ORIGINAL ARTICLE

Meropenem versus imipenem/cilastatin as empirical monotherapy for

serious bacterial infections in the intensive care unit
C.Verwaest, on behalf of the Belgian Multicenter Study Group

Department of Intensive Care Medicine, University Hospital Gasthuisberg, Leuven, Belgium

Objective To compare the e¤cacy and tolerability of meropenem and imipenem/cilastatin as empirical
monotherapy in intensive care unit (ICU) patients with serious bacterial infections.
Methods A multicenter, open-label, randomized, parallel-group trial was conducted in Belgium,
evaluating empirical monotherapy with meropenem or imipenem/cilastatin (both 1 g  /8  h intravenously) in
ICU patients with one or more of the following infections caused by sensitive pathogens: lower respiratory
tract infection (LRTI) in ventilated patients, intra-abdominal infection or sepsis.
Results The overall satisfactory clinical response rate at the end of randomized treatment was 77.0% (67/
87) with meropenem and 68.1% (62/91) with imipenem/cilastatin (di¡erence 8.9%; 95% con¢dence
interval ÿ4.2% to 21.9%; P  ˆ  0.185).The two drugs produced similar satisfactory clinical response rates
against LRTIs: 68.3% (41/60) with meropenem versus 68.6% (35/51) with imipenem/cilastatin.
Meropenem appeared to be slightly more e¡ective against intra-abdominal infections: 95.5% (21/22) versus
76.7% (23/30), respectively. All ¢ve meropenem recipients with sepsis had a satisfactory clinical response,
compared to 40.0% (4/10) of those who received imipenem/cilastatin.The overall satisfactory bacteriologic
response rate was 67.1% (49/73) with meropenem and 60.3% (44/73) with imipenem/cilastatin (di¡erence
6.9%; 95% con¢dence interval ÿ8.7% to 22.4%; P  ˆ  0.389).The predominant pathogens were Escherichia
coli, Enterobacter spp. and Pseudomonas aeruginosa. No incidences of drug-related nausea and vomiting were
reported, but one probable drug-related seizure occurred in the imipenem/cilastatin group.
Conclusions Meropenem is at least as e¤cacious (clinically and bacteriologically) as imipenem/cilastatin
for the empirical monotherapy of serious bacterial infections in ICU patients, and it can therefore be
considered a useful option in this setting. Moreover, meropenem is well tolerated and o¡ers several potential
advantages, including greater in vitro activity against Gram-negative pathogens and the option of bolus
administration.
Keywords bacterial infections, meropenem, imipenem/cilastatin, intensive care, resistance

Accepted 17 September 1999

Clin Microbiol Infect 2000: 6: 294^302

infections (LRTIs) were the most prevalent infections

INTRODUCTION
(64.7%), followed by urinary tract (17.6%) and bloodstream
Serious bacterial infections are common in patients in the infections (12.0%).The most frequently reported bacterial iso-
intensive care unit (ICU) and they result in considerable mor- lates were Enterobacteriaceae (mainly Escherichia coli, Klebsiella
bidity and mortality. Approximately 45% of the 10,038 ICU spp. and Enterobacter spp.), Staphylococcus aureus, Pseudomonas aer-
patients included in the European Prevalence of Infection in uginosa and coagulase-negative staphylococci. Fifty-¢ve per
Intensive Care (EPIC) Survey of 1992 had at least one bacterial cent of ICU-acquired infections were polymicrobial.
infection [1]. Pneumonia and other lower respiratory tract In this setting, empirical antibiotic treatment must therefore
have a broad spectrum of antibacterial activity, covering
Corresponding author and reprint requests: C.Verwaest, Department of Inten- Gram-positive and Gram-negative aerobes and anaerobes.
sive Care Medicine, University Hospital Gasthuisberg, Herestraat 49, 3000 Leu- Traditionally, this has been achieved by using combinations of
ven, Belgium
antibiotics, usually a b-lactam agent (e.g. a third-generation
Tel:  ‡   32   016344021
Fax:  ‡   32   016344015 cephalosporin) plus an aminoglycoside or antianaerobic agent.
E-mail: Charles.Verwaest@uz.kuleuven.ac.be However, the usefulness of many b-lactams, including the

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases
 Verwaest Meropenem versus imipenem/cilastatin in serious infections áñä

third-generation cephalosporins, is threatened by the spread Patients were excluded if they were pregnant or breastfeed-
of plasmid-mediated, extended-spectrum b-lactamases ing, if they had a hypersensitivity to any b-lactam, if they suf-
(ESBLs) in Enterobacteriaceae [2]. Up to 23% of Klebsiella fered from hepatic impairment, neutropenia or cystic ¢brosis,
isolates from European ICUs may produce these enzymes [3]. or if they had a history of CNS disease or other disorder likely
Furthermore, over-expression of inducible group I b-lacta- to cause convulsions. Patients with severe underlying disease
mases by Enterobacter, Citrobacter, Serratia and Pseudomonas spp. who were unlikely to complete at least 48  h of study treatment
compromises the activity of cephalosporins and b-lactam/b- were excluded, as were those receiving probenecid. Patients
lactamase inhibitor combinations, while constitutive hyper- with secondary infections at sites other than those speci¢ed
producers are readily selected during therapy with these agents above were also ineligible. Patients could only enter the trial
[4]. once.
The carbapenems possess the widest antibacterial spectra
and greatest b-lactamase stability of all b-lactams, and there-
fore these agents o¡er a realistic option for monotherapy in Therapeutic regimen
serious bacterial infections. The only two carbapenems avail- Patients were randomized to receive either meropenem or imi-
able outside Japan, meropenem and imipenem/cilastatin, are penem/cilastatin, each administered intravenously at a dosage
active against most clinically important pathogens and are of 1 g  every 8  h. Randomization was strati¢ed according to the
stable to the vast majority of serine-based b-lactamases three types of primary infection. Meropenem was adminis-
(including ESBLs and Bush Group I enzymes) [5, 6]. How- tered either as a bolus injection or as an infusion over a period
ever, important di¡erences exist between the two compounds of 20^30  min. Imipenem/cilastatin was infused over 40^60  
[7], which may favor the use of meropenem in the ICU set- min (or at a slower rate if the patient experienced nausea or
ting. These include the greater in vitro activity of meropenem vomiting). Since both study drugs are chie£y eliminated ren-
against the predominant Gram-negative pathogens [5] and its ally, their dosages were adjusted according to the degree of
stability to renal dehydropeptidase-I (DHP-I), which permits renal impairment. An additional dose of study drug was given
its administration without a DHP-I inhibitor such as cilastatin after hemodialysis.The recommended duration of therapy was
(which can accumulate in renal failure [8]). In addition, mero- 5^10  days (maximum 28  days). Treatment for a minimum of
penem is well tolerated by the central nervous system (CNS) 48  h was required for the assessment of e¤cacy. No other anti-
with regard to seizures [7] and can be administered by bolus bacterials were allowed during the study, except in the event
intravenous injection. of surgical prophylaxis or treatment failure.
This study compared the e¤cacy and tolerability of mero-
penem and imipenem/cilastatin as empirical monotherapy in
ICU patients with serious bacterial infections. Clinical assessment
The overall condition of the patient at entry was assessed
using the Acute Physiology and Chronic Health Evaluation
MATERIALS AND METHODS
(APACHE II) scoring system [9]. Clinical signs and symptoms
This was a multicenter, open-label, randomized, parallel- associated with infection were evaluated within 3  days prior to
group trial conducted in Belgium to compare intravenous commencing treatment and daily during treatment.
meropenem and imipenem/cilastatin in ICU patients with Sepsis was de¢ned as a clinical entity characterized by one
severe or life-threatening infections caused by sensitive patho- or more of the following signs and symptoms [10]: fever (>  
gens. 38.3   C), chills, leukocytosis, hyperventilation, hypothermia,
skin lesions, septic embolism, change in mental status, hypo-
tension, disseminated intravascular coagulation (DIC) and
Patients organ failure.
These comprised ICU adults (aged r18  years) requiring par- Intra-abdominal infections were described as either
enteral antibacterial therapy for one or more of the following abscesses or peritonitis originating from the following intra-
infections: LRTI (in mechanically ventilated patients), intra- abdominal organs: stomach, duodenum, biliary tract, pan-
abdominal infections and sepsis. For eligibility, at least one creas, appendix, small intestine and colon.
responsible pathogen isolated at study entry had to be suscepti- For LRTI to be diagnosed, the following signs and symp-
ble to both study drugs. In patients with multiple infection toms had to be present: pulmonary in¢ltration thought to be
sites, the most serious was designated as the primary site, and due to infection on chest X-ray and at least two of the follow-
additional sites were ranked in order of importance. This trial ing criteriaöpurulent sputum (< 1  0 squamous epithelial
gained local Ethics Committee approval, and informed con- cells,  >  25 white blood cells, and a pathogen should be cul-
sent was obtained from all patients or their next of kin. tured), fever and leukocytosis. The respiratory sample should

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
áñå Clinical Microbiology and Infection, Volume 6 Number 6, June 2000

be obtained by endotracheal aspirate, bronchoalveolar lavage, for treatment failure or new infections) and at follow-up, if
protected specimen brush, pleural £uid or lung biopsy. possible.
Hospital-acquired infections were de¢ned as late-onset The response was considered satisfactory when the original
infections (>  72  h) or any postoperative infections, whereas pathogen(s) were eradicated or presumed eradicated (i.e. when
community-acquired infections were of early onset (<  72  h). further sampling was not considered justi¢ed because of clini-
The clinical response was assessed at the end of treatment cal cure/improvement). The response was considered unsatis-
or at the time when additional antibiotics or antifungals were factory if the primary pathogen persisted (or was presumed to
added in the event of treatment failure and/or upon the devel- have persisted) or if a new pathogen isolated at the original
opment of new infections or superinfections. Clinical infection site during study therapy required antibacterial
response was classi¢ed as cured (complete remission of local treatment (i.e. superinfection occurred). At follow-up, the
and systemic signs and symptoms), improved (improvement bacteriologic response could also be classi¢ed as relapse.
of local and systemic signs and symptoms but without com-
plete resolution) or unchanged/worse (no improvement or
deterioration of signs and symptoms). If relevant, clinical Statistical methods
response was reassessed at a follow-up examination 2^4  weeks The primary endpoint was the percentage of patients in each
post-treatment. At follow-up, clinical response could also be treatment group with a satisfactory clinical response at the end
classi¢ed as relapse (initial cure/improvement followed by a of randomized treatment. A satisfactory response rate of 85%
general decline and worsening of the clinical condition). was assumed for imipenem/cilastatin. To demonstrate equiva-
The clinical response was considered satisfactory if the lence between meropenem and imipenem/cilastatin, the
patient was cured or improved, and unsatisfactory if the upper 95% con¢dence interval (CI) (a ˆ 0.05) for the di¡er-
patient was unchanged/worse or experienced a relapse. ence between response rates was not to exceed 15%. In order
The tolerability of the study drugs was assessed by monitor- to ensure a power of 80% (b  ˆ  0.20), 90 fully evaluable
ing adverse events and routine clinical laboratory tests. patients were required in each group. Since only 85% of ran-
domized patients were expected to be evaluable for response,
we aimed to recruit a total of approximately 210 patients.
Bacteriologic assessment
Secondary endpoints included the satisfactory clinical
Cultures were obtained from sites appropriate to the infection response rate at the end of treatment in bacteriologically eva-
immediately (or within 3  days) prior to starting treatment, luable patients, and the bacteriologic response at the end of
during treatment if clinically indicated, preferably immedi- treatment for evaluable patients. In addition, a subgroup ana-
ately after study treatment was discontinued or changed, and lysis according to the type of infection was performed.
at 2^4  weeks follow-up, if possible. For each endpoint, the proportions of patients with a satis-
Acceptable specimens from patients with intra-abdominal factory response in each group were compared using a chi-
infections included peritoneal £uid, drain £uid, abscess mate- square test. The di¡erence in proportions was estimated,
rial or pus. For patients with LRTI, endotracheal aspirates, together with 95% CI (calculated using a normal approxima-
bronchoalveolar lavage or protected specimen brush samples, tion), and statistical signi¢cance determined using the chi-
abscess material, pleural £uid and hemocultures or blood cul- square test.
tures were accepted.Two blood cultures, preferably from sepa-
rate sites or from an intra-arterial line (taken at least 30  min
apart), were collected before treatment from patients with sus- RESULTS
pected sepsis. Susceptibility testing was performed according
to standard accepted disk sensitivity criteria [11]. Patients
Disks loaded with meropenem (10  mg) or Rosco tablets of A total of 212 patients entered the study, 107 in the meropenem
meropenem were provided by AstraZeneca Pharmaceuticals group and 105 in the imipenem/cilastatin group. The demo-
for sensitivity testing using Kirby ^ Bauer (or equivalent) graphic and baseline clinical characteristics of the two groups
determinations. Commercially available disks or tablets were were similar (Table 1  ). The mean APACHE II score was
used for sensitivity testing of imipenem for assessment of slightly higher in the meropenem group than in the imipe-
comparative activity. nem/cilastatin group, as was the number of patients who had
Resistance to meropenem was de¢ned as a zone size of received previous antibacterial therapy. The majority of
R11 m   m, and resistance to imipenem as a zone size of R13   patients (66.9%) were mechanically ventilated. LRTI was the
mm (R13  mm and R16  mm, respectively, for Rosco tablets). most common infection diagnosis, followed by intra-abdom-
The bacteriologic response was assessed at the end of the inal infection and sepsis. The mean (þ Standard Deviation
therapy (or at the time when additional antibiotics were given (SD)) durations of treatment in the meropenem and imipe-

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
 Verwaest Meropenem versus imipenem/cilastatin in serious infections áñæ

Table  1 Demographic and baseline clinical characteristics of clinically evaluable patients

Meropenem Imipenem/cilastatin
Criteria (n  ˆ  87) (n  ˆ  91)

Gender (M/F) 59/28 59/32

Mean age (years) [range] 64.2 [20±87] 56.4 [6±87]
Aged r65  years 57 (65.5%) 48 (52.7%)
Mean APACHE II score (standard deviation) 18.4 (8.6) 15.7 (8.1)
APACHE II score
0±10 9 (10.3%) 18 (19.8%)
11±20 43 (49.4%) 46 (50.5%)
21±30 24 (27.6%) 23 (25.3%)
>  31 8 (9.2%) 4 (4.4%)
Not recorded at study entry 3 (3.4%) 0 (0.0%)
Previous antibiotic therapy 56 (64.4%) 46 (50.5%)
Mechanically ventilated 63 (72.4%) 56 (61.5%)
Recent surgery 20 (23.0%) 27 (29.7%)
Active neoplasm 4 (4.6%) 4 (4.4%)
Primary infection diagnosis
Lower respiratory tract 60 (69.0%) 51 (56.0%)
Pneumonia 49 (56.3%) 37 (40.7%)
Intra-abdominal 22 (25.3%) 30 (33.0%)
Peritonitis 19 (21.8%) 23 (25.3%)
Abscess 4 (4.6%) 11 (12.1%)
Sepsis 5 (5.7%) 10 (11.0%)
Primary infection source
Hospital 69 (79.3%) 64 (70.3%)

nem/cilastatin groups were 9.8  þ  4.9  days and 9.8  þ  5.3  days, clinical response, compared with 40.0% (4/10) of those in the
respectively. Almost 60% of the meropenem patients received imipenem/cilastatin group (Table  2 ).
drug by bolus injection. The overall satisfactory clinical response at the end of treat-
Twenty patients were excluded from the analysis due to ment in bacteriologically evaluable patients was 76.7% (56/73)
protocol violations (Figure 1), leaving 192 evaluable patients in the meropenem group and 72.6% (53/73) in the imipenem/
(94 meropenem, 98 imipenem/cilastatin). Of these, 178 (87 cilastatin group. The respective response rates for meropenem
meropenem, 91 imipenem/cilastatin) were evaluable for clini- and imipenem/cilastatin in the following infections were
cal response at the end of treatment, and 146 (73 meropenem, 69.2% (36/52) and 73.8% (31/42) for LRTI, 94.4% (17/18) and
73 imipenem/cilastatin) were evaluable for bacteriologic 82.6% (19/23) for intra-abdominal infections, and 100% (3/3)
response at the end of treatment. The reasons for clinical non- and 37.5% (3/8) for sepsis.
evaluability are outlined in Figure 1. The overall satisfactory bacteriologic response rates at the
end of treatment were 67.1% (49/73) in the meropenem group
and 60.3% (44/73) in the imipenem/cilastatin group (di¡er-
ence 6.9%; 95% CI, ÿ8.7% to 22.4%; P  ˆ  0.389). The two
Ef®cacy drugs showed similar bacteriologic e¤cacy against LRTIs and
Among evaluable patients, the overall satisfactory clinical intra-abdominal infections, although meropenem produced a
response rates at the end of randomized treatment were 77.0% higher response rate in the small number of patients with sep-
(67/87) in the meropenem group and 68.1% (62/91) in the imi- sis (Table  2 ).
penem/cilastatin group (di¡erence 8.9%; 95% CI ÿ4.2% to Similar percentages of isolated pathogens were eradicated
21.9%; P  ˆ  0.185). Meropenem and imipenem/cilastatin pro- (or presumed to be eradicated) by meropenem (88/108; 81.5%)
duced similar satisfactory clinical response rates against and imipenem/cilastatin (94/128; 73.5%). The predominant
LRTIs, while meropenem appeared to be slightly more e¡ec- pathogens were Enterobacteriaceae (Table  3), against which
tive against intra-abdominal infections (Table  2 ). All ¢ve meropenem appeared to be slightly more e¡ective than imipe-
patients with sepsis treated with meropenem had a satisfactory nem/cilastatin.

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
áñð Clinical Microbiology and Infection, Volume 6 Number 6, June 2000

Figure  1 Summary of patient numbers and reasons for exclusion/non-evaluability.aReasons for protocol violations necessitating
exclusion from analysis: secondary infections (two patients with skin/skin structure infections and three with urinary tract infec-
tions), concomitant antibiotics at entry (6), fever of unknown origin at entry (1), and no ventilatory support LRTI patients (1) in the
meropenem group; and secondary infection (2 patients with skin/skin structure infections), concomitant antibiotics at entry (3),
fever of unknown origin at entry (1), and no ventilatory support LRTI patients (1) in the imipenem/cilastatin group.
b
Reasons for clinical non-evaluability: death during therapy (2 patients),  <  48  h monotherapy (2), known resistant pretherapy
pathogens (1), amoebic infection (1), and vancomycin added for superinfection (1) in the meropenem group; and < 48  h monother-
apy (5, of which 3 were due to death), known resistant pretherapy pathogens (1) and incorrect treatment regimen (1 day of mero-
penem) (1) in the imipenem/cilastatin group.

Bacterial superinfection occurred in nine bacteriologically 60 adverse events were noti¢ed in 44 patients (41.9%) treated
evaluable meropenem recipients and 11 of those who received with imipenem/cilastatin. Adverse events considered to be
imipenem/cilastatin (Table  4). related to the study drug were reported in 3.7% (4/107) of
patients treated with meropenem and 2.9% (3/105) of those
who received imipenem/cilastatin. No incidences of drug-
Safety and tolerability related nausea and vomiting were reported in either group,
Both drugs were well tolerated. In the meropenem group, 66 but one probable drug-related seizure was reported in a patient
adverse events were reported in 50 patients (46.7%), whereas treated with imipenem/cilastatin.

Table  2 Satisfactory clinical and bacteriologic response rates at end of treatment in evaluable patients

Response ratea Meropenem Imipenem/cilastatin

Clinical
Overall 77.0% (67/87) 68.1% (62/91)
Lower respiratory tract infection 68.3% (41/60) 68.6% (35/51)
Intra-abdominal infection 95.5% (21/22) 76.7% (23/30)
Sepsis 100.0% (5/5) 40.0% (4/10)
Bacteriologic
Overall 67.1% (49/73) 60.3% (44/73)
Lower respiratory tract infection 61.5% (32/52) 57.1% (24/42)
Intra-abdominal infection 77.8% (14/18) 69.6% (16/23)
Sepsis 100.0% (3/3) 50.0% (4/8)

a
Percentages of patients cured or improved (clinical) or with pathogens eradicated or presumed eradicated (bacteriologic).

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
 Verwaest Meropenem versus imipenem/cilastatin in serious infections áññ

Table  3  Organism eradication and presumed eradication rates at end of treatment according to causative pathogen

Eradicated/total Eradicated/total
Clinical isolate Meropenem Imipenem/cilastatin

Gram-negative aerobes
Escherichia coli 14/17 13/20
Enterobacter spp. 11/13 6/14
Pseudomonas aeruginosa 7/10 5/10
Klebsiella pneumoniae 6/6 3/3
Proteus mirabilis 7/7 0/2
Haemophilus spp. 2/2 5/5
Morganella morganii 1/1 5/7
Acinetobacter spp. 2/5 4/7
Serratia marcescens 2/2 4/5
Klebsiella oxytoca 1/1 3/4
Citrobacter freundii 1/1 2/3
Proteus vulgaris 1/2 ±
Others 4/5 4/6
Total 59/72 (81.9%) 54/86 (62.8%)
Gram-positive aerobes
Staphylococcus aureus 6/10 6/6
Enterococcus spp. 4/6 7/8
Others 11/12 10/10
Total 21/28 (75.0%) 23/24 (95.8%)
Anaerobes
Total 8/8 (100.0%) 17/18 (94.4%)
All clinical isolates 88/108 (81.5%) 94/128 (73.5%)

Table  4  Superinfections and colonizations that occurred in bacteriologically evaluable patients

Meropenem (n  =  73) Imipenem/cilastatin (n  =  73)

Organisms
No. of patients with No. of patients with No. of patients No. of patients with
superinfection colonizationa with superinfection colonizationa

Gram-negative aerobes
Enterobacter cloacae ± ± 1 0
Proteus mirabilis ± ± 2 5
Providencia stuartii ± ± 1 0
Pseudomonas aeruginosa 2 1 4 0
Stenotrophomonas maltophilia 1 3 0 1
Total 3 4 8 6
Gram-positive aerobes
Corynebacterium xerosis 1 0 ± ±
Enterococcus faecalis 0 2 0 1
Methicillin-resistant 2 0 ± ±
Staphylococcus aureus
Staphylococcus epidermidis 3 6 3 3
Staphylococcus spp. ± ± 0 3
Total 6 8 3 7
Fungi/yeasts
Aspergillus spp. ± ± 0 1
Candida albicans 3 4 2 2
Candida spp. 1 1 1 1
Yeast (unspeci®ed) 0 1 1 0
Total 4 6 4 4

a
Colonization = any new organism isolated at the primary site but not requiring any additional antimicrobial therapy.

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
âòò Clinical Microbiology and Infection, Volume 6 Number 6, June 2000

There were nine deaths during therapy in the meropenem rate with meropenem. However, the number of patients in
group (8.4% of recruited patients) and 14 (13.3%) in the imi- some infection subgroups was small, particularly for those
penem/cilastatin group. with sepsis.
In the meropenem group there were six deaths (6/9; The most common infection diagnosis was LRTI, and all
66.7%) considered to be probably or possibly related to infec- of these patients were mechanically ventilated. Pneumonia is
tion, versus 10 (10/14; 71.4%) in the imipenem/cilastatin the hospital-acquired infection most likely to lead to the death
group. of critically ill patients [19, 20]. Controversy remains about the
No deaths in either treatment group were considered to be most accurate way to diagnose this infection and the appropri-
related to study therapy. Death appeared to be related to pre- ate timing of antibiotic therapy in relation to the clinical suspi-
therapyAPACHE II score, although the data are limited by the cion of pneumonia [21]. Of all bronchoscopic and non-
small numbers of patients involved [9]. bronchoscopic diagnostic tests, endotracheal aspirate culture
During the follow-up period, after the end of treatment, might be the most practical, and was mostly used in this study.
there were 22 deaths (20.6%) in the meropenem group and 14 Additional research is needed to determine the applicability of
in the imipenem/cilastatin group (13.3%). None of these other diagnostic procedures in daily clinical practice. Some
deaths was considered to be drug-related. As most patients at studies have shown that antibiotic therapy may have a favor-
this stage of their critical illness are heavily colonized and/or able impact on the outcome of pneumonia, with lower mor-
infected in one or several organ systems, it seems hazardous to tality rates being observed if appropriate therapy is prescribed
directly attribute death to infection. [22^24]. Recent trials have con¢rmed that timely and appro-
In the meropenem group, six deaths (27.3%) were probably priate empirical antibiotic therapy is crucial and can improve
not infection-related (acute vascular injury (n  ˆ 1  ), cardiac survival in patients with pneumonia [25, 26]. The two carba-
arrest (n  ˆ 1  ), brain damage (n  ˆ  2 ), withdrawal of all therapy penems proved similarly e¤cacious in these patients. Interest-
(n  ˆ  2 )); four deaths (18.1%) were probably infection-related ingly, another recent study performed in seriously ill patients
(refractory septic shockötwo within 3  days post trial); and 12 [18] noted a slightly better response rate with meropenem
deaths (54.6%) were possibly infection-related (single organ compared to imipenem/cilastatin in patients with nosocomial
failureöacute respiratory distress syndrome (ARDS), acute LRTIs (89% versus 76%). Unlike in the present study, how-
renal failure (ARF)öor multiple organ failure (MOF)). ever, not all patients with LRTIs in the study of Garau et al [18]
In the imipenem group, six deaths (42.9%) were probably were mechanically ventilated, which may explain the di¡er-
not related to infection (cerebral damage (n  ˆ 1  ), cardiac arrest ence in response rates. The severity of the underlying disease
(n  ˆ  2 ), acute respiratory failure (n  ˆ  3)); three deaths (21.4%) and acute illness of the a¡ected patients largely account for the
were probably infection-related (refractory septic shocköall poor outcome in the present study. Most infections are caused
within 5  days post trial); and ¢ve deaths (35.7%) were possibly by Gram-negative bacilli, especially Pseudomonas species, and
related to infection (ARDS, MOF). up to 55% of cases are caused by polymicrobial infection [27].
Prior antimicrobial therapy is a risk factor both for pneumonia
and for infections with di¤cult-to-treat organisms, leading to
DISCUSSION
poor response to therapy and a poor outcome [28].
Meropenem monotherapy (1 g  /8  h) has been shown to be at Di¡erences between treatment groups in the satisfactory
least as e¤cacious as combination antibacterial therapy in the bacteriologic response rates generally mirrored the clinical
treatment of a variety of infections, including nosocomial ¢ndings, with meropenem proving at least as e¤cacious as
LRTI [12], intra-abdominal infection [13, 14] and sepsis [15]. imipenem/cilastatin overall and for each infection subgroup.
Previous randomized studies have shown meropenem and Eradication rates according to pathogen were generally as
imipenem/cilastatin to be similarly e¤cacious at this dosage would be predicted from the in vitro spectra of the two carba-
for the empirical treatment of serious infections in ICU penems. Notable di¡erences, however, were the apparently
patients [16^18]. more favorable results with meropenem against important
In the present study, meropenem produced a slightly higher Gram-negative ICU pathogens, particularly Escherichia coli,
overall satisfactory clinical response than imipenem/cilastatin Enterobacter spp. and Pseudomonas aeruginosa. According to a
(77.0% versus 68.1%), although the di¡erence did not reach similar study, both meropenem and imipenem/cilastatin
statistical signi¢cance. This ¢nding does not appear to be proved highly e¡ective against the anaerobes, while a some-
related to any di¡erence between the patients in the two what higher eradication rate of Gram-positive aerobes was
groups. Rather, analysis of the infection subgroups indicates observed in the imipenem/cilastatin group [16].
that apparent di¡erences between the drugs in terms of their Both drugs were well tolerated with regard to serious
e¤cacy against intra-abdominal infection and sepsis were adverse events. The risk of seizures is a concern with higher
responsible for the slightly better overall satisfactory response doses of imipenem/cilastatin, especially in the presence of

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
 Verwaest Meropenem versus imipenem/cilastatin in serious infections âò"

impaired renal function and/or CNS disease [29]. However, Our thanks also go to Thierry Van Hedent, Steven Huysse,
patients with CNS disease were excluded from this study, and Geert Byttebier, Freya De Beurme, Peggy Van Wijmeersch,
dosage reduction was performed according to the degree of Philippe Peeters, Hilde De Roeck and Kristof Daems, (Astra-
renal impairment. This may explain why only one probable Zeneca, Brussels, Belgium) for their logistic support.
drug-related seizure was reported with imipenem/cilastatin. This study was supported by a grant from AstraZeneca
This compares with a seizure rate of 6% of patients receiving Pharmaceuticals.
the same dosage of imipenem/cilastatin for the treatment of
severe pneumonia [30]. The absence of seizures with merope-
nem is in line with the documented low seizurogenic activity REFERENCES
of this agent in animal studies [31, 32] and in clinical trials that
1. Vincent JL, Bihari DJ, Suter PM et al. The prevalence of nosoco-
used doses up to 6  g /day [7], some of which included patients mial infection in intensive care units in Europe. Results of the
with meningitis [33, 34]. European Prevalence Infection Intensive Care (EPIC) Study.
No drug-related instances of nausea or vomiting were JAMA 1995; 274: 639^44.
reported in this study. Previous studies, mostly (but not exclu- 2. Sirot D. Extended spectrum plasmid-mediated b-lactamases. J
Antimicrob Chemother 1995; 36(suppl A): 19^34.
sively) in neutropenic patients, have found nausea and vomit-
3. Livermore DM,Yuan M. Antibiotic resistance and production of
ing to be a problem with imipenem/cilastatin [35^39]. In extended-spectrum b-lactamases amongst Klebsiella spp. from
contrast, nausea and vomiting is not a signi¢cant problem intensive care units in Europe. J Antimicrob Chemother 1996; 38:
with meropenem, even after bolus administration of the drug 409^24.
[7]. This adverse e¡ect, which appears to be related to the 4. Sanders CC, Sanders WE. b-lactam resistance in Gram-negative
bacteria: global trends and clinical impact. Clin Infect Dis 1992; 15:
dosage and rate of infusion of imipenem/cilastatin [35, 40],
824^39.
was avoided in this study by administering this agent slowly 5. Edwards JR. Meropenem: a microbiological overview. J Antimi-
over 40^60  min. Meropenem was administered by bolus crob Chemother 1995; 36(suppl A): 1^17.
injection in almost 60% of patients in the present study. The 6. Balfour JA, Bryson HM, Brogden RN. Imipenem/cilastatin: an
option of bolus administration is particularly useful in ICU update of its antibacterial activity, pharmacokinetics and thera-
peutic e¤cacy in the treatment of serious infections. Drugs 1996;
patients, in whom renal failure and left ventricular failure are
51: 99^136.
common and £uid overload/hypervolemia must be avoided. 7. Norrby SR, Gildon KM. Safety pro¢le of meropenem: a review
In summary, meropenem is at least as e¤cacious (clinically of nearly 5000 patients treated with meropenem. Scand J Infect Dis
and bacteriologically) as imipenem/cilastatin for the empirical 1999; 31(1): 3^10.
monotherapy of serious bacterial infections in ICU patients, 8. Verpooten GA, Verbist L, Buntinx AP, Entwistle LA, Jones KH,
De Broe ME. The pharmacokinetics of imipenem (thienamycin-
and it can therefore be considered a useful option in this set-
formanmidine) and the renal dehydropeptidase inhibitor cilasta-
ting. Moreover, meropenem is well tolerated and o¡ers several tin sodium in normal subjects and patients with renal failure. Br J
potential advantages over imipenem/cilastatin, including Clin Pharmacol 1984; 18: 183^93.
greater in vitro activity against Gram-negative pathogens and 9. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE
the option of bolus administration. II: a severity of disease classi¢cation system. Crit Care Med 1985; 13:
818^29.
10. Bone RC, Sprung CL, Sibbald WJ. De¢nitions for sepsis and
organ failure. Crit Care Med 1992; 20(6): 724^6.
11. National Committee for Clinical Laboratory Standards. Perfor-
ACKNOWLEDGMENTS
mance standards for antimicrobial disk susceptibility tests, 4th edn.
This work was presented in part at the 17th International Sym- Approved Standard M2-A4.Villanova: NCCLS, 1990.
12. Sieger B, Berman SJ, Geckler RW, Farkas SA, for the Meropenem
posium on Intensive Care and Emergency Medicine, Brussels,
Lower Respiratory Infection Group. Empiric treatment of hospi-
Belgium, 18^21 March 1997, and at the 8th European Congress tal-acquired lower respiratory tract infections with meropenem or
of Clinical Microbiology and Infectious Diseases, Lausanne, ceftazidime with tobramycin: a randomized study. Crit Care Med
Switzerland, 25^28 May 1997. The Belgian Multicenter Study 1997; 25: 1663^70.
Group comprised the following (number of patients 13. Condon RE, Walker AP, Sirinek KR et al. Meropenem versus
tobramycin plus clindamycin for treatment of intraabdominal
enrolled): M. Aoun, Brussels (3); J. Bockaert, Mechelen (7); F.
infections: results of a prospective, randomized, double-blind
Colardyn, Gent (22); R. De Jongh, Genk (5); G. Willems, B. clinical trial. Clin Infect Dis1995; 21: 544^50.
De Waele, Brussels (9); M. Genard, Mons (12); G. Huylen- 14. HuizingaWKJ,Warren BL, Baker P et al. Antibiotic monotherapy
broeck, Aalst (7); E. Installë, Mont-Godinne (7); P. Jamaer, with meropenem in the surgical management of intra-abdominal
Hasselt (3); L. Janssen, Geel (28); M. Laurent, La Louvie© re (2); infections. JAntimicrob Chemother 1995; 36(suppl A): 179^89.
15. Solberg CO, Sjursen H. Safety and e¤cacy of meropenem in
P. Lauwers, Leuven (40); J. C. Legrand, Charleroi (18); H. Les-
patients with septicaemia: a randomised comparison with ceftazi-
sire, Brussels (22); D. Noirot, P. Paris, Lie© ge (9); G. Nollet, J. dime, alone or combined with amikacin. J Antimicrob Chemother
Verbeke, Aalst (18). 1995; 36(suppl A): 157^66.

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302
âòá Clinical Microbiology and Infection, Volume 6 Number 6, June 2000

16. Colardyn F, Faulkner KL, on behalf of the Meropenem Serious antibiotics: experience with imipenem/cilastatin. Am J Med 1988;
Infection Study Group. Intravenous meropenem versus imipe- 84: 911^18.
nem/cilastatin in the treatment of serious bacterial infections in 30. Fink MP, Snydman DS, Niederman MS et al. Treatment of severe
hospitalized patients. JAntimicrob Chemother 1996; 38: 523^37. pneumonia in hospitalized patients: results of a multicenter, ran-
17. Hartenauer U, Kljucar S, Bender HJ, Weilemann S, Nusser H, domized double-blind trial comparing intravenous cipro£oxacin
Bodmann KF. Meropenem versus imipenem/cilastatin for the with imipenem/cilastatin. Antimicrob Agents Chemother 1994; 38:
treatment of serious bacterial infections at ICU. Antiinfect Drugs 547^57.
Chemother 1997; 15: 65^70. 31. Patel JB, Giles RE. Meropenem: evidence of lack of proconvulsive
18. Garau J, Blanquer J, Cobo L et al. Prospective, randomised, multi- tendency in mice. J Antimicrob Chemother 1989; 24(suppl A): 307^
centre study of meropenem versus imipenem/cilastatin as empiric 9.
monotherapy in severe nosocomial infections. Eur J Clin Microbiol 32. De Sarro A, Ammendola D, Zappala M, Grasso S, De Sarro GB.
Infect Dis 1997; 16: 789^96. Relationship between structure and convulsant properties of
19. Niederman MS, Craven DE, Fein AM, Schultz DE. Pneumonia some b-lactam antibiotics following intracerebroventricular
in the critically ill hospitalized patient. Chest 1990; 97(1): 170^81. microinjection in rats. Antimicrob Agents Chemother 1995; 39: 232^
20. Fagon JY, Chastre J, Hance AJ, Montravers P, Novara A, Gibert 7.
C. Nosocomial pneumonia in ventilated patients: a cohort study 33. Schmutzhard E, Williams KJ, Vukmirovits G et al. A randomised
evaluating attributable mortality and hospital stay. AmJ Med 1993; comparison of meropenem with cefotaxime or ceftriaxone for the
94: 281^8. treatment of meningitis in adults. J Antimicrob Chemother 1995;
21. Fagon JY, Chastre J, Domart Y et al. Nosocomial pneumonia in 36(suppl. A): 85^98.
patients receiving continuous mechanical ventilation. Am Rev 34. Klugman KP, Dagan R, the Meropenem Meningitis Study
Respir Dis 1989;139: 877^84. Group. A randomised comparison of meropenem with cefotax-
22. Celis R, Torres A, Gatell JM, Almela M, Rodriguez-Roisin R, ime for the treatment of bacterial meningitis. Antimicrob Agents
Agusti-Vidal A. Nosocomial pneumonia. A multivariate analysis Chemother 1995; 39: 1140^6.
of risk and prognosis. Chest1988; 93 (2): 318^24. 35. Norrby SR,Vandercam B, LouieT. Imipenem/cilastatin vs amika-
23. Torres A, Aznar R, Gatell JM et al. Incidence, risk, and prognosis cin plus piperacillin in the treatment of infections in neutropenic
factors of nosocomial pneumonia in mechanically ventilated patients: a prospective, randomized, multi-clinic study. Scand J
patients. Am Rev Respir Dis 1990; 142(3): 523^8. Infect Dis Suppl 1987; 52: 65^78.
24. Alvarez-Lerma F. Modi¢cation of empiric antibiotic treatment in 36. Winston DJ,Winston G, Bruckner DA, Champlin RE. Beta-lac-
patients with pneumonia acquired in the intensive care unit. ICU- tam antibiotic therapy in febrile granulocytopenic patients. Ann
Acquired Pneumonia Study Group. Intensive Care Med 1996; 22(5): Intern Med 1991; 115: 849^59.
387^94. 37. Freifeld AG,Walsh T, Marshall Tet al. Monotherapy for fever and
25. Luna CM, Vujacich P, Niederman MS et al. Impact of BAL data neutropenia in cancer patients: a randomized comparison of cefta-
on the therapy and outcome of ventilator-associated pneumonia. zidime versus imipenem. J Clin Oncol 1995; 13: 165^76.
Chest1997; 111(3): 676^85. 38. Norwegian Study Group. Imipenem/cilastatin as monotherapy
26. Kollef MH, Sherman G,Ward S, Fraser VJ. Inadequate antimicro- in severe infections: comparison with cefotaxime in combination
bial treatment of infections: a risk factor for hospital mortality with metronidazole and cloxacillin. Scand J Infect Dis 1987; 19:
among critically ill patients. Chest1999;115(2): 462^74. 667^75.
27. Rello J, Quintana E, Ausina V et al. Incidence, etiology, and out- 39. Mandell LA, Turgeon PL, Ronalds AR, Canadian Trials Group.
come of nosocomial pneumonia in mechanically ventilated A prospective randomized trial of imipenem/cilastatin vs. clinda-
patients. Chest1991; 100(2): 439^44. mycin-tobramycin in the treatment of intra-abdominal and pelvic
28. Kollef MH, Silver P, Murphy DM, Trovillion E. The e¡ect of infections. CanJ Infect Dis1993; 4: 279^87.
late-onset ventilator-associated pneumonia in determining 40. Wang C, Calandra GB, Aziz MA, Brown KR. E¤cacy and safety
patient mortality. Chest 1995; 108(6): 1655^62. of imipenem/cilastatin: a review of world-wide clinical experi-
29. Calandra G, Lydick E, Carrigan J, Weiss L, Guess H. Factors pre- ence. Rev Infect Dis 1985; 7(suppl 3): 528S ^36S.
disposing to seizures in seriously ill infected patients receiving

= 2000 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 6, 294±302