Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy D. Kuang and C. Ronco

Introduction Infection is a common problem in the intensive care unit (ICU). Severe sepsis and septic shock are conditions at the end of the spectrum of human response to infection. Acute renal failure is increasingly seen as part of the multiple organ dysfunction syndrome, which is the most frequent cause of death in patients admitted to the ICU. However, severe sepsis and septic shock are the primary causes of the multiple organ dysfunction syndrome. In the past decades, continuous renal replacement therapy (CRRT) has been widely employed as an extracorporeal blood purification method in the management of septic patients with or without acute renal failure in the ICU, because it offers several advantages over conventional intermitant hemodialysis and peritoneal dialysis [1]. Antimicrobial therapy poses one of the greatest challenges to the intensivist involved in the management of septic patients with persisting high mortality and morbidity rates in ICU [2]. The goal of antimicrobial prescription is to achieve effective active drug concentrations that result in clinical cure while avoiding drug-associated toxicity [3]. However, sepsis, acute renal failure, and CRRT may have profound effects on the pharmarcokinetic and pharmarcodynamic properties of various antimicrobials used in the ICU. The purpose of this chapter is to discuss the impact of these factors on the pharmacological processes and dosing adjustment of antimircobials used in critically ill patients.

Basic Pharmacological Parameters of Antimicrobial Agents: Pharmacokinetics and Pharmacodynamics A constellation of pathophysiological changes can occur in patients with sepsis, which, along with the influence of sepsis-related organ failure and organ-supportive therapy, can complicate antimicrobial dosing. Knowledge of the pharmacokinetic and pharmacodynamic properties of the antimicrobials used for the management of sepsis is essential for selecting the antibacterial dosage regimens [2, 4]. Pharmacokinetics refers to the study of concentration changes of a drug over a given time period. The important pharmacokinetic parameters include plasma protein binding (PPB), volume of distribution (Vd), clearance (Cl), half-life (t½), peak serum drug concentration achieved by a single dose (Cmax), minimum serum drug concentration during a dosing period (Cmin), and area under the serum concentration-time curve (AUC). These factors can be used to determine whether appropriate concentrations of the antimicrobial agent are being delivered to the target area.

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy

Pharmacodynamics is the study of the biochemical and physiological effects of drugs and their mechanisms of action. The primary pharmacodynamic parameters include the time for which the serum concentration of a drug remains above the minimum inhibitory concentration for a dosing period (T 8 MIC), the ratio of the antibacterial Cmax to MIC (Cmax/MIC), the ratio of the AUC during a 24-hour time period to MIC (AUC24/MIC), and the post-antibiotic effect. The rate and extent of the bactericidal activity of an antimicrobial agent is dependent on the interaction between drug concentration at the site of infection, bacterial load, phase of bacterial growth, and the MIC of the pathogen. Different antimicrobial classes appear to have different types of kill characteristics on bacteria. The q -lactam group of antimicrobials has a time-dependent kill characteristic with T 8 MIC as the best predictor of efficacy. These agents appear to have improved antibiotic efficacy when the exposure time rather than concentrations are maximized. Other representative agents in this group include aztreonam, carbapenems, macrolides, clindamycin, and vancomycin. In contrast, aminoglycosides, metronidazole, and daptomycin have a concentration-dependent kill characteristic. More effective killing is observed with higher drug concentrations. Cmax/ MIC and AUC/MIC ratios are the parameters correlating with clinical efficacy with this group of agents. Fluoroquinolones are more complex and were initially reported to be Cmax/MIC dependent, although subsequent studies have also found that AUC24/MIC is important.

Impact of Sepsis on Pharmacological Characteristics of Antimicrobial Agents Based on Pathophysiological Changes The appropriate prescription of antimicrobials requires a detailed knowledge of the pathophysiological and subsequent pharmacokinetic changes that occur throughout the course of sepsis [2]. Concomitant patient factors that may influence the pharmacological characteristics of the antimicrobials include changes in total body water, albumin and acute phase protein levels, muscle mass, blood pH, bilirubin concentration, renal, hepatic, and cardiac function. Various endogenous inflammatory mediators are produced during the development of sepsis. These mediators may affect the vascular endothelium directly or indirectly, resulting in maldistribution of blood flow, endothelial damage, and increased capillary permeability, which cause fluid shifts from the intravascular compartment to the interstitial space. This would increase the volume of distribution of water-soluble antimicrobials. In addition, it is interesting to note that the serum concentration of albumin may change in these critically ill patients, as acute phase reactant proteins are preferentially synthesized. The binding affinity of drugs to albumin may also decrease due to uremia. Hypotension is very common due to the inflammatory response associated with sepsis. Inotropic agents are often prescribed in septic patients who fail to respond to administration of intravenous fluids. Therefore, an increased cardiac index and renal preload can always be found. Consequently, serum creatinine and drug clearance are increased in patients with absent kidney and/or liver dysfunction. As a serious complication, multiple organ dysfunction syndrome often occurs and results in a consequent decrease in antimicrobial clearance, which prolongs the elimination t1/2 and may increase antimicrobial concentrations and/or lead to the accumulation of metabolites. These pathophysiological changes will reverse with

593

594

D. Kuang and C. Ronco

recovery from sepsis. Furthermore, since the physiology of these patients may change over a relatively short period of time, ongoing evaluations are indicated to allow timely adjustment of antimicrobial dosing.

Impact of Acute Kidney Injury on Pharmacological Characteristics and Dosage Adjustment of Antimicrobial Agents: Difference from Chronic Renal Disease Impaired renal function may have profound effects on the pharmacokinetics and pharmacodynamics of renally excreted antimicrobials, necessitating modification of the dosage regimen in order to avoid toxicity through accumulation of the parent and/or its metabolites. The most universal pharmacokinetic equation is: t1/2 = 0.693 × Vd/Cl

(Eqn 1)

Since the t1/2 is reciprocal to the clearance, an interpolation for any degree of renal impairment can be made from the extreme values for normal kidney function and anuria. Although the volume of distribution of most drugs rises slightly, these discrepancies are considerably smaller than the total volume of body fluid and as a result the influence is limited. However, great inter- and intraindividual variations of actual volumes of distribution have been described in critically ill patients with acute renal failure. A difference was also found in the PPB due to the reduction in net protein content and accumulation of various uremic toxins. Neither the presence of renal failure nor of CRRT requires adjustment of a loading dose which depends solely on the volume of distribution. However, maintenance doses that undergo considerable renal excretion should be adapted to the reduced renal clearance. There are two approaches to adjusting drug dosage in non-dialyzed patients with impaired renal function: the Dettli rule and the Kunin rule. The Dettli rule adjusts the maintenance dosage in proportion to the reduced clearance. Alternatively, Kunin’s rule is derived from the elimination t1/2. The normal starting dose is given, and one-half of the starting dose is repeated at an interval corresponding to one t1/2. The Dettli rule results in an AUC that is the same as in normal individuals. With the Kunin rule, the peak levels (Cmax) are identical, but the AUC and the Cmin are higher than in normal individuals [5]. Adjustment of the microbial regimen mainly depends on a precise estimate of the glomerular filtration rate (GFR) of the patient. However, if a patient is anuric or in acute renal failure, the Cockcroft-Gault equation or the Modification of Diet in Renal Disease (MDRD) equation will not give a true reflection of GFR. It is still difficult for clinicians to measure the accurate GFR in patients with acute renal failure and a non-steady-state condition. Convenient biochemical markers such as serum creatinine or blood-urea-nitrogen (BUN) are not accurate reflections of renal function due to the lag time in rapidly fluctuating renal function. Urine output may also be misleading in view of the different phases in the natural progression of acute renal failure. Among newer markers, serum cystatin C has not yet been well validated as an early and reliable GFR indicator in patients with acute renal failure. Several pharmacotherapeutic recommendations on adjustment of antimicrobial regimens according to renal impairment are available [6 – 8]. However, the variation between these sources is remarkable, including drugs for which no adjustment was recommended in one source while another marked them as contraindicated in renal failure. We should clearly identify the categories of renal impairment for dose or

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy

interval adjustment and be prudent in choosing a regimen with different recommendations in different sources for the same condition. Nevertheless, we have to emphasize that all the recommendations currently available are based on chronic renal dysfunction and pharmacokinetic data from studies conducted among healthy volunteers. They cannot, therefore, be extrapolated readily to critically ill patients with acute renal failure due to sepsis. Consequently, drug handling in such patients remains largely unpredictable and calculations based on data in the literature only yield rough estimates of drug dosage adaptation.

Impact of CRRT on Pharmacological Characteristics and Dosage Adjustment of Antimicrobial Agents: How to Achieve an Accurate Regimen? CRRT would have a profound effect upon the pharmacokinetics of antimicrobial agents with multiple variables affecting drug clearance. The factors governing the extent of drug removal from the extracorporeal system can be broadly classified into two major categories:

Pharmacological Factors of Antimicrobial Agent Molecular weight Most antimicrobial agents have a molecular weight (MW) up to 500 Da. Generally, it is easier for smaller drugs (MW ‹ 500 Da) to pass through a membrane and be removed. However, large molecular drugs, such as vancomycin at 1448 Da, can easily pass through typical high-flux membranes. Only cuprophane and some other cellulose-based membranes with small pores create a significant filtration barrier to unbound drugs. Volume of distribution The removal of agents with a large volume of distribution by CRRT is minimal despite efficient clearance due to the small proportion of total body drug present in the systemic circulation. Since total body water constitutes approximately 67 % of the body weight, a drug that distributes well in all fluid compartments would have a volume of distribution of close to 0.7 l/kg. As a result, any drug with a volume of distribution 8 0.8 l/kg would likely signify tissue binding, and therefore, probably not be efficiently removed by CRRT. Plasma protein binding Only unbound drug present in plasma water is pharmacologically active and can be removed by extracorporeal processes. Therefore, antimicrobials with a high degree of PPB ( 8 80 %) will be poorly cleared by CRRT [9]. Many factors may alter the fraction of unbound drug such as systemic pH, heparin therapy, hyperbilirubinemia, concentration of free fatty acids, relative concentration of drug, and proteins that may act as competitive displacers. Thus, the reported unbound fraction in healthy volunteers and in patients with chronic renal insufficiency may differ substantially from that in critically ill patients [10].

595

596

D. Kuang and C. Ronco

Fractional extracorporeal clearance The total body clearance of an antimicrobial agent is the sum of clearances from different sites in the body which may include hepatic, renal, other metabolic pathways, and extracorporeal therapy. But extracorporeal elimination is only considered clinically significant if its contribution to total body clearance exceeds 25 – 30 % [11]. This also explains why extracorporeal elimination is not clinically relevant for drugs with overwhelming non-renal clearance. A patient’s residual renal function also needs to be taken into account for total body clearance, because significant residual renal function may reduce the fraction that is removed by extracorporeal procedures and may render extracorporeal elimination negligible. It should be emphasized that extracorporeal elimination only replaces glomerular filtration. However, renal drug clearance includes glomerular filtration, tubular secretion, and reabsorption. Therefore, any attempt to determine the extracorporeal creatinine clearance and use the same dosage guidelines as in patients with reduced renal function cannot be recommended, especially with drugs largely eliminated by tubular secretion [9]. Drug charge The Gibbs-Donnan effect may have a significant effect on polycationic drugs. Since large anionic molecules such as albumin do not pass through membranes readily, and retained proteins on the blood side of the membrane make the membrane negatively charged, they may partially retard the transmembrane movement of polycationic drugs (e.g., aminoglycosides). This drug charge and membrane interaction may help to explain the discrepancy between PPB and the observed sieving coefficient.

Technical Factors of Extracorporeal Blood Purification Therapies Membrane The surface area and the pore size of the dialytic membrane or the hemofilter are considered as two crucial factors determining the extent of drug removal. In general, the pore size of conventional dialytic membranes made up of natural substances (cellulose or cuprophane) is relatively small, permitting passage of fluid and small solutes ( ‹ 500 Da) only. High-flux dialytic membranes are usually made up of biosynthetic material (polysulfone, polyacrylonitrile, polyamide) with relatively larger pore sizes (5,000 – 20,000 Da). Even larger pore sizes are used in hemofilters (20,000 – 50,000 Da). Diffusion (hemodialysis) The efficiency of solute removal based on diffusion in hemodialysis is determined by the concentration gradient in addition to the porosity and surface area of the dialytic membrane. Compared with convective clearance, diffusive clearance will decrease as MW increases. Due to the lower diffusive permeability, MW has a greater influence on diffusive clearance with conventional dialysis membranes than with the synthetic membranes used in CRRT [9]. In continuous veno-venous hemodialysis (CVVHD), diffusive clearance of small unbound solutes will equal the dialysate flow rate (Qd). Dialysate saturation (Sd) represents the capacity of a drug to diffuse through a dialysis membrane and saturate the dialysate, which is calculated by dividing drug concentration in the dialysate (Cd) by its plasma concentration (Cp): S d = Cd / Cp

(Eqn 2)

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy

Consequently, diffusive drug clearance (ClHD) is calculated by multiplying Qd by Sd: ClHD = Qd × Sd

(Eqn 3)

Since a higher MW decreases the speed of diffusion or a higher Qd decreases the time available for diffusion, an increase in each of these parameters will produce a decrease in Sd [11]. Sd can theoretically be influenced by drug-membrane interactions and by protein adsorption to the membrane. When extracorporeal drug clearance is calculated, Sd can be replaced approximately by the unbound fraction. However, it should be emphasized that Sd does not remain constant, and it would be a serious mistake to use the same Sd in different Qds. Convection (hemofiltration) Convective solute removal used in hemofiltration is not affected by MW up to the cut-off value of the membrane. Continuous hemofiltration usually uses highly permeable membranes, with high cut-off values (20,000 – 50,000 Da), so the MW of antimicrobials will have little impact on drug sieving with hemofiltration. The capacity of a drug to pass through the membrane of a hemofilter is expressed mathematically as the sieving coefficient, which is the relation between drug concentration in the ultrafiltrate (Cuf ) and in the plasma (Cp): Sieving coefficient = Cuf / Cp

(Eqn 4)

For most antimicrobials, the sieving coefficient can be estimated by the extent of the unbound fraction (sieving coefficient » 1 – PPB). However, the sieving coefficient is a dynamic parameter and is dependent on the age of the membrane and on the filtration fraction (Quf/Qb). There are two basic dilutional modes (predilution and postdilution) for the substitution fluid which may influence the efficiency of solute removal. In the postdilution mode, the convective clearance (Clpost-HF) of an antimicrobial agent can thus be easily obtained by multiplying the ultrafiltration rate (Quf ) by its sieving coefficient: Clpost-HF = Quf × sieving coefficient

(Eqn 5)

However, if hemofiltration is used in predilution mode, the patient’s blood is diluted with a substitute fluid prior to entry into the dialyzer. So the correction of the predilutional effect must be integrated in the clearance equation: Clpre-HF = Quf × sieving coefficient × [Qb / (Qb + Quf )]

(Eqn 6)

As a newer technical evolution in CRRT, high volume hemofiltration (HVHF) or pulse-HVHF are increasingly used in critically ill patients in the ICU. In order to achieve the balance between greater solute clearance and fewer associated complications, such as circuit clotting, predilution and postdilution are often used simultaneously each with a certain percent. This makes it more complicated to calculate the drug clearance. Combination with diffusion and convection (hemodiafiltration) In hemodiafiltration, calculation of the drug clearance during this combination therapy is extremely difficult, especially at different Quf and Qd rates. Drug clearance with continuous veno-venous hemodiafiltration (CVVHDF) in postdilution may be estimated by calculating the convective clearance and diffusive clearance from the following equation: ClHDF = Quf × sieving coefficient + Qd × Sd

(Eq 7)

597

598

D. Kuang and C. Ronco

However, available data demonstrate that a greater overestimation will be induced if Sd is replaced by the unbound fraction, especially with a high Qd. Since an interaction between diffusive and convective solute transfer has been demonstrated in intermittent high-flux hemodiafiltration by protein layer formation on the blood side of the capillary, the two processes may interact in such a manner in CVVHDF that solute removal is significantly less than what is expected if the individual components are simply added together. In CVVHDF, as the presence of convection-derived solute in the dialysate decreases the concentration gradient, the driving force for diffusion, the Sd can be lowered even further. The diffusive clearance of a drug during CVVHDF is difficult to predict and will depend on its MW, Qb, Qd, Quf, and the membrane used. In continuous arteriovenous hemodiafiltration (CAVHDF), Vos and Vincent [12, 13] found a close exponential correlation of a drug’s diffusive mass transfer coefficient (Krel) through membranes: Krel = Kd / Kcr = (MW / 113)-0.42

(Eqn 8)

where Kd and Kcr are the diffusive mass transfer coefficients for the drug and creatinine, respectively, and 113 is the MW of creatinine. The drug clearance of CAVHDF (ClCAVHDF) may be estimated as: ClCAVHDF = Quf × sieving coefficient + Qd × Sd × Krel

(Eqn 9)

When using the above equation to estimate CVVHDF clearance, Kroh et al. [14] found very good correlations between observed and estimated clearances (y = 0.004 + 0.96x). However, whether this method is also suitable for all antimicrobial agents has not yet been investigated. Adsorption to membrane Adsorption to filter membranes leads to increased drug removal from plasma and the various filters have different absorptive capacities. Some dialysis membranes, such as polyacrylonitrile (PAN), may adsorb a substantial amount of drug to their surface. For example, PAN membranes are described to have a high adsorbent capacity to bind aminoglycosides and levofloxacin. However, adsorption is a saturant process, and the influence on drug removal will depend on the frequency of filter changes [10]. Although dosing adjustment will not account for adsorption effects, using drug-adsorbing membranes for CRRT is not usually recommended [15]. High volume CRRT (HV-CRRT) High volume CRRT (HV-CRRT), like HVHF, is increasingly used in septic patients with acute renal failure in the ICU. Nevertheless, the different effects of HV-CRRT and low volume CRRT (LV-CRRT) on the pharmacological characteristics of antibiotic removal have been understated [16]. Pharmacokinetic experiments have found that many antimicrobials exhibit two and three compartment characteristics. In standard LV-CRRT, the rate-limiting step of drug clearance has been Qd and/or Quf because Qb greatly exceeds Qd or Quf. Consequently, no appreciable rebound occurs after LV-CRRT stops because drugs transfer to the central compartment at least as fast as the drug is being removed by the CRRT. On initiation of HV-CRRT, the central compartment becomes rapidly stripped of unbound drug. The rate-limiting step of any further drug removal becomes the rate at which the drug can transfer from the peripheral compartments into the center compartment for removal by HVCRRT.

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy

Available data indicate that an increase in Quf from 14 ml/min to 28 ml/min will decrease the sieving coefficient for drugs like vancomycin by approximately 30 %. However, as Qd increased from 8.3 ml/min up to 33.3 ml/min, a 30 % decline in the sieving coefficient for vancomycin was seen with AN69 hemodiafilters. Doubling Qd from standard low-volume flows to higher dialysate flows may result in substantially less than a doubling of solute dialytic clearance, particularly for larger solutes. Increasing Qd (= 2,000 ml/hr) should result in decreasing Sd, but the rate of Sd decline is filter dependent. Therefore, the drug clearance calculation during HVCRRT is rather complex, and the change in sieving coefficient and Sd should be further considered.

Adjustment of Drug Regimen In patients with concomitant renal failure on CRRT, underdosing may lead to inadequate anti-infective therapy while overdosing may lead to unnecessary toxicity. Drug dosing adjustments during CRRT can be guided by using available drug-dosing recommendations, by measuring or estimating CRRT drug clearance, or by monitoring drug serum concentrations. Available drug-dosing recommendations Drug-dosing recommendations for patients with acute renal failure being treated with CRRT have not kept pace with the advances in CRRT technology and the speedy development of newer antimicrobial agents. Nonetheless, published drugdosing recommendations for acute renal failure patients on CRRT are becoming available [7, 8, 11, 15, 17, 18]. We have searched and reviewed the recent clinical investigations and referred to some of these recommendations, then summarized the pharmacokinetic characteristics and dosing recommendations for 60 antimicrobials most commonly used in critically ill patients undergoing CRRT into a complete dosing guide (Table 1). However, all these recommendations have unavoidable, inherent shortcomings that influence their clinical practicability. First, all these dosing recommendations are based on low Quf and Qd with old dialysis membranes or hemofilters. Second, pharmacokinetic data are based on data obtained mainly from healthy persons or stable chronic kidney disease patients. Third, some recommendations were derived from CRRT conducted in arteriovenous mode. Fourth, the filters, Quf and Qd, and treatment time vary considerably among these recommendations. Finally, most of these recommendations are based on very limited clinical data. Further clinical data are urgently needed to support such extrapolations, and these recommendations should not supercede sound clinical judgment [18]. It is widely recognized that the extent of drug removal during CRRT in critically ill patients with acute renal failure is dependent on numerous factors involving the patient, the illness, the drug, and the operational mode of CRRT. These parameters vary widely among different patients, or even at different moments of time in the same patient. CRRT does not always yield a stable condition, as Qb and Quf are quite variable during the therapeutic process. Moreover, renal function and sepsis state may also improve during the course of disease with effective treatment. Therefore, it is extremely difficult and almost impossible to devise a comprehensive dosing guide for various antimicrobial agents that encompasses all of the potentially changing variables involved in CRRT for all patients. Therapy must be individualized to tailor to the needs of each patient.

599

600

D. Kuang and C. Ronco Table 1. Adjustment of antimicrobial regimen in patients with acute renal failure undergoing continuous renal replacement therapy Drug

MW (Da)

PPB

Vd (l/kg)

T1/2normal (h)

T1/2anuria (h)

Normal Dosage

Dosage adjustment on CRRT

0.25 – 0.4

2.0 – 3.0

30 – 90

7.5 mg/kg q12h

7.5 mg/kg q24h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Aminoglycoside Antibiotics Amikacin

585.6

0 – 11 %

Gentamycin

477.6

‹

5%

0.26 – 0.4

2.0 – 3.0

20 – 60

1.7 mg/kg q8h

2.0 mg/kg q24h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Netilmicin

475.6

‹

5%

0.25 – 0.4

2.0 – 3.0

35 – 70

2.0 mg/kg q8h

2.0 mg/kg q12h (CVVHF: Qd 0.5 – 1.8 l/h, Quf 100 – 400 ml/h, predilution, 0.6 m2 AN69)

Tobramycin

467.5

‹

5%

0.26 – 0.4

2.0 – 3.0

30 – 60

1.7 mg/kg q8h

2.0 mg/kg q24h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

13 – 21 %

0.23

1

4

0.25 – 1.0 g 0.5 g q6-q12h (CVVHD/CVVH/ q6h CVVHDF: Qd 1 l/h, Quf 1 l/h, 0.9 m2 AN69)

2%

0.35

1

7

0.5 – 1.0 g q6h

Carbapenem Antibiotics Imipenem

299.3

Meropenem 383.5

1.0 g q12h (CVVHD: Qd 1 l/h) 1.0 g q12h (CVVH: Quf 1 – 2 l/h, postdilution, 0.9 m2 AN69) 1.0 g q8h (CVVH: Quf 2.6 l/h, postdilution, 0.43 m2 high-flux PS) 1.0 g q12h (CVVHDF: Qd 1 – 1.5 l/h, Quf 1 – 1.5 l/h, pre-/postdilution, 0.9 m2 AN69)

Cephalosporin Antibiotics 23.50 %

0.24 – 0.35

1

3

250 – 500 mg 500 mg q8 – 12h (CVVHD/CVVH: q8h Qd 1.5 l/h, Quf 1.5 l/h)

Cefamandole 475.6

75 %

0.16 – 0.25

0.5

6.0 – 11

0.5 – 1.0 g q4 – 8h

1.0 g q18h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Cefazolin

84 %

0.13 – 0.22

2

40 – 70

0.5 – 1.5 g q6h

2.0 g q12h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 l/h)

Cefaclor

367.8

454.5

1.0 – 2.0 g q12h (CVVH: Quf 1 – 2 l/h) 1.0 g q8h (CVVH: Quf 3 l/h) ‹

0.71

4.6

8.1

0.25 – 2.0 g 1.0 – 2.0 g q12h (CVVHD/CVVH/ q8h CVVHDF: Qd 1 l/h, Quf 1 l/h, postdilution, 0.6 m2 AN69)

45 – 75 %

0.27 – 0.37

1

6.0 – 12

1.0 g q6h

1.0 g q24h (CVVH: Quf 1 l/h)

645.7

90 %

0.14

2

3

1.0 – 2.0 g q12h

1.0 – 2.0 g q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

455.5

37 %

0.35

2

15 – 35

1.0 g q6h

2.0 g q12h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 l/h)

Cefepime

480.6

Cefmenoxime

511.6

Cefoperazone Cefotaxime

20 %

1.0 g q6 – 8h (CVVH: Quf 1 – 2 l/h) 2.0 g q8h (CVVH: Quf 3 l/h) Cefoxitin

427.4

Cefpirome

512

Cefradine

349.4

40.75 % ‹

10 %

8 – 17 %

0.31

1

13 – 23

1.0 – 2.0 g q6 – 8h

1.0 g q18h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

0.32

2

14.5

2.0 g q12h

2.0 g q8h (CVVH: Quf 3 l/h, postdilution, 0.7 m2 high-flux PS)

0.25 – 0.46

0.7 – 1.3

6.0 – 15

1.0 – 2.0 g q6h

1.0 g q12h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy Table 1. (cont.) Drug

MW (Da)

PPB

Ceftazidime

546.6 17 %

Vd (l/kg)

T1/2normal T1/2anuria (h) Normal Dosage (h)

Dosage adjustment on CRRT

0.28

2

2.0 g q12h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 l/h)

13 – 25

1.0 – 2.0 g q8h

1.0 – 2.0 g q12h (CVVH: Quf 1 l/h) 2.0 g q8h (CVVH: Quf 3 l/h, postdilution, 0.7 m2 high-flux PS) Ceftriaxone

554.6 95 %

0.12 – 0.18 6.0 – 9.0 12.0 – 24.0 0.5 – 1.0 g q12h

2.0 g q12 – 24h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Cefuroxime

424.4 50 %

0.19

1.5

17

0.75 – 1.5 g 0.5 g q8h (CVVHD/CVVH: Qd 1.5 l/h, q8h Quf 1.5 l/h)

Cephalexin

347.4 14 %

0.35

1

16

250 – 500 mg 0.5 g q12h (CVVHD/CVVH: Qd 1.5 l/h, q6h Quf 1.5 l/h)

4.4

8.7

400 mg q12h

Fluoroquinolone Antibiotics Ciprofloxacin 331.3 20 – 40 % 1.9 – 2.8

200 mg q8 – 12h (CVVHD: Qd 1 – 2 l/h, 0.43 m2 AN69) 200 mg q12h (CVVH: Quf 1 l/h, postdilution, 0.6 m2 AN69) 200 mg q12h (CVVHDF: Qd 1 l/h, Quf 1 l/h, postdilution, 0.6 m2 AN69) 200 mg q8h (CVVHDF: Qd 1 l/h, Quf 2 l/h, predilution, 0.6 m2 AN69)

Enoxacin

320.3 40 %

Levofloxacin 361

1.6

3.0 – 6.0 15 – 25

24 – 38 % 1.09 – 1.26 6.3

Moxifloxacin 401.4 47 %

3.3

12

400 mg q12h

400 mg q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

76

500 – 750 mg 250 mg q24h (CVVHD/CVVH/ q24h CVVHDF: Qd 1 l/h, Quf 1 l/h, postdilution, 0.6 m2 AN69)

12

400 mg q24h

400 mg q 24h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h, pre-/postdilution, 0.6 m2 AN69)

Ofloxacin

361.4 20 – 25 % 1.5 – 2.5

4.0 – 7.0 40 – 50

Pefloxacin

333.4 20 – 30 % 1.8

8.6

12.0 – 15.0 400 – 800 mg 400 – 800 mg q24h (CVVHD/CVVH: q24h Qd 1.5 l/h, Quf 1.5 l/h)

0.9

1.5

6

150 – 300 mg 250 – 500 mg q12 – 24h (CVVHD/ q6h CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

0.25

2

6.0 – 8.0

1.0 – 2.0 g q8 – 12h

200 – 400 mg 400 mg q24h (CVVH: Quf 3 l/h, q12h postdilution, 0.7 m2 high-flux PS)

Macrolide Antibiotics Erythromycin 734

84 %

Miscellaneous Antibiotics Aztreonam

435.4 55 %

2.0 g q12h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 l/h) 1.0 g q8h (CVVH: Quf 1 l/h) 2.0 g q12h (CVVH: Quf 2 l/h) 2.0 g q8h (CVVH: Quf 3 l/h)

Chloramphenicol

323.1 53 %

0.9

4

3.0 – 7.0

12.5 mg/kg 12.5 mg/kg q6h (CVVHD/CVVH: q6h Qd 1.5 l/h, Quf 1.5 l/h)

Clindamycin 425

60 – 95 % 0.7

2.5

4

150 – 300 mg 600 – 900 mg q8h (CVVHD/CVVH/ q6h CVVHDF: Qd 1 l/h, Quf 1 l/h)

Colistin

55 %

2

7.5

2.5 mg/kg q24hr

1750

0.34

2.5 mg/kg q48h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

601

602

D. Kuang and C. Ronco Table 1. (cont.) Drug

MW (Da)

Daptomycin

PPB

Vd (l/kg)

T1/2normal T1/2anuria (h) Normal Dosage (h)

Dosage adjustment on CRRT

1619.7 92 %

0.13

8

4 – 6 mg q24h

4 – 6 mg q48h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Linezolid

337.3

31 %

0.6 – 0.8

4.4 – 5.5 7.0 – 8.0

600 mg q12h

600 mg q12h (CVVHD/CVVH/ CVVHDF: Qd 1 – 2 l/h, Quf 1 – 2 l/h, pre-/postdilution)

Metronidazole

171.2

20 %

0.8

6.0 – 14

7.5 mg/kg q6h

7.5 mg/kg q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Teicoplanin

1879.7 8 90 %

400 mg q24h

200 mg q48h (CVVHD/CVVH/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Trimethoprim

290.3

Vancomycin

1449.3 10 – 55 % 0.64

29.3

7.0 – 21

0.34 – 0.89 30 – 140 157 – 567

30 – 50 % 1 – 2.2

11

20 – 50

100 – 200 mg 100 – 200 mg q12h (CVVHD/CVVH: q12h Qd 1.5 l/h, Quf 1.5 l/h)

6

200 – 250

500 mg q6h/1.0 g q12h

1.0 g q24h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 – 2 l/h) 1.0 g q48h (CVVH: Quf 1 – 1.5 l/h)

Penicillins 15 – 25 % 0.37

1

5.0 – 20

250 – 500 mg 1.0 g q12h (CVVHD/CVVH: Qd 1.5 l/h, q8h Quf 1.5 l/h)

349.4 Ampicillin (/Sulbactam 2:1)

20 %

1

7.0 – 20

1.5 – 3.0 g q6h

Azlocillin

20 – 46 % 0.29

1.3 – 1.5 6

2.0 – 3.0 g q4h

3.0 g q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Flucloxacillin 453.9

95 %

0.54

1

3

1.0 – 2.0 g q6 – 8h

2.0 g q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Mezlocillin

539.6

20 – 46

0.26

1.3

3.0 – 5.0

1.5 – 4.0 g q4 – 6h

2.0 g q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Nafcillin

414.5

85 %

0.35

1

2

1.0 – 2.0 g q4 – 6h

2.0 g q4 – 6h (CVVH/CVVHD/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Oxacillin

435.9

92 – 96 % 0.19 – 0.33 0.5

?

0.25 – 1.0 g 1.0 g q8h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h) q4 – 6h

Penicillin G

334.4

6 – 20 %

0.3

0.5

6.0 – 20

0.8 – 4.0 million U q4 – 6h

Piperacillin (/Tazobactam 8:1)

516.5

30 %

0.3

1

3.0 – 5.1

3.375 g q6h 2.25 g q6 – 8h (CVVHD, Qd 1 – 1.5 l/h, 0.9 m2 AN69)

Amoxicillin

365.4

461.5

0.22

3.0 g q8h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 l/h) 3.0 g q12h (CVVH: Quf 1 l/h)

2.0 million U q12h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

2.25 g q4 – 6h (CVVH/CVVHDF, Qd 1 l/h, Quf 1 – 2 l/h, postdilution, 0.7 m2 PS)

384.4 Ticarcillin (/Clavulanate 30:1)

45 – 60 % 0.14 – 0.22 2.2

11.0 – 17.0 3.1 g q4 – 6h

3.1 g q6h (CVVHD/CVVHDF: Qd 1 l/h, Quf 1 l/h) 2.0 g q6 – 8h (CVVH: Quf 1 l/h)

Tetracycline Antibiotics Doxycycline

8

444.4

90 %

0.75

15 – 20

18 – 25

100 mg q24h

100 mg q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

90 %

1.7 – 3.9

173

173

5 mg/kg q24h

3.0 – 5.0 mg/kg q24h (CVVHD/ CVVH/CVVHDF: Qd 1 l/h, Quf 1 – 2 l/h)

Antifungal Antibiotics Amphotericin B lipid complex

924.1

8

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy Table 1. (cont.) Drug

MW (Da)

PPB

Vd (l/kg) T1/2normal (h)

T1/2anuria (h)

Normal Dos- Dosage adjustment on CRRT age

Fluconazole

306.3

12 %

0.7

37

100

200 – 400 mg q24h

400 – 800 mg q24h (CVVHD/ CVVHDF: Qd 1 l/h, Quf 1 l/h)

Flucytosine

129.1

10 %

0.6

3.0 – 6.0

75 – 200

37.5 mg/kg q6h

37.5 mg/kg q12h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Itraconazole, 705.6 i.v.

99.80 %

10

21

25

100 – 200 mg 100 – 200 mg q12h (CVVHD/CVVH: q12h Qd 1.5 l/h, Quf 1.5 l/h)

Voriconazole, 349.3 i.v.

58 %

4.6

12

13.7

6 mg/kg q12h twice then 4 mg/ kg q12h

200 – 400 mg q24h (CVVH: Quf 1l/h) ‹

6 mg/kg q12h twice then 4 mg/kg q12h (CVVHDF: Qd 1 l/h, Quf 0.5 l/h, predilution, 0.9 m2 AN69)

Antituberculous Antibiotics Ethambutol

204.3

20 – 30

1.6

4

20

15 – 25 mg/ 10 – 15 mg/kg q24 – 48 h (CVVHD/ kg q24h CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Isoniazid

137.1

4 – 30 %

0.75

1.0 – 4.0

1.0 – 17

300 mg q24h

300 mg q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Rifampin

823

89 %

0.9

3.5

9

600 mg q24h

600 mg q24h (CVVHD/CVVH: Qd 1.5 l/h, Quf 1.5 l/h)

Acyclovir, i.v. 225.2

9 – 33 %

0.7

2.5

20

5.0 mg/kg q8h

5.0 – 7.5 mg/kg q24h (CVVHD/ CVVH/CVVHDF: Qd 1 l/h, Quf 1 l/h)

Amantadine 151.2

67 %

4.0 – 5.0 10.0 – 14.0 7 – 10 days 100 mg q12h

100 – 200 mg q60h (CVVHD/CVVH: Qd 1.5l/h, Quf 1.5 l/h)

Ganciclovir

1–2%

0.47

5.0 mg/kg q48h (CVVHD/CVVHDF: Qd 1 l/h, Quf 0.3 l/h, postdilution)

Antiviral Agents

256.2

3

30

5.0 mg/kg q12h

MW: molecular weight (Da); PPB: plasma protein binding ( %); Vd: apparent volume of distribution (l/kg); Quf: ultrafiltration rate; Qd: dialysate flow rate; T1/2normal: normal plasma half-life (h); T1/2anuria: plasma half-life in anuric nondialyzed patients (h); CRRT: continuous renal replacement therapy; CVVHD: continuous venovenous hemodialysis; CVVH: continuous venovenous hemofiltration; CVVHDF: continuous venovenous hemodiafiltration; i.v. intravenous; ?: data not available

Estimation by mathematical equation Making these estimations is time consuming, requiring a careful search for basic pharmacokinetic data. Drug clearance must be calculated to determine a maintenance dose. The serum concentration at steady state (Cpss) multiplied by the extracorporeal clearance (ClEC) provides the clinician with the amount of drug specifically removed by ultrafiltration per hour under steady-state conditions. Thereupon, one can calculate the amount of drug removed by CRRT (DEC) with the following equation: DEC = Cpss × ClEC × Tdur

(Eqn 10)

where Tdur is the duration of CRRT. The extracorporeal clearance can be calculated by equations 3, 5, 6, 7, and 9 according to the type of operational mode. The total amount of a drug during CRRT (D) may be calculated using the following equation including the typical anuric dose (Danur) in addition to DEC [19]: D = Danur + DEC = Danur + Cpss × ClEC × Tdur

(Eqn 11)

603

604

D. Kuang and C. Ronco

The drug dose during CRRT in an anuric patient may also be estimated from the following equation [15]: D = Danur × [1 + ClEC / ClNR / 2(Interval/Halflife)]

(Eqn 12)

where ClNR is the non-renal clearance, Halflife is the t1/2 of the drug in an anuric non-dialyzed patient, and Interval is the dose interval in an anuric non-dialyzed patient. At present, there is increasingly a tendency to start CRRT earlier in the course of illness, and renal replacement therapy may contribute to drug clearance. According to Dettli’s equation and the related investigation by Keller et al., the estimated dose during CRRT in a patient with RRT may be [5]: DEC = Dn × [Px + (1 – Px) × ClCRtot / ClCRn]

(Eq 13)

where Dn is the normal dose, ClN is the normal drug clearance, Px = ClNR/ClN, ClCRtot is the sum of renal and extracorporeal creatinine clearance, and ClCRn is the normal creatinine clearance. However, complex mathematical models have been proposed, but an accurate calculable equation remains unavailable, because drug dose data in patients with acute renal failure are rare and the calculation of drug clearance in various modalities of CRRT is also complicated. Most mathematical models have only been demonstrated to be suitable for certain drugs; their application in clinical practice is still limited. Whether it may be more appropriate to increase the drug dose or to shorten the dosing interval in critically ill patients during CRRT is dependent on the mechanisms of action and the kill characteristics of the various classes of antimicrobial agents. For concentration-dependent kill characteristic antimicrobial agents such as aminoglycosides, it is better to increase the drug dose because their antibiotic effects correlate with the Cmax. In contrast, for time-dependent kill characteristic antimicrobial agents such as q -lactam antibiotics, it is better to shorten the drug dosing interval because their antibiotic effects correlate with the T 8 MIC. The shorter dosing interval during CRRT may be estimated from the following equation: IvEC = Ivanu × [ClNR / (ClEC + ClNR)]

(Eqn 14)

where IvEC is the interval during CRRT and Ivanu is the interval in an anuric patient Drug serum concentration monitoring Not only are pharmacokinetics and pharmacodynamics often less predictable in critically ill patients, but it has not been sufficiently documented that doses may be accurately adjusted according to current drug-dosing recommendations or available mathematical equations. Therefore, monitoring of plasma concentrations is highly recommended whenever possible, and especially for drugs with a narrow therapeutic index, such as vancomycin and aminoglycosides. Although monitoring of drug concentrations is considered reasonable to enhance optimal dosing and minimize toxic side effects, it is not readily available for all medications. The following formula is often used to estimate the dose requiring (Dreq) to achieve the desired peak concentration (Cpeak) from the actual trough (or any) concentration (Cactual): Dreq = (Cpeak – Cactual) × Vd × Body weight

(Eqn 15)

Adjustment of Antimicrobial Regimen in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy

Conclusion Appropriate anti-infective therapy remains essential in decreasing the persistent high morbidity and mortality rates in the ICU. Considerable data are available to demonstrate that sepsis, acute renal failure, and CRRT may each have profound effects on the pharmacokinetic and pharmacodynamic characteristics of various antibmicrobial agents commonly used in the ICU. The extent of the alteration is dependent on multiple mechanical and drug factors during the treatment of sepsis. Understanding of these interactions, fundamental pharmacological principles, and drug clearance during CRRT is important to adjust antibiotic regimens in critically ill patients. Awareness of the kill characteristics of the antibiotic in question helps determine the optimum mode of administration. Meanwhile, monitoring the drug serum concentration is still mandatory whenever clinically feasible. More pharmacokinetic simulation modeling and clinical studies are needed to provide accurate guidance on the appropriate dosage adjustment under different circumstances. Acknowledgment: This work was made possible through funding of Dr Dingwei Kuang’s Fellowship by the International Society of Nephrology Fellowship.

References 1. Kellum JA, Bellomo R, Ronco C, Mehta R, Clark W, Levin NW (2005) The 3rd International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI). Int J Artif Organs 28:441 – 444 2. Roberts JA, Lipman J (2006) Antinacterial dosing in intensive care: pharmacokinetics, degree of disease and pharmacodynamics of sepsis. Clin Pharmacokinet 45:755 – 773 3. Pinder M, Bellomo R, Lipman J (2002) Pharmarcological principles of antibiotic prescription in the criticallly ill. Anaesth Intensive Care 30:134 – 144 4. Nicolau DP (2003) Optimizing outcomes with antimicrobial therapy through pharmacodynamic profiling. J Infect Chemother 9:292 – 296 5. Keller F, Giehl M, Frankewitsch T, Zellner D (1995) Pharmacokinetics and drug dosage adjustment to renal impairment. Nephrol Dial Transplant 10:1516 – 1520 6. Sweetman SC (2004) Martindale: The Complete Drug Reference. Pharmaceutical Press, London 7. Aronoff GR, Berns JS, Brier ME, et al (1999) Drug Prescribing in Renal Failure: Dosing Guidelines for Adults, 4th Ed. American College of Physicians, Philadelphia 8. Ashley C, Currie A The Renal Drug Handbook, 2nd Ed. Radcliffe Medical Press, Oxford 9. Schetz M, Ferdinande P, Van den Berghe G, Verwaest C, Lauwers P (1995) Pharmacokinetics of continuous renal replacement therapy. Intensive Care Med 21:612 – 620 10. Bugge JF (2001) Pharmacokinetics and drug dosing adjustments during continuous venovenous hemofiltration or hemodiafiltration in critically ill patients. Acta Anaesthesiol Scand 45:929 – 934 11. Reetze-Bonorden P, Bohler J, Keller E (1993) Drug dosage in patients during continuous renal replacement therapy: pharmacokinetic and therapeutic considerations. Clin Pharmacokinet 24:162 – 179 12. Vos MC, Vincent HH (1991) Continuous arteriovenous hemodiafiltration: Predicting the clearance of drugs. Contrib Nephrol 93:143 – 145 13. Vincent HH, Vos MC, Akcahuseyin E, Goessons WH, van Duyl WA, Schalekamp MA (1993) Drug clearance by continuous haemodiafiltration (CAVHD). Analysis of sieving coefficients and mass transfer coefficients of diffusion. Blood Purif 11:99 – 107 14. Kroh UF, Holl TJ, Steinhausser W (1996) Management of drug dosing in continuous renal replacement therapy. Semin Dial 9:161 – 165 15. Bohler J, Donauer J, Keller F (1999) Pharmacokinetic principle during continuous renal replacement therapy: drugs and dosage. Kidney Int Suppl 72:s24-s28

605

606

D. Kuang and C. Ronco 16. Mueller BA, Pasko DA, Sowinski KM (2003) Higher renal replacement therapy dose delivery influences on drug therapy. Artif Organs 27:808 – 814 17. Kroh UF (1995) Drug administration in critically ill patients with acute renal failure. New Horiz 3:748 – 759 18. Trotman RL, Williamson JC, Shoemaker DM, Salzer WL (2005) Antibiotic dosing in critically ill adult patients receiving continuous renal replacement therapy. Clin Infect Dis 41:1159 – 1166 19. Scheetz MH, Scarsi KK, Ghossein C, Hurt KW, Zembower TR, Postelnick MJ (2006) Adjustment of antimicrobial dosages for continuous venovenous hemofiltration based on patientspecific information. Clin Infect Dis 42:436 – 437