Continuous-Infusion β-Lactam Antibiotics During Continuous Venovenous Hemofiltration for the Treatment of Resistant Gram-Negative Bacteria

Abstract

OBJECTIVE:

To describe the rationale, principles, and dosage calculations for continuous-infusion β-lactam antibiotics to treat multidrug-resistant bacteria in patients undergoing continuous venovenous hemofiltration (CVVH).

DATA SOURCES:

A MEDLINE search (1968–November 2008) of the English-language literature was performed using the terms continuous infusion and Pseudomonas or Acinetobacter, hemofiltration or CVVH or hemodiafiltration or CVVHDF or continuous renal replacement therapy or pharmacokinetics; and terms describing different β-lactam antibiotics.

STUDY SELECTION AND DATA EXTRACTION:

In vitro, in vivo, and human studies were evaluated that used continuous-infusion β-lactam antibiotics to treat Pseudomonas aeruginosa and Acinetobacter baumannii infections. Studies were reviewed that described the pharmacokinetics of β-lactam antibiotics during CVVH as well as other modalities of continuous renal replacement therapy.

DATA SYNTHESIS:

Continuous infusion of β-lactam antibiotics, maintaining drug concentrations 4–5 times higher than the minimum inhibitory concentration, is a promising approach for managing infections caused by P. aeruginosa and A. baumannii. Safe yet effective continuous infusion therapy is made difficult by the occurrence of acute renal failure and the need for renal replacement therapy. Case series and pharmacokinetic properties indicate that several β-lactam antimicrobials that have been studied for continuous infusion, such as cefepime, ceftazidime, piperacillin, ticarcillin, clavulanic acid, and tazobactam, are significantly cleared by hemofiltration. Methodology and formulas are provided that allow practitioners to calculate dosage regimens and reach target drug concentrations for continuous β-lactam antibiotic infusions during CVVH based on a literature review, pharmacokinetic principles, and our experience at the National Institutes of Health Clinical Center.

CONCLUSIONS:

Continuous infusion of β-lactam antibiotics may be a useful treatment strategy for multidrug-resistant gram-negative infections in the intensive care unit. Well-established pharmacokinetic and pharmacodynamic principles can be used to safely reach and maintain steady-state target concentrations of β-lactam antibiotics in critical illness complicated by acute renal failure requiring CVVH.

Recent years have witnessed a striking increase in the incidence of multidrug-resistant (MDR) gram-negative bacteria including Pseudomonas aeruginosa, Acinetobacter baumannii, and organisms that produce extended-spectrum β-lactamases, type 1 β-lactamases, and metaloenzymes.^1–3 Many of the infections caused by these bacteria commonly occur in intensive care units (ICUs) among critically ill patients who have a high risk of acute kidney injury. Strategies to treat these resistant bacterial infections have relied on new agents, the use of older antibiotics associated with greater toxicity, and administration of existing drugs in novel ways. Unfortunately, there has been a decline in research and development of new antibiotics with activity against resistant gram-negative bacteria.^1,2 The Infectious Diseases Society of America has highlighted these concerns in several position statements.^1,4 An alternative treatment strategy is exploiting innovative administration methods of existing agents, such as continuous infusion, to maximize the pharmacodynamics of β-lactam antibiotics.

The objective of this review is to provide a rational approach to dosing continuous-infusion β-lactam antibiotics in patients undergoing continuous venovenous hemofiltration (CVVH) for the treatment of MDR gram-negative bacteria. We focus on P. aeruginosa and Acinetobacter spp. because these organisms have emerged as life-threatening pathogens with the potential for resistance against multiple classes of antimicrobial agents. Although other gram-negative pathogens, such as extended-spectrum β-lactamase–producing Escherichia coli and type 1 β-lactamase–producing Enterobacter spp., are also important, multidrug resistance is less frequent and there is a paucity of data on continuous infusion of β-lactam antibiotics for these organisms. While there is limited published literature, recommendations are based on well-established pharmacokinetic and pharmacodynamic principles of both β-lactam antibiotics and of hemofiltration. We have used these methods to treat several patients infected with MDR gram-negative bacteria in the medical ICU of the National Institutes of Health Clinical Center. The following case illustrates this method.

Clinical Scenario

A 35-year-old male with a history of severe aplastic anemia and profound neutropenia developed septic shock and was transferred to the ICU. The patient developed acute respiratory failure requiring intubation and acute renal failure, with urine output 5 mL/h or less. The microbiology laboratory reported growth of P. aeruginosa from both blood and a tracheal aspirate culture. The isolates were intermediately resistant to ceftazidime (minimum inhibitory concentration [MIC] = 16 mg/L) and amikacin (MIC = 32 mg/L) but fully resistant to piperacillin, ticarcillin, imipenem, meropenem, aztreonam, ciprofloxacin, levofloxacin, gentamicin, and tobramycin. The patient was started on CVVH with the following settings: AN69 M100 filter, blood flow rate 200 mL/min, predilution replacement fluid rate 2100 mL/h, and ultrafiltration rate 2100 mL/h, with no net fluid removal. The patient’s weight was 70 kg and height was 170 cm. The infectious diseases consultants recommend amikacin, colistin, and continuous-infusion ceftazidime. They page you, the ICU pharmacist, for drug dosing recommendations for continuous-infusion ceftazidime. What do you recommend? (Recommendation from clinical scenario is shown in Appendix I.)

Overview of Continuous-Infusion β-Lactam Antibiotics

PRINCIPLES AND THEORY

β-Lactam pharmacodynamics are described by time-dependent killing (ie, bacterial killing is dependent on the amount of time that the drug concentration is above the MIC of the causative bacteria).^5,6 Time-kill studies in aggregate indicate that maximal bactericidal activity occurs for most β-lactam antibiotics when drug concentrations are 4 times the MIC, with no additional effect at higher concentrations.^7–11 Continuous-infusion β-lactam antibiotics can ensure time above MIC within the range of safely achievable concentrations for the entire dosing interval. This may be particularly critical for the treatment of MDR bacteria, as intermittent dosing may allow drug concentrations to fall below the MIC of the organism, thereby permitting survival and regrowth. While continuous infusion of β-lactam antibiotics has been shown to be effective in a number of clinical trials,^12,13 there are limited data in patients receiving continuous renal replacement therapy (CRRT) and for the treatment of infections caused by P. aeruginosa and Acinetobacter spp. The following sections summarize the pharmacodynamic rationale of continuous infusion of β-lactam antibiotics for the treatment of P. aeruginosa infections. This rationale is based upon in vitro and in vivo studies as well as the current literature for treatment of P. aeruginosa infections in humans. Although in vitro, in vivo, and clinical data for Acinetobacter spp. are more limited, similar concepts may also be applied for the treatment of these organisms with continuous-infusion β-lactam antibiotics.

FORMULAS

A continuous infusion dosing regimen for β-lactam antibiotics that follows linear pharmacokinetics can be calculated from the following equation^14,15:

$$\text{loading}\text{dose}\left(\mathrm{m}\mathrm{g}\right)={\mathrm{C}}_{\text{peak}}(\mathrm{m}\mathrm{g}/\mathrm{L})\times {\mathrm{V}}_{\mathrm{d}}(\mathrm{L}/\mathrm{k}\mathrm{g})\times \text{weight}\left(\mathrm{k}\mathrm{g}\right)$$ Eq. 1

$$\text{maintenance}\text{infusion}\text{rate}(\mathrm{m}\mathrm{g}/\mathrm{h})=\mathrm{Css}(\mathrm{m}\mathrm{g}/\mathrm{L})\times {\mathrm{C}\mathrm{l}}_{\text{total}}(\mathrm{L}/\mathrm{h})$$ Eq. 2

$${\mathrm{C}\mathrm{l}}_{\text{total}}(\mathrm{L}/\mathrm{h})={\mathrm{K}}_{\mathrm{e}}\left({\mathrm{h}}^{-1}\right)\times {\mathrm{V}}_{\mathrm{d}}(\mathrm{L}/\mathrm{k}\mathrm{g})\times \text{weight}\left(\mathrm{k}\mathrm{g}\right)$$ Eq. 3

$${\mathrm{K}}_{\mathrm{e}}\left({\mathrm{h}}^{-1}\right)=0.693/{\mathrm{t}}_{\mathrm{1/2}}\left(\mathrm{h}\right)$$ Eq. 4

where ${\mathrm{C}}_{\text{peak}}=$ target peak concentration; $\mathrm{C}\mathrm{s}\mathrm{s}$ = target mean steady-state concentration; ${\mathrm{C}\mathrm{l}}_{\text{total}}=$ total body clearance; ${\mathrm{K}}_{\mathrm{e}}=$ elimination rate constant $;{\mathrm{t}}_{1/2}=$ half-life; and ${\mathrm{V}}_{\mathrm{d}}=$ volume of distribution.

The dosing regimen consists of an initial loading dose to rapidly achieve therapeutic concentrations (Equation 1). The loading dose is calculated from the ${\mathrm{V}}_{\mathrm{d}}$ and ${\mathrm{C}}_{\text{peak}}$ of the β-lactam antibiotic. The continuous infusion rate, which maintains the target concentration, is calculated from the Css and the ${\mathrm{C}\mathrm{l}}_{\text{total}}$ of the drug (Equation 2).

PSEUDOMONAS SPP. AND ACINETOBACTER SPP.

In Vitro and In Vivo Studies

Several in vitro and in vivo studies of continuous-infusion β-lactam antibiotics suggest that maximal bactericidal activity against P. aeruginosa occurs when drug concentrations are 4–5 times the MIC (Table 1).^16–23 An in vitro study by Mouton et al.¹⁶ compared intermittent with continuous infusion of ceftazidime for the treatment of P. aeruginosa infections. At 32 hours for strains with MICs of 1 mg/L and 4 mg/L, respectively, bacterial concentrations decreased approximately 10¹ cfu/mL and 10¹ cfu/mL with intermittent dosing and 10³ cfu/mL and 10⁴ cfu/mL with continuous infusion. The authors concluded that continuous infusion of ceftazidime was more effective than intermittent dosing if concentrations were maintained at greater than or equal to 4 times MIC. Tessier et al.¹⁸ also suggested a target concentration of 4 times MIC for continuous-infusion cefepime. In this study there was a reduction of approximately 10² cfu/mL for the P. aeruginosa strains treated with continuous-infusion cefepime alone at 24 hours. Other studies are summarized in Table 1.^17,19

Table 1.

Selected In Vitro and In Vivo Studies of Continuous-Infusion β-Lactam Antibiotics for the Treatment of P. aeruginosa and A. baumannii

Microorganism	Drug (model)	MICa (mg/L)	Dosage Regimen	Drug Concentrations	Conclusion	Target Drug Concentration as Multiple of MICb	Reference
P. aeruginosa	ceftazidime (in vitro, dialyzer unit model)	ceftazidime: 1, 4, 16	300 mg/L/24 h by intermittent dosing (every 8 h) or continuous infusion for 36 h	intermittent dosing peak:  92.3 mg/L  trough: 1.4 mg/L continuous infusion 19.8 mg/L   concentrations at 24 h	continuous infusion more effective (in reduction log10 cfu/mL) than intermittent dosing with concentrations ≥4 times MIC	ceftazidime 4–5 times MIC	16
P. aeruginosa	ceftazidime ± amikacin (in vitro pharmacodynamic model)	ceftazidime: 1.56, 50	simulated dosing of ceftazidime: 2 g iv q12h or q8h or continuous infusion (loading dose 2 g, then concentrations of 5 mg/L, 10 mg/L, or 20 mg/L) for 48 h simulated dosing of amikacin: 15 mg/kg/day	ceftazidime intermittent  dosing troughs:   2 g hr every 12 h: 2.6 mg/L   2 g iv every 8 h: 9.8 mg/L ceftazidime continuous infusion  expected: observed   5 mg/L: 5.6 mg/L   10 mg/L: 11.9 mg/L   20 mg/L: 30.8 mg/L (concentrations in central compartment at 24 h)	continuous infusion 10 and 20 mg/L as effective (in reduction log10 cfu/mL) as 2 g iv every 12 h and 2 g iv every 8 h, respectively, for sensitive strain amikacin synergistic with all regimens with sensitive strain and synergistic with ceftazidime 20 mg/L or 2 g iv q12 or q8h with resistant strain	ceftazidime 10–20 times MIC	17
P. aeruginosa	cefepime ± tobramycin (in vitro pharmacodynamic model)	cefepime: 2.8	simulated dosing of cefepime: 1 g iv q12h or continuous infusion (loading dose 1 g, then 2 g/24 h) for 48 h simulated dosing of once-daily tobramycin: peak 10 mg/L	intermittent dosing: peak 106.6 mg/L continuous infusion: 11.48 mg/L	continuous infusion more effective (in reduction log10 cfu/mL) than intermittent dosing addition of tobramycin to continuous infusion cefepime increased bactericidal activity	cefepime 4 times MIC	18
P. aeruginosa	ceftazidime (in vitro, computer-controlled model)	ceftazidime: 8 16 (N = 2) 32	simulated dosing of 2 g iv q8h or continuous infusion (loading dose 1 g, then 6 g/24 h) for 32 h	intermittent dosing:  peak: 119.97 mg/L  trough: 9.17 mg/L continuous infusion:  40.38 mg/L	continuous infusion and intermittent dosing produced ≥103 reduction in cfu/mL up to 32 h for strains with MIC 8 and 16 mg/L		19
A. baumannii	CMS ± ceftazidime (in vitro pharmacodynamic model)	colistin sulfate: 0.5 ceftazidime: ≥64	regimen 1: CMS bolus for concentrations of 3, 6, 12, or 24 mg/L at time zero; in 1 experiment, additional 24 mg/L CMS bolus at 12 h regimen 2: CMS bolus of 24 mg/L + continuous-infusion ceftazidime (concentration 50 mg/L) for 24 h; 3 experiments with different antibiotic starting times:  CMS 0 h, ceftazidime 2 h  CMS 2 h, ceftazidime 0 h  ceftazidime 0 h	not determined	CMS bolus with continuous-infusion ceftazidime produced a 103 reduction in cfu/mL, prevented bacterial regrowth, and prevented the development of resistance to colistin sulfate		20
P. aeruginosa	ceftazidime, amikacin, sulbactam (in vivo, rabbit endocarditis model)	ceftazidime: 16 sulbactam: >256	continuous infusion over 24 h starting 12 h or 48 h after infection:  ceftazidime 800 mg/kg, amikacin 400 mg/kg  ceftazidime 800 mg/kg + amikacin 400 mg/kg  ceftazidime 800 mg/kg + sulbactam 400 mg/kg  ceftazidime 800 mg/kg + amikacin 400 mg/kg + sulbactam 400 mg/kg	ceftazidime continuous infusion: 127 mg/L	continuous-infusion ceftazidime effective (101 decrease in cfu/g) when started 12 h after infection continuous-infusion ceftazidime effective when started 48 h after infection only with combination therapy with amikacin, sulbactam, or both drugs		21
P. aeruginosa	ceftazidime ± amikacin (in vivo, rabbit endocarditis model)	ceftazidime: 1 8 (oxacillinase producer) 4 (penicillinase producer) 8 (cephalosporinase producer)	simulated dosing of ceftazidime: 2 g iv every 8 h or continuous infusion (4, 6, or 8 g/24 h) for 24 h amikacin 15 mg/kg/day	ceftazidime intermittent dosing  peak: 160 mg/L ceftazidime continuous infusion  4 g: 22.7 mg/L  6g: 34.8 mg/L  8 g: 79.6 mg/L	continuous-infusion effective as intermittent dosing with concentrations ≥4 times MIC, but efficacy depends on bacterial strain no effect with addition of amikacin to continuous-infusion ceftazidime	ceftazidime 4–5 times MIC	22
P. aeruginosa	imipenem or cefepime ± tobramycin (in vivo, rabbit endocarditis model)	imipenem: 2 cefepime: 1	continuous-infusion imipenem (10, 15, 25, 100 mg/kg/day) over 24 h continuous-infusion cefepime (10, 25, 40, 100 mg/kg/day) over 24 h simulated dosing of tobramycin 3 mg/kg/day	imipenem continuous infusionc   15 mg/kg/day: 0.5 mg/L   25 mg/kg/day: 0.8 mg/L   100 mg/kg/day: 2.3 mg/L cefepime continuous infusion  10 mg/kg/day: 1.6 mg/L  25 mg/kg/day: 3.8 mg/L  40 mg/kg/day: 5.7 mg/L  100 mg/kg/day: 15.1 mg/L	lowest effective steady-state concentration (102 decrease in cfu/g) with concentrations 3–4 times MIC with cefepime and 0.25 times MIC for imipenem no effect with addition of tobramycin	cefepime 4–6 times MIC	23

CMS = colistin methanesulfonate; MIC = minimum inhibitory concentration.

^aUnless otherwise noted, each study used one strain.

^bAuthor’s conclusion, see reference.

^cThe clinical significance and mechanism of subinhibitory concentrations of imipenem in this study are uncertain.

Ceftazidime intermittent infusion was compared with continuous infusion with or without amikacin in a rabbit model of P. aeruginosa endocarditis.²² Continuous infusion was as effective as intermittent dosing if concentrations were 4 or more times MIC, and there was no additional effect of amikacin. The authors recommended a target concentration of 4–5 times MIC when using continuous-infusion ceftazidime. In another study of P. aeruginosa endocarditis in rabbits, Navas et al.²³ compared the effect of continuous-infusion imipenem or cefepime with or without tobramycin. A decrease in 10² cfu/g of P. aeruginosa within the vegetations occurred with concentrations 3–4 times MIC for cefepime and 0.25 times MIC for imipenem, with no additional effect of tobramycin. However, the clinical significance and mechanism of subinhibitory concentrations of imipenem in this study are uncertain. Cefepime concentrations 4–6 times MIC were suggested as a reasonable goal.

A critical issue when treating patients with continuous-infusion β-lactam antibiotics is whether concentrations 4–5 times the MIC are actually bactericidal against intermediate and resistant strains of P. aeruginosa. The bactericidal activity of ceftazidime against sensitive, intermediate, and resistant strains of P. aeruginosa has been studied in a time-kill assay.⁷ Reductions of greater than or equal to 10³ cfu/mL occurred in 11 of 13 strains and 12 of 13 strains exposed to ceftazidime at 4 times and 8 times MIC, respectively. There was no increase in bactericidal activity at concentrations 8 times MIC. Bactericidal activity against intermediate strains of P. aeruginosa has also been shown in vitro with continuous-infusion ceftazidime. In a study by Alou et al.,¹⁹ continuous-infusion ceftazidime (mean steady-state concentration 40.38 mg/L) produced reductions greater than or equal to 10³ cfu/mL in 2 strains of P. aeruginosa (MIC = 16 mg/L). However, the effectiveness of continuous-infusion β-lactam antibiotic treatment may depend on factors other than the MIC of the bacterial strain and therefore may not be entirely dependent on just achieving appropriate drug concentrations. In the study by Robaux et al.,²² continuous-infusion ceftazidime was not bactericidal at concentrations up to 79.6 mg/L against a cephalosporinase-producing strain of P. aeruginosa (MIC = 8 mg/L). Likewise, in a time-kill study by Xiong et al.,²¹ ceftazidime (concentration 64 mg/L) failed to achieve bactericidal activity against a cephalosporinase-producing strain of P. aeruginosa (MIC = 16 mg/L). In a rabbit model of endocarditis, this same strain of P. aeruginosa was reduced by only 10¹ cfu/g in vegetations by continuous-infusion ceftazidime (concentration of 127 mg/L) started 12 hours after infection.²¹

In comparison with continuous-infusion β-lactam antibiotic studies with P. aeruginosa, in vitro and in vivo studies with A. baumannii are limited (Table 1). In a study by Kroeger et al.,²⁰ the combination of colistin methanesulfonate and continuous-infusion ceftazidime produced a 10³ cfu/mL reduction in bacterial load and prevented regrowth of A. baumannii for 24 hours.

Clinical Studies

Only a few studies have been published in which continuous-infusion β-lactam antibiotics were used for the treatment of infections caused by P. aeruginosa (Table 2).^24–31 Continuous-infusion ceftazidime appears to be effective for treating pulmonary infections in cystic fibrosis and skin lesions in neutropenic patients.^24–28 In addition, one case report described the use of continuous-infusion meropenem for the treatment of MDR P. aeruginosa pneumonia.²⁹

Table 2.

Selected Studies of Continuous-Infusion β-Lactam Antibiotics for the Treatment of P. aeruginosa and A. baumannii in Humans

Microorganism	Drug	MIC (mg/L)	Dosage Regimen	Drug Concentrations	Pts.	Type of Infection	Conclusiona	Reference
P. aeruginosa	ceftazidime	NR	pt. 1  LD: 2 g iv  MD: 6 g iv per day infused over 24 h pt.2  LD: 2 g iv  MD: NR	NR	30-y-old male, aplastic anemia, neutropenia	persistent skin lesions following bacteremia	continuous-infusion ceftazidime may be more effective than intermittent dosing in pseudomonal infections in granulocytopenic pts.	24
				28 mg/L	46-y-old male, leukemia, neutropenia
P. aeruginosa	ceftazidime	NR	regimen A  LD: 10 mg/kg iv  MD: 4.5 mg/kg/h for 7 days regimen B  LD: 7.5 mg/kg iv  MD: 3.4 mg/kg/h for 7 days	regimen A (mean) 38.3 mg/L, (n = 3) regimen B (mean) 21.3 mg/L (n = 2) concentrations on day 7, at 8–9 h	cystic fibrosis, 9–25 y old, (n = 6)	pulmonary infection	pts. improved on continuous-infusion ceftazidime	25
P. aeruginosa	ceftazidime	NR	101.5 mg/kg/day infused over 24 h (average dose)	28.4 mg/L (mean) (n = 10)	cystic fibrosis, 15–52 y old (N = 12)	exacerbation of acute lower respiratory tract	home iv therapy with continuous-infusion ceftazidime clinically effective	26
P. aeruginosa	ceftazidime + tobramycin	1.5–6	ceftazidime 50 mg/kg iv every 8 h (maximum 6 g) for 10 days ceftazidime continuous infusion 6.6 times MIC (maximum 6 g, average 78 mg/kg/day) for 10 days	NR	cystic fibrosis, 19–32 y old (N = 5)	exacerbation of acute lower respiratory tract	continuous infusion as effective as intermittent dosing with changes in WBC count, sputum density, and pulmonary function test	27
P. aeruginosa	ceftazidime + amikacin	0.5–4	ceftazidime 200 mg/kg/day 3 divided doses for 14 days ceftazidime continuous infusion 100 mg/kg/day × 14 days amikacin 20 mg/kg/day	ceftazidime intermittent dosing trough (mean) 6.1 mg/L ceftazidime continuous infusion (mean) day 3: 29.7 mg/L day 10: 27.4 mg/L	cystic fibrosis 5–16.8 y old (N = 14)	chronic pulmonary colonization	continuous infusion as effective as intermittent dosing with changes in pulmonary, inflammatory, and nutritional status	28
P. aeruginosa	meropenem	32	LD: 2 g iv MD: 8 g iv per day infused over 24 h	NR	58-y-old male, emphysema, double lung transplant	pneumonia	MDR P. aeruginosa pneumonia successfully treated with continuous-infusion meropenem	29
P. aeruginosa	aztreonam + tobramycin	4	aztreonam continuous infusion 200 mg/kg/day tobramycin 5 mg/kg iv every 12 h	aztreonam continuous infusion 160 mg/L	cystic fibrosis, 3-mo-old female	tracheobronchial colonization	P. aeruginosa successfully eradicated with continuous-infusion aztreonam and tobramycin	30
A. baumannii	imipenem + amikacin	NR	imipenem 1 g iv per day infused over 24 h amikacin 500 mg iv per day	NR	40-y-old male, motorcycle accident	pneumonia	MDR A. baumannii pneumonia successfully treated with continuous-infusion imipenem after rapid desensitization procedure	31

LD = loading dose; MD = maintenance dose; MDR = multidrug resistant; MIC = minimum inhibitory concentration; NR = not reported; WBC = white blood cell.

^aAuthor’s conclusion; see reference.

Only a single case report described the use of continuous-infusion β-lactam antibiotics for the treatment of infections caused by A. baumannii (Table 2).³¹ Continuous-infusion imipenem was reported to successfully treat A. baumannii ventilator-associated pneumonia. However, that regimen was employed to prevent allergic reactions after a rapid desensitization protocol and not specifically to provide a constant level of bactericidal activity.

Overview of Continuous Renal Replacement Therapy

A number of excellent articles on drug dosing in CRRT have been published; we recommend that the reader consult these references for a more in-depth review.^32–43 Here, we briefly describe the different modes of CRRT and review critical concepts related to drug dosing in hemofiltration.

The common modes of CRRT include slow continuous ultrafiltration, CVVH, continuous venovenous hemodialysis (CVVHD), and continuous venovenous hemodiafiltration (CVVHDF).^44,45 Slow continuous ultrafiltration and CVVH are forms of hemofiltration, CVVHD is a form of hemodialysis, and CVVHDF is a combination of hemofiltration and hemodialysis. In hemofiltration, solutes are removed by convection (ie, solutes move with water flow). In hemodialysis, solutes are removed by diffusion (ie, solutes move from an area of high concentration to an area of low concentration). Slow continuous ultrafiltration is used only for fluid removal (solute removal is negligible), while CVVH, CVVHD, and CVVHDF are used to remove solutes as well as fluid.

A diagram of a typical CVVH system is illustrated in Figure 1.^{32,35,36,42,43} Both access and return lines are connected to a double-lumen catheter that, for adults, is commonly inserted into the femoral or internal jugular vein. Computerized external pumps control the blood flow rate and the replacement fluid rate, which can be given prefilter and/or postfilter. The ultrafiltration rate is the sum of the replacement fluid rate and any additional fluid removed from the patient.

Figure 1.

Schematic of a CVVH system.^{32,35,36,12,43}

1. $\mathrm{S}=\mathrm{C}\mathrm{u}\mathrm{f}/\mathrm{C}\mathrm{p}$

2. ${\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}=\mathrm{U}\mathrm{F}\mathrm{R}\times \mathrm{S}$ (if all replacement fluid is postfilter).

3. ${\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}=\mathrm{U}\mathrm{F}\mathrm{R}\times \mathrm{S}\times [\text{blood}\text{flow}\text{rate}/(\text{blood}\text{flow}\text{rate}+\text{replacement}\text{fluid}\text{rate})]\left(\text{if}\text{all}\text{replacement}\text{fluid}\text{is}\text{prefilter}\right)$.

4. ${\mathrm{C}\mathrm{l}}_{\text{total}}={\mathrm{C}\mathrm{l}}_{\text{renal}}+{\mathrm{C}\mathrm{l}}_{\text{nonrenal}}+{\mathrm{C}\mathrm{l}}_{\mathrm{E}\mathrm{C}}$.

5. ${\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}={\mathrm{C}\mathrm{l}}_{\mathrm{E}\mathrm{C}}/{\mathrm{C}\mathrm{l}}_{\text{total}}$.

$\mathrm{C}\mathrm{p}=$ drug concentration in plasma; $\mathrm{C}\mathrm{u}\mathrm{f}=$ drug concentration in ultrafiltrate; $\mathrm{C}\mathrm{V}\mathrm{V}\mathrm{H}=$ continuous venovenous hemofiltration; ${\mathrm{C}\mathrm{l}}_{\mathrm{E}\mathrm{C}}=$ extracorporeal clearance; ${\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}=$ hemotiltration clearance; ${\mathrm{C}\mathrm{l}}_{\text{nonrenal}}=$ nonrenal clearance; ${\mathrm{C}\mathrm{l}}_{\text{renal}}=$ renal clearance; ${\mathrm{C}\mathrm{l}}_{\text{total}}=$ total body clearance; ${\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}=$ fractional extracorporeal clearance; $\mathrm{S}=$ sleving coefficient; UFR $=$ ultrafiltration rate.

Two important pharmacokinetic concepts related to drug dosing in hemofiltration include the sieving coefficient and hemofiltration clearance (Figure 1). The sieving coefficient is calculated by dividing the solute concentration in the ultrafiltrate (Cuf) by the solute concentration in plasma (Cp) (Figure 1, Equation 1). A sieving coefficient of 0 indicates that a drug does not pass across the hemofilter, while a sieving coefficient of 1 indicates that a drug freely passes across the hemofilter.^35,36,42 The primary factor that affects the sieving coefficient is protein binding and there is a reasonable correlation between the free fraction and the sieving coefficient of a drug.^{36–38,42,43} Hemofiltration clearance is calculated from the ultrafiltration rate (UFR) and the sieving coefficient (Figure 1; Equations 2 and 3). Clearance is affected by the location of fluid replacement, as prefilter administration dilutes the solute concentration entering the hemofilter, thereby decreasing solute clearance.⁴⁶ The correction factor for the effect of predilution on hemofiltration clearance is the extracorporeal blood flow rate divided by the blood flow rate plus the prefilter replacement fluid rate (Figure 1; Equation 3).

Hemofiltration clearance is thought to be clinically significant, requiring a change in maintenance dose if the fractional extracorporeal clearance $\left({\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}\right)$ of a drug is greater than 25–30% (Figure 1; Equation 5).⁴² The major drug properties affecting removal during hemofiltration include ${\mathrm{V}}_{\mathrm{d}}$, protein binding, and molecular weight. A ${\mathrm{V}}_{\mathrm{d}}$ less than 1 L/kg, protein binding less than 70–80%, and molecular weight below the cutoff of the hemofilter suggest that a drug will be removed by hemofiltration.

Continuous-Infusion β-Lactam Antibiotics and CVVH

PHARMACOKINETICS

The pharmacokinetic properties of selected β-lactam antibiotics are summarized in Table 3.^47–69 In general, these agents have low molecular weights, ranging from 237.3 to 636.6 daltons, low ${\mathrm{V}}_{\mathrm{d}}$ around 0.3–0.4 L/kg, and relatively low protein binding varying between 10% and 68%. With the exception of ampicillin, aztreonam, and penicillin G, these antibiotics are excreted primarily unchanged by the kidney via glomerular filtration (GFR) with a small, clinically insignificant component of tubular secretion.

Table 3.

Pharmacokinetic Properties of Selected β-Lactam Antibiotics in Healthy Volunteers and Patients with Renal Failure

Drug	Molecular Weight (daltons)	Vd (L/kg)	Protein Binding (%)	t1/2 (h)	Percent Excreted Unchanged by Kidney	Cltotal (mL/min)	Clrenal (mL/min)	Clnonrenal (mL/min)	Notes	Assays Lab/Type	Suggested Maximum Tolerated Concentrations (mg/L)
Ampicillin	371.3947	0.28–0.3348	2847	147	7549		203–31948		glomerular filtration and tubular secretion48	ampicillin	ampicillin
						218.650	158.550	6050		Focus BA	100
						3150,a	0.3850,a	30.650,a		Mayo HPLC
Sulbactam	255.2247	0.24–0.448	3847	147	8949		169–20448		glomerular filtration and tubular secretion48
					75.551	216.650	167.750	5050
						266.351	203.851	65.351
						45.350,a	0.550,a	44.850,a
Aztreonam	435.4452	12.6L52	5652	1.752	6853	9152	5652		glomerular filtration and tubular secretion49	Focus BA	128
		0.2254				10754		26.7554
						2954,b
Cefepime	571.5155	18L56	2056	256	82.957	12056			glomerular filtration49	Mayo HPLC	100
		0.2858				13157	11057	20.257
						18.757,a	4.2257,a	16.957,a
Ceftazidime	636.659	0.18–0.3148	<1059	1.959	80–9059	11559	10059		glomerular filtration49	Focus BA	128
		0.2360				130 mL/min/1,73 m2 60	103.9 mL/min/1,73 m2 60
						6.8 mL/min/1.73 m2 60,a
Penicillin Gd	372.4861	0.53–0.6748	45–6848	0.4–0.948	60–8562	42.763,a	not widely reported	not widely reported	glomerular filtration and tubular secretion61	Focus BA	20
										Mayo HPLC
Piperacillin	539.564	0.14–0.3148	3064	0.7–1.264	6864	153–297 mL/min/1.73 m2 48			glomerular filtration and tubular secretion64	piperacillin Focus BA	piperacillin
											100
						22565	16865	5765
						62.865,a		6265,a
Tazobactam	322.364	0.2162	3064	0.7–1.264	8064	21965	14865	7265	glomerular filtration and tubular secretion64
						25.165,a		2565,a
Ticarcillin	428.455	0.167–0.17348	4566	1.166	80–9348	132–25348			glomerular filtration and tubular secretion49,67	Focus BA	ticarcillinc
					78.267	97.867	78.267	19.767			100
						15.868,b	1.668,b
Clavulanic acid	237.369	0.315–0.34248	2566	1.166	5049	21067	10267	10867	extensively metabolized, glomerular filtration48
					47.567	51.568,b	0.568,b

BA = bioassay; Cl_nonrenal = nonrenal clearance; Cl_renal = renal clearance; Cl_total = total body clearance; Focus = Focus Diagnostics; HPLC = high-performance liquid chromatography; Mayo = Mayo Medical Laboratories; t_1/2 = half-life; V_d = volume of distribution.

^aAnuric patients or on maintenance dialysis.

^bCrCl <10 mL/min.

^cAlthough ticarcillin concentrations are not currently measured at the NIH Clinical Center, we would suggest 100 mg/L as a maximum therapeutic concentration.

^dPenicillin G included in paper as an example of a β-lactam antibiotic that undergoes extensive tubular secretion.

Clinical Laboratory Improvement Amendment–approved drug concentration assays are available by high-performance liquid chromatography (HPLC) for ampicillin, cefepime, and penicillin G from Mayo Clinic Medical Laboratories (Rochester, MN), while the concentrations of other β-lactam antibiotics are available by bioassay from Focus Diagnostics (Cypress, CA). Unfortunately, bioassays are not reliable when other antibiotics are being used concurrently and the number of drug concentration assays available by HPLC has declined over the past several years, making it more difficult to monitor concentrations. Our suggested maximum recommended drug concentrations for the β-lactam antibiotics based on our clinical experience and review of the literature are listed in Table 3. We recommend using total rather than unbound drug concentrations for several reasons: the measurement of unbound drug concentrations is difficult and has not been standardized; serum protein levels vary considerably in ICU patients⁷⁰; and unbound versus bound drug is unpredictable.

The low ${\mathrm{V}}_{\mathrm{d}}$, low protein binding, low molecular weight, and high renal clearance of the β-lactam antibiotics shown in Table 3 suggest that they will be removed by hemofiltration. As noted in the previous section, hemofiltration clearance is clinically significant when extracorporeal clearance exceeds 25–30% of total clearance $\left({\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}\ge 25-30\mathrm{\%}\right)$.⁴² Therefore, cefepime, ceftazidime, piperacillin, ticarcillin, clavulanic acid, and tazobactam (which have a ${\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}\ge 25\mathrm{\%}$) would all require dosage adjustments when CVVH or other forms of renal replacement therapy are initiated for acute renal failure (Table 4).^3,38,71–78 There are limited pharmacokinetic data describing the effects of hemofiltration on ampicillin, aztreonam, and penicillin G clearance.^79–81 The pharmacokinetics of sulbactam have been reported only with CVVHD; ${\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}$ ranged from 22.9% to 46.4%.⁸²

Table 4.

Sieving Coefficient and Fractional Extracorporeal Clearance for Selected β-Lactam Antibiotics

Drug	CRRT Type	Clinical Setting (pts.)	S	Filter	UFR	Cltotal	ClCRRT	FrEC (%)	Reference
Ampicillin	CAVH		069	Amicon (polysulfone)					35
Cefepime			0.72						38
	CVVH	adult ICU (N = 5)	0.86	Hospal Multiflow 60 (AN69)	16 mL/min	36 mL/min	13 mL/min	40	71
	CVVH	adult ICU (N = 2)	pt. 1  0.76 (Cr Cl 29 mL/min)	Prisma M100 (AN69)	35 mL/min	70.4 mL/min	29.1 mL/min	41.3	72
			pt. 3  0.47 (CrCl not determined)		16.7 mL/min	172.2 mL/min	7.8 mL/min	4.5
Ceftazidime			0.9						38
	CVVH	adult ICU (N = 2)	pt. 3  1.01 (CrCl 80 mL/min)	Prisma M100 (AN69)	16.7 mL/min	333.8 mL/min	16.8 mL/min	5	73
			pt. 4  0.91 (CrCl 75 mL/min)		16.7 mL/min	154.3 mL/min	17.4 mL/min	11.3
	CVVH	adult ICU (N = 7)	0.87	CT-190G (cellulose triacetate)	25 mL/kg/h				74
	CVVH	adult ICU (N = 12)	0.69	Diafilter-30 (polysulphone)	47 mL/min	98.7 mL/min	32.1 mL/min	32.5a	75
	CVVH	adult non-ICU, end-stage renal disease (N = 8)	0.8	Filtryzer B1–2.1U (PMMA)					76
			0.97	Hospal Multiflow 60 (AN69)
			0.97	Fresenius F40 (polysulfone)
Penicillin G			0.68						38
Piperacillin			0.82						38
	CVVH	adult ICU (N = 14)	0.42 (CrCl <10 mL/min)	Prisma M100 (AN69)	27.1 mL/min	50 mL/min	11.45 mL/min	37	77
			0.38 (CrCl 10–50 mL/min)		30.3 mL/min	90.6 mL/min	12.2 mL/min	12.7
			0.23 (CrCl > 50 mL/min)		20 mL/min	265.2 mL/min	4.8 mL/min	2.8
Tazobactam			0.76 (CrCl <10 mL/min)		27.1 mL/min	50.4 mL/min	20.9 mL/min	62.5	77
			0.73 (CrCl 10–50 mL/min)		30.3 mL/min	68.2 mL/min	21.9 mL/min	35.4
			0.86 (CrCl >50 mL/min)		20 mL/min	180.1 mL/min	19.6 mL/min	13.1
Ticarcillin			0.83						38
	CVVH	pediatric ICU (N = 3)	0.83	Amicon D-20	890 mL/h	0.038 L/kg/h	0.022 L/kg/h	57.9a	78
Clavulanic acid			1.69			0.184 L/kg/h	0.049 L/kg/h	26.6a	78

CAVH = continuous arteriovenous hemofiltration; Cl_CRRT = CRRT clearance; Cl_total = total body clearance; CrCl = creatinine clearance; CRRT = continuous renal replacement therapy; CVVH = continuous venovenous hemofiltration; Fr_EC = fractional extracorporeal clearance; ICU = intensive care unit; S = sieving coefficient; UFR = ultrafiltration rate.

^aFr_EC calculated from clearance data in sstudy.

Studies investigating the removal of piperacillin by hemofiltration are conflicting. In a study by Capellier et al.⁸³ a minimal amount of piperacillin was found in ultrafiltrate, suggesting that the drug was not significantly removed by hemofiltration. In this study, hemofiltration clearance and the sieving coefficient were not reported. However, in a study by Arzuaga et al.,⁷⁷ piperacillin was removed by hemofiltration, with an ${\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}$ of 37% in patients with creatinine clearance less than 10 mL/min. Notably, the mean sieving coefficient was only 0.34, while the mean unbound drug fraction was 78.8%. The reason for this discrepancy is not clear, as the molecular weight of piperacillin suggests that unbound drug should freely cross hemofilter membranes.

Results of studies investigating the removal of tazobactam by hemofiltration are also conflicting. Two studies suggest that tazobactam may accumulate during CVVH, but the authors did not report hemofiltration clearance and the sieving coefficient.^84,85 In the study by Arzuaga et al.,⁷⁷ tazobactam was significantly cleared by CVVH, with an ${\mathrm{F}\mathrm{r}}_{\mathrm{E}\mathrm{C}}$ of 62.5% and 34.5% in patients with creatinine clearance values of less than 10 mL/min and between 10 and 50 mL/min, respectively. No accumulation of tazobactam was reported. These conflicting findings underscore the importance of monitoring concentrations of β-lactam antibiotics in serum of patients receiving CVVH.

STUDIES WITH CONTINUOUS-INFUSION β-LACTAM ANTIBIOTICS IN CRRT

Only 2 studies have been published investigating the pharmacokinetics of continuous-infusion β-lactam antibiotics during CRRT. In a study by Mariat et al.,⁸⁶ 7 adults in ICU received a ceftazidime loading dose of 2 g, followed by a continuous infusion of 3 g/day for 72 hours. The CVVHDF settings were a blood flow rate of 150 mL/min, dialysis flow rate of 1 L/h, UFR of 1.5 L/h, and prefilter fluid replacement. The ceftazidime dosing calculations were based on a ${\mathrm{V}}_{\mathrm{d}}$ of 0.3 L/kg, ${\mathrm{t}}_{1/2}$ of 4 hours, and estimated target concentration of 30–40mg/L. The mean steady-state concentration with continuous-infusion ceftazidime was 33.5 mg/L.

The pharmacokinetics of continuous infusion and intermittent dosing of meropenem on CVVHDF were compared in a randomized crossover study.⁸⁷ Six adults in ICU received either (1) meropenem loading dose of 0.5 g followed by a continuous infusion of 2 g /day for 2 days or (2) meropenem 1 g intravenously every 12 hours for 2 days. After crossover to the opposite arm of the study, patients continued to receive the study drug for another 2 days. The reported CVVHDF settings were a blood flow rate of 150 mL/min and UFR of 25 mL/kg/h. The median meropenem concentration was 19.1 mg/L with continuous infusion; with intermittent dosing, the median peak was 62.8 mg/L and the median trough was 8.2 mg/L.

CALCULATING A DOSAGE REGIMEN DURING CVVH

Methods for calculating a continuous-infusion dosage regimen during CVVH for β-lactam antibiotics with either renal clearance $\left({\mathrm{C}\mathrm{l}}_{\text{renal}}\right)$ by GFR alone or renal clearance by GFR and tubular secretion, are listed below. Estimating or even measuring residual creatinine clearance is difficult and often inaccurate in ICU patients with changing renal function. Furthermore, nonrenal clearance for drugs metabolized by the liver is difficult to determine in ICU patients.

Vigilant monitoring is required, as changes in CRRT settings (UFR, location of fluid replacement, blood flow rate) as well as system down-times due to procedures and filter clotting, may require major dose adjustments. Frequent drug concentrations obtained at or near steady-state should be measured if available (Table 3). Prior to collecting samples for drug concentrations of β-lactam antibiotics, clinicians should note any interruptions in the infusions, infusion rate changes, and the administration of repeat loading doses.

The following 2 methods for dosage calculation during CVVH are recommended.

$\mathrm{\beta }$-Lactam Antiblotics with ${\mathrm{C}\mathrm{l}}_{\text{renal}}$ by Glomerular Filtration (Cefepime, Ceftazidime, Ticarcillin)

1. Obtain MIC of bacteria from microbiology laboratory.

2. Select target mean Css concentration at least 4 times MIC (Table 3).

3. Calculate residual CrCl with a timed urine collection, if possible, before initiation of CVVH. During the course of CVVH, a decline in urinary output, often to an anuric state, may occur. In such instances, residual CrCl = 0 mL/min.

4. Calculate ${\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}$ (Figure 1).

5. Calculate estimated clearance.

   $${\mathrm{C}\mathrm{l}}_{\text{estimated}}(\mathrm{m}\mathrm{L}/\mathrm{m}\mathrm{i}\mathrm{n})={\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}+\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{i}\mathrm{d}\mathrm{u}\mathrm{a}\mathrm{l}\mathrm{C}\mathrm{r}\mathrm{C}\mathrm{l}$$
6. Calculate loading dose (Table 3).

   $$\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g}\mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e}\left(\mathrm{m}\mathrm{g}\right)={\mathrm{C}}_{\text{peak}}(\mathrm{m}\mathrm{g}/\mathrm{L})\times {\mathrm{V}}_{\mathrm{d}}(\mathrm{L}/\mathrm{k}\mathrm{g})\times \text{weight}\left(\mathrm{k}\mathrm{g}\right)$$

   ^a

   $$\mathrm{S}\mathrm{u}\mathrm{g}\mathrm{g}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{f}\mathrm{a}\mathrm{u}\mathrm{l}\mathrm{t}{\mathrm{V}}_{\mathrm{d}}=0.3\mathrm{L}/\mathrm{k}\mathrm{g}$$
7. Calculate maintenance infusion rate for a patient with normal renal function (Table 3).

   $${\mathrm{K}}_{\mathrm{e}}\left({\mathrm{h}}^{-1}\right)=0.693/{\mathrm{t}}_{1/2}\left(\mathrm{h}\right)$$

   $${\mathrm{C}\mathrm{l}}_{\text{total}}(\mathrm{L}/\mathrm{h})={\mathrm{K}}_{\mathrm{e}}\left({\mathrm{h}}^{-1}\right)\times {\mathrm{V}}_{\mathrm{d}}(\mathrm{L}/\mathrm{k}\mathrm{g})\times \text{weight}\left(\mathrm{k}\mathrm{g}\right)$$

   ^a

   $$\mathrm{S}\mathrm{u}\mathrm{g}\mathrm{g}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{f}\mathrm{a}\mathrm{u}\mathrm{l}\mathrm{t}{\mathrm{V}}_{\mathrm{d}}=0.3\mathrm{L}/\mathrm{k}\mathrm{g}$$

   $$\mathrm{M}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{f}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}(\mathrm{m}\mathrm{g}/\mathrm{h})=\mathrm{C}\mathrm{s}\mathrm{s}(\mathrm{m}\mathrm{g}/\mathrm{L})\times {\mathrm{C}\mathrm{l}}_{\text{total}}(\mathrm{L}/\mathrm{h})$$
8. Calculate maintenance infusion rate on CVVH

   $$\mathrm{m}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{f}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}(\mathrm{m}\mathrm{g}/\mathrm{h})=\left(\mathrm{m}\mathrm{a}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{f}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{f}\mathrm{r}\mathrm{o}\mathrm{m}\mathrm{E}\mathrm{q}\mathrm{u}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}7\right)\times \left({\mathrm{C}\mathrm{l}}_{\text{estimated}}/100\right)$$
9. Adjust maintenance infusion rate based on drug concentrations

   $${\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\text{new}}/{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\text{old}}={\mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e}}_{\mathrm{n}\mathrm{e}\mathrm{w}}/{\mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e}}_{\mathrm{o}\mathrm{l}\mathrm{d}}$$

$\mathrm{\beta }$-Lactam Antibiotics with ${\mathrm{C}\mathrm{l}}_{\text{renal}}$ by Glomerular Filtration and Substantial Tubular Secretion (Ampicillin, Aztreonam, Penicillin G, Piperacillin)

The calculations below should only be used in patients with anuria.

• 1–2.Same as above.
• 3.Calculate ${\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}$ (Figure 1).
• 4.
  Calculate estimated clearance (Table 3)^b

  $${\mathrm{C}\mathrm{l}}_{\text{estimated1}}(\mathrm{m}\mathrm{L}/\mathrm{m}\mathrm{i}\mathrm{n})={\mathrm{C}\mathrm{l}}_{\mathrm{H}\mathrm{F}}+{\mathrm{C}\mathrm{l}}_{\text{nenrenal}}$$
• 5.
  Calculate loading dose (Table 3).

  $$\begin{gathered} \text{loading}\text{dose}\left(\mathrm{m}\mathrm{g}\right)= \\ {\mathrm{C}}_{\text{peak}}(\mathrm{m}\mathrm{g}/\mathrm{L})\times {\mathrm{V}}_{\mathrm{d}}(\mathrm{L}/\mathrm{k}\mathrm{g})\times \text{weight}\left(\mathrm{k}\mathrm{g}\right) \end{gathered}$$

  ^a

  $$\mathrm{S}\mathrm{u}\mathrm{g}\mathrm{g}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{f}\mathrm{a}\mathrm{u}\mathrm{l}\mathrm{t}{\mathrm{V}}_{\mathrm{d}}=0.3\mathrm{L}/\mathrm{k}\mathrm{g}$$
• 6.
  Calculate maintenance infusion rate on CVVH.

  $$\begin{gathered} \text{maintenance}\text{infusion}\text{rate}(\mathrm{m}\mathrm{g}/\mathrm{h})= \\ \mathrm{C}\mathrm{s}\mathrm{s}(\mathrm{m}\mathrm{g}/\mathrm{L})\times {\mathrm{C}\mathrm{l}}_{\text{estimated}1}(\mathrm{L}/\mathrm{h}) \end{gathered}$$
• 7.
  Adjust maintenance infusion rate based on drug concentrations.

  $${\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\text{new}}/{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{o}\mathrm{l}\mathrm{d}}={\mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e}}_{\mathrm{n}\mathrm{e}\mathrm{w}}/{\mathrm{d}\mathrm{o}\mathrm{s}\mathrm{e}}_{\mathrm{o}\mathrm{l}\mathrm{d}}$$

Summary

While the incidence of antibiotic resistance in gram-negative bacteria is increasing, there has been a decline in the development and approval of new antibacterial agents. ICU practitioners and their infectious diseases consultants commonly need to treat patients infected with MDR gram-negative bacteria, such as P. aeruginosa and A. baumannii, that are resistant to the majority of available antibiotics and who are at high risk for acute renal failure requiring renal replacement therapy. Diligent monitoring of the CRRT system, as well as monitoring of drug concentrations, is recommended when following the concepts described in this paper. Although the number of β-lactam antibiotic concentration assays available by HPLC has declined over the past several years, we hope that this article will stimulate a renewed interest in the measurement and availability of these concentrations.

Appendix I. Recommendation from Clinical Scenario

You call the infectious diseases consultants with suggested drug dosing recommendations for continuous-infusion ceftazidime.

1. P. aeruginosa MIC = 16 mg/L

2. Target mean Css concentration = 64 mg/L

3. Residual CrCl = 0 mL/min

4. ClHF = UFR × S × [blood flow rate/(blood flow rate + replacement fluid rate)]

ClHF = 2100 mL/h × 0.9 × [200 mL/min/(200 mL/min + 35 mL/min)]

ClHF = 1608.5 mL/h × 1 h/60 min = 26.8 mL/min

5. Clestimated = ClHF + residual CrCl

Clestimated = 26.8 mL/min + 0 mL/min = 26.8 mL/min

6. Loading dose = Cpeak (mg/L) × Vd (L/kg) × weight (kg)

loading dose = 64 mg/L × 0.3 L/kg × 70 kg = 1344 mg

7. Maintenance infusion rate for a patient with normal renal function

Ke = 0.693/t1/2 (h)

Cltotal = Ke (h−1) × Vd (L/kg) × weight (kg)

Maintenance infusion rate = Css (mg/L) × Cltotal (L/h)

Ke = 0.693/1.9 h = 0.3647 h−1

Cltotal = 0.3647 h−1 × 0.3 L/kg × 70 kg = 7.659 L/h

Maintenance infusion rate = 64 mg/L × 7.659 L/h = 490.21 mg/h

8. Maintenance infusion rate during CVVH = (maintenance infusion rate from equation 7) × (Clestimated/100)

Maintenance infusion rate = 490.21 mg/h × (26.8 mL/min/100 mL/min) = 131 mg/h

C_peak = peak concentration; Css = steady-state concentration; Cl_estimated = estimated clearance; Cl_HF = hemofiltration clearance; Cl_total = total clearance; CrCl = creatinine clearance; CVVH = continuous venovenous hemofiltration; K_e = elimination rate constant; MIC = minimum inhibitory concentration; S = sieving coefficient; UFR = ultrafiltration rate; V_d = volume of distribution.