key: cord-0006713-tampgv7c
authors: Böhme, Angelika; Bergmann, L.; Hoelzer, D.; Just-Nübling, G.; Shah, P. M.; Stille, W.
title: A randomized study of imipenem compared to cefotaxime plus piperacillin as initial therapy of infections in granulocytopenic patients
date: 1995
journal: Infection
DOI: 10.1007/bf01713564
sha: f4e12125fec5404fc07f4ce12ea7f65beb38f58d
doc_id: 6713
cord_uid: tampgv7c

The objective of the presented, randomized study was to compare the efficacy of antimicrobial monotherapy with imipenem (3×0.5g/d) to a combination therapy with cefotaxime (3×2g/d) plus piperacillin (3×4g/d) for empirical treatment of infections in neutropenic patients. In 165 patients, 237 infectious episodes were evaluable. The overall response rate of patients treated with cefotaxime plus piperacillin was 67/115 (58%), of those treated with impienem 66/122 (54%). In patients not responding to the initial therapy regimen within 2 or 3 days, the antimicrobial therapy was modified. After therapy modification 85/100 patients were cured. Fever of unknown origin (FUO) showed the most favourable course compared to other infection types, with a response in 46/59 (78%) and in 35/50 (70%) cases, respectively. In comparison, pneumonias were successfully treated in only 3/21 (14%) and 7/37 (19%) cases. Even including patients with modified therapy, only 66% (21/32) of pneumonia episodes responded. The unfavourable results in pneumonias is mainly due to the high rate of 13 systemic mycoses in this group (22%). Overall, a similar response was observed in patients treated with cefotaxime plus piperacillin in comparison with imipenem. In primary bacteremias however, an advantage was observed in patients treated with imipenem (20/27; 74%) compared with cefotaxime plus piperacillin (11/23; 48%).

Severe, life-threatening infections are a major risk for patients with hematologic malignancies undergoing intensive chemotherapy with its long-term granulocytopenia [1, 2] . Up to now empiric broad-spectrum antibiotic combination therapy is the accepted practice of treatment in febrile granulocytopenic patients. Combinations usually include an aminoglycoside plus an extended-spectrum cephalosporin or penicillin or two [3-1actam antibiotics. In the last decade, gram-positive bacteria and especially fungi have emerged as increasingly frequent pathogens resulting in empirical several-step treatment regimens [1, [3] [4] [5] [6] [7] . But also cost-effective, less toxic monotherapies, e.g. with imipenem or ceftazidime, have been investigated in febrile neutropenic patients, compared to combination therapy or to one another. The study results (Table 1 ) have indicated that combination and monotherapy are similarly effective [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] . Monotherapy, however, has been controversially discussed especially because of resistance problems. Ceftazidime is more effective than imipenem in the treatment of Pseudomonas sp., but less active against gram-positive cocci in comparison to other cephalosporins; about 20% of Pseudomonas aeruginosa and more than 80% of Stenotrophomonas maltophilia strains are resistant against imipenem, whereas imipenem is the most effective [3-1actam against gram-positive cocci (except oxacillin-resistant staphylococci) compared with other 13-1actams [18, 19] . Furthermore, imipenem has an acceptable spectrum of side-effects and no essential organospecific toxicity. Mostly, imipenem has been used at a dose of 4×0.5 g to 4×1.0 g ( Table 1) per day with success rates ranging from 70% to 90%. A non-comparative, open pilot study performed by our group with imipenem at a dose of 1.5g per day resulted in a satisfactory response of 65% [20] . Thus a larger open prospective, randomized study was planned to compare the efficacy of monotherapy with imipenem to a "standard" combination therapy (cefotaxime plus piperacillin) as initial treatment of infections in grannlocytopenic patients. Interim results of this study have been reported at the Annual Congress of the German and Austrian Society of Hematology and Oncology in Essen, October 10-13, 1993 [21] .

This was an open, prospective randomized study in which patients were enrolled after informed consent was obtained. Patients were centrally randomized by the physician on duty in the emergency room not participating in the study. Patient selection: Patients with the following diagnoses were entered into the study: acute leukemia, chronic myelogenous leukemia with blast excess, Hodgkin's disease, myelodysplastic syndrome RAEB and RAEB-T, non-Hodgkin's lymphoma. Inclusion criteria were fever of -> 38.0°C and granulocytopenia < 1,000/ixl. Patients with suspected non-infectious fever (for example transfusion reaction, drug fever), age < 18 and > 80 years, known hypersensitivity against [3-1actams, intravenous antibiotic therapy < 7 days prior to randomization were excluded. Patients 65  57  66  54  ALL  14  12  16  13  NHL  30  26  33  27  others  6  5  7  6   Remission induction  40  35  34  28  Remission  48  42  52  43  Progression  12  10  12  10  Relapse  15  13  24  19 Neutropenia from start of fever (days)  1-12  68  59  62  51  > 12  23  20  24  19  Not evaluable  24  21  36 30 N: number; CEF/PIP: cefotaxime/piperacinin; IMI: imipenem.

were allowed to reenter the study for a second infectious episode provided the episodes were more than 4 weeks apart.

Procedures: Before the start of antibiotic therapy all patients underwent a thorough physical examination and chest x-ray. Specimens for cultures, at least two blood cultures from peripheral veins and central venous catheters, cultures from urine, throat and other suspected sites were collected. In patients with no response to the initial therapy the procedures were repeated before modification of antibiotic therapy. In patients with pneumonia and in those not responding after 6 days of antibiotic treatment, a thoracic CT-scan was taken for establishing a pulmonary aspergillosis even at an early stage of disease [22] . If possible, a bronchoalveolar lavage was obtained in patients with pneumonia. Moreover, the following serologic tests were performed: herpes simplex virus, varicella-zoster virus, cytomegalovirus, Epstein Barr virus, Aspergillus, Candida, and in pneumonia patients, additionally Chlamydia, Mycoplasma and Legionella. Treatment: The patients were randomized to receive daily either cefotaxime 3×2g plus piperacillin 3x4g or imipenem/cilastatin 3×0.5g as a short infusion [20] . Patients who did not respond after 48-72 h to initial therapy were additionally treated with gentamicin 3 × 40 mg daily. In patients with microbiologically proven documented infection, the antibiotic therapy was modified according to susceptibility testing. In responders with persistent granulocytopenia the regimen was continued for at least 5 days, with increasing granulocytes for 3 days after defervescence. In patients with pneumonia the antibiotic treatment was given until complete regression of the pulmonary infiltrates.

The following types of infection were defined: 1) FUO (fever of unknown origin) -suspected infection without pathogen identification or infectious loci, 2) pneumonia without detectable pathogen, 3) pneumonia with evidence of an infectious agent proven by microbiological, serological or radiological methods or by autopsy, 4) primary bacteremia: positive blood culture or catheter luminal cultures without infectious focus, 5) others -clinically documented infections, except pneumonias, with or without proven pathogens.

Evaluation of response: Response was defined as stable defervescence and complete regression of infectious loci without any modification of initial antibiotic therapy. A non-response was defined as initial transient defervescence with recurrence of fever, or clinical progression and lack of defervescence within the first 3 days, which resulted in therapy modification. In some patients the initial therapy without modification was continued in spite of persistent fever > 38.0°C. These patients showed improvement of other clinical features or had a causative organism identified which proved to be susceptible to initial treatment. These episodes were defined as delayed response. The infectious episodes were followed until remission or death of patients even if the study was finished earlier.

Statistics: Statistical significance was calculated by Fisher's exact test.

F r o m July 1990 to July 1993,195 neutropenic patients with 267 infectious episodes entered the study; 237 episodes were evaluable. In 115 episodes cefotaxime plus piperacillin were administered, 122 were treated with imipenem/cilastatin. There were no differences b e t w e e n the randomization arms in age, sex, diagnosis, remission status of malignancy, duration of granulocytopenia ( Table 2) , central venous catheters (29% vs. 36%), corticosteroids (31% vs. 28%) or therapy with hematopoietic growth factors G-CSF or G M -C S F (37% vs. 43%). In 50% vs. 48% of the episodes, the patients received antibacterial prophylaxis with co-trimoxazole, in 23% vs. 28% with ciprofloxacin.

The classification of infection is listed in Table 3 . The majority of the patients was diagnosed as FUO. In 16 episodes (7%) an abdominal, in six (3%) a skin infection was diagnosed. Three patients suffered from infection of ENT and two from urinary tract infections.

Forty-one/one hundred fifteen and 51/122 episodes in the combination and imipenem group, respectively, were microbiologicalty documented (Table 4 ). In both groups gram-positive organisms (only episodes with at least two positive cultures from peripheral vein or central venous catheter were accepted) were isolated more frequently than gram-negative bacteria. Thirteen patients were diagnosed to have systemic fungal infection, in most cases Aspergillus fumigatus was found. Viral infections and infections caused by Chlamydia, Mycoplasma, mycobacteria and anaerobes were rare. The pathogens were isolated from peripheral veins in 90% (in 21% of them additionally from central venous catheters). Out of 58 pneumonias only 34 were microbiologically documented. 19 gram-positive, ten gram-negative, 13 fungi and four other microorganisms (mycobacteria, anaerobes, Mycoplasma and adenovirus) were diagnosed. Twelve patients had mixed infections. Sixty-four percent of the pathogens were diagnosed on the first day of fever: 85% of those found in primary bacteremias, 36% of organisms causing pneumonias and 75% of those in other infections. In 45% of the pneumonia episodes the pathogen was diagnosed after the fifth febrile day. Most of these pathogens were fungi, but also grampositive bacteria.

Fifty-eight percent of the episodes treated with cefotaxime plus piperaciUin and 54% treated with imipenem responded (n.s., Table 5 ). The cure rate of the FUO episodes was 78% vs. 70%, of pneumonias 14% vs. 19%, respectively (only clinically documented 33% vs. 40%, with proven pathogens 0% vs. 5%). The patients with primary bacteremias responded in 48% vs. 74% (p=0.08) and other infections in 58% vs. 50% of the episodes. Though not statistically significant, more patients with positive blood cultures responded to empirical treatment with imipenem (75%) than the combination arm (48%). Table 6 shows the better response of the pathogens in primary bacteremias than in pneumonias and other clinically documented infections. One hundred episodes without response of initial therapy regimens were additionally treated with gentamicin or according to susceptibility, resulting in an 85% response rate. Twenty-three (10%) patients died during the infectious episode. In 14 cases an autopsy was performed. The causes of death were defined according to autopsy results or clinical course: Infection as cause of death was seen in 70%, the other patients died from their malignancy and/or 

The increase of granulocytes was the most important prognostic factor resulting in response rates of 96% and 98% (cefotaxime/piperacillin and imipenem, respectively) compared to 28% and 33% with persistent granulocytopenia (Table 7) . Prolonged granulocytopenia from start of fever was correlated with decreasing response: 88% and 87% with neutropenia of 1-3 days, 37% and 34% with neutropenia > 7 days. The different responses of AML, ALL and NHL are caused by the different duration of granulocytopenia due to various chemotherapy regimens. Central venous catheters had no influence on response rates (removal of catheters was not necessary as all patients with positive catheter cuttures improved soon after start of antibiotic therapy). Primary p/aeumonias or other clinically documented infections diagnosed within the first 3 days of fever had an overall response rate of 50% compared to 7% when diagnosed later.

Hypersensitivity occurred in four (3%) of the patients treated with cefotaxime plus piperaciUin, in three (2%) of the patients treated with imipenem. Seven (6%) of the patients in the imipenem group suffered from nausea (WHO grade 1-2), in two patients nausea improved when infusion time was increased, in one patient by prophylaxis with alizapride. Nephro-or hepatotoxicity did not occur; there were no signs of CNS-symptoms under imipenem.

The initial monotherapy with imipenem resulted in similar efficacy compared to the combination of cefotaxime with piperacillin (response rate 58% vs. 54%, n.s.). Both therapy groups were identical with respect to time until defervescence. Similar response rates of monotherapy compared to combinations have been reported in previous trials. In the first trials carried out in the early eighties predominantly ceftazidime was investigated. The response rates with ceftazidime were 89% compared to 78% of cephalothin/gentamicin/carbenicillin [15] and 79% vs. 61% with ceftazidime/piperacillin [3] . In recent years, imipenem has been investigated as monotherapy. Imipenem treatment resulted in cure rates of 81% in comparison to 63% with amikacin/piperacillin [13] , 81% compared to 78% of cefotaxime/piperacillin and 82% with ceftazidime/piperacillin [17] . Imipenem was as effective as cefuroxime/gentamicin or cephalothin/gentamicin (91% vs. 74%) [9] . In comparison with ceftazidime, there was no significantly different cure rate (56% vs. 77%) [11] . Prospective, randomized trials have not reported any essential difference of efficacy in monotherapy with imipenem or ceftazidime compared to the combination treatments. The lower total response rates in our study may be due to the inclusion of patients with predominantly acute leukemias who have longer periods of granulocytopenia than patients with solid tumors.

Although study results have not reported significant dif-ferences of efficacy between combination and monotherapy, combination therapy has been widely recommended, especially in documented infections [15] . Recent studies using imipenem, however, suggest at least similar response rates compared to combination therapy even in documented infections. In primary bacteremias, therapy with imipenem appears to be more advantageous than the combination therapy applied (89% vs. 53%), as reported by

Cornelissen [9] and by our own study (74% vs. 48%).

Whether the high response rate to imipenem is due to lower endotoxin release, as reported by Dofferhoff and colleagues, can be debated [23] . Likewise, a more favourable outcome of only clinically documented infections with imipenern compared to combination has been found (88% vs. 58-60%) [171.

In various infection types, we noted essential differences of cure rates: F U O 74%, bacteremias 62%, pneumonias 17%, which responded most unfavourably. Similar findings have been reported in the study of the Paul Ehrlich Society (PEG) with response of 65-71% in F U O episodes and about 60% in pneumonias [6] . Other authors also reported an unfavourable outcome in pneumonias compared to F U O with 41%-63% vs. 77%-81% [10] or 43%-60% vs. 67%-74% [11]. In our study, pneumonias with documented pathogens had especially low response rates, caused by the high incidence of fungi, partially As-pergilIus, which was seen in 10/58 pneumonia episodes.

The results of the PEG study have drawn attention to the problem of pulmonary mycoses [6] reporting a high incidence of fungi especially in pneumonias occurring in the late infection stage. Thus an early empirical treatment with antimycotics in pneumonia episodes is recommended and is under investigation in ongoing trials using amphotericin B (present study of PEG [24] ) or fluconazole [25] .

But not only fungal pneumonias caused worse response in pneumonia patients: 5/10 gram-negative, all streptococcal infections, 2/3 S. aureus and 4/5 Staphylococcus epidermidis infections with pulmonary infiltrates failed to respond in the first 72 h and were rated as failures. None of the strains was resistant in vitro to the study drugs and thus the susceptibility pattern cannot explain the poor response rate in pneumonia. The microorganisms, mostly found in our study patients, were gram-positive bacteria (23%) followed by gram-negatives and fungi (15% and 5%, respectively). The most frequent fungus was Aspergillus (ten episodes), systemic Candida infections were rarely seen (three episodes). These results disagree with those of other studies observing Candida sp. to be the predominant fungi [1, 6] . Resistance against imipenem especially in Pseudomonas sp. and coagulase-negative staphylococci is one of the main objections to initial monotherapy with imipenem. An analysis of resistance patterns at the Centre of Internal Medicine, University Frankfurt, between 1990 and 1992 has not shown an increase of imipenem-resistant P. aeruginosa strains (17% to 16%). Moreover, P. aeruginosa bacteremias were rarely seen in our study patients. Over the year there has been an increase from 17% to 23% in imipenem-resistant coagulase-negative staphylococci. However, as these bacteremias are not fulminant, treatment modification was initiated after the results of sensitivity testing became available. None of the patients with a bacteremia with S. epidermidis died. In contrast to favourable courses of most infections due to gram-positive bacteria, infections caused by streptococci of the viridans group, particularly Streptococcus mitis, have been reported to be associated with a high rate of lethal courses by development of ARDS [26] . The empirical treatment with vancomycin-containing combinations has been widely discussed. The PEG study has not observed any advantage of additional treatment with vancomycin in non-responders of initial therapy [6] . In contrast, several randomized studies reported more favourable results in patients with initial treatment with vancomycin [7, 27] . The duration of granulocytopenia was found to be the main prognostic factor for the outcome of infections, as also described by others [3, 6, 10] . The importance of the duration of granulocytopenia on occurrence and course of infection is reflected by the recommendation of colony-stimulating factors (G-CSF or GM-CSF) [14] . Using G-CSF a shortening of granulocytopenic period and activation of granulocyte function can be achieved [28] [29] [30] . In patients with acute lymphoblastic leukemia simultaneously receiving chemo-radiotherapy and randomized G-CSF vs. placebo severe infections such as sepsis or pneumonia occurred in 14% of the G-CSF group, but in 30% of the placebo group [29] .

In conclusion, the initial monotherapy with imipenem can be recommended in infections of granulocytopenic patients. Monotherapy with a drug such as imipenem can be more cost-effective in initial empirical treatment in such patients. Increasing incidence of systemic mycoses, especially in patients with pulmonary lesions, would warrant an early empiric antimycotic treatment. Colony-stimulating factors for earlier hematopoietic reconstitution and activation of neutrophil function should be considered in febrile patients with long-term granulocytopenia. 

Infection in cancer patients

Approaching the controversies in antibacterial management of cancer patients

Randomized trial of beta-lactam regimens in febrile neutropenic cancer patients

Guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever

and the EORTC Antimierobial Therapy Cooperative Group: Empiric antimicrobial therapy for febrile granulocytopenic cancer patients: lessons from four EORTC trials

Empirische antimikrobielle Therapie bei nentropenischen Patienten. Ergebnisse einer multizentrischen Studie der Arbeitsgruppe Infektionen in der H~imatologie der Paul-Ehrlich-Gesellschaft

Vancomycin, ticarcillin and amikacin compared with ticarcillin-clavulanate and amikacin in the empirical treatment of febrile, neutropenic children with cancer

Ja A prospective, randomized comparison of ceftazidime and ciprofloxacin as initial empiric therapy in neutropenic patients with fever

Imipenem versus gentamicin combined with either cefuroxime or cephalothin as initial therapy for febrile neutropenic patients

A randomized study of ceftazidime compared to ceftazidime and tobramycin for the treatment of infections in cancer patients

Ceftazidime versus imipenem-cilastatin as initial monotherapy for febrile neutropenic patients

Monotherapy in serious hospital-acquired infections: a clinical trial of ceftazidime versus imipenem/cilastatin

Imipenem/cilastatin versus amikacin plus piperacillin in the treatment of infections in neutropenic patients: a prospective, randomized multi-clinic study

Management of fever in patients with cancer and treatment-induced neutropenia

A randomized trial comparing ceftazidime alone with combination antibiotic therapy in cancer patients with fever and nentropenia

Monotherapy for empiric treatment of fever in granulocytopenic cancer patients

Betalactam antibiotic therapy in febrile granulocytopenic patients: a randomized trial comparing cefoperazone plus piperacillin, ceftazidime plus piperacillin, and imipenem alone

Incidence of multi-resistance in gram-negative aerobes from intensive care units of 10 German hospitals. Scand

Therapie yon Infektionen bei abwehrgeschw~chten Patienten mit lmipenem/Cilastatin

Cefotaxime/piperacillin versus imipenem/cilastatin for initial antimicrobial therapy and early treatment with itraconazole in febrile neutropenic patients

Diagnostisehe Validit~t bildgebender Verfahren bei Detektiou und Verlanfsbeobachtung des gef~ginvasiven AspergiUus-Befalls der Lunge

Effects of different types and combinations of antimicrobial agents on endotoxin release from gram-negative bacteria: an in-vitro and in-vivo study

Interventionstherapie yon Infektionen mad Fieber unklarer Genese bei neutropenischen Patienten mit malignen hfimatologisehen Grunderkrankungen

Interventional treatment of unexplained fever (FUO) and documented infections in neutropenic patients with hematologic malignancies: fluconazole versus ampho B/5-FC

Streptococcal bacteremia in adult patient with leukemia undergoing aggressive chemotherapy. A review of 55 Cases

Empiric use of vancomycin during prolonged treatment-induced granutocytopenia. Randomized, double-blind, placebo-controlled clinical trial in patients with acute leukemia

Protective effect of recombinant human granulocyte colony-stimulating factor (rG-CSF) against various microbial infections in neutropenic mice

M-Simultaneous r-metHuG-CSF and chemoradiotherapy during induction treatment in adult ALL: a randomized trial

Protective effect of human granulocyte colony-stimulating factor (hG-CSF) on cryptococcus and aspergillus infections in normal and immunosuppressed mice

© MMV lVledizin Verlag GmbH Mttnchen

A. Biihme et aI.: Imipenem vs. Cefotaxime/Piperacillin in Neutropenic Fever