The incidence of acute renal failure (ARF) in critically ill patients ranges from 1% to 25%.1,2 This may be explained by variations in the definitions used when reporting ARF. It has been suggested that more than 35 different definitions exist in the literature.3 Regardless of definition, morbidity and mortality rates associated with ARF are high and increase with critical illness.1,4

Historically, intermittent hemodialysis (IHD) has been the renal replacement therapy (RRT) of choice for critically ill patients. However, many patients cannot tolerate the rapid removal of fluid required during this process.5 Also, a typical IHD session (performed three to five times per week) may be inadequate to remove metabolic toxins seen in these hypercatabolic patients.6 Continuous forms of RRT were introduced in 1977 when Kramer et al. reported their experience with continuous arteriovenous hemofiltration (CAVH) in 12 critically ill patients.7 Since then continuous therapies have evolved to include:

• Ultrafiltration (slow continuous ultrafiltration, used primarily
for fluid removal)

• Continuous venovenous hemofiltration (CVVH)

• Continuous venovenous hemodialysis (CVVHD)

• Continuous venovenous hemodiafiltration (CVVHDF)

Collectively, these are termed continuous renal replacement therapies
(CRRT).

Properties affecting drug removal during CRRT are different from those seen with IHD. Therefore, dosing recommendations
applicable for IHD would not be appropriate with these forms of therapy. In general, clearance properties of CRRT are dependent on the therapy utilized and on the pharmacokinetic/pharmacodynamic properties of each medication.

Clearance Mechanisms
Drug removal during CRRT may occur by convection, diffusion and adsorption. Convection and diffusion have the greatest influence on drug removal, while adsorption probably is less important, though it has not been studied widely. Drug removal is inversely proportional to the percentage of drug that is protein bound. If a drug is ≥ 80% bound, little will be removed. This principle holds true for convection and diffusion. Ultrafiltration and dialysis flow rates (UFR/ DFR) also affect drug clearance.

Because CRRT uses highly permeable membranes, the molecular weight (MW) of most drugs has little impact on overall clearance. During convection, clearance of an unbound drug can be dramatic since CVVH can remove easily compounds with MW <15,000 dalton (Da). The impact of MW on drug removal during CVVHD is greater than the impact seen during CVVH. Solute clearance during CVVHD is dependent on diffusion and, given that diffusion is inversely proportional to MW, the greatest impact is seen with drugs having a MW of <500 Da. Because many drugs used in the critical care setting have a MW of <500 Da, CVVHD can impact their removal significantly.

Volume of distribution, although important during IHD, plays a limited role in CRRT. The slow continuous nature of CRRT allows drugs to equilibrate between body compartments, making them more susceptible to extracorporeal elimination. A volume of distribution <0.6 L/kg suggests limited tissue binding and greater concentrations exposed to the removal techniques of the extracorporeal therapies. Blood flow and filter type play a lesser role, but data are limited. Although some differences do exist, it is common to interchange the drug clearance data of CVVH and CVVHD at low flow rates.

Continuous Venovenous Hemofiltration
Clearance during CVVH is accomplished through the process of convection. The ultrafiltrate produced is replaced, either in part or completely, with appropriate replacement solutions to purify the blood and control volume.8 Clearance of unbound drugs during CVVH can be dramatic, and dose adjustments are required to prevent underdosing.

The sieving coefficient (Sc) of a solute represents its ability to cross a membrane (via convection) and ranges from zero to one (zero representing no convective clearance and one representing free movement). This is represented as Sc = UF/A9 (UF = ultrafiltrate drug concentration, and A = filter inlet drug concentration) and expresses the clinical and mathematical concepts associated with drug clearance during CVVH. When estimating drug clearance, a measured Sc should be used and total clearance taken into consideration (ClTotal = ClCRRT + Clresidual renal + Clnon-renal).

Because all drugs do not have a measured Sc and ClTotal rarely is reported, clearance can be estimated using a concept commonly called the creatinine clearance method (Cl = Sc x UFR). Because unbound drugs are removed during CVVH, the Sc should approximate that value. This method, although useful, does not include Clresidual renal or Clnon-renal, and inherent problems exist.

Published Sc data should be used if available. Protein binding (unbound percent) can be used if these data are not available. For example, if the unbound fraction of cefepime is approximately 79%, the expected Sc should be 0.79. Isla et al. measured the Sc of cefepime during both CVVH and CVVHD10; they found it to correlate well with a measured Sc of 0.76 during the in vivo portion of the study.

In addition, non-renal clearance can change with acute illness.11 Mueller et al measured the total and non-renal clearance of imipenem during CVVH in patients with acute and chronic renal failure.12 Both groups were found to have an increase in patients with acute renal failure. These patients probably would have been underdosed if the creatinine clearance method was used.

Continuous Venovenous Hemodialysis
CVVHD utilizes diffusion to remove solute. A dialysate is infused countercurrent to the flow of blood and solute moves from an area of higher concentration (e.g., blood) to that of lower concentration (e.g., dialysate). This process occurs until an equilibrium is established. A number of commercially prepared dialysates are available for use in CVVHD. Although their use is becoming less common, most pharmacy departments can prepare these solutions.

The ability of a drug to cross a filter during CVVHD is represented by the saturation coefficient (SA). In practice, dosing is treated much the same as with CVVH. This is possible because many drugs used in the critical care setting have a molecular weights of <500 Da, and as with CVVH, drug removal is dependent on the unbound fraction. Keep in mind that the clearance principles are different and this practice may not be appropriate for all drugs. In addition, at lower flow rates, the dialysate is almost completely saturated. If dialysis flow rates are increased (for example, > 2 L/h), this principle may not hold true and the amount of small solute clearance is increased.

Continuous Venovenous Hemodiafiltration
This form of therapy utilizes diffusion, convection and ultrafiltration to remove solute and water. Both a dialysis solution and replacement fluids are used. Many drug elimination studies exclude CVVHDF, making adjustments problematic.

Extended Daily Dialysis/Slow Low-Efficiency Dialysis
Even less is known about drug dosing during extended daily dialysis/slow low-efficiency dialysis. These are intermittent therapies prescribed for periods greater than those typically seen during IHD. No current recommendations for drug dosing are available.

High-Volume Hemofiltration
High-volume hemofiltration is a variant of CVVH and requires larger surface area hemofilters with ultrafiltration volumes typically in excess of >35 mL/kg/h. Even though early studies suggest a benefit in patients with septic shock,13 no data exist describing drug clearance, and a complex model has been suggested.14

Drug Dosing Strategies
Therapeutic drug monitoring should be performed when possible. Unfortunately, this is not always feasible when treating the critical care patient. The next step should be to review the literature for drug specific recommendations. Individual reports are useful for newer therapies, and review articles can summarize therapeutic classes of medications.15 In addition, drug references, such as Drug Prescribing in Renal Failure, provide useful guidelines in a concise tabular format. 16 If therapeutic drug monitoring or published data are not available or lacking, the creatinine clearance method can be used as a last resort to estimate drug elimination.