Journal of Antimicrobial Chemotherapy (1999) 43, 523–527

Introduction

The optimal mode of administration of -lactam antibioticsin the treatment of bacterial infections remains controver-sial. Unlike aminoglycosides, which exhibit concentration-dependent killing rates, -lactam antibiotics achievemaximal killing at concentrations four or five times theMIC.1,2 Serum concentrations of -lactam antibioticsexceeding these values show no additional benefit. Thebactericidal activity does not increase. In addition, the significance of a carbapenem post-antibiotic effect (PAE)against Gram-negative bacteria is highly disputed.3,4 Thebactericidal effect of -lactams is closely related to the timeduring which the serum concentration of the antibioticremains above the MIC (T MIC).5

Intermittent administration of a drug results in highpeak and low trough serum levels. For -lactam antibioticsthis method of administration could result in concentra-tions below the MIC for the target organism over a longperiod of the dosing interval. During the last decade con-

tinuous infusion of -lactam antibiotics has been studied inorder to exploit these pharmacodynamic properties. Somein-vitro and in-vivo studies of continuous infusion of ceftazidime6–8 and meropenem published recently havedemonstrated the effectiveness of continuous infusion.9

The present study was performed firstly to compare thepharmacokinetic parameters of meropenem by continuousinfusion (CI) and intermittent administration (IA) in criti-cally ill patients and secondly to determine the possibilityof achieving therapeutic meropenem concentrations with a3 g iv CI over 24 h. The third aim was to study the applica-bility and side-effects of the IA regimen.

Materials and methods

Patients

The study was conducted in the intensive care unit of ateaching hospital. In accordance with ethical requirements,

523

Continuous infusion versus intermittent administration of meropenemin critically ill patients

Florian Thalhammera*, Friedericke Traunmüllera, Ibrahim El Menyawia, Michael Frassb,

Ursula M. Hollensteina, Gottfried J. Lockerb, Brigitte Stoiserb, Thomas Staudingerb,

Renate Thalhammer-Scherrerc and Heinz Burgmanna

aDepartment of Internal Medicine I, Division of Infectious Diseases; bDepartment of Internal Medicine I,Intensive Care Unit; cDepartment of Laboratory Medicine, University of Vienna, Waehringer Guertel 18–20,

A-1090 Vienna, Austria

This prospective crossover study compared the pharmacokinetics of meropenem by continuousinfusion and by intermittent administration in critically ill patients. Fifteen patients were ran-domized to receive meropenem either as a 2 g iv loading dose, followed by a 3 g continuousinfusion (CI) over 24 h, or by intermittent administration (IA) of 2 g iv every 8 h (q8h). Each regimen was followed for a period of 2 days, succeeded by crossover to the alternative regimenfor the same period. Pharmacokinetic parameters (mean SD) of CI included the following: con-centration at steady state (CSS) was 11.9 5.0 mg/L; area under the curve (AUC) was 117.5 12.9 mg/L·h. The maximum and minimum serum concentrations of meropenem (Cmax, Cmin) andtotal meropenem clearance (Cltot) for IA were 110.1 6.9 mg/L, 8.5 1.0 mg/L and 9.4 1.2 L/h,respectively. The AUC during the IA regimen was larger than the AUC during CI (P < 0.001). Inboth treatment groups, meropenem serum concentrations remained above the MICs for themost common bacterial pathogens. We conclude that CI of meropenem is equivalent to the IAregimen and is therefore suitable for treating critically ill patients. Further studies are neces-sary to compare the clinical effects of CI and IA in this patient group.

*Corresponding author. Tel: 43-1-40400-4440; Fax: 49-89-66617-18696; E-mail: florian.thalhammer@akh-wien.ac.at

© 1999 The British Society for Antimicrobial Chemotherapy

JAC

F. Thalhammer et al.

informed consent was obtained from patients or their nextof kin. Fifteen critically ill patients (four females, 11 males;Table I), admitted to the intensive care unit suffering fromsuspected or proven severe community or hospitalacquired infection, were eligible for enrolment in the study.Inclusion criteria demanded at least two of the following:(i) elevated C-reactive protein of 10 mg/dL (normal

0.5 mg/dL), (ii) at least one positive blood culture (Gram-negative or Gram-positive bacteria) or two positive bloodcultures growing coagulase-negative staphylococci, (iii)clinical signs of infection, (iv) respiratory tract infection(new infiltrate on a chest X-ray), or (v) positive urine cul-ture. Predicted duration of treatment had to be 4 days.Patients with known hypersensitivity to meropenem,imipenem/cilastatin, or other -lactam antibiotics, bleedingdisorders, a history of convulsions, or a decreased creatin-ine clearance were excluded from the study. Concomitantantimicrobial therapy with vancomycin was permitted toinclude cover for methicillin-resistant strains of Staphylo -coccus spp.

Study design

The study was performed as a prospective, randomized,crossover trial. All patients were randomized to receiveeither a 2 g iv loading dose of meropenem followed by adaily 3 g continuous infusion (CI; group 1) over 48 h or

intermittent administration (IA; group 2) of 2 g of mero-penem iv every 8 h for 2 days. After 2 days the patientsreceived the alternative dose regimen. If necessary, vanco-mycin was added to cover methicillin-resistant staphylo-cocci or enterococci. Steady-state concentrations ofmeropenem were expected to be achieved on day 2 ofadministration of CI or IA therapy.

Drug administration

For group 1 meropenem was administered via an infusionpump (Braun Melsungen, Melsungen, Germany). Onegram of meropenem (Optinem; Zeneca, Macclesfield, UK)was reconstituted according to the manufacturer’s guide-lines and then diluted in 50 mL of isotonic saline solution.New solutions were prepared every 8 h. The doses of 2 g ofmeropenem in group 2 were diluted in 100 mL of isotonicsaline solution and administered over 15 min.

Blood and urine sampling

Blood samples were taken at 0, 0.25, 0.5, 1, 2, 3, 6, 12, 24 and48 h after the start of CI and at 0, 0.5, 8, 8.5, 16, 16.5, 24, 24.5,32, 32.5, 40, 40.5 and 48 h after the start of IA. All bloodsamples were drawn from indwelling arterial cathetersafter discarding the first 10 mL of blood. After centrifuga-tion, serum was stored at –70°C until assayed.

Routine laboratory parameters (e.g. leucocyte andplatelet counts, renal and liver function tests) were deter-mined daily by the institution’s clinical chemistry depart-ment. Creatinine clearance was calculated by standardmethods.

Determination of meropenem concentration

The concentration of meropenem in serum was determinedby HPLC as described previously.10 The limit of detectionin serum was defined as the lowest concentration of mero-penem resulting in a signal-to-noise ratio of 3:1. The lowestdetection limit was 0.1 mg/L serum. The percentage recov-eries from sera were 94.6 3.1%, 92.4 4.3%, 95.2 3.0%and 91.9 4.0% with coefficients of variation (CV) of2.7%, 3.0%, 2.5% and 5.5% for assays of 5.0, 10.0, 50.0 and100.0 mg/L serum, respectively. The intra-assay repro-ducibility characterized by CV was 4.3%, 3.57% and 5.0%for assays of 5, 100 and 250 mg/L, respectively. The inter-assay reproducibility precision values calculated by CVwere 3.5%, 4.7% and 5.6% for assays of 5.0, 100.0 and 250.0 mg/L, respectively.

Interference studies were carried out with many sub-stances that might be administered with meropenem: -lactam antibiotics (penicillins, imipenem), aminoglycosides(gentamicin, tobramycin). None of these compounds wascoeluted with meropenem during chromatography. Duringspecificity studies all chromatograms were carefullychecked for skewed shouldering, or tailing peaks.

524

Table I. Patient’s characteristics

Patient Age Body weightno. (years) (kg) Diagnosis

Group 11 60 61 sepsis2 62 75 CPR, pneumonia3 67 80 CPR, sepsis4 55 93 CPR, pneumonia5 73 95 SIRS6 54 70 SIRS7 69 77 pneumonia

Group 28 63 120 ARDS, SIRS9 60 80 CPR, pneumonia

10 17 85 CPR11 60 60 sepsis12 45 85 pneumonia13 63 100 MCI, pneumonia14 50 75 SIRS15 31 100 SIRS

Mean 55.3 83.6SD 14.3 15.4

CPR, cardiopulmonary resuscitation; MCI, myocardial infarction;ARDS, acute respiratory distress syndrome; SIRS, systemicinflammatory response syndrome.

Meropenem and continuous infusion

Pharmacokinetic analysis

Meropenem data were analysed with a curve-fitting com-puter program, KINETICA 2.0 (MicroPharm Inter-national, Champs sur Marne, France). The volume ofdistribution at steady state (VSS), elimination rate constant(kel), concentration at steady state (CSS), serum half-life(t½), total meropenem clearance (Cltot) and the area underthe concentration–time curve over the dosing interval(AUC) were calculated for each patient.

Statistical analysis

Results are given as mean values standard deviation.Pharmacokinetic parameters were compared with the two-tailed Student’s t-test. Significance was defined as P 0.05.

Results

Fifteen patients (age 55.3 14.3 years, body weight 83.7 15.4 kg) suffering from pneumonia (n 7), sepsis (n 3),or systemic inflammatory distress syndrome (n 5) wereenrolled in the study (Table II). Their creatinine clearancewas 83.7 53.1 mL/min, the white blood cell count was 16.5

11.5 G/L (normal range 4–10 G/L) and the C-reactiveprotein level was 19.8 7.0 mg/100 mL. No pathogens, withthe exception of a coagulase-negative staphylococcus inone blood culture, were isolated during the study; nounderlying infection was detected by microbiological tests.

The pharmacokinetic parameters for continuous andintermittent administration of meropenem are presentedin Table II and the Figure. AUC and total meropenemclearance (Cltot) showed statistically significant differencesbetween groups 1 and 2. The IA group achieved an AUC of193.8 21.1 mg/L·h compared with 117.5 12.9 mg/L·h

for the CI group (P 0.001). The Cltot in the two patientgroups was 9.4 1.2 L/h (IA) and 7.7 1.4 L/h (CI),respectively (P 0.01). A continuous infusion of mero-penem 3 g/24 h achieved a steady-state concentration of11.9 5.0 mg/L (Figure). The minimum concentration inthe IA group was 8.5 1.0 mg/L. In both treatment groupsthe T MIC was 100% for the most common bacterialstrains found in ICU patients, throughout the observationperiod.

No adverse effects were observed in any patient duringthe study period.

Discussion

Constant concentrations of -lactam antibiotics above theMICs for target organisms can be achieved in vivo byadministering these compounds via a continuous infusion.Concentration-independent bacterial killing is also a property of -lactams. The pharmacodynamic parametercorrelating best with clinical outcome for these antibioticsis the time of serum concentration above the MIC (T

MIC).5,11 Intermittent administration is associated withvery high peak concentrations, which provide little addi-tional bactericidal activity but may be associated with moreside-effects. In between the doses, serum concentrationsoften fall below the MIC for the specific pathogen for aconsiderable time. Most authors agree that T MIC has tobe at least 40–50% of the dosing interval to achieve clinicaleffectiveness. Maximum killing is seen when T MIC is atleast 60–70%.12

525

Table II. Pharmacokinetic parameters

Intermittent Continuous Parameter administration infusion(unit) (mean S.D.) (mean S.D.)

kel (h–1) 0.32 0.12 NDt½ (h) 2.4 0.7 NDVSS (L) 26.6 3.2 25.9 5.7Cmax (mg/L) 110.1 6.9 NDCmin (mg/L) 8.5 1.0 NDCss (mg/L) ND 11.9 5.0AUCa (mg/L · h) 193.8 21.1 117.5 12.9Cltot

a (L/h) 9.4 1.2 7.7 1.4

aP 0.05.kel, elimination rate constant; t½, half-life; VSS, volume of distribution atsteady state; Cmax, maximum concentration of drug in serum; Cmin,minimum concentration of drug in serum; CSS, steady-stateconcentration in serum; AUC, area under the concentration–time curve;Cltot, total body clearance of meropenem; ND, not done.

Figure. Mean serum meropenem concentrations with SD (bars).( ), intermittent administration; ( ), continuous infusion.

F. Thalhammer et al.

The most effective administration mode of parenteralantibiotics remains controversial. Administration of -lactam antibiotics by CI results in constant serum levelswhich can be maintained above the MIC for the targetorganisms to promote maximal bactericidal activity. 12 Theextent of tissue penetration following CI appears to be similar to that following intermittent administration.13

Although the advantages of such a therapy are often discussed, only a few clinical trials have been per-formed.6,7,14–17 The main objectives of this study were toinvestigate the pharmacokinetics of continuous infusionversus intermittent administration of meropenem, in ICUpatients with severe infections, to determine the applica-bility of CI and to compare the side-effects and cost-effectiveness of the two treatment regimens.

In this study, a loading dose of 2 g of meropenem wasused followed by a CI of 36 8.4 mg/kg per day. Meanmeropenem serum concentrations in the CI group at steadystate were 11.9 5.0 mg/L, compared with trough levels of8.5 1.0 mg/L in the IA group (P 0.001). Serum levelswell above the MIC for most pathogens were achieved inboth groups (Table III).18 In addition, continuous bacteri-cidal activity during CI was achieved with only 50% of themeropenem dose used for the intermittent regimen. There-fore, the use of CI could lower the costs of antimicrobialtherapy.

A recent study by Benko et al.7 compared a 3 g cefta-zidime CI with 2 g q8h IA in critically ill patients, with similar results. In both treatment groups serum concentra-tions exceeded the MIC by four to five times. The serumbactericidal titres were equal in both regimens. The latter

study and other clinical trials with ceftazidime, demon-strated the effectiveness of CI and confirmed ourresults.6,14,19 Furthermore, Keil & Wiedemann described anin-vitro dynamic model to compare the antimicrobialeffects of CI versus IA of carbapenem antibiotics.9 Theevaluation of the corresponding kill curves for the twoadministration regimens showed improved antimicrobialactivity of CI (1 g per 24 h) compared with IA (1 g q8h).

The stability of an antibiotic is an important considera-tion if CI administration is to be used. At room tempera-ture most dissolved antimicrobials are stable for 24 h.20

However, the manufacturer’s guidelines state that oncemeropenem is reconstituted in isotonic saline solution it isstable at room temperature for 8 h. Thus, in this study, theantibiotic solution was changed every 8 h for the CI group.No problems with stability occurred during the studyperiod. This characteristic reduces the applicability ofmeropenem for CI for outpatient parenteral antibiotictherapy.

In conclusion, the rationale for using meropenem as CIis supported by the pharmacokinetic data of our study.Serum concentrations remained above the MIC for mostlikely target pathogens in all patients. A loading dose of 2 gof meropenem should be given initially to attain bacteri-cidal drug concentrations as rapidly as possible. Duringcontinuous infusion, no major adverse events related to theuse of CI were observed. Thus, meropenem can be admin-istered safely by CI. Additionally, a CI regimen can savecosts, bactericidal serum levels being achieved with only50% of the amount of drug used for IA. This study did notevaluate the clinical efficacy of the two different antibiotictreatment schedules. Further investigations are required toevaluate pharmacodynamic and economic perspectives inthe clinical setting.

References

1. Craig, W. A. & Ebert, S. C. (1990). Killing and regrowth of bac-teria in vitro: a review. Scandinavian Journal of Infectious Diseases74, Suppl., 63–70.

2. Soriano, F. (1992). Optimal dosage of -lactam antibiotics: timeabove the MIC and inoculum effect. Journal of Antimicrobial Chemo-therapy 30, 566–9.

3. Odenholt-Tornqvist, I. (1993). Studies on the postantibiotic effectand the postantibiotic sub-MIC effect of meropenem. Journal ofAntimicrobial Chemotherapy 31, 881–92.

4. Hanberger, H., Svensson, E., Nilsson, L. E. & Nilsson, M. (1995).Pharmacodynamic effects of meropenem on Gram-negative bac-teria. European Journal of Clinical Microbiology and Infectious Diseases 14, 383–90.

5. Drusano, G. L. (1988). Role of pharmacokinetics in the outcomeof infections. Antimicrobial Agents and Chemotherapy 32, 289–97.

6. Daenen, S., Erjavec, Z., Uges, D. R. A., De Vries-Hospers, H. G.,De Jonge, P. & Halie, M. R. (1995). Continuous infusion of cefta-zidime in febrile neutropenic patients with acute myeloid leukemia.

526

Table III. In-vitro activity of meropenem for clinicallyimportant Gram-positive and Gram-negative pathogens18

Organism MIC90 Range(no. tested) (mg/L) (mg/L)

Gram-positive organismsEnterococcus faecalis (377) 8.0 1.0–16.0Staphylococcus aureus (MS; 791) 0.25 0.016–4.0Staphylococcus epidermidis (MS; 70) 0.5 0.06–8.0Streptococcus pneumoniae (165) 0.03 0.008–1.0

Gram-negative organismsAcinetobacter anitratus (149) 1.00 0.06–4.0Citrobacter freundii (107) 0.06 0.03–0.25Enterobacter cloacae (409) 0.12 0.008–0.5Escherichia coli (1458) 0.03 0.008–0.06Haemophilus influenzae (275) 0.06 0.016–0.25Klebsiella pneumoniae (210) 0.03 0.016–0.06Proteus mirabilis (424) 0.12 0.03–0.5Pseudomonas aeruginosa (778) 4.0 0.06–8.0Serratia marcescens (218) 0.12 0.016–1.0

MIC90, minimal inhibitory concentration for 90% of isolates; MS,methicillin-susceptible.

Meropenem and continuous infusion

European Journal of Clinical Microbiology and Infectious Diseases14, 188–92.

7. Benko, A. S., Cappelletty, D. M., Kruse, J. A. & Rybak, M. J.(1996). Continuous infusion versus intermittent administration of ceftazidime in critically ill patients with suspected Gram-negativeinfections. Antimicrobial Agents and Chemotherapy 40, 691–5.

8 . Mouton, J. W., Vinks, A. A. T. M. M. & Punt, N. C. (1997). Pharma-cokinetic–pharmacodynamic modeling of activity of ceftazidime during continuous and intermittent infusion. Antimicrobial Agentsand Chemotherapy 41, 733–8.

9. Keil, S. & Wiedemann, B. (1997). Antimicrobial effects of contin-uous versus intermittent administration of carbapenem antibiotics inan in vitro dynamic model. Antimicrobial Agents and Chemotherapy41, 1215–9.

10. Thalhammer, F., Schenk, P., Burgmann, H., El Menyawi, I.,Hollenstein, U. M., Rosenkranz, A. R. et al. (1998). Single-dosepharmacokinetics of meropenem during continuous venovenoushemofiltration. Antimicrobial Agents and Chemotherapy 42,2417–20.

11. Hatano, K., Wakai, Y., Watanabe, Y. & Mine, Y. (1994). Simula-tion of human plasma levels of -lactams in mice by multiple dosingand the relationship between the therapeutic efficacy and pharma-codynamic parameters. Chemotherapy 40, 1–7.

12. Craig, W. A. (1998). Pharmacokinetic/pharmacodynamicparameters: rationale for antibacterial dosing of mice and men. Clin-ical Infectious Diseases 26, 1–12.

13. Craig, W. A. & Ebert, S. C. (1992). Continuous infusion of -lactam antibiotics. Antimicrobial Agents and Chemotherapy 36,

2577–83.

14. Visser, L. G., Arnouts, P., van Furth, R., Mattie, H. & van denBroek, P. J. (1993). Clinical pharmacokinetics of continuous intra-venous administration of penicillins. Clinical Infectious Diseases 17,491–5.

15. Nicolau, D. P., Nightingale, C. H., Banevicius, M. A., Fu, Q. &Quintiliani, R. (1996). Serum bactericidal activity of ceftazidime: con-tinuous infusion versus intermittent injections. Antimicrobial Agentsand Chemotherapy 40, 61–4.

16. James, J. K., Palmer, S. M., Levine, D. P. & Rybak, M. J. (1996).Comparison of conventional dosing versus continuous-infusion vancomycin therapy for patients with suspected or documentedGram-positive infections. Antimicrobial Agents and Chemotherapy40, 696–700.

17. Di Filippo, A., De Gaudio, A. R., Novelli, A., Paternostro, E.,Pelagatti, C., Livi, P. et al. (1998). Continuous infusion of vancomycinin methicillin-resistant staphylococcus infection. Chemotherapy 44,63–8.

18. Pryka, R. D. & Haig, G. M. (1994). Meropenem: a new car-bapenem antimicrobial. Annals of Pharmacotherapy 28, 1045–54.

19. Kuzemko, J. & Crawford, C. (1989). Continuous infusion of ceftazidime in cystic fibrosis. Lancet ii, 385.

20. Craig, W. A. (1995). Antibiotic selection factors and descriptionof a hospital-based outpatient antibiotic therapy program in the USA.European Journal of Clinical Microbiology and Infectious Diseases14, 636–42.

Received 9 July 1998; returned 8 October 1998; revised 26 October1998; accepted 1 December 1998

527