Continuous Renal Replacement Therapies for Acute Kidney Injury

Basic Principles and Operational Characteristics

CRRT is actually an umbrella term for four different continuous modalities: slow continuous ultrafiltration (SCUF), continuous venovenous hemofiltration (CVVH), continuous venovenous hemodialysis (CVVHD), and continuous venovenous hemodiafiltration (CVVHDF). The type of modality chosen depends on the goal of therapy and experience of the center. Treatment may be used for solute removal, fluid removal, or both. Solute clearance through a dialysis filter occurs by diffusion, convection, or combination, called hemodiafiltration ( Tables 67.1–67.3 and Fig. 67.1 ).

Convective techniques (ultrafiltration [UF] and hemofiltration) rely on solvent drag, whereby dissolved molecules are dragged along with ultrafiltered plasma water across a semipermeable membrane in response to a hydrostatic or osmotic force. The process of forcing fluid against a membrane is called ultrafiltration; the fluid collected after it passes through a membrane is the ultrafiltrate. Depending on the amount of UF, fluid replacement is required to prevent hemodynamic instability associated with fluid removal. Solute removal depends on the size of the pores in the membrane, the size and weight of the molecule, the transmembrane pressure, and the UF rate (UFR; Q _uf ). Middle- and large-molecular-weight solutes are more effectively cleared by convection. SCUF and CVVH only use convective transport for removal of solutes.

Diffusive techniques (dialysis) rely on a solute concentration gradient between the blood and the dialysate for clearance across a semipermeable membrane. Solute removal depends on the size of the pores in the membrane, the size and weight of the molecule, and the magnitude of the concentration gradient. The gradient is affected by the dialysate, dialysate infusion rate (Q _d ), and blood flow rate (Q _b ). The dialysate runs countercurrent to the blood and can be customized to promote the diffusion of specific molecules. This is the same technique used in intermittent hemodialysis (IHD), but Q _d is much slower than Q _b , so complete saturation of the dialysate can be reached. Therefore, the Q _d is the rate-limiting factor for solute removal. The smaller the size and weight of the solute and the greater the gradient, the more efficiently solute clearance occurs. CVVHD uses diffusion to remove solutes.

CVVHDF combines both convection and diffusion to clear solutes, utilizing both dialysate and replacement fluids. This modality allows for removal of small- and middle-sized molecules. Solute clearance for this combination technique is equal to the sum of the convective and diffusive clearances. The clearance is the product of a solute's sieving coefficient (SC; ratio of solute concentration in filtrate to solute concentration in plasma) and effluent flow rate (dialysate plus ultrafiltrate). A solute with an SC of 1 means that it can pass freely through a filter; if the SC is 0, then a solute cannot pass through the filter at all.

Technical Issues

Treatment Dosing

The surrogate of solute removal in CRRT is the effluent rate that can be expressed in milliliters per kilogram per hour. Based on randomized controlled trials (RCTs), KDIGO guidelines recommend delivering an effluent dose of 20 to 25 mL/kg/h for patients with AKI; other references recommend a dose of 25 to 30 mL/kg/h. In 2000, Ronco and colleagues performed a prospective, randomized trial of CRRT UF dosing that compared doses of 20 mL/kg/h, 35 mL/kg/h, and 45 mL/kg/h in AKI ICU patients. They demonstrated that a dose of at least 35 mL/kg/h improved survival at 15 days after stopping CRRT compared with the lower dose; there was no significant difference in survival between patients receiving 35 mL/kg/h and those receiving 45 mL/kg/h. After this study, a dose of 35 mL/kg/h became the standard dose widely used.

These concepts changed after two large studies were published. The Veterans Affairs/National Institutes of Health Acute Renal Failure Trial Network (ARFTN) performed a large, multicenter study comparing standard-dose CVVHDF of 20 mL/kg/h with high-intensity CVVHDF of 35 mL/kg/h in critically ill AKI patients; it demonstrated no significant difference in 60-day all-cause mortality, rate of recovery of renal function, and duration of RRT. The Randomized Evaluation of Normal vs. Augmented Level (RENAL) Renal Replacement Therapy Study was another large, multicenter study comparing a CVVHDF effluent flow of 40 mL/kg/h to 25 mL/kg/h to see if there was a difference in 90-day mortality and continued need for RRT in ICU patients with AKI; there was no significant difference in either outcome between the two groups.

An international survey of intensivists reported that use of CRRT hemofiltration doses of at least 45 mL/kg/h was prescribed for septic AKI patients despite little recent evidence supporting this. As stated earlier, there is no difference in outcomes when using high-volume hemofiltration (HVHF) versus standard-volume hemofiltration in these patients.

An important caveat in assessing dose based on effluent volume is that during therapy, there is a progressive filter fouling and clotting that can lower the efficacy of solute removal. Thus, the true dose delivered may be significantly less than that estimated from the effluent volume. Measuring solute removal in the effluent and calculating clearance based on the mass extracted should be the gold-standard method to assess delivered dose. When considering the results of the studies on dialysis dose improving outcomes, it is important to remember that greater than or equal to 85% of the prescribed dose was delivered to AKI patients. This is usually much more than what is delivered in clinical practice (see Table 67.6 for factors that affect dose delivery). Therefore, when prescribing a treatment dose, clinicians should have at least a 25% safety margin; prescribing a dose of 30 to 35 mL/kg/h may then deliver an adequate dose.

Factors That Impact Dialysis Dose

As noted earlier, filter permeability decreases over time; thus, effluent volume (derived from blood-based kinetics) can overestimate delivered clearance. Exposure of the membrane to plasma leads to adsorption and deposition of proteins on the membrane. Concentration polarization results from the accumulation of solute rejected by the UF membrane. Both phenomena result in a concentrated layer along the membrane, causing resistance to mass transfer. Higher filtration fraction and high volume of UF accelerate this process. Transmembrane pressure needs to be increased to maintain an adequate Q _uf to overcome this and lower the concentration of important solutes in the effluent. A second factor is progressive filter clotting, which decreases the SC and filters permeability. The decrement in permeability can be measured by the ratio of effluent fluid urea nitrogen (FUN) to blood urea nitrogen (BUN), which at the beginning of the filter life is equal to 1 and progressively decreases with filter clotting. The CRRT machine measures effluent volume but cannot take into account changes in filter permeability ( Fig. 67.2 ). The FUN-to-BUN ratio should be measured at least daily to assess the dose in the CRRT. Table 67.7 shows the calculations to assess dialysis dose in CRRT. These calculations help correct dialysis prescriptions with regard to small solute clearance, but changes in filter permeability also affect middle molecule clearance. It has been suggested that future studies include measured clearances of small and middle molecules to determine the actual delivered dose. Because the effluent rate provides only an inaccurate estimate of the delivered dose, in 2012, Claure-Del Granado et al. actually recommend measuring and expressing the delivered dose as urea clearance (K _D or K _urea ), derived from dialysate-side kinetics. This is the ratio of mass removal rate to blood concentration and is calculated using the formula: K delivered = (FUN × EV)/BUN, where FUN is urea nitrogen in the effluent (mg/dL), BUN is urea nitrogen in the plasma (mg/dL), and EV is the effluent volume (mL/min). K _D takes into account filter function, duration, and effective time of treatment. They also suggest that the equivalent renal urea clearance (EKR), based on urea kinetic modeling, provides a good estimate of delivered dialysis dose (please see reference for calculating EKR) and can be used as a tool to compare different types of therapies. These recommendations are based on findings evaluating six different methods of assessing and expressing CRRT delivered dose (three equations from blood-side kinetics and three equations from dialysate-side kinetics) in critically ill AKI patients treated with predilution CVVHDF and regional citrate anticoagulation. The dialysate-side measurement ensures delivery of dose, and the blood-side measurement helps determine if changes need to be made in the prescription.

Other factors can also affect the delivered dose. The use of predilution is a third factor, which may decrease urea clearance by as much as 15%. Using replacement fluid prefilter reduces the concentration of solutes in the plasma and decreases solute clearance. Table 67.6 lists other factors impacting dose delivery.

Fluid Management

CRRT is highly effective for fluid control. Fluid balance can be maintained without compromising metabolic–solute balance because they can be dissociated from each other. Fluid removal in CRRT is achieved through manipulating the UFR, which can be tailored to individual needs. The net ultrafiltrate is the difference between total ultrafiltrate (the plasma water removed) and total substitution (the fluid given to the patient) through the CRRT machine. CRRT machines balance all the fluids removed and replaced across the dialysis circuit in order to generate a net amount of fluid removal. All intakes and outputs need to be included to achieve patient fluid balance and integrated into machine and patient fluid prescription and delivery.

Bouchard and Mehta describe three techniques for achieving fluid balance with CRRT ( Table 67.8 ). The level 1 technique is to vary the net UFR (Q _uf ) to meet the anticipated fluid balance needs over 8 to 24 hours; for example, a patient may have a total anticipated fluid intake of 3 L with desired 1-L net loss over 24 hours; the Q _uf would be set at − 170 mL/h (3 L + 1 L = 4 L/24 h). This may not be the best technique to use because there may be unanticipated changes in clinical status and fluid needs, leading to a different net ultrafiltrate that is different from the desired fluid balance. In addition, effluent volume and treatment dose will vary because net Q _uf is not constant with this method.

The level 2 method of maintaining fluid balance is to vary the amount of postdilution replacement fluid administered; the net ultrafiltrate stays the same and exceeds the anticipated hourly intake. With this technique, the postdilution fluid is not given through the CRRT pump but through a separate pump. A patient can be maintained in negative fluid balance by decreasing the amount of postdilution fluid received to be less than the total output, in positive fluid balance by increasing postdilution replacement to be greater than all output, or in even balance by having equal postdilution replacement and total output. This method allows for variation in intake and a predetermined convective clearance because net ultrafiltrate does not vary as in the first technique.

The level 3 technique is similar to the second, but fluid balance is tailored to achieve a targeted hemodynamic parameter every hour. Predefined targets are set for parameters, such as central venous pressure (CVP), mean arterial pressure (MAP), or pulmonary arterial wedge pressure, and algorithms are used to achieve these targets. For example, if the CVP is to be maintained between 8- and 12-mm Hg when this is achieved, the algorithm would determine that net fluid balance be set to zero. If the CVP is above target, the algorithm would call for fluid removal; if CVP is below target, then fluid would be added. This technique allows for greater flexibility and maximally uses CRRT as a fluid regulatory device.