This chapter is related to the aims of SSection H2(i) from the 2023 CICM Primary Syllabus, which expects the exam candidate to "describe the principles of dialysis and filtration". This having never been examined, one might instead say that realistically it is the bedside nurse, junior colleague or supervisor of training that expects the CICM primary exam candidate to be able to describe these principles. The origin of these notes in their earliest form has its source in the need to answer basic questions from other staff, questions which the author was at that stage unable to field.
This chapter explores the differences between two possible directions of blood and dialysate flow across the dialysis circuit. In summary:
- Dialysate can be managed in several ways:
- It remain in a fixed position like a water bath, as in the earliest dialysers
- It can flow concurrently with blood flow (i.e. in the same direction)
- It can run counter-current with the blood flow (i.e. in the opposite direction)
- The counter-current method is the most efficient because it maintains the same concentration gradient along the entire length of the circuit.
- The difference in efficiency (in terms of urea clearance) is approximately 20% in modern filters, when comparing concurrent and countercurrent arrangements.
Consider a hypothetical circuit where blood and dialysate run concurrently.
Blood with a high concentration of solute and dialysate with a zero concentration of solute would enter the circuit at the same level in the filter. Let's say the blood concentration of the solute is 30mmol/L, thinking of urea. Therefore the concentration gradient between the blood and the dialysate would be roughly 30mmol/L.
As both fluids progress further in the filter, some solute would have been exchanged, and half-way though the filter the concentration in the dialysate is 10mmol/L, and 20mmol/L in the blood. Now the concentration gradient is only 10mmol/L. Because the concentration gradient is such an influential factor in determining the solvent diffusive flux, the rate of solute movement into the dialysate is massively affected by this. Let's say that at the end of the circuit the concentration in the blood and in dialysate have equilibrated - then there is no gradient, and therefore no diffusion - the maximum removal of the solute has occurred.
This is the sort of problem which plagued early experimental dialysis designs. For instance, the fragile colloideon dialysis tubes made by Abel, Rowntree and Turner (the first "proper" dialysis circuit) were bathed in a solution of saline which sat still in a cylindrical tank for the duration of the two hour procedure. The objective of this dialysis was not to cleanse the blood but to collect interesting organic molecules in the dialysate for later analysis. The clearance of urea by this method was quite slow; little of it was recovered from the fluid. In short, this would not be a satisfactory method of clearing uraemic toxins from the patients.
The design of dialysis machines was optimised massively with the advent of the Allwall dialyser in the 1960s, which - instead of a stationary lake of fluid or some sort of rotating drug arrangement- circulated the dialysate fluid over the blood compartment. The effects is best illustrated with a similar diagram:
As you can see, the concentration gradient is smaller than with concurrent methods, but it is sustained across the entire filter length.
Does this translate into any sort of non-theoretical benefit? Turns out, yes. A group at the Austin hospital in Melbourne (Baldwin et al, 2016) demonstrated with a modern efficient circuit that countercurrent flow of dialysate increases the clearance of urea and creatinine by 20% when compared to a concurrent circuit. The reason for why the magnitude of this difference is so small (i.e. smaller than in the above diagrams is because the movement of even something like urea is very sluggish, and in concurrent flow the concentration of blood urea and dialysate urea would never equilibrate 50:50, In other words, there is always a driving gradient along the entire circuit, even with concurrent flow.