Skip Navigation

A Multicenter, Double-Blind, Placebo-Controlled Trial Comparing Piperacillin-Tazobactam with and without Amikacin as Empiric Therapy for Febrile Neutropenia

  1. Albano Del Favero1,
  2. Francesco Menichetti2,
  3. Pietro Martino3,
  4. Giampaolo Bucaneve1,
  5. Alessandra Micozzi3,
  6. Giuseppe Gentile3,
  7. Paolo Furno1,
  8. Domenico Russo4,
  9. Domenico D'Antonio5,
  10. Paolo Ricci6,
  11. Bruno Martino7,
  12. Franco Mandelli3, and
  13. Gruppo Italiano Malattie Ematologiche dell'Adulto (GIMEMA) Infection Program
  1. 1Istituto Medicina Interna e Scienze Oncologiche, Università di Perugia, Perugia
  2. 2Unità Operativa Malattie Infettive, Ospedale Cisanello, Pisa
  3. 3Istituto di Ematologia, Dipartimento di Biopatologia, Università “La Sapienza,”, Rome
  4. 4Clinica Ematologica, Policlinico Università di Udine, Udine
  5. 5Divisione di Ematologia, Ospedale Civile Spirito Santo, Pescara
  6. 6Istituto di Ematologia, Policlinico S. Orsola, Bologna
  7. 7Divisione di Ematologia, Azienda Ospedaliera, Reggio Calabria, Italy
  1. Reprints or correspondence: Albano Del Favero, Sezione di Medicina Interna e Scienze Oncologiche, Dipartimento di Medicina Clinica e Sperimentale, Università degli Studi di Perugia, Policlinico “Monteluce,” 06122 Perugia, Italy (delfa{at}unipg.it).
 
Next Section

Abstract

In a prospective, multicenter, double-blind, randomized clinical trial, we compared the efficacy of piperacillin-tazobactam (4.5 g 3 times daily intravenously) plus placebo versus piperacillin-tazobactam plus amikacin (7.5 mg/kg twice daily intravenously) for the treatment of 760 febrile, adult patients with cancer with chemotherapy-induced profound (<500 neutrophils/mm3) and prolonged (>10 days) neutropenia. A total of 733 patients were assessable for efficacy of the drug regimens, and an overall successful outcome was reported in 49% (179 of 364) of the patients who received monotherapy, compared with 53% (196 of 369) of patients who received combination therapy (P = .2). Response rates were similar with both regimens, as were incidences of bacteremia and clinically documented and possible infections. In our epidemiological setting, the initial empiric combination therapy was not associated with improved outcomes when compared with initial monotherapy.

The combination of an antipseudomonal β-lactam with an aminoglycoside has been considered the standard empiric treatment of neutropenic febrile patients [1–4]. Broad-spectrum bactericidal antibiotics have been tested as monotherapy for febrile neutropenia and have shown an efficacy similar to the combination therapy [5–13]. However, these studies have 2 main shortcomings: first, the lack of comparison of the same β-lactam in the 2 treatment arms does not allow the evaluation of the real advantage brought about by the addition of an aminoglycoside; and second, because any modification of initial empiric therapy is defined as failure, the unblinded design introduces a bias in the decision of the investigator to modify the initial empiric regimen because monotherapy may be perceived as inadequate. Furthermore, a formal comparison of short versus regular courses of amikacin in combination with the same β-lactam antibiotic in febrile neutropenia showed the advantage of the classical regimen in controlling gram-negative bacteremia [9]. Therefore, the issue of monotherapy and, in particular, the advantage of including an aminoglycoside in combination with a β-lactam antibiotic at the start of empiric therapy remains controversial. This prompted us to perform a double-blind, placebo-controlled, randomized clinical trial to compare piperacillin-tazobactam with or without amikacin for the empiric treatment of febrile neutropenia.

Previous SectionNext Section

Methods

We evaluated patients aged >14 years who were admitted for cancer chemotherapy or autologous bone marrow transplantation to 34 Italian oncologic and hematologic centers. Patients were eligible for randomization if they had fever (⩾38.5°C on 1 occasion, or ⩾38.0°C on ⩾2 occasions within 24 h) likely to be due to an infectious cause and had neutropenia (<500 neutrophils/mm3) anticipated to last longer than 10 days.

Patients were excluded from the study if they had a history of allergy to the protocol's antibiotics, if they had serum creatinine levels >1.5 mg/dL, or if they had received iv administered antibiotics during the preceding 96 h. Orally administered fluoroquinolones and cotrimoxazole were allowed as prophylaxis, but they were stopped at randomization.

Randomization. Patients were randomized centrally only once, according to a computer-generated random-number program accessible 24 h daily. Patients were stratified according to the center and the underlying disease (acute leukemia and allogeneic bone marrow transplantation vs. lymphoma and solid tumors).

Primary end-point and sample size calculation. The primary objective of the study was to compare the success rates of both regimens. An equivalence trial design was chosen. Assuming from experience elsewhere [10] that the expected overall response rate to the piperacillin-tazobactam plus amikacin regimen was 61%, a required sample size of 712 assessable patients was estimated to ensure a 80% chance of rejecting, at α = 0.05, the null hypothesis of equivalence when the true difference between treatments was ∼10%.

Therapeutic regimens. Piperacillin-tazobactam was administered in a dose of 4.5 g q8h by infusion over the course of 30 min to both treatment groups. Amikacin (7.5 mg/kg q12h) or matching placebo (100 mL of 0.9% NaCl) were infused over the course of 15 min. Calculated renal creatinine clearance was used to adjust maintenance doses of piperacillin-tazobactam. Maintenance doses of amikacin or placebo were modified according to the dosing method of Sarubbi and Hull [14].

Classification of febrile episodes. Febrile episodes were classified as microbiologically documented infection with or without bacteremia, clinically documented infection, and unexplained fever. In case of isolation of skin saprophytes, 2 positive blood cultures were required.

Evaluation of response. Response was defined as follows. We considered the response a success if fever and clinical signs of infection resolved and if the infecting microorganisms were eradicated without change of the initial allocated treatment. Response was defined as “failure” if the patient died as a result of primary infection; if bacteremia persisted beyond the first 24 h of therapy; if breakthrough bacteremia was documented; if the isolated pathogen was resistant to piperacillin-tazobactam, regardless of the clinical response of the patient; if a temperature of >38°C persisted for at least 5 days and prompted modification of empiric antibacterial treatment, either in a deteriorating or in an otherwise stable patient without any other documentation of infection; if infection relapsed within 7 days of discontinuation of treatment; and if toxicity occurred that required interruption of treatment. Response was also evaluated by assessing the time to defervescence, the time to failure, and the survival at day 30.

National Committee for Clinical Laboratory Standards methods were used to define resistance. Piperacillin-tazobactam resistance was defined for all microorganisms as a zone diameter <17 mm, according to diffusion susceptibility testing, and as an MIC ⩾128/4 µg/mL for gram-negative microorganisms and an MIC ⩾16/4 µg/mL for staphylococci, according to MIC testing.

Statistical analysis. All case report forms were centrally reviewed, and data analysis (SAS, SAS Institute [15]) was blinded to the assigned treatment. A by-protocol analysis on the assessable patients and an intent-to-treat analysis on all eligible patients were performed.

The χ2 test with a correction for continuity and Fisher's exact test were used, when appropriate, to compare differences in proportions. The Wilcoxon test was used to compare the medians. A logistic regression model was used to assess the relative importance of the various prognostic factors assessable at the time of randomization. The OR of success and its 95% CI were calculated for each factor included in the multivariate analysis model. The distributions of time-to-event variables were evaluated by the Kaplan-Meier method and were compared by the log-rank test. The 95% CIs for the difference between proportions are given, when appropriate.

Previous SectionNext Section

Results

From December 1997 through May 1999, out of a total 760 patients randomized, 6 were not eligible and 21 were considered “not assessable” for the following reasons (monotherapy/combination therapy): protocol violation (2/12); early discontinuation of protocol therapy (0/1); noninfectious fever (1/2); and nonbacterial infection (3/0). Therefore, we were able to assess the data from 733 patients (364/369) for therapy efficacy. Patient characteristics are illustrated in table 1. There were no significant differences between the 2 treatment groups in any important characteristic.

Table 1
Table 1

Characteristics of the 733 patients assessable for response to therapy.

Sites of infection were equally distributed and were represented by the following (monotherapy/combination therapy): central venous catheter (19/17); lung (14/19); skin and soft tissue (15/13); oral cavity (13/12); gastrointestinal (6/4); urinary tract (2/1); and others (4/6). The classification of febrile episodes and response to therapy are illustrated in table 2.

Table 2
Table 2

Classification of febrile episodes, response to therapy by protocol, and susceptibility of bacterial isolates.

The overall success rate in the 2 treatment arms (monotherapy vs. combination therapy) was similar: 49% versus 53% (P = .2; 95% CI, −11 to 3). The response rate of microbiologically documented infections accounted for 32% in both regimens (P = .9; 95% CI, −10 to 10), whereas that of clinically documented infections was 53% versus 43% (P = .5; 95% CI, −13 to 32). The response rate of patients with fever of unexplained origin was 63% with monotherapy and 70% with combination therapy (P = .1; 95% CI, −17 to 2).

The causes of failure were similar for both treatment groups and were mainly represented by persistent or relapsing fever and by the isolation of a resistant pathogen (table 3).

Table 3
Table 3

Reasons for modification of empiric antibiotic treatment (per-protocol analysis) for patients who failed to respond to therapy.

Time to defervescence and time to failure. The overall distribution of time to defervescence and that of time to failure were found to be similar with the 2 treatments (log-rank test, P = .6 and P = .9, respectively). We also found no difference in subgroups of patients, such as those with gram-negative bacteremia, including polymicrobial infections (time to defervescence: log-rank test, P = .48; time to failure: log-rank test, P = .24); those with single-agent gram-positive bacteremia (time to defervescence: log-rank test, P = .49; time to failure: log-rank test, P = .35); and those with pneumonia, including those with bacteremia (time to defervescence: log-rank test, P = .67; time to failure: log-rank test, P = .85).

Response of bacteremia. Pathogens responsible for bacteremia (277 episodes) and response to therapy are shown in table 4. Gram-positive microorganisms were responsible for ∼60% of the bacteremic episodes and gram-negative bacilli for 30%, whereas 10% of bacteremias were polymicrobial. All pathogens were equally distributed in the 2 treatment arms. Methicillin-resistant strains (both coagulase-negative staphylococci and Staphylococcus aureus) represented 64% of all staphylococci isolated from blood.

Table 4
Table 4

Success rates, by infecting organism, in patients with bacteremia.

The response rate achieved by the 2 regimens (monotherapy vs. combination therapy) in single-agent, gram-positive bacteremia was low and similar: 27% versus 32% (P = .6; 95% CI, −18 to 9). Patients with bacteremia due to coagulase-negative staphylococci had a success rate of 17% versus 18% (P = .8; 95% CI, −14 to 13). The response rate of streptococcal and enterococcal bacteremia was better than that obtained in staphylococcal bacteremia and was similar for the 2 treatment groups: 60% versus 71% (P = .7; 95% CI, −42 to 20).

The response rate achieved by the 2 regimens (monotherapy vs. combination therapy) in single-agent, gram-negative bacteremia also was low and similar: 36% versus 34% (P = .9; 95% CI, −17 to 22). Reasons for failure during gram-negative bacteremia were similar for both treatment groups and are represented by the following: resistant pathogens, 26% versus 45% (P = .2; 95% CI, −42 to 6); persistent fever, 26% versus 17% (P = .5; 95% CI, −12 to 31); and deterioration of clinical conditions, 48% versus 38% (P = .6; 95% CI, −15 to 36). Patients with bacteremia due to Escherichia coli had a similar success rate of 42% versus 53% (P = .6; 95% CI, −42 to 19); strains resistant to piperacillin-tazobactam accounted for 19% of all E. coli isolated from blood in both treatment groups. The response rate of Pseudomonas aeruginosa bacteremia was poor and was similar for the 2 treatment groups: 14% versus 12% (P = .6; 95% CI, −27 to 32). Fifty percent and 60% of blood isolates of P. aeruginosa were resistant, respectively, to piperacillin-tazobactam and amikacin. No specific resistance problems were found at any center.

Multivariate analysis. In order to estimate predictive factors that could influence the failure of empiric therapy, data from all assessable patients were fitted with a multivariate logistic regression model. Factors included in the model were as follows: monotherapy versus combination therapy (OR, 0.80; 95% CI, 0.63–1.14); male versus female (OR, 0.68; 95% CI, 0.50–0.91); age <65 versus age >65 years (OR, 1.39; 95% CI, 0.86–2.25); performance status (World Health Organization 5-degree scale) <2 versus >2 (OR, 1.60; 95% CI, 0.81–3.34); acute leukemia versus solid tumor or lymphoma (OR, 0.47; 95% CI, 0.34–0.64); neutrophil count <100 neutrophils/mm3 versus >100 neutrophils/mm3 (OR, 0.87; 95% CI, 0.39–1.27); and duration of neutropenia <10 days versus >10 days (OR, 1.78; 95% CI, 1.27–2.48). The antibiotic regimen was not a predictive factor of failure, and only the diagnosis of acute leukemia represented an independent risk factor for failure.

Modification of the allocated antibiotic regimen. No difference was found between the 2 groups in the type of antibiotics prescribed as rescue treatment. Overall, a triple antibiotic combination (β-lactam plus an aminoglycoside and a glycopeptide) was chosen in 49% of patients. A combination of 2 compounds was prescribed in 36% of patients (β-lactam plus an aminoglycoside, 12%; β-lactam plus a glycopeptide, 24%), whereas monotherapy, most frequently a carbapenem, was chosen to treat 15% of patients.

Mortality. The overall mortality at the end of the febrile episode was 4% (14 patients) in the monotherapy group and 6% (22 patients) in the combination-therapy group (P = .2; 95% CI, −5 to 1). Deaths considered to be related to infection, with or without a noninfectious concomitant cause, were 9 of 14 in patients treated with monotherapy (7 with bacteremia) and 19 of 22 in those receiving combination therapy (10 with bacteremia).

Mortality in patients with single-agent bacteremia was similar in the 2 treatment groups and was greater in patients with gram-negative sepsis (monotherapy vs. combination therapy, 5 [12%] of 41 vs. 7 [16%] of 45) than in those with gram-positive sepsis (1 [1%] of 85 vs. 1 [1%] of 82). Microorganisms causing single-agent bacteremia and death were E. coli (4 vs. 3), P. aeruginosa (1 vs. 4), and methicillin-resistant staphylococci (1 vs. 1).

Among the 12 patients who died as a result of gram-negative bacteremia, only 3 deaths were caused by a microorganism resistant to piperacillin-tazobactam (1 P. aeruginosa in the monotherapy group and 2 E. coli in the combination-therapy group), and only 1 case of shock occurred in a patient treated with monotherapy (E. coli, susceptible to piperacillin-tazobactam). Mortality as a result of polymicrobial bacteremia was also similar between the 2 groups: 7% (1 of 15) versus 18% (2 of 11).

An early death (within 4 days and while still on unmodified initial empiric therapy, before cultures became positive) in bacteremic febrile episodes occurred in 2 (0.5%) of 364 patients treated with monotherapy (2 E. coli) and in 3 (0.8%) of 369 patients treated with combination therapy (1 P. aeruginosa, 1 methicillin-resistant coagulase-negative Staphylococcus, and 1 polymicrobial). The overall mortality rate at day 30 from the start of empiric antibiotic therapy was 6% in the monotherapy group (22 patients) and 9% in the combination-therapy group (32 patients; P = .2; 95% CI, −6 to 1).

Adverse effects. Both treatments were well tolerated, and they were discontinued because of toxicity only in 2 patients treated with monotherapy (1 patient with skin rash and 1 with gastrointestinal disturbance) and in 5 patients receiving combination therapy (2 patients with skin rashes, 2 with nephrotoxic events, and 1 with hypersensitivity reaction).

Intent-to-treat analysis. An intent-to-treat analysis was also performed after the inclusion of 21 of 27 unassessable patients (6 patients, 3 in each treatment group, were excluded from the analysis because they were not eligible). The results were almost identical to those obtained in the efficacy analysis (table 5).

Table 5
Table 5

Intent-to-treat analysis.

Previous SectionNext Section

Discussion

Our study addresses 2 important questions related to the empiric antibiotic therapy for febrile, high-risk, neutropenic patients with cancer: whether an aminoglycoside is to be added to the β-lactam antibiotic in the initial empiric therapy and whether piperacillin-tazobactam can be used as empiric antibiotic monotherapy.

In this trial, piperacillin-tazobactam monotherapy was as effective as piperacillin-tazobactam plus amikacin according to all efficacy parameters, such as clinical and microbiological success, time to defervescence, and time to failure. No difference in these parameters was also found in those subgroups of patients for whom a benefit from the combined treatment could be reasonably hypothesized (i.e., patients with gram-negative bacteremia, pneumonia, or both). However, the overall response rate was low in both treatment groups, which was mainly due to the poor response of patients with bacteremia.

The inadequate response is not unexpected in patients with gram-positive bacteremia because of the large number of methicillin-resistant, and therefore piperacillin-tazobactam-resistant, staphylococci. The microbiological failure of empiric therapy was not always accompanied by clinical failure. Patients remained in stable clinical conditions during empiric therapy, and they eventually improved when the therapy was modified by adding a glycopeptide. Only 1 early death was caused by coagulase-negative staphylococci. The response rate observed in gram-positive bacteremia in our study (27% with monotherapy and 32% with combination therapy) is similar to that obtained in 2 large studies in febrile neutropenic patients who received a piperacillin-tazobactam and amikacin combination (response rate, 38%) [6] or meropenem versus ceftazidime plus amikacin (response rate, 31% and 26%, respectively) [10].

The response rate of patients with gram-negative bacteremia was also poor in our study (36%) and lower in comparison with the results of the above-mentioned trials. This may have 3 explanations: the low susceptibility and high prevalence of P. aeruginosa in our study; the lower daily dosage of piperacillin-tazobactam used in our study; and the double-blind design of the trial. In our study, only 50% and 43% of bacteremia due to P. aeruginosa were susceptible to piperacillin-tazobactam and amikacin, respectively, in comparison with the European Organization for Research and Treatment of Cancer (EORTC) trial that used piperacillin-tazobactam [10], in which the susceptibility rate was 73% and 100%, and in comparison with the EORTC trial that used meropenem [6], in which 100% and 75% of P. aeruginosa were susceptible to meropenem and amikacin, respectively. Moreover, the prevalence of P. aeruginosa infections was also higher, compared with the EORTC trials. In fact, in our study, P. aeruginosa infections accounted for 28% (24 of 85) of total single-agent, gram-negative bacteremia, in comparison with the EORTC trial that used piperacillin-tazobactam, in which P. aeruginosa infections accounted for 19% (10 of 53) of total single-agent gram-negative bacteremia, or the EORTC trial that used meropenem, in which P. aeruginosa infections accounted for only 13% (8 of 61) of total gram-negative bacteremia.

The double-blind design of the study can also be blamed for the low response rate of treatments in gram-negative bacteremia, in that the documentation of a gram-negative bacteremia may have prompted the clinician to modify an initial treatment perceived to be inadequate against these serious infections. In fact, combination therapy containing an aminoglycoside was used in 70% of rescue treatments.

Finally, the strict definition of failure adopted in our study may have further contributed to the observed response rate. Overall, the reasons accounting for the modification of the initial allocated empiric antibiotic treatment were similar for the 2 treatment groups and were mainly represented by persistent fever (45%) and by isolation of a resistant pathogen (30%). Only in the remaining 25% of patients were the reasons for changing therapy clearly justified by signs or symptoms of clinical deterioration or toxicity.

Although according to our definitions the response to empiric treatment was quite low, overall mortality was limited and was similar with the 2 regimens. A higher rate of death occurred in patients with gram-negative and polymicrobial infections than in those with gram-positive infections, but early deaths were similar and were not influenced by the use of amikacin.

The lower response rate found in patients with acute leukemia, compared with that in patients with lymphoma or solid tumors, demonstrates once again that neutropenic oncohematologic patients are not a homogeneous group and stresses the importance of planning randomized clinical trials with patient population stratified according to risk for infection.

Our study showed that, for the treatment of high-risk, febrile neutropenic patients with cancer, empiric monotherapy with piperacillin-tazobactam represents an option as efficacious as the combination of piperacillin-tazobactam with an aminoglycoside. This option must be evaluated in the context of local epidemiological setting, with special reference to the rate of resistant microorganisms and patient risk factors. In an epidemiological context of the high prevalence of gram-negative strains resistant to piperacillin-tazobactam, the use of ceftazidime or a carbapenem should be considered.

The postulated advantage of adding an aminoglycoside to an antipseudomonal β-lactam antibiotic, assessed on the basis of greater bactericidal activity and possible synergism against selected gram-negative and gram-positive microorganisms, has not been confirmed in our study, and this approach should be therefore reconsidered.

Previous SectionNext Section

Study Group Members

The centers and participants who assisted this study, listed according to the number of patients enrolled in the trial (provided parenthetically), are as follows: Study Coordinators, A. Del Favero, F. Menichetti, and P. Martino; Data Review Committee, G. Bucaneve and A. Micozzi; Data Managing, M.S. Dionisi, G. Barbabietola, P. Pincardini, and G. Calabrò, Istituto Medicina Interna e Scienze Oncologiche, Università di Perugia, Perugia. Istituto di Ematologia, Dipartimento di Biopatologia, Università “La Sapienza,” Roma (114), C. Girmenia; Istituto Medicina Interna e Scienze Oncologiche ed Istituto di Ematologia, Università di Perugia, Perugia (86), S. Ballanti; Clinica Ematologica, Policlinico Università di Udine, Udine (67), D. Russo; Divisione di Ematologia, Ospedale Civile Spirito Santo, Pescara (54), G. Fioritoni; Istituto di Ematologia, Policlinico S. Orsola, Bologna (31), P. Ricci; Divisione di Ematologia, Azienda Ospedaliera, Reggio Calabria (31), F. Nobile; Divisione di Oncoematologia, Istituto Europeo di Oncologia, Milano (29), S. Cinieri; Clinica Ematologica, Ospedale Torrette di Ancona, Torrette di Ancona (24), G. Leone; Ematologia, Ospedali Riuniti di Bergamo, Bergamo (21), M. Buelli; Centro Trapianti Midollo, Ospedale Maggiore, Milano (21), G. Lambertenghi Deliliers; Divisione di Ematologia, Nuovo Policlinico, Napoli (20), M. Picardi; Istituto di Ematologia, Ospedale S. Eugenio, Roma (20), L. Cudillo; Divisione di Ematologia, Ospedale S. Martino, Genova (19), A. Isaza; Ematologia, Ospedale Niguarda, Milano (19), A.M. Nosari; Divisione di Ematologia, Ospedale S. Francesco, Nuoro (19), A. Gabbas; Divisione di Ematologia, Ospedale “Casa Sollievo della Sofferenza,” S. Giovanni Rotondo (19), N. Cascavilla; Divisione di Ematologia, Ospedale di Siena, Siena (17), P. Galieni; Divisione di Oncologia Medica, Centro di Riferimento Oncologico, Aviano (15), V. Zagonel; Divisione di Ematologia, Ospedale S. Maria Goretti, Latina (14), A. Chierichini; U.O. di I Medicina Interna, Arcispedale S. Maria Nuova, Reggio Emilia (14), A. Dall'Asta; U.O.A. Ematologia, Ospedale Molinette, Torino (13), R. Calvi; Centro Trapianto Midollo Osseo, Azienda Ospedaliera, Reggio Calabria (11), P. Jacopino; Ematologia, Ospedale S. Carlo, Potenza (10), F. Ricciuti; Divisione di Ematologia, Ospedale Civile SS. Giovanni e Paolo, Venezia (10), G. Capnist; Divisione di Ematologia, Università Cattolica, Roma (9), L. Pagano; Ematologia, Università di Bari, Bari (8), V. Liso; Divisione di Ematologia, Ospedale S. Bortolo, Vicenza (8), A. D'Emilio; Divisione di Ematologia, Ospedale Oncologico “Businco,” Cagliari (7), W. Deplano; U.O.A. Ematologia, Azienda Osp. S. Croce e Carle, Cuneo (7), D. Mattei; U.O. Ematologia, Azienda Ospedaliera Careggi, Firenze (7), R. Fanci; Divisione Clinicizzata di Ematologia, Ospedale Ferrarotto, Catania (7), G. Milone; Patologia Medica, Azienda Ospedaliera S. Luigi Gonzaga, Orbassano (4), F Vischia; Medicina Interna e Oncologia Medica, Policlinico S. Matteo, Pavia (3), R. Invernizzi; Ematologia, Università di Parma, Parma (2), C. Caramatti. All listed institutions are located in Italy.

Previous SectionNext Section

Footnotes

  • This clinical research was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, and the protocol was approved by the ethics committees of each participating institution. Written informed consent was obtained from all patients.

  • Financial support: Consiglio Nazionale Ricerca (progetto ACRO n. 00551 39/115 19159), Wyeth-Lederle, Italy.

  • Received November 20, 2000.
  • Revision received March 30, 2001.
Previous Section
 

References

  1. ↵
    1. Elting LS,
    2. Rubenstein EB,
    3. Rolston KVI,
    4. et al
    . Outcomes of bacteremia in patients with cancer and neutropenia: observations from two decades of epidemiological and clinical trials. Clin Infect Dis 1997;25:247-59.
    Abstract/FREE Full Text
    1. Hughes WT,
    2. Armstrong D,
    3. Bodey GP,
    4. et al
    . 1997 Guidelines for the use of antimicrobial agents in neutropenic patients with unexplained fever. Infectious Diseases Society of America. Clin Infect Dis 1997;25:551-73.
    Abstract/FREE Full Text
    1. Klastersky J,
    2. Meunier-Carpentier F,
    3. Prevost JM
    . Significance of antimicrobial synergism for the outcome of gram-negative sepsis. Am J Med Sci 1977;273:157-61.
    MedlineWeb of Science
  4. ↵
    1. Klastersky J,
    2. Zinner SH
    . Synergistic combinations of antibiotics in gram-negative bacillary infections. Rev Infect Dis 1982;4:294-8.
    MedlineWeb of Science
  5. ↵
    1. Pizzo PA,
    2. Hathorn JW,
    3. Hiemenz J,
    4. et al
    . A randomized trial comparing ceftazidime alone with combination antibiotic therapy in cancer patients with fever and neutropenia. N Engl J Med 1986;315:552-8.
    MedlineWeb of Science
  6. ↵
    1. Cometta A,
    2. Calandra T,
    3. Gaya H,
    4. et al
    . Monotherapy with meropenem versus combination therapy with ceftazidime plus amikacin as empiric therapy for fever in granulocytopenic patients with cancer. Antimicrob Agents Chemother 1996;40:1108-15.
    Abstract/FREE Full Text
    1. De Pauw BE,
    2. Deresinski SC,
    3. Feld R,
    4. et al
    . Ceftazidime compared with piperacillin and tobramycin for the empiric treatment of fever in neutropenic patients with cancer: a multicenter randomized trial. Ann Intern Med 1994;120:834-44.
    Abstract/FREE Full Text
    1. Yamamura D,
    2. Gucalp R,
    3. Carlisle P,
    4. et al
    . Open randomized study of cefepime versus piperacillin-gentamicin for treatment of febrile neutropenic cancer patients. Antimicrob Agents Chemother 1997;41:1704-8.
    Abstract/FREE Full Text
  9. ↵
    1. EORTC International Antimicrobial Therapy Cooperative Group
    . Ceftazidime combined with a short or long course of amikacin for empirical therapy of gram-negative bacteremia in cancer patients with granulocytopenia. N Engl J Med 1987;317:1692-8.
    MedlineWeb of Science
  10. ↵
    1. Cometta A,
    2. Zinner S,
    3. de Bock R,
    4. et al
    . Piperacillin/tazobactam plus amikacin versus ceftazidime plus amikacin as empiric therapy for fever in granulocytopenic patients with cancer. Antimicrob Agents Chemother 1995;39:445-52.
    Abstract/FREE Full Text
    1. Fainstein V,
    2. Bodey GP,
    3. Elting L,
    4. et al
    . A randomized study of ceftazidime compared to ceftazidime and tobramycin for the treatment of infections in cancer patients. J Antimicrob Chemother 1983;12(Suppl A):101-10.
    Abstract/FREE Full Text
    1. Rolston KV,
    2. Berkey P,
    3. Bodey GP,
    4. et al
    . A comparison of imipenem to ceftazidime with or without amikacin as empiric therapy in febrile neutropenic patients. Arch Intern Med 1992;152:283-91.
    Abstract/FREE Full Text
  13. ↵
    1. Marie JP,
    2. Pico J,
    3. Lapierre V,
    4. et al
    . Traitement empirique des episodes febriles survenant chez les patients cancereux presentant une neutropenie prolongee: essai comparatif ceftazidime seule, ceftazidime + amikacine et ceftazidime + vancomycine. Med Maladies Infect 1991;33:87-91.
  14. ↵
    1. Sarubbi FA Jr,
    2. Hull JH
    . Amikacin serum concentrations: prediction of levels and dosage guidelines. Ann Intern Med 1978;89:612-8.
    Abstract/FREE Full Text
  15. ↵
    1. SAS Institute
    . SAS/STAT guide for personal computers. Cary, NC: SAS Institute; 1987. version 6.
| Table of Contents

This Article

  1. Clin Infect Dis. (2001) 33 (8): 1295-1301. doi: 10.1086/322646
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most