interstitial fluid pressure at these sites is much lower than the microvascular pressure. We suggest that fluid reabsorption takes place at the thick part of the air-blood barrier, which, most likely, has a higher fluid pressure.

Transvascular fluid shift and redistribution of blood in hypothermia.

The effect of hypothermia on circulating blood volume was studied using 9 dogs subjected to body surface and blood cooling. To analyze the movement of water between the vascular and interstitial fluid spaces, Ringer solution was infused during normothermia and hypothermia (30 degrees C core temperature). During body cooling and infusion experiments, blood volume, hematocrit, and colloid osmotic pressure were monitored continuously. When body temperature was lowered to about 30 degrees C within 2 hr, plasma volume decreased from 39.0 +/- 0.1 to 36.2 +/- 5.3 ml/kg and colloid osmotic pressure increased from 16.4 +/- 0.2 to 17.0 +/- 3.1 mmHg. The changes in blood volume, colloid osmotic pressure, and central venous pressure during the infusion of Ringer solution and recovery periods, were used for a simulation analysis based on a two-compartment model. The instantaneous and delayed compliances were determined on the vascular space and on the interstitial fluid space together with the transvascular filtration coefficient of water. Hypothermia to 30 degrees C increased vascular compliance, but decreased interstitial fluid space. The transvascular filtration coefficient of water decreased in hypothermia. These results suggest that the redistribution of blood due to hypothermia causes a reduction of capillary area available for fluid exchange. The transvascular filtration coefficient determined on the whole body decreases due to reduction of the compliances of the "effective" interstitial fluid space.

Transvascular fluid shift after blood volume modification in relation to compliances of the total vascular bed and interstitial fluid space.

The transvascular fluid shift after blood volume modification was analyzed by means of continuous blood volume determination and stimulation analysis. On splenectomized dogs, 15% of the circulating blood was withdrawn and retransfused, while the circulating blood volume, hematocrit and colloid osmotic pressure were continuously monitored. Based on the results, the transvascular fluid shift was stimulated by a two-compartment model. The compliance of the intravascular and interstitial fluid spaces as well as the transvascular filtration coefficient was determined on the whole body, simultaneously. No statistically significant differences were found between any of the values for blood withdrawal and retransfusion. The compliance of the vascular space, 4.62 +/- 0.82 ml/mmHg . kg of body weight, was almost identical with that of the interstitial fluid space, 4.75 +/- 1.52 ml/mmHg . kg of body weight. The whole body capillary filtration coefficient was 0.74 +/- 0.39 ml/min . mmHg . kg of body weight, which is within the range of the reported values of the capillary filtration coefficient for muscle and intestine. It is suggested that the rapid transvascular fluid shift after blood volume modification is provided with the fact that the compliances of intravascular and interstitial fluid spaces are almost identical.

Determinants of transvascular fluid shifts in zone I lungs.

We studied the fluid shifts in isolated, plasma-perfused rabbit lungs kept completely within zone I. The rate of fluid filtration or reabsorption was determined gravimetrically. A rise in alveolar pressure at constant pleural and vascular pressures reduced th rate of filtration or increased the rate of reabsorption in seven of eight lungs. In seven of seven lungs a reduction in pleural pressure at constant alveolar and vascular pressures increased the rate of filtration or decreased the rate of reabsorption. Thus, a given rise in lung volume had opposite effects depending on whether this rise was caused by an increased alveolar or reduced pleural pressure. Therefore, the exchange vessels studied cannot be true extra-alveolar vessels, which always expand (reflecting a rise in transmural pressure) with a rise in lung volume. When alveolar and pleural pressures were equally increased at constant vascular pressure, the rate of filtration was reduced in four of four lungs. The results can be explained through the existence of exchange vessels situated neither in the alveolar septae proper nor among the true extra-alveolar vessels. The vessels in the alveolar junctions are the most likely candidates.