Anae20p242

Acta Anaesthesiol Scand 2001; 45: 929–934
Copyright C Acta Anaesthesiol Scand 2001 Printed in Denmark. All rights reserved ACTA ANAESTHESIOLOGICA SCANDINAVICA
Pharmacokinetics and drug dosing adjustments during
continuous venovenous hemofiltration or hemodiafiltration
in critically ill patients
J. F. BUGGEDepartment of Anesthesia, Rikshospitalet, Oslo, Norway Continuous renal replacement therapy (CRRT) in critically ill gested, but it has the same limitations in overestimating drug patients with renal failure may significantly increase drug clear- clearance when dialysis is combined with filtration. For non- ance, requiring drug dosing adjustments. Drugs significantly toxic drugs, doses can safely be increased 30% above actual esti- eliminated by the kidney often undergo substantial removal mates to ensure adequate dosing. For drugs with a narrow during CRRT, and a supplemental dose corresponding to the therapeutical margin, monitoring plasma concentrations are amount of drug removed by CRRT should be administered.
mandatory. When appropriate, the use of a readily available ref- Clearance by CRRT can either be measured or estimated. The erence for drug dosing is recommended.
high-flux membranes used in CRRT make no filtration barrierto most drugs, and the filtrate concentration can be estimated Key words: Acute renal failure; continuous renal replacement
by the unbound fraction of the drug in plasma. When adding therapy; drug dosing; hemofiltration; hemodiafiltration; pharm- dialysis to filtration, this approach overestimates drug clearance, and a correcting factor should be used. A method for estimatingdrug clearance as a function of creatinine clearance is also sug- c Acta Anaesthesiologica Scandinavica 44 (2001) INTHEINTENSIVEcareunit(ICU)bothintermittentand CRRT, and to provide some practical guidance for continuous renal replacement therapies (CRRTs) are used for treatment of acute renal failure (ARF). Duringthe last two decades there has been an evolution of con- Drug properties
tinuous therapies from the initial arteriovenous hemo-filtration (CAVH) driven by the patient’s arterioven- Total body clearance of a drug is the sum of clearances ous pressure difference, to the more sophisticated from different sites in the body which may include he- pumpdriven devices performing continuous venoven- patic, renal, and other metabolic pathways. It is the con- ous hemofiltration (CVVH), hemodialysis (CVVHD) or tribution of renal clearance to total body clearance that hemodiafiltration (CVVHDF). In critically ill patients, is the major determinant of making drug dosing adjust- CRRT is superior to intermittent hemodialysis (IHD) in ments in renal failure. If the renal clearance of a drug is maintaining hemodynamic stability (1). Together with normally less than 25–30% of total body clearance, im- better volume, electrolyte, and acid-base control, this is paired renal function is unlikely to have clinically sig- probably the main reason for it becoming the therapy nificant influence on drug removal (3). Similarily, drug removal by CRRT will have little influence on total Critically ill patients with ARF often have multi- body clearance and dosing adjustments do not have to organ dysfunction, sepsis, or other conditions that re- be considered. If patients develop hepatic failure, the quire complex drug therapy, and which may influence extent to which CRRT contributes to total body clear- drug concentrations through changes in absorption, ance may increase, and dose adjustments may become distribution, metabolism, and elimination. The ad- necessary. Drugs significantly eliminated by the kidney dition of CRRT may further complicate drug therapy, often undergo substantial removal during CRRT, and and the purpose of this review is to outline the gen- dosing adjustments are frequently required.
eral principles determing whether a dose adjustment However, there are other drug properties affecting is required to compensate for drug elimination during clearance by CRRT, including protein binding, vol- J. F. Bugge
ume of distribution (Vd), molecular weight, and drug IHD and CRRT. A drug with a large Vd and high charge. Only the unbound fraction of a drug is avail- clearance during high-flux IHD will rapidly be re- able for filtration, and drugs with a high protein bind- moved from plasma, but only a small amount of the ing are poorly cleared by CRRT. Many factors may body’s drug content is removed during one dialysis alter the fraction of unbound drug such as systemic session, and plasma concentration will be restored be- pH, heparin therapy, hyperbilirubinemia, concen- tween therapies. CRRT by its continuous and slower tration of free fatty acids, relative concentration of action has much less influence on plasma concen- drug and protein, as well as presence of uremic prod- trations of drugs with large Vds, because there is time ucts and other drugs that may act as competetive dis- for continuous redistribution of the drug from the placers (4–6). Most of our knowledge about alter- tissues to the blood. Although drug elimination dur- ations in protein binding and pharmacokinetics is de- ing CRRT is much slower for drugs with large Vds rived from studies on chronic renal insufficiency. The than for drugs with small, the same is true for endoge- influence of ARF on protein binding is not well de- nous (hepatic) elimination which has to clear the same scribed. Critically ill patients often have low albumin Vd. As a consequence, drug dosing adjustments to be values which may increase the unbound fraction of made during CRRT are much more dependent on the many drugs with possible deleterious effects, as docu- relative contribution of CRRT to total body clearance mented for phenytoin (7). These patients also often of the drug than on the drug’s Vd (9).
have increased levels of acid a1-glycoprotein, which Most drugs have a molecular weight Æ500 Da, and may increase protein binding of some drugs. Thus, very few are greater than 1500 Da (vancomycin at the reported unbound fraction in healthy volunteers 1448 Da). Conventional dialysis membranes favor and in patients with chronic renal insufficiency may diffusive clearance of low molecular weight solutes differ substantially from the unbound fraction of below 500 Da, whereas the typical high-flux mem- drugs in critically ill patients receiving CRRT.
branes used for CRRT have larger pores (20 000– Drug charge affects clearance by the Gibbs-Donnan 30 000 Da), making no significant filtration barrier to effect: Retained proteins on the blood side of the membrane make the membrane negatively chargedand the filtered fraction of cationic drugs will be a Methods of drug removal
little less than expected from the unbound fraction,whereas the opposite is true for anionic drugs.
Essential to rational drug prescribing in patients un- A large Vd reflects a drug that is highly tissue dergoing CRRT is an understanding of the different bound, and consequently only a small proportion ac- methods of solute removal that occur with the various tually resides in the vascular compartment available types of treatments. Diffusion, convection, and ad- for clearance by endogenous or extracorporal routes.
sorption are the three mechanisms for solute removal The Vd is a mathematical reflection of the volume in during CRRT. The process of moving a solute through which a drug would need to be dissolved to obtain a membrane from an area of high concentration to an the observed blood concentration, assuming homoge- area of lower concentration is called diffusion, and it neous mixing in the body. For intravenously adminis- is the primary method of solute removal during dialy- tered drugs, the Vd determines the dose (D) needed sis. Diffusive clearances varies between filter mem- to achieve the desired plasma concentration (C): branes and are greater for polyacrylonitrile (PAN,AN-69) than for polyamide membranes (10). In a study by Morabito et al. (10), diffusive equilibrium In critically ill patients the actual Vd may differ between dialysate and plasma for urea could only be from values obtained from pharmacological tables, obtained with the AN-69 filter at a dialysate flow rate and it shows great inter- and intraindividual vari- of 1.5 l ¡ hª1 during CAVHDF. Similarily, the diffusive ations (8). This may increase the error when using Vd clearances of drugs will vary depending on the filter in estimating drug dosing. The Vd of aminoglycosides increases approximately 25% in the critically ill, Convection is the removal of solutes along with the whereas vancomycin, metronidazole, and most b-lac- solvent in which they are present, and the rate of ul- tam antibiotics show near normal values, but with in- trafiltration determines the convective clearance of a dividual variations (8). A drug with a small Vd (Æ1 solute during CRRT. It is not influenced by the con- l ¡ kgª1) is more likely to be cleared by extracorporal centration gradients across the membrane and is only therapies than a drug with a large Vd (Ø2 l ¡ kgª1).
However, there is a significant difference between The diffusion of a solute is inversely proportional Drug dosing during CRRT
to its molecular weight, and the relative importance placement fluid is administered distal to the filter, and of convection increases with increasing molecular mass removal per unit volume of effluent fluid is rela- weight. Diffusion favors clearance of small solutes tively high. For small solutes the effluent concen- whereas middle- and large-sized molecules mainly tration may be the same as in plasma water. However, are removed by convection. As an example: for the the ultrafiltration rate is limited by the operating 1448 Da molecule, vancomycin, convection is quanti- characteristics of the system, mainly the hematocrit tatively more important than diffusion for its removal and the blood flow rate. The upper limit of ultrafil- tration rate is approximately 25–30% of plasma flow Adsorption to filter membranes leads to increased rate (17). In pre-dilution procedures, the replacement removal from plasma (12) and the various filters have fluid is administered proximal to the filter, diluting different absorptive capacity. Some filter membranes, the concentration of solutes in the blood, and thus re- such as the commonly used polyacrylonitrile, may ad- ducing solute clearances approximately 15–20% (16).
sorb a substantial amount of drug to its surface (13, This modest decrease in efficiency can be overcome 14). However, adsorption is a saturant process, and by increasing the ultrafiltration rate which in this set- the influence on drug removal will depend on the fre- ting is not restricted by hematocrit and blood flow quency of filter changes. In general, with filters lasting approximately 18–24 h, adsorption probably has mi-nor influence on drug removal, but at present, infor-mation about the various filters’ adsorptive capacity Drug dosing adjustments
for most drugs is lacking. Filter adsorption is not ac-counted for in drug dosing guidelines (9, 15). CRRT The critically ill patient with renal failure is at risk for drug accumulation and overdose, but also for under- In CVVH, solutes are removed only by convection, dosing that may be life threatening, such as in the case and drug removal is limited by the ultrafiltration of insufficient antibiotic treatment. Drug dosing ad- rate. By adding dialysis to filtration, as in CVVHDF, justments can be performed by reducing the dose in solutes are removed by both convection and dif- proportion to the reduction in total body drug clear- fusion, increasing the removal of small molecules more than middle-sized and large molecules. It also gives the possibility for the two processes to interact in such a manner that solute removal is significantly where DN is the normal dose, ClANUR is drug clear- less than what is expected if the individual compo- ance in anuric patients, and ClN is normal drug clear- nents are simply added together. In a study by Bru- ance. ClANUR and ClN are retrieved from pharmaco- net et al. (16), the clearance during CVVHDF of logical tables. If CRRT contributes significantly to the small molecules, such as urea and creatinine, was total body clearance of a drug (Ø25–30% of total body approximately the sum of the two clearances ob- clearance), a supplemental dose, corresponding to the tained when CVVHD and CVVH were performed amount of drug removed by CRRT, should be ad- separately. Even with the highest flow rates of 2.5 and 2.0 l ¡ hª1 of dialysate and ultrafiltration, re- spectively, the differences between actual and pre- dicted clearances from addition of the two pro- where ClCRRT is the CRRT drug clearance. Clearance cedures differed by less than 10%, indicating mini- mal interaction between diffusion and convection.
2-microglobulin, however, combining diffusion and convection did not increase clearance compared where, CE and CP are drug concentrations in effluent to convection alone, indicating that this small pro- fluid and plasma, respectively. QE is the effluent flow tein was only removed by convection. This means rate which is the sum of ultrafiltration flow rate (QUF) that the diffusive clearance of a drug during and dialysate flow rate (QD). Substitution into the CVVHDF is difficult to predict and will depend on its molecular weight, blood and dialysate flow rates, In CVVH and CVVHDF, the method by which re- For most drugs, measurements are not available, and placement fluids are administered influences solute CRRT clearances have to be estimated. The sieving co- removal efficiency. In post-dilution procedures the re- efficient (S) of a drug is the concentration in ultrafil- J. F. Bugge
trate (CUF) devided by the concentration in plasma, was underestimated in the range of 30%. This means that the actual non-CRRT clearance was greater thanestimated from P correlations between doses calculated from measured The exact formula for the sieving coefficient is SΩ2 kinetic data and doses estimated by the above equa- CUF/(CPinπCPout ), but the differences between CPin tion during CVVH, but they did not tell how PX was and CPout are negligible, making the above equation estimated. For CVVHDF data are lacking, but when almost correct. For readily filtrable molecules CUF ap- using the Kdrel on CAVHDF clearances reported in the proximates the concentration of unbound drug in literature, Kroh et al. (22) found very good corre- plasma, and S can be estimated by the unbound frac- lations between observed and estimated clearances For non-toxic drugs, doses can safely be increased ClCRRTΩfu ¡ QUF or ClCRRTΩfu ¡ (QUFπQD) beyond actual estimates and a 30% increase is recom- during CVVH or CVVHDF, respectively. The value of mended to ensure adequate dosing (8).
fu is retrieved from pharmacological tables, but as out- Making these estimates is time consuming, requir- lined above, the unbound fraction in the critically ill ing a careful search for basic pharmacokinetic data, may differ from these values. However, with some ex- and they are based on totally non-functioning kid- ceptions and individual variations, Golper and Marx neys. Today there is a tendency to start CRRT earlier found that for most of the 60 drugs measured, the in the course of illness, and residual renal function filtered fraction during CRRT correlated well with the may contribute to drug clearance. According to unbound fraction known from studies on healthy sub- Dettli’s equation as quoted by Keller and Czock (24), jects (19). If CRRT is performed in a predilution mode, actual total body clearance (ClACTUAL) of a drug is a fu has to be corrected by the dilution factorΩplama linear function of creatinine clearance (ClCR): flow rate/(plasma flow rateπreplacement fluid flow rate). The dialysate flow rate during CVVHDF is low (750–2500 ml ¡ hª1), allowing almost diffusive equilib- where aΩ(ClNªClANUR)/ClCRn. ClCRn is normal crea- rium to occur between dialysate and plasma concen- tinine clearance. To simplify and individualize drug trations for small molecules, making fu ¡ QD an ac- dosing estimation, Keller et al. (25) suggested to apply ceptable estimate of diffusive clearance. However, it this equation during CRRT by introducing the total will always be overestimated, and increasingly over- ClCR concept (ClCRtot) as the sum of renal ClCR estimated with increasing molecular weight and di- (ClCRren) and extracorporal ClCR (ClCRfilt): alysate flow rate. Vos and Vincent (20, 21) found a close exponential correlation of a drug’s diffusive mass transfer coefficient through membranes in ClCRtot can easily be calculated from creatinine meas- urements in urine, effluent fluid and blood. These cal-culations can then be substituted into dosing adjust- ment equations. Using the same approach as above, where Kd and Kdc are the diffusive mass transfer co- and setting PXΩClANUR/ClN, the estimated dose will efficients for the drug and creatinine, respectively, and MW is the drug’s molecular weight (113 is the mol- ecular weight of creatinine). The limiting factor of dif- fusive clearance can be approximated by Kdrel (22), This equation uses an individual’s actual creatinine clearance to estimate drug clearance and drug dosing,and dose estimates will automatically be adjusted as changes occur in the patient’s renal function. How- ever, extracorporal creatinine clearance may overesti-mate extracorporal drug clearance, especially diffus- DEΩDN ¡ [PXπfu ¡ (QUFπQD ¡ Kdrel)/ClN] ive clearance of high molecular drugs during where PX is the extrarenal clearance fraction of the CVVHDF. PX is not measured, and patients with se- drug (ΩClANUR/ClN). Using this approach for anti- vere ARF have a residual clearance that is sometimes microbiological agents during CVVH (QDΩ0), Joos et remarkable (22), leading to underestimating of drug al. (23) found that CCVH clearance deviated less than dosing. The applicability of this approach to drug 15% from estimated, whereas total body clearance dosing has not been clinically evaluated yet, and it Drug dosing during CRRT
remains to be seen whether it represents any improve- recommended to ensure adequate dosing. For drugs ment in making drug dosing adjustments during with a narrow therapeutical margin, drug monitoring Another and strongly suggested approach to drug dosing during CRRT is to utilize a readily available References
reference. In a publication by Kroh (8), there is a tableof normal kinetic data of many actual drugs, and sug- 1. Manns M, Sigler MH, Teehan BP. Intradialytic renal hemo- gested dosing adjustment factors for different ultra- dynamics. Potential consequences for the management of filtrate rates during CVVH. If using references that the patient with acute renal failure. Nephrol Dial Transplant
1997: 12: 870–872.
utilize glomerular filtration rate (GFR) in its specific 2. Bellomo R, Ronco C. Continuous versus intermittent renal recommendations, such as the Bennett tables (26), it is replacement therapy in the intensive care unit. Kidney Int useful to regard CRRT as a GFR of 10–50 ml ¡ minª1 1998: 53 (Suppl 66): S125–S128.
depending on dialysate and hemofiltration flow rates.
3. Levy G. Pharmacokinetics in renal disease. Am J Med 1977: 62: 461–463.
However, these tables (26) are based on measure- 4. Suh B, Craig WA, England AC, Elliott RL. Effect of free fatty ments from patients with chronic renal insufficiency acids on protein binding of antimicrobial agents. J Infect Dis and do not take into account the pharmacokinetic 1981: 143: 609–610.
changes induced by acute critical illness as discussed 5. Gulyassy PE, Depner TA. Impaired binding of drugs and endogenous ligands in renal disease. Am J Kidney Dis 1983: 2: 578–601.
All predictions have shortcomings, and for toxic 6. McNamara PJ, Lalka D, Gibaldi M. Endogenous accumu- drugs and for drugs with a narrow therapeutic range lation products and serum protein binding in uremia. J Lab
Clin Med 1981: 98: 730–740.
such as aminoglycosides, drug monitoring with meas- 7. Driscoll DF, McMahon M, Blackburn GL, Bistrain BR.
urements of plasma concentrations is mandatory for Phenytoin toxicity in a critically ill, hypoalbuminemic pa- safe drug dosing. This method addresses the actual tient with normal serum drug concentrations. Crit Care Med total body clearance in which CRRT is only one part, 1988: 16: 1248–1249.
8. Kroh UF. Drug administration in critically ill patients with and adjustments can be made as total body clearance acute renal failure. New Horizons 1995: 3: 748–759.
is changed by the clinical condition and/or by CRRT 9. Bohler J, Donauer J, Keller F. Pharmacokinetic principles adjustments. The following dosing formula is often during continuous renal replacement therapy: Drugs and used to achieve the desired peak concentration dosage. Kidney Int 1999: 56 (Suppl 72): S24–S28.
10. Morabito S, Pierucci A, Marinelli R, Cicciarelli A, Guzzo I, (CPEAK) from the actual trough (or any concentration) Cinotti GA et al. Efficiency of different hollow-fiber hemo- filters in continuous arteriovenous hemodiafiltration. Am J
Nephrol 2000: 20: 116–121.
DΩ(CPEAKªCACTUAL) ¡ Vd ¡ Body weight 11. Jeffrey RF, Khan AA, Prabhu P, Todd N, Goutcher E, Will EJ et al. A comparison of molecular clearance rates during The estimated Vd in this equation may differ from the continuous hemofiltration and hemodialysis with a novel actual Vd, but with repeated measurements this error volumetric continuous renal replacement system. Artif Or- is self-correcting. As outlined above, the Vd of amino- gans 1994: 18: 425–428.
glycosides is increased in critically ill patients, indi- 12. Rumpf KW, Rieger J, Ansorg R, Dohl B, Scheler F. Binding of antibiotics by dialysis membranes and its clinical rel- cating higher doses to achieve the desired CPEAK and evance. Proc EDTA 1977: 14: 677–688.
longer intervals than expected from actual renal and 13. Kronofol NO, Lau AH, Barakat MM. Aminoglycoside bind- CRRT clearances to reach acceptable trough concen- ing to polyacrylonitrile hemofilter during continuous hemo-
filtration. Trans ASAIO 1987: 33: 300–303.
trations. There has been a discussion lately about 14. Oborne CA, Buckton G, Barker N. The interaction of albu- aminoglycoside dosing and dosing intervals, and it min and drugs with two hemofiltration membranes. J Clin seems that giving the total daily dose as one single Pharm Ther 1994: 19: 119–125.
bolus is at least as effective and less toxic than divided 15. Joy MS, Matzke GR, Armstrong DK, Marx MA, Zarowitz BJ. A primer on continuous renal replacement therapy for doses in patients with normal renal function (27). The critically ill patients. Ann Pharmacother 1998: 32: 362–375.
consequences for adjusting aminoglycoside dosing 16. Brunet S, Leblanc M, Geadah D, Parent D, Courteau S, Car- during CRRT are not clear, but probably one should dinal J. Diffusive and convective solute clearances during be more concerned about avoiding high trough than continuous renal replacement therapy at various dialysate
and ultrafiltration flow rates. Am J Kidney Dis 1999: 34: 486–
In conclusion, drug dosing adjustments during 17. Kaplan AA. Continuous arteriovenous hemofiltration and CRRT can be guided by measuring or estimating related therapies. In: Winchester J, ed. Replacement of renal CRRT drug clearance, by using available references, function by dialysis. Dortdrecht: Kluwer Academic, 1996: 392–400.
or by monitoring drug plasma concentrations. For 18. Clark WR, Ronco C. CRRT efficiency and efficacy in relation non-toxic drugs, doses 30% above actual estimates are to solute size. Kidney Int 1999: 56 (Suppl 72): S3–S7.
J. F. Bugge
19. Golper TA, Marx MA. Drug dosing adjustments during con- with renal impairment. Int J Clin Pharmacol Ther 1998: 36:
tinuous renal replacement therapies. Kidney Int 1998: 53
25. Keller F, Bohler J, Czock D, Zellner D, Mertz AKH. Individu- 20. Vos MC, Vincent HH. Continuous arteriovenous hemodi- alized drug dosage in patients treated with continuous afiltration: Predicting the clearance of drugs. Contrib Nephrol hemofiltration. Kidney Int 1999: 56 (Suppl 72): S29–S31.
1991: 93: 143–145.
26. Aronoff GR, Berns JS, Brier ME, Golper TA, Morrison G, 21. Vincent HH, Vos MC, Akcahuseyin E, Goessons WH, van Singer I, et al. Drug prescribing in renal failure: dosing guide- Duyl WA, Schalekamp MA. Drug clearance by continuous lines for adults. Philadelphia: Am College Physicians, 1999.
haemodiafiltration (CAVHD). Analysis of sieving coef- 27. Swan SK. Aminoglycoside nephrotoxicity. Semin Nephrol ficients and mass transfer coefficients of diffusion. Blood Pur- 1997: 17: 27–33.
if 1993: 11: 99–107.
22. Kroh UF, Holl TJ, Steinhausser W. Management of drug dos- ing in continuous renal replacement therapy. Semin Dial 1996: 9: 161–165.
23. Joos B, Scmidli M, Keusch G. Pharmacokinetics of anti- microbial agents in anuric patients during continuous venovenous hemofiltration. Nephrol Dial Transplant 1996: 11:
24. Keller F, Czock D. Pharmacokinetic studies in volunteers e-mail: jan.fredrik.bugge/rikshospitalet.no

Source: http://gambramedical.ru/assets/files/Pharmacogenetics%20of%20drugs%20during%20CRRT.pdf