Quantifying the Effect of Plasma Viscosity on In Vivo Dialyzer Performance

Dialyzer manufacturers characterize performance of their devices based on measurements of clearance using crystalloid solutions. Typically, in vitro dialyzer mass transfer area coefficients for urea (K0A) are substantially higher than values measured in vivo. The reason for this reduction has not been clearly determined. We hypothesized that the known effect of viscosity on reducing solute diffusivity might partially or fully account for this reduction. In vitro dialyzer clearances of urea, glucose, and lactate were measured using crystalloid solutions as well as bovine blood with varying hematocrit and plasma viscosity under a wide range of operating conditions. Viscosities of crystalloid solutions, of blood plasma, and of whole blood were measured at 37°C at a shear rate of 100/s. Diffusivity and relative K0A values (K0Arel) in eight dialyzers were computed for each solute under these different conditions. Plasma was 1.84 times more viscous compared with crystalloid solution (ηrel = 1.84), suggesting a correction multiplier of 1/ηrel = 0.54 for in vivo K0A relative to the in vitro value. Experimental K0Arel at that ηrel was on average reduced to 52% of crystalloid in vitro K0A values. The multiplier 0.52 measured in this study is close to the multiplier 0.55 predicted for average plasma viscosities and also close to the multiplier 0.54 assumed for urea kinetic modeling to provide reasonable urea distribution volumes. The known effect of viscosity on solute diffusivity is therefore sufficient to explain the reduction in dialyzer K0A for urea and glucose in vivo compared with in vitro measurements.