Skip Navigation


NDT Advance Access originally published online on February 27, 2006
Nephrology Dialysis Transplantation 2006 21(5):1402-1406; doi:10.1093/ndt/gfl048
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
21/5/1402    most recent
gfl048v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (5)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Fiaccadori, E.
Right arrow Articles by Cabassi, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fiaccadori, E.
Right arrow Articles by Cabassi, A.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2006]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Brief Report

Does haemodialysis significantly affect serum linezolid concentrations in critically ill patients with renal failure? A pilot investigation

Enrico Fiaccadori, Umberto Maggiore, Carlo Rotelli, Roberto Giacosa, Elisabetta Parenti, Edoardo Picetti1, Paola Manini, Roberta Andreoli and Aderville Cabassi

Terapia Intensiva, Dipartimento di Clinica Medica, Nefrologia & Scienze della Prevenzione, Università degli Studi di Parma and 1 Servizio di Anestesia e Rianimazione, Azienda Ospedaliera – Universitaria, Parma, Italy

Correspondence and offprint requests to: Enrico Fiaccadori, MD, PhD, Terapia Intensiva, Dipartimento di Clinica Medica, Nefrologia & Scienze della Prevenzione, Università degli Studi di Parma, Via Gramsci 14–43100 Parma, Italy. Email: enrico.fiaccadori{at}unipr.it



   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 References
 
Background. Previous studies have shown that a single haemodialysis (HD) session removes about one-third of the linezolid dose administered, but it is unknown whether in critically ill patients with renal failure on intermittent HD, this removal adversely affects serum antibiotic concentrations.

Methods. Five male critically ill patients (mean age 75 years, range 68–82; APACHE II score 26.4, range 23–29; survival 2/5) with sepsis and renal failure on haemodialysis, were administered i.v. linezolid, 600 mg every 12 h. Serum antibiotic levels were measured by high-performance liquid chromatography/mass spectrometry. We classified trough concentrations as ‘with HD’ when a HD session was performed after linezolid infusion, and ‘without HD’ otherwise. We also computed population pharmacokinetics while patients were on-dialysis and off-dialysis.

Results. A total of 222 serum linezolid concentrations were available over 36 days of antibiotic therapy, during which patients underwent 31 HD sessions. Trough serum linezolid levels averaged 5.83 mg/l (range 1.48–15.84), exceeding 4.0 mg/l in 68.9% of the samples; however, the trough levels ‘with HD’ were lower than those ‘without HD’ (4.68 mg/l [range 1.48–9.07] vs 6.74 mg/l [range 2.04–15.84], P<0.001). Clearance and half-life were 6.0 l/h and 4.0 h, respectively, while patients were on-dialysis, and 4.4 l/h and 7.3 h, respectively, when they were off-dialysis.

Conclusions. HD can significantly reduce serum linezolid levels in critically ill patients with renal failure.

Keywords: critical illness; dosage guidelines; haemodialysis; linezolid; renal failure



   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 References
 
Linezolid is a novel oxazolidinone antimicrobial for clinically relevant Gram-positive pathogens (Staphylococcus aureus, S. epidermidis, and Enterococcus species), including multiresistant strains [1–3]. In a previous study on critically ill patients, it was shown that with the currently recommended dosage for adults (600 mg i.v. twice daily) peak concentrations average 14.0 mg/dl, and trough concentrations were between 3.9 and 4.5 mg/l after 48–72 h [4]. These values are very close to the minimal inhibitory concentration against 90% (MIC90) of clinically relevant strains [1]. As linezolid has time-dependent antibacterial activity, time above MIC90 for free drug concentrations is significantly associated with increased efficacy of the antibiotic, and is predictive of a more favourable outcome [5]. Thus, it is of critical importance to modulate antibiotic administration in such a way as to maintain optimal levels throughout the treatment cycle.

Even in the presence of renal or hepatic dysfunction no dosage reduction is currently recommended [6]; rather, owing to the pharmacokinetic properties of the antibiotic (molecular weight 337 Da, protein-bound fraction 31%, distribution volume 0.5–0.8 l/kg), it is likely that the antibiotic might be removed by renal replacement therapy (RRT) in amounts varying according to the modality chosen [mainly diffusive such as haemodialysis (HD), or convective such as continuous haemofiltration, etc.], the filter membranes used (low- vs high-flux) and duration of therapy.

Previous studies in critically ill patients with acute renal failure have suggested that continuous RRT modalities, namely continuous venovenous haemofiltration and continuous venovenous haemodiafiltration, do not remove significant amounts of the drug [7–9]. More recently, however, a multiple dose pharmacokinetic study has demonstrated that linezolid removal can be clinically relevant during continuous haemofiltration with highly permeable polysulphone membranes [10]. Also, some widely used diffusive modalities, namely standard intermittent HD and sustained low-efficiency HD, can remove important amounts of the drug (from 32 to 34% when low-flux polysulphone membranes are employed [11]), but it remains uncertain whether such removals may lower serum linezolid concentrations to the levels close to the MIC90 in critically ill patients with renal failure receiving multiple doses of the antibiotic.

To fill this gap, we designed the present pilot investigation in which we assessed serially intradialytic and interdialytic the linezolid serum levels in patients receiving multiple doses of the drug.



   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 References
 
We included in the study all critically ill patients (APACHE II ≥23) consecutively admitted to our unit over a period of 12 months (from January to December 2004) who presented the following features: suspected or proven Gram-positive infection requiring parenteral wide-spectrum antibiotic therapy including linezolid, oliguria (urinary output less than 200 ml in the 24 h) due to acute or chronic renal failure, need of RRT. All patients received linezolid as a part of the empirical therapy for sepsis, together with piperacillin-tazobactam or ceftazidime to cover Gram-negative organisms. The study was approved by the Human Studies Ethics Committee of the Parma University Medical School, and written informed consent was obtained from patients and/or their relatives.

For HD, we used an AK200 Ultra type 1 monitor (Gambro, Felino, Italy) in all patients, along with low-flux polysulphone filters (BLS 514G, Bellco, Mirandola, Italy, ultrafiltration coefficient 12, 1.4 m2). The dialysis fluid was online-generated, ultrapure, bicarbonate-based. The HD sessions typically lasted 4 h without the use of antihaemostatic agents, with nominal blood flow of 300 ml/min, dialysate flow of 500 ml/min and dialysis fluid temperature kept at 35°C.

Antibiotic administration was scheduled at fixed hours (8 a.m. and 8 p.m.), and patients received i.v. linezolid (Zyvoxid, Pfizer, Italy) twice a day at the dose of 600 mg in 100 ml of saline, by parenteral pump-driven infusion over 60 min, followed by a saline flush. Time of the start of the RRT was flexible due to the patient's and ICU needs, and it was not possible, for either logistic or ethical reasons, to standardize the time of drug administration with regard to the start of the next dialysis.

Blood samples were collected by direct venipuncture, before and after linezolid infusion, and from the arterial line of the extracorporeal circulation, before and at the end of each HD session; further blood samples were collected at 60, 120 and 180 min during a single haemodialysis session, as well as during a single interdialytic period at 30, 60, 120, 240, 300, 420 and 540 min after the end of antibiotic administration. In one patient, we also obtained dialysis fluid samples from the outlet of the dialysate/ultrafiltrate compartment of filters. To this purpose, we used a micropump-based partial collection system allowing continuous sampling of constant aliquotes of the total dialysate for the measurement of the total removal of the antibiotic. Total removal per session was calculated as the product of concentration in the spent dialysate and the amount of fluid collected [11].

We measured serum linezolid levels by a high-performance liquid chromatography/mass spectrometry method, as previously described in detail [11].

Trough linezolid levels were classified as ‘with HD’ when an HD session followed the preceding linezolid administration, and ‘without HD’ otherwise.

The difference between levels ‘with HD’ and those ‘without HD’ were examined by means of a mixed model for repeated measurements using the REML directive of the statistical software GenStat release 8.0 (VSN International Ltd, Waterhouse Street, Hemel Hempstead, UK) which allows the analysis of highly correlated and unbalanced data [12].

Population pharmacokinetics of multiple-dose regimen was estimated by fitting a two-compartment non-linear mixed-effect model with the package WinNonMix Version 2.0 (Pharsight Corporation, Cary, NC, USA). Clearance and volume of distribution were included as random parameters. Goodness of fit was checked by inspecting residual plots (observed and weighted residuals vs predicted, weighted residuals and observed/predicted vs time).



   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 References
 
We enrolled five critically ill patients (Table 1) with a severe sepsis or septic shock and renal failure (four with acute renal failure, one with chronic renal failure already on RRT) receiving intermittent 4 h HD with a daily (4/5) or alternate-day schedule (1/5), as well as linezolid administration as part of empiric antibiotic coverage. Their mean age was 75 years (range 68–82), the APACHE II score was 26.4 (range 23–29); in-hospital mortality was 3/5 (cause of death: septic shock in two, pulmonary haemorrhage in one). Piperacillin-clavulanate was administered to two patients and ceftazidime to the others as concomitant antibiotic therapy. All patients received vasoactive drugs (4/5 noradrenaline, dobutamine 2/5). No major side effects directly related to linezolid administration were observed.


View this table:
[in this window]
[in a new window]
  		Table 1. Patient demographics and clinical characteristics

 
Figure 1 displays the concentration vs time profile of linezolid in the serum of the patients. A total of 222 serum linezolid concentrations were available over 36 days of antibiotic therapy, during which patients underwent 31 HD sessions. Notably, linezolid concentrations were lower in patients with higher body weight (P<0.001), as can be visually appreciated from Figure 1 by comparing, for instance, patient A (body weight = 84.5 kg) with patient E (body weight 60 kg).


Figure 1
View larger version (20K):
[in this window]
[in a new window]
  		Fig. 1. Serum linezolid levels of the five patients from A to D. The horizontal dotted line represents the MIC90 (the 90% minimum inhibitory concentration) for S. aureus. Each of the upper panels report a series of timed samples (between day 2 and 3 in Panel A, between day 4 and 5 in Panel C, between day 3 and 4 in Panel E): the actual shape of the curve between two subsequent measures can be appreciated from these series of measurements. The data in Panel C are truncated at day 8.

 
Trough serum linezolid levels averaged 5.83 mg/l (range 1.48–15.84), exceeding 4.0 mg/l in 68.9% of the samples; however, the trough levels ‘with HD’ were lower than those ‘without HD’ (4.68 mg/l [range 1.48–9.07] vs 6.74 mg/l [range 2.04–15.84], P<0.001) (Figure 2).


Figure 2
View larger version (27K):
[in this window]
[in a new window]
  		Fig. 2. This figure reports the same data as Figure 1 but, additionally, hightlights haemodialysis data. The thick continuous lines connect pre- and post-HD linezolid concentrations. To make it easier, the visual comparison between trough levels measured ‘with HD’ and ‘without HD’ (i.e. whether or not HD was performed after linezolid administration), we artificially connected the relevant points using dashed red and blue lines, respectively. The trough levels ‘with HD’ were significantly lower than dose ‘without HD’ (P<0.001). HD, haemodialysis.

 
This phenomenon was mirrored by the population pharmacokinetics. Central distribution volume was 45.9 l (SE 6.5, between-subjects coefficient of variation 14.2%); when patients were off-dialysis, clearance was 4.4 l/h (SE 0.6, between-subjects coefficient of variation 12.7%), the elimination rate constant was 0.09 h–1 and the half-life 7.3 h; in contrast, when patients were on-dialysis, clearance was 6.0 l/h (SE 0.7, between-subjects coefficient of variation 11.0%), the elimination rate constant was 0.17 h–1 and the half-life 4.0 h. In the only patient in whom it was measured, a single HD session removed 221 mg of the antibiotic, i.e. 36.8% of the dose administered during a single dosing period.

The linezolid concentration decreased on average by 59.4% (SE 19.2) from the beginning to the end of dialysis treatment. The between-patient and within-patient coefficient of variation of the percent reduction was 13.1 and 21.7%, respectively. The relatively high value of the within-patient coefficient of variation was at least in part related to the varying time lags between the beginning of linezolid infusion and the beginning of the subsequent dialysis sessions. In fact, when the lag times are short, it is more likely for the dialysis treatment to take place during the distribution phase, thus enhancing the percent reduction in linezolid concentration. This relation can be visually appreciated from Figure 2, where the absolute reduction in linezolid levels is represented by the height of the thick black lines. For instance, Patient E, who had the lowest range in time lag (namely, from 1 to 1.5 h) clearly shows the lowest variability (percent reduction ranging from 68.8 to 71.7), whereas Patient B, who had the widest range in time lags (namely, from 1 min to 7 h) shows a high variability, with the percent reduction ranging from 28.5 to 68.8. Further studies are clearly needed to specifically address this issue. It is also worth noting that Patient E, who had the most consistent percent reduction, had also the largest difference between trough levels with and without HD (Figure 2). Although we did not specifically investigate the rebound of linezolid concentrations after dialysis, in three occasions (two patients), a sample was taken as soon as 1 h following the end of the dialysis session. With respect to post-dialysis concentrations, it was found to be increased by 0.14 (+10%), 0.50 (+9%) and 0.57 mg/l (+13%), respectively.



   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
References
 
The main finding of the present study is that HD significantly affects serum linezolid concentrations in the critically ill patients with renal failure on RRT.

In fact, in our patients trough concentrations ‘with HD’ were lower than ‘without HD’. Pharmacokinetic analysis confirmed this finding.

Because linezolid has a time-dependent antibacterial activity [5] and the clinical efficacy is maximal with drug exposure concentrations corresponding to t>MIC90 values of ≥85% [13], HD has the potential to impair treatment efficacy by reducing serum antibiotic levels for an appreciable length of time. This drawback might be particularly relevant in patients displaying trough antibiotic concentrations close to the MIC90 values for the clinically relevant pathogens.

We acknowledge the limitations of the present study.

First, we did not use a control group not requiring HD; thus, we could not examine the effect of severe renal failure on the time necessary to reach the steady state. The lack of a control group could have prevented us from controlling for the various confounding factors. Many factors can in fact profoundly alter the drug volume of distribution of the critically ill patient such as increased capillary leak, decreased plasma proteins, aggressive fluid resuscitation, altered haemodynamic status, organ dysfunction, complex drug–drug interactions [4,7,10].

Second, linezolid levels were lower in patients with higher body weight. We did not adjust the linezolid dose for patient weight, as concerning this issue there are specific recommendations in paediatric patients only. However, it is likely that matching the antibiotic dose for weight could have led to the achievement of more adequate concentrations with reduced time lag.

Third, the timing of the drug administration was fixed every 12 h; instead, haemodialysis timing was not fixed, but it changed according to the patient's and ICU needs. Although we recognize that this is not the optimum for a pharmacokinetic study, such an approach was mandatory in our patients for both logistical and ethical reasons, as it happens all too often in the ICU.

Fourth, it was not possible to modify antibiotic doses during the course of each patient's progress, as there was a time delay between drawing the samples and getting the results. As a matter of fact, linezolid concentration measurement is not routine in our hospital but only a research tool. Again, in every day clinical practice, linezolid concentration values are not used to adjust the doses.

Fifth, we collected data on linezolid removal by HD in one patient only. However, his data fully agree with the results we recently published in a single-dose study carried out on several patients, showing that fractional removal by a single haemodialysis session is about 32% of the last dose administered [11].

Sixth, our results cannot be extrapolated to intermittent renal replacement modalities based on different depurative mechanisms (convective or mixed convective and diffusive techniques, etc.), filters (membrane, surface, etc.), operational parameters (dialysis fluid rate and blood flow rates) and duration of the RRT.

Even with all these limitations, our study, besides providing further evidence that haemodialysis reduces the linezolid concentrations, highlights the important point that this modality of intermittent RRT has the potential to delay the achievement of the steady state. We recommend that, at least in some clinical circumstances, such as in patients with high body weight and dialysis sessions performed early after antibiotic infusion, a further antibiotic supplementation (say, at least a third of the full dose) should be given after dialysis. One might also consider the possibility of a loading dose at the start (say, 1200 mg) to achieve a steady state earlier.

In conclusion, on the basis of this pilot investigation, intermittent haemodialysis might negatively affect serum linezolid levels in critically ill patients with renal failure. It can also be expected to delay the time needed to achieve steady state levels. Further studies are needed to define better how the antibiotic dosage should be adjusted to achieve adequate serum antibiotic concentrations at an earlier time point.

Financial support for the study was provided by the institutional departmental funds.



   Acknowledgments
 
The authors thank Dante Tagliavini and Bruno Cermaglia for their technical assistance.

Conflict of interest statement. None declared.



   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 References
 

  1. Moellering RC. Linezolid: The first oxazolidinone antimicrobial. Ann Int Med 2003; 138: 135–142[Abstract/Free Full Text]
  2. MacGowan AP. Pharmacokinetic and pharmacodynamic profile of linezolid in healthy volunteers and patients with gram-positive infections. J Antimicrob Chemoter 2003; 51 [Suppl S2]: ii17–ii25
  3. Cepeda JA, Whitehouse T, Cooper B et al. Linezolid versus teicoplanin in the treatment of gram-positive infections in the critically ill: a randomised, double-blind, multicentre study. J Antimicrob Chemother 2004; 53: 345–355[Abstract/Free Full Text]
  4. Whitehouse T, Cepeda JA, Shulman R et al. Pharmacokinetic studies of linezolid and teicoplanin in the critically ill. J Antimicrob Chemother 2005; 5: 333–340
  5. Craig WA. Basic pharmacodynamics of antibacterials with clinical applications to the use of beta-lactams, glycopeptides and linezoid. Infect Dis North America 2003; 17: 479–501
  6. Brier ME, Stalker DJ, Aronoff GR et al. Pharmacokinetics of linezolid in subjects with renal dysfunction. Antimicrob Agents Chemother 2003; 47: 2775–2780[Abstract/Free Full Text]
  7. Pea F, Viale P, Lugano M et al. Linezolid disposition after standard dosages in critically ill patients undergoing continuous venovenous hemofiltration: a report of 2 cases. Am J Kidney Dis 2004; 44: 1097–102[Medline]
  8. Kraft MD, Pasko DA, DePestel DD et al. Linezolid clearance during continuous venovenous hemodiafiltration: a case report. Pharmacotherapy 2003; 23: 1071–1075[CrossRef][Web of Science][Medline]
  9. Mauro LS, Sharief I, Assaly R, Malhoura D. Determination of linezolid clearance via continuous hemodiafiltration [abstract]. Am J Kidney Dis 2001; 37: A77
  10. Meyer B, Kornek GV, Nikfardjam M et al. Multiple-dose pharmacokinetics of linezolid during continuous venovenous hemofiltration. J Antimicrob Chemother 2005; 56: 172–179[Abstract/Free Full Text]
  11. Fiaccadori E, Maggiore U, Rotelli C et al. Removal of linezolid by conventional intermittent hemodialysis, sustained low-efficiency dialysis, or continuous venovenous hemofiltration in patients with acute renal failure. Crit Care Med 2004; 32: 2437–2442[CrossRef][Medline]
  12. Paine RW. REML analysis of mixed models, in The Guide to Genstat, Release 8. Part 2, Statistic VSN International. 2005. Oxford, UK
  13. Stalker DJ, Jungbluth GL, Hopkins NK, Batts DH. Pharmacokinetics and tolerance of single- and multiple dose oral or intravenous linezolid, an oxazolidinone antibiotic, in healthy volunteers. J Antimicrob Chemother 2003; 51: 1239–1246[Abstract/Free Full Text]
Received for publication: 2. 7.05
Accepted in revised form: 26. 1.06


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Nephrol Dial TransplantHome page
J. T. Kielstein, J. Lorenzen, V. Kaever, H. Burhenne, M. Broll, C. Hafer, and O. Burkhardt
Risk of underdosing of ampicillin/sulbactam in patients with acute kidney injury undergoing extended daily dialysis--a single case
Nephrol. Dial. Transplant., July 1, 2009; 24(7): 2283 - 2285.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
O. Burkhardt, C. Hafer, A. Langhoff, V. Kaever, V. Kumar, T. Welte, H. Haller, D. Fliser, and J. T. Kielstein
Pharmacokinetics of ertapenem in critically ill patients with acute renal failure undergoing extended daily dialysis
Nephrol. Dial. Transplant., January 1, 2009; 24(1): 267 - 271.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
C. Ponticelli
New recommendations in the treatment of Gram-positive bacteraemia in dialysis patients
Nephrol. Dial. Transplant., January 1, 2008; 23(1): 27 - 32.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
21/5/1402    most recent
gfl048v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (5)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Fiaccadori, E.
Right arrow Articles by Cabassi, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fiaccadori, E.
Right arrow Articles by Cabassi, A.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?