Pharmacokinetic considerations for antimicrobial therapy in patients receiving renal replacement therapy

Continuous renal replacement therapy (CRRT), particularly continuous venovenous haemofiltration (CVVH) and continuous venovenous haemodiafiltration (CVVHDF), are gaining increasing relevance in routine clinical management of intensive care unit patients. The application of CRRT, by leading to extracorporeal clearance (CL(CRRT)), may significantly alter the pharmacokinetic behaviour of some drugs. This may be of particular interest in critically ill patients presenting with life-threatening infections, since the risk of underdosing with antimicrobial agents during this procedure may lead to both therapeutic failure and the spread of breakthrough resistance. The intent of this review is to discuss the pharmacokinetic principles of CL(CRRT) of antimicrobial agents during the application of CVVH and CVVHDF and to summarise the most recent findings on this topic (from 1996 to December 2006) in order to understand the basis for optimal dosage adjustments of different antimicrobial agents. Removal of solutes from the blood through semi-permeable membranes during RRT may occur by means of two different physicochemical processes, namely, diffusion or convection. Whereas intermittent haemodialysis (IHD) is essentially a diffusive technique and CVVH is a convective technique, CVVHDF is a combination of both. As a general rule, the efficiency of drug removal by the different techniques is expected to be CVVHDF > CVVH > IHD, but indeed CL(CRRT) may vary greatly depending mainly on the peculiar physicochemical properties of each single compound and the CRRT device's characteristics and operating conditions. Considering that RRT substitutes for renal function in clearing plasma, CL(CRRT) is expected to be clinically relevant for drugs with dominant renal clearance, especially when presenting a limited volume of distribution and poor plasma protein binding. Consistently, CL(CRRT) should be clinically relevant particularly for most hydrophilic antimicrobial agents (e.g. beta-lactams, aminoglycosides, glycopeptides), whereas it should assume much lower relevance for lipophilic compounds (e.g. fluoroquinolones, oxazolidinones), which generally are nonrenally cleared. However, there are some notable exceptions: ceftriaxone and oxacillin, although hydrophilics, are characterised by primary biliary elimination; levofloxacin and ciprofloxacin, although lipophilics, are renally cleared. As far as CRRT characteristics are concerned, the extent of drug removal is expected to be directly proportional to the device's surface area and to be dependent on the mode of replacement fluid administration (predilution or postdilution) and on the ultrafiltration and/or dialysate flow rates applied.Conversely, drug removal by means of CVVH or CVVHDF is unaffected by the drug size, considering that almost all antimicrobial agents have molecular weights significantly lower (<2000Da) than the haemofilter cut-off (30,000-50,000Da). Drugs that normally have high renal clearance and that exhibit high CL(CRRT) during CVVH or CVVHDF may need a significant dosage increase in comparison with renal failure or even IHD. Conversely, drugs that are normally nonrenally cleared and that exhibit very low CL(CRRT) during CVVH or CVVHDF may need no dosage modification in comparison with normal renal function. Bearing these principles in mind will almost certainly aid the management of antimicrobial therapy in critically ill patients undergoing CRRT, thus containing the risk of inappropriate exposure. However, some peculiar pathophysiological conditions occurring in critical illness may significantly contribute to further alteration of the pharmacokinetics of antimicrobial agents during CRRT (i.e. hypoalbuminaemia, expansion of extracellular fluids or presence of residual renal function). Accordingly, therapeutic drug monitoring should be considered a very helpful tool for optimising drug exposure during CRRT.