Determination of Total and Unbound Meropenem, Imipenem/Cilastatin, and Cefoperazone/Sulbactam in Human Plasma: Application for Therapeutic Drug Monitoring in Critically Ill Patients.

BACKGROUND: Critically ill patients show several pathophysiological alterations that can complicate antibiotic dosing. Hence, there is a strong rationale to individualize anti-infective dosing in these patients by using therapeutic drug monitoring (TDM). The current study aimed to develop and validate a liquid chromatography-tandem mass spectrometry method for the simultaneous determination of total and unbound plasma concentrations of 3 commonly used antibiotics (meropenem, imipenem/cilastatin, and cefoperazone/sulbactam) in the treatment of infections in critically ill patients in China, which could be suitable for routine TDM in hospital laboratories. METHODS: The unbound drug was separated from the bound drug by ultrafiltration. Simple protein precipitation was used for sample preparation. Meropenem, imipenem/cilastatin, cefoperazone/sulbactam, and their corresponding internal standards were then resolved using the Waters CORTECS C18 column. All the compounds were detected using electrospray ionization in the positive/negative ion-switching mode. RESULTS: The calibration curves were linear for all compounds, with correlation coefficients (R) above 0.99 for total concentrations in human plasma and unbound concentrations in the plasma ultrafiltrate. For both total and unbound drugs, the relative errors and intra-assay/interassay relative standard deviations were below 15%. The limit of quantification was 0.05 mcg/mL for both total plasma concentrations and plasma ultrafiltrate concentrations of all compounds. CONCLUSIONS: The method was simple, rapid, and reliable and is currently being used to provide a TDM service to enhance the efficacious use of the 3 antibiotics.

Effect of elevated imipenem/cilastatin minimum inhibitory concentrations on patient outcomes in Gram-negative bloodstream infections.

OBJECTIVES: Carbapenem minimum inhibitory concentration (MICs) are known to predict outcomes for patients with Gram-negative bacteraemia. However, limited data exist on how MICs influence such outcomes when organisms are classified as carbapenem-resistant. The purpose of this study was to evaluate the effect of increasing imipenem/cilastatin MICs on mortality in patients with Gram-negative bloodstream infection (BSI). METHODS: Patients with an imipenem/cilastatin-resistant (MIC>4mg/L) monomicrobial Gram-negative BSI were eligible for inclusion in the study and were assessed for baseline characteristics, organ function, microbiological data, timing and type of therapeutic treatment, and in-hospital mortality. RESULTS: A total of 62 patients with imipenem/cilastatin-resistant bacterial isolates (MIC>4mg/L) were retrospectively studied. Time to event analyses found no difference between patients who received carbapenem therapy and those who did not (P=0.10). After adjustment, patients receiving directed therapy were less likely to die (adjusted hazard ratio=0.35, 95% confidence interval 0.15-0.83; P<0.01), whereas higher modified Acute Physiology and Chronic Health Evaluation (APACHE) II score and days to positive culture were associated with non-survival. CONCLUSION: This study did not demonstrate a relationship between receipt of a carbapenem and mortality in patients with carbapenem-resistant Gram-negative BSI.

Impact of the introduction of real-time therapeutic drug monitoring on empirical doses of carbapenems in critically ill burn patients.

PURPOSE: Adequate empirical antibiotic dose selection for critically ill burn patients is difficult due to extreme variability in drug pharmacokinetics. Therapeutic drug monitoring (TDM) may aid antibiotic prescription and implementation of initial empirical antimicrobial dosage recommendations. This study evaluated how gradual TDM introduction altered empirical dosages of meropenem and imipenem/cilastatin in our burn ICU. METHODS: Imipenem/cilastatin and meropenem use and daily empirical dosage at a five-bed burn ICU were analyzed retrospectively. Data for all burn admissions between 2001 and 2011 were extracted from the hospital's computerized information system. For each patient receiving a carbapenem, episodes of infection were reviewed and scored according to predefined criteria. Carbapenem trough serum levels were characterized. Prior to May 2007, TDM was available only by special request. Real-time carbapenem TDM was introduced in June 2007; it was initially available weekly and has been available 4 days a week since 2010. RESULTS: Of 365 patients, 229 (63%) received antibiotics (109 received carbapenems). Of 23 TDM determinations for imipenem/cilastatin, none exceeded the predefined upper limit and 11 (47.8%) were insufficient; the number of TDM requests was correlated with daily dose (r=0.7). Similar numbers of inappropriate meropenem trough levels (30.4%) were below and above the upper limit. Real-time TDM introduction increased the empirical dose of imipenem/cilastatin, but not meropenem. CONCLUSIONS: Real-time carbapenem TDM availability significantly altered the empirical daily dosage of imipenem/cilastatin at our burn ICU. Further studies are needed to evaluate the individual impact of TDM-based antibiotic adjustment on infection outcomes in these patients.

Nonvolatile salt-free stabilizer for the quantification of polar imipenem and cilastatin in human plasma using hydrophilic interaction chromatography/quadrupole mass spectrometry with contamination sensitive off-axis electrospray.

A hydrophilic interaction chromatography/mass spectrometry (HILIC-MS)-based assay for imipenem (IMP) and cilastatin (CIL) was recently reported. This orthogonal electrospray ion source-based (ORS) assay utilized nonvolatile salt (unremovable) to stabilize IMI in plasma. Unfortunately, this method was not applicable to conventional MS with off-axis spray (OAS-MS) because MS sensitivity was rapidly deteriorated by the nonvolatile salt. Therefore, we aimed to find a nonvolatile salt- and ion suppression-free approach to stabilize and measure the analytes in plasma using OAS-MS. Acetonitrile and methanol were tested to stabilize the analytes in the plasma samples. The recoveries, matrix effects and stabilities of the analytes in the stabilizer-treated samples were studied. The variations in MS signal intensities were used as the indicator of the assay ruggedness. The results show that a mixture of methanol and acetonitrile (1:1) is best for the storage and measurement of IMP and CIL in human plasma. Utilization of this precipitant not only blocked the hydrolysis of the analytes in plasma but also resulted in an ion suppression-free, fast (120 s per sample) and sensitive detection. The sensitivity obtained using the less sensitive OAS-MS (API3000, 4 pg on column) is much greater than that of the published ORS-MS-based assay (API4000, 77 pg on column). The ruggedness of the assay was demonstrated by its constant MS signal intensity. In conclusion, an improved HILIC/MS-based assay for IMP and CIL was established. The approach presented here provides a simple solution to the challenge of analyzing hydrolytically unstable ß-lactam antibiotics in biological samples.

Hydrophilic interaction chromatography/tandem mass spectrometry for the simultaneous determination of three polar non-structurally related compounds, imipenem, cilastatin and an investigational beta-lactamase inhibitor, MK-4698, in biological matrices.

A method coupling hydrophilic interaction chromatography (HILIC) with tandem mass spectrometry (MS/MS) has been developed for the simultaneous determination of three polar non-structurally related compounds--a carbapenem antibiotic, imipenem (IMP), a renal dehydropeptidase inhibitor, cilastatin (CIL), and an investigational beta-lactamase inhibitor, MK-4698 (BLI), in rat plasma, monkey plasma and mouse blood. The analytes were extracted through protein precipitation, chromatographed on a Waters Atlantis HILIC column, and detected on a Sciex API4000 mass spectrometer using a Turbo-Ion Spray ion source in positive ionization mode following multiple-reaction monitoring (MRM). The assay dynamic range was 0.1-100 microg/mL for IMP, CIL and BLI, respectively, using a total of 20-25 microL biologic samples, and the total HPLC/MS/MS run time was 4 min/injection. The assay was found to be sensitive, selective and reproducible. The challenges, namely, sample stability, blood sample processing, matrix effect in monkey study samples, and dilution re-assays for the limited mouse blood samples, are resolved and discussed. This technique allowed rapid analysis of polar compounds in biologic matrixes with satisfactory chromatographic retention and increased throughput.

Pharmacokinetics of imipenem-cilastatin following intravenous administration in healthy adult horses.

In two studies, six healthy adult horses were given imipenem-cilastatin by slow intravenous (i.v.) infusion at an imipenem dosage of 10 mg/kg (study 1) and 20 mg/kg (study 2). The same horses were used in each dosage schedule, with a 2-week washout period between studies. In each dosage group, serial blood and synovial fluid samples were collected for 6 h after completion of the infusion. HPLC was used to determine the imipenem concentration in all samples. Imipenem was well tolerated by all horses at both dosages; no adverse effects were noted during the study period or during the 24-hour postinfusion observation period. The pharmacokinetic profiles of imipenem in the plasma and synovial fluid indicate that an imipenem dosage of 10-20 mg/kg by slow i.v. infusion q6h (every 6 h) is appropriate for most susceptible pathogens.

Comparative therapeutic efficacy of clinafloxacin in a Pseudomonas aeruginosa mouse renal abscess model.

A deep-seated Pseudomonas aeruginosa mouse kidney abscess model was used to compare the therapeutic efficacy of clinafloxacin, a fluoroquinolone in clinical trials, with that of clinically relevant standard drugs. Following 50 mg/kg oral doses, twice daily for five consecutive days, clinafloxacin produced a 4 log decrease in mean bacterial count, the greatest decrease of all drugs tested. The same dosage regimen resulted in complete bacterial eradication in 88% of the kidneys. No other compound produced total bacterial clearance in 50% of the kidneys at the highest dose tested.

Pharmacokinetics of imipenem-cilastatin in critically ill patients undergoing continuous venovenous hemofiltration.

The pharmacokinetics of imipenem-cilastatin were investigated in 12 critically ill patients with acute renal failure (ARF) managed by continuous veno-venous hemofiltration (CVVH) while receiving a fixed combination of 500 mg of imipenem-cilastatin intravenously three or four times daily. No adverse drug reactions were observed. Plasma and hemofiltrate samples were taken at specified times during one dosing interval, and the concentrations of imipenem and cilastatin were determined by high-performance liquid chromatography. Pharmacokinetic variables were calculated by a first-order, two-compartment pharmacokinetic model for both substances. Total clearances of imipenem and cilastatin (mean +/- standard deviations) were 122.2 +/- 28.6 and 29.2 +/- 13.7 ml/min, respectively, with hemofiltration clearances of 22.9 +/- 2.5 and 16.1 +/- 3.1 ml/min, respectively, and nonrenal, nonhemofiltration clearances of 90.8 +/- 26.3 and 13.2 +/- 13.9 ml/min, respectively. Mean imipenem dosage requirements were approximately 2,000 mg/24 h (2,111.8 +/- 493.4 mg/24 h). They were calculated in order to achieve an average steady-state concentration of 12 mg/liter to ensure that concentrations in plasma exceeded the MICs at which 90% of intermediately resistent bacteria are inhibited (8 mg/liter) during the majority of the dosing interval. By contrast, the recommended dosage for patients with end-stage renal failure (ESRF) and infections caused by intermediately resistant bacteria is 1,000 mg/24 h. This remarkable difference may be due (i) to differences in the nonrenal clearance of imipenem between patients with ARF and ESRF and (ii) to the additional clearance by the hemofilter. Since the total clearance of cilastatin was low, marked accumulation occurred, and this was particularly pronounced in patients with additional liver dysfunction. Thus, in patients with ARF managed by CVVH, rather high imipenem doses are required, and these inevitably result in a marked accumulation of cilastatin. The doses of imipenem recommended for patients with ESRF, however, will lead to underdosing and inadequate antibiotic therapy.

Serum bactericidal activities and comparative pharmacokinetics of meropenem and imipenem-cilastatin.

The pharmacokinetics and serum bactericidal activities (SBAs) of imipenem and meropenem were investigated in a randomized crossover study. Twelve healthy male volunteers received a constant 30-min infusion of either 1 g of imipenem plus 1 g of cilastatin or 1 g of meropenem. The concentrations of the drugs in serum and urine were determined by bioassay and high-pressure liquid chromatography. Pharmacokinetic parameters were based on an open two-compartment model and a noncompartmental technique. At the end of infusion, the mean concentrations of imipenem and meropenem measured in serum were 61.2 +/- 9.8 and 51.6 +/- 6.5 mg/liter, respectively; urinary recoveries were 48.6% +/- 8.2% and 60.0% +/- 6.5% of the dose in 12 h, respectively; and the areas under the concentration-time curve from time zero to infinity were 96.1 +/- 14.4 and 70.5 +/- 10.3 mg.h/liter, respectively (P < or = 0.02). Imipenem had a mean half-life of 66.7 +/- 10.4 min; that of meropenem was 64.4 +/- 6.9 min. The volumes of distribution at steady state of imipenem and meropenem were 15.3 +/- 3.3 and 18.6 +/- 3.0 liters/70 kg, respectively, and the mean renal clearances per 1.73 m2 were 85.6 +/- 17.6 and 144.6 +/- 26.0 ml/min, respectively. Both antibiotics were well tolerated in this single-dose administration study. The SBAs were measured by the microdilution method of Reller and Stratton (L. B. Reller and C. W. Stratton, J. Infect. Dis. 136:196-204, 1977) against 40 clinically isolated strains. Mean reciprocal bactericidal titers were measured 1 and 6 h after administration. After 1 and 6 h the median SBAs for imipenem and meropenem, were 409 and 34.9 and 97.9 and 5.8, respectively, against Staphylococcus aureus, 19.9 and 4.4 and 19.4 and 4.8, respectively, against Pseudomonas aeruginosa, 34.3 and 2.2 and 232 and 15.5, respectively, against Enterobacter cloacae, and 13.4 and 2.25 and 90.7 and 7.9, respectively, against Proteus mirabilis. Both drugs had rather short biological elimination half-lives and a predominantly renal route of elimination. Both carbapenems revealed high SBAs against clinically important pathogens at 1 h; meropenem had a higher SBA against E. cloacae and P. mirabilis, and the SBA of imipenem against S. aureus was greater than the SBA of meropenem.

Comparison of simulated and in-vivo plasma levels of cilastatin following intravenous in-line drug administration.

The primary objective of this work was to establish a method to simulate the plasma levels of cilastatin, a model drug, following an intravenous in-line delivery scheme. In-vivo data in dogs obtained from this work were used to demonstrate the validity of the proposed approach. The in-line drug delivery system consists of a drug containing device which is placed between a large volume parenteral and a patient. Numerous advantages have been identified for this automatic in-line reconstitution delivery system. The numerical convolution integral algorithm was used in this work to perform plasma profile simulation. The results indicated that the simulated cilastatin plasma profile following in-line delivery closely agreed with the in-vivo data.

Removal of imipenem and cilastatin by hemodialysis in patients with end-stage renal failure.

The removal of imipenem and cilastatin by hemodialysis was studied in 14 (for imipenem) and 6 (for cilastatin) subjects. Following intravenous infusion of imipenem and cilastatin at a combined concentration of 10 mg/kg of body weight, drug levels in plasma were determined serially during off- and on-hemodialysis periods, which were 2 and 4 h, respectively. The biexponential decay of the drug levels in plasma was evident in each subject for both imipenem and cilastatin. Hemodialysis accelerated the elimination of both imipenem and cilastatin: the mean elimination-phase half-life of imipenem was shortened from 200 to 78 min, and that of cilastatin was shortened from 445 to 115 min. Hemodialysis clearance of imipenem and cilastatin was calculated by five different methods, each with intrinsic assumptions. The mean hemodialysis clearance of imipenem was estimated to be 74.08 +/- 13.29 ml/min, and that of cilastatin was estimated to be 65.0 +/- 8.6 ml/min, after consideration of various methodological limitations. It was estimated that in a hypothetical anephric patient weighing 60 kg, a 4-h hemodialysis treatment would remove 54.8% of the imipenem and 62.9% of the cilastatin present in the body at the start of dialysis.

Pharmacokinetics of parenteral imipenem/cilastatin in patients on continuous ambulatory peritoneal dialysis.

We investigated the pharmacokinetics of two intravenous (iv) dose regimens of imipenem/cilastatin in Chinese patients on chronic ambulatory peritoneal dialysis (CAPD), who had an average creatinine clearance of 3.2 ml/min/1.73 m2. Doses of 0.5 and 1.0 g produced mean peak serum imipenem concentrations of 30 and 70 mg/l respectively, about 60% of cilastatin. Peritoneal dialysis fluid (PDF) imipenem concentrations reached 20-30% of the serum peak 4-5 h after iv injection, and the lowest maximum PDF concentrations were 2 mg/l after the 0.5 g dose and 14 mg/l after 1.0 g. Thus both regimes produced PDF imipenem concentrations above the MICs of susceptible pathogens. The half-life of imipenem was 6.4 h and the plasma clearance 66 ml/min; serum and PDF imipenem were in equilibration after about 5 h. Cilastatin had a prolonged half-life of 19 h and a plasma clearance of 10 ml/min, and accumulated in both serum and PDF. With a 0.5 g dose, the pharmacokinetics of imipenem/cilastatin suggest that the combination may prove an effective treatment for peritonitis associated with CAPD.

High-performance liquid chromatographic determination of cilastatin and its major metabolite N-acetylcilastatin in rat plasma, urine and bile.

A new high-performance liquid chromatographic method coupled with solid-phase (C8) sample extraction has been developed for the simultaneous quantification of cilastatin and its major metabolite N-acetylcilastatin in rat plasma, urine and bile. The method is linear, reproducible and reliable with a detection limit of 1 microgram/ml in all three fluids. Plasma concentrations of cilastatin and N-acetylcilastatin at selected time intervals and biliary and urinary recoveries of cilastatin and N-acetylcilastatin following an intravenous dose of 10 mg/kg cilastatin are presented.

Imipenem pharmacokinetics in patients with burns.

The pharmacokinetics of imipenem were studied in 11 adult patients with severe burns who were receiving a therapeutic regimen of imipenem-cilastatin 500 mg intravenously every 6 hours. Serial blood samples for measuring imipenem and 24-hour urine collections for creatinine clearance (CrCl) were obtained after the initial dose and after multiple dosing. Plasma was assayed for imipenem by use of HPLC. A two-compartment model provided a superior fit to the data compared with a one-compartment model in 9 of the 11 patients. There was no significant difference in any pharmacokinetic parameter between the initial dose and after multiple dosing (p greater than 0.05). Combined mean (+/- SD) parameter estimates for the two dosing periods were as follows: VC, 0.11 +/- 0.06 L/kg; Vss, 0.22 +/- 0.06 L/kg; CL, 12.5 +/- 3.6 L/hr/1.73 m2; t1/2 alpha, 0.18 +/- 0.13 hr; t1/2 beta, 1.12 +/- 0.44 hr. Mean clearance in two patients with creatinine clearance values greater than 150 ml/min/1.73 m2 was 17.7 L/hr/1.73 m2. Mean clearance in two patients with creatinine clearance values less than 50 ml/min/1.73 m2 was 8.5 L/hr/1.73 m2. No pharmacokinetic parameter was significantly different from previously reported parameters in normal volunteers (p greater than 0.05). Creatinine clearance ranged from 17 to 218 ml/min/1.73 m2. Imipenem clearance was significantly related to creatinine clearance (CL = 63 + 0.059 CLCR; r2 = 0.60, p = 0.001). No significant association was found between total body surface area burns and imipenem clearance (p greater than 0.05). Our data suggest imipenem pharmacokinetics in patients with burns are comparable to those in normal volunteers although substantial intersubject variability exists.(ABSTRACT TRUNCATED AT 250 WORDS)

[Study of the intraperitoneal penetration of imipenem/cilastatin in acute peritonitis with perforation]. / Etude de la pénétration intrapéritonéale de l'imipénème/cilastatine au cours de péritonites aiguës avec perforation.

A study was performed to investigate the intraperitoneal penetration of Imipenem/cilastatin into inflammatory peritoneal fluid. Six patients undergoing abdominal surgery (acute peritonitis), were treated with Imipenem/cilastatin (4 perfusions of 0.5 g/day) during 5 days after the intervention. Plasma samples were obtained on day 1 and 4 at the pic and at the valley; peritoneal samples were obtained every days for 4 days, 1, 3 and 6 hours after the end of a perfusion. The samples were immediately stabilised following the manufacturer instructions and quick freezed at -80 degrees C. Dosages were performed using a microbiological assay. The mean peritoneal levels are above the MIC 90 of the more frequent bacteria which cause infection in abdominal surgery. Moreover none of the patients showed relapse of infection or complication during this treatment.

[Clinical pharmacokinetics of an imipenem-cilastatin combination]. / Pharmacocinétique clinique de l'association imipénème-cilastatine.

In the combination of imipenem (antibiotic of the thienamycin class, in the beta-lactam family) with cilastatin (renal dehydropeptidase I inhibitor) the ratio is 1:1. The urinary excretion of imipenem shows considerable interindividual variations due to the activity of renal dehydropeptidase I which degrades the compound. This makes it difficult to determine a mean dosage for therapeutic uses. However, when the dehydropeptidase I inhibitor cilastatin is given concomitantly with imipenem the urinary excretion and clearance of imipenem reach similar values in all subjects; in addition, the antibiotic is better tolerated by the kidney. Moreover, cilastatin has been shown to increase by 15 to 20 per cent the area under the imipenem plasma concentration curve. Following an intravenous infusion of imipenem 500 mg, the main pharmacokinetic values for the antibiotic are: area under the plasma concentration curve 43.2 +/- 4.7 h.mg/l; plasma clearance 195 +/- 25 ml/min; half-life 1.0 +/- 0.1 h; apparent volume of distribution 10.4 +/- 1.7 l. With repeated infusions of 500 ml every 6 to 8 hours, imipenem and cilastatin do not accumulate. In patients with renal impairment (creatinine clearance below 30 ml/min), cilastatin is excreted more slowly than imipenem and dosage must be adjusted accordingly.

Dose-dependent kinetics of cilastatin in laboratory animals.

Cilastatin, a potent inhibitor of renal dehydropeptidase I, was specifically designed to inhibit renal metabolism of the antibiotic imipenem in order to achieve therapeutically relevant imipenem concentrations in the urinary tract. In this study the elimination kinetics of cilastatin in rats at doses of 5, 10, 20, 50, 100, and 200 mg/kg iv were demonstrated to be dose dependent, with total plasma clearance and non-renal clearance falling from 20.2 +/- 3.1 ml/min/kg and 17.7 +/- 3.3 ml/min/kg (mean +/- S.D.) at the 5 mg/kg dose to 11.4 +/- 1.2 ml/min/kg and 5.30 +/- 1.2 ml/min/kg, respectively, at the 200 mg/kg dose, whereas the volume of distribution of the drug remained unchanged. Since cilastatin is mainly eliminated by renal excretion as well as by N-acetylation, the non-renal clearance may reasonably reflect the N-acetylation process. Thus, the dose-dependent kinetics of cilastatin might be explained, at least partly, by the saturation of the N-acetylation of the drug. The dose-related decrease in the fraction (fm) of cilastatin converted to its N-acetylated metabolite provided further evidence for the saturable N-acetylation. The fm values decreased from 0.915 at the 10 mg/kg dose to 0.626 at the 100 mg/kg dose. Although both the total plasma clearance and non-renal clearance decreased with increasing dose, the dose had an opposite effect on the renal clearance of cilastatin. The renal clearance of cilastatin increased from 2.50 +/- 0.40 ml/min/kg at the lowest dose to 6.10 +/- 0.50 ml/min/kg at the highest dose as the dose increased.(ABSTRACT TRUNCATED AT 250 WORDS)

[Clinical and pharmacokinetic study of imipenem/cilastatin in children and newborn infants]. / Etude clinique et pharmacocinétique de l'imipénème/cilastatine chez l'enfant et le nouveau-né.

Imipenem, a new carbapenem (thienamycin) beta lactam antibiotic which is clinically used in a 1:1 combination with cilastatin, an inhibitor or renal metabolism of imipenem, was evaluated in 25 patients; 11 children and 14 neonates. A mean daily dose of 60 mg/kg was given to children and the dose in neonates was 50 mg/kg. Clinically, 21 patients were cured, two failed to respond to treatment and two were not evaluable. Pharmacokinetic studies were performed in the 11 children and in 10 of the neonates. The mean elimination half-life of imipenem was 0.87 h in children and 2.1 h in neonates. The mean cilastatin elimination half-life was 0.73 h in children and 5.1 h in neonates. This difference in half-life between children and neonates is similar to the one noted between healthy adults and adults with renal insufficiency. No accumulation of imipenem was seen in neonates studied on the first and fifth days of treatment.

Single-dose kinetics of imipenem/cilastatin during continuous arteriovenous haemofiltration in intensive care patients.

The single i.v. dose kinetics of a fixed combination of imipenem/cilastatin were investigated in ten critically ill patients treated by continuous arteriovenous haemofiltration (CAVH). Eight patients suffered from acute renal failure and two had normal renal function. Both drugs were measured in plasma and ultrafiltrate by high-performance liquid chromatography. While the pharmacokinetics of both drugs are almost identical in patients with normal renal function, we found the following dissociation of pharmacokinetic parameters in our patients with renal failure: for imipenem the total clearance and elimination half-life was 104 +/- 12 ml/min and 2.2 +/- 0.1 h, respectively, and for cilastatin 29 +/- 10 ml/min and 13.8 +/- 4.5 h. The pharmacokinetics of imipenem and cilastatin differ from each other in renal failure because imipenem, unlike cilastatin, undergoes marked elimination by non-renal pathways. Our results did not differ from previously reported data in healthy volunteers and patients with impaired renal function. Elimination of imipenem by CAVH was low (7% of the dose). As a consequence of the unsatisfactory non-renal clearance of cilastatin, however, the fraction of the dose removed by CAVH was significantly greater (approximately 30%) than that of imipenem. This did not, however, correct the dissociation of the pharmacokinetic profiles of the two drugs. In conclusion, the dose of imipenem/cilastatin in critically ill patients with renal failure treated by CAVH should be modified according to renal function but elimination by CAVH does not need to be considered.