Dynamics of fluid movement between intravascular and interstitial spaces

Simulation study of the intercompartmental fluid shifts during hemodialysis

Hypotension is the most frequent complication during hemodialysis. An important cause of hypotension is a decrease in the intravascular volume. In addition, a decrease in plasma osmolality may be a contributing factor. Modeling of sodium and ultrafiltration (UF) may help in the understanding of underlying relationships. We therefore simulated, in a mathematical model, the intercompartmental fluid shifts during standard hemodialysis (SHD), diffusive hemodialysis (DHD), and isolated ultrafiltration (IU). We analyzed the relative theoretical effect of hydration status, dialysate sodium concentration, the initial plasma concentrations of sodium and urea, and tissue permeation to solutes on the magnitude and direction of intracellular and intravascular volume changes. This theoretical analysis shows that the transcellular fluid shifts taking place during hemodialysis treatment are, to a great part, due to inhomogeneous distribution of regional blood flow and tissue fluid volumes. During hemodialysis treatment, the cellular fluid shifts in tissue groups with relatively high perfusion and small volume occur from the intra- to the extracellular spaces. However, the fluid shift in tissue groups with a low perfusion and large volume takes place in the opposite direction. The UF volume and rates, and the size of the sodium (Na+) gradient between the dialysate and blood side of the dialyzer membrane are the most important factors influencing the fluid shifts. Higher UF volumes and flow rates cause an increasing decline in the plasma volume in both SHD and IU. High dialysate sodium concentration (150 mEq L(-1)) helps plasma refilling slightly when compared with a normal dialysate sodium concentration (140 mEq L(-1)). However, a high dialysate sodium concentration is associated with a high plasma sodium rebound, which in turn may lead to interdialytic water intake resulting from thirst and may cause increased weight gain and hypertension.

Nature and rate of vascular refilling during hemodialysis and ultrafiltration

The change of blood volume, of blood and plasma density (rho b, rho p) following a short ultrafiltration pulse (duration: 20 min; mean rate -35 ml/min) within the first hour of hemodialysis was analyzed in 13 hemodynamically stable patients (30 single measurements). Protein concentration of refilling volume (7 g/liter) was calculated from its density (1009.25 +/- 3.7 kg/m3, at 20 degrees C) and from the linear relationship between plasma density and protein concentration (cp) of uremic plasma samples (rho p = 1007.46 + 0.2422 x cp). The filtration coefficient (Lp,calc) determined from a relation derived from Starling's hypothesis was 5.6 +/- 1.4 ml/(min.mm Hg.50 kg lean body mass); N = 13, mean +/- SD, minimum 3.2, maximum 8.0. A model describing the dynamics of blood and plasma volume was developed. It was fit to on-line measurements of relative blood volume changes by variation of the filtration coefficient and of initial blood volume (Lp,fit, Vb,fit). The linear regression between Vb,fit and blood volume determined from anthropometry (Vb,calc) was highly significant (r = 0.79, N = 30, P < 0.001). Compared to Vb,calc, Vb,fit was typically increased by 21 +/- 11%, reflecting a fluid overload at the beginning of the treatment. Lp,fit was not different from Lp,calc. Lp,fit significantly increased with blood volume excess. Due to the small but definite protein content of refilling volume, the model accounts for increased blood volume recovery and occasional overshoot of blood and plasma volumes following ultrafiltration.

The nonlinear characteristics of soft gels and hydrated connective tissues in ultrafiltration

A one-dimensional ultrafiltration problem of fluid flow through a soft permeable tissue or gel under high pressure and compressive strain is solved. A finite deformation biphasic theory is used to model the behavior of the soft porous permeable solid matrix. This theory includes a Helmholtz free energy function which depends on the three principal invariants (I, II, III) of the right Cauchy-Green tensor and which satisfies the Baker-Ericksen inequalities on the principal stresses and strains. The dependence of the porosity phi f and the solidity phi s on deformation is deduced and a generalization of the exponential strain-dependent functional form for the permeability, k = k0 exp (M epsilon), of Lai and Mow (Biorheology 103, 111-123, 1980) is proposed. In this one-dimensional problem, we show that the dependence of the permeability on phi f, phi s, and III is equivalent to its dependence on hydration as proposed by Fatt and Goldstick (J. Colloid Sci. 20, 962-988, 1965). The exact solution of the ultrafiltration problem is derived and asymptotic and numerical methods are used to evaluate it. For high pressures and finite strains, the solution provides some surprising effects. The theory predicts that a material starting with a homogeneous porosity will have a strongly non-homogeneous porosity throughout the column during ultrafiltration. The resulting change in pore size through the filtration column may be very important in understanding its filtration characteristics. It is also found that there is a long delay time, up to 10 to 15 min, before the filtration velocity reaches an equilibrium. In filtration experiments where the rate of mass transport across the tissue or column of gel is important, sufficient time must be allowed for the steady state to be reached.

Changes in blood volume and vascular compliance during body heating in rats

The effects of hyperthermia on blood volume and effective vascular compliance were studied in control and heat acclimated rats (three weeks at 32 degrees C and 50% R.H.). Experiments were performed on conscious rats whose abdominal aorta and both jugular veins were cannulated. Continuous changes in blood volume (BV) were monitored by measuring 51Cr tagged erythrocyte dilution, using an arterio-venous extracorporeal shunt passing through a gamma counter. Total vascular compliance was calculated from the relation between changes in BV and central venous pressure during 10 min of infusion of saline at a rate of 1.6% body wt/10 min. Hyperthermia induced a significant blood volume expansion. This expansion was more pronounced in non-acclimated rats. Effective vascular compliance was similar in the normothermic, both non-acclimated and acclimated rats. However, while hyperthermia did not affect the vascular compliance of non-acclimated rats, it was decreased significantly in the acclimated hyperthermic rats. The data suggest that changes in vascular compliance play a role when rapid blood volume changes take place, especially in acclimated hyperthermic animals. The relations between changes in vascular compliance and heat induced redistribution of cardiac output are discussed.

Vascular responses to hypercapnia in anesthetized dogs

To evaluate the vascular responses to systemic acute mild hypercapnia (Pa(CO)(2) = 65 mmHg), we determined the vascular compliance with the relation between the change in circulating blood volume and the change in central venous pressure during and after fluid infusion in dogs anesthetized with halothane in normocapnia and hypercapnia. Circulating blood volume was measured continuously by (51)Cr-labeled erythrocyte dilution method together with hemodynamic variables. Small reduction in vascular compliance (8.1 +/- 1.0 ml.mmHg(-1).kg(-1) in normocapnia, 5.8 +/- 0.5 ml.mmHg(-1).kg(-1) in hypercapnia), large reduction in delayed compliance, which were quantitated by computer simulation using Maxwell's viscoelastic model, and significant increase in blood volume in central circulation were observed in hypercapnia. The essential change in hypercapnia was concluded as the vasoconstriction in capacitance vessels. Simultaneously, the reduction of total peripheral resistance (1.09 +/- 0.08 mmHg.min.kg.ml(-1) in normocapnia, 0.98 +/- 0.07 mmHg.min.kg.ml(-1) in hypercapnia) with no change in transvascular filtration coefficient (0.14 +/- 0.02 ml.mmHg(-1).min(-1).kg(-1)) suggests the increase in shunt flow in peripheral circulation.