The Impact of CRRT Modality on Filter Life

Continuous renal replacement therapy (CRRT) is a dialysis modality used in critically ill intensive care unit (ICU) patients because it is associated with less hemodynamic instability than traditional hemodialysis. CRRT can be conducted using one of two different modalities, which are continuous venovenous hemofiltration (CVVH) and continuous venovenous hemodialysis (CVVHD). CVVH used conductive clearance, whereas CVVHD uses diffusive clearance (CVVHD).

In CVVH, plasma water is removed across the filter using transmembrane pressure which forces fluid into the effluent space. Solutes are removed via "solvent drag", which is the process by which solutes follow fluid flow. The rate of fluid removal is high (usually around 2-3 L/hr), and so hemodynamic collapse would quickly occur in the fluid were not replaced. The patient is therefore given "replacement fluid" which can be given pre-filter or post-filter, and usually nearly matches the fluid removal rate. For instance, if 2 L are removed per hour, 2 L are given back in the form of replacement fluid, if volume neutrality is desired. If fluid removal is desired, 1.9 L (e.g) may be given back, resulting in the net loss of 0.1 L of fluid per hour. The composition of the replacement fluid determines the serum concentration. As an example, if 2 L of fluid with a potassium concentration of 6.0 mmol/L (a high amount, 4.0 is normal) is removed and replaced with 2 L of replacement fluid with a potassium concentration of 2 mmol/L, then 8 mmol of potassium will be "removed" per hour, and the potassium concentration of the serum will fall.

In CVVHD, dialysate flows in the effluent space, creating a gradient which leads to movement of solutes either into or out of the blood, depending on the gradient direction. For instance, if serum has a potassium concentration of 6 mmol/L, and the dialysate has a concentration of 2 mmol/L, potassium will move from the blood into the dialysate, until the dialysate concentration is also equal to 6 mmol/L. At that point, net potassium movement will be 0, as the concentration will be equal. As with CVVH, with a dialysate flow rate of 2 L/hr will result in the removal of 8 mmol of potassium per hour. While these technical details concern only the specialist, the important take away is that both of these modalities achieve equivalent clearance, via different mechanisms. Notably, there are no differences in clinical outcomes with use of one or the other modality, and these modalities are currently thought to have complete equipoise. The decision is generally based on provider comfort and preference. There is no clinical scenario where one modality would be preferable to the other, based on current evidence.

A major issue when using CRRT is filter clotting, which occurs on average every 30 hours. Clotting renders the circuit unusable, and requires that the machine be reset, which is a process that takes 1-2 hours on average, and potentially longer depending on the availability of the dialysis nurses who change the circuits. This results in less delivery of dialysis than what is prescribed, and clotting is often also associated with the loss of the blood that is clotted in the circuit. This could be up to 150 ccs, which is a potentially significant amount in critically ill patients. As noted within the figures, post-filter CVVH is associated with hemoconcentration, and is thought to be associated with more clotting than CVVHD, where no hemoconcentration occurs. The situation is more complicated in pre-filter CVVH, however. As in Figure 3, the fluid is initially diluted, and then the fluid is removed over the course of flow through the filter. Some Nephrologists believe that this initial dilution of the blood results in decreased risk of clotting. However, other nephrologists believe that due to differential flow rates in the blood and replacement fluid that hemoconcentration still occurs within the filter, albeit to a much lesser degree than in post-filter CVVH (this argument is difficult to explain conceptually to a lay audience, but an attempt is made within Figure 3). It is therefore unclear currently whether pre-filter CVVH and CVVHD have equal or unequal rates of clotting. Because centers tend to use exclusively one modality or the other (or a combination, which is beyond the scope of this summary), retrospective analyses are often not possible due to lack of a comparison group.

Figure 3. In pre-filter CVVH, replacement fluid is administered prior to the blood passing through the filter. Nephrologists argue about the effect this has on the rates of hemoconcentration. Some would argue that after the initial dilution, the blood only returns to its initial concentration by the end of the filter, and thus over the course of the filter, there is actually hemodilution that decreases likelihood of clotting. Other nephrologists argue like so: With a blood flow rate of 12 L/hr (standard), and a fluid flow rate of 2 L/hr, then 14 L/hr enter the fluid, and 2 L/hr are removed, which is 14% (2/14) of the fluid that entered. If the fluid flow rate is increased to 4 L/hr, then 16 L /hr enter the filter, and 4L/hr are removed, which is 25% (4/16). This would therefore theoretically result in higher degrees of clotting. The question of pre-filter CVVH and its impact on clotting relative to CVVHD is therefore unanswered, and is the reason for the proposed study."

In short, the investigators plan to start patients who need CRRT on either CVVH or CVVHD by block randomization, and then to measure how often the filters on the machine need to be replaced.

The investigators plan to exclude minors, prisoners, and patients undergoing a certain procedure called "extracorporeal membrane oxygenation" (ECMO). ECMO is designed to pump blood and provide oxygen to the patient when the heart and lungs are not working correctly (ECMO attempts to do the work of the heart and lungs in instances where the patient's organs have failed). CRRT can be added to an ECMO circuit if renal failure is also present, but this is a specialized instance with unique risk factors for clotting, and is not representative of the general CRRT population.