-
PDF
- Split View
-
Views
-
Cite
Cite
Ron Ram, Yael Halavy, Odelia Amit, Yael Paran, Eugene Katchman, Bruria Yachini, Svetlana Kor, Irit Avivi, Ronen Ben-Ami, Extended vs Bolus Infusion of Broad-Spectrum β-Lactams for Febrile Neutropenia: An Unblinded, Randomized Trial, Clinical Infectious Diseases, Volume 67, Issue 8, 15 October 2018, Pages 1153–1160, https://doi.org/10.1093/cid/ciy258
Close -
Share
Abstract
Febrile neutropenia may be a sign of severe infection and is associated with significant morbidity and mortality in high-risk patients with hematologic malignancies. Extended infusion of β-lactam antibiotics is associated with greater clinical response than is bolus infusion in nonneutropenic critically ill patients, but data are lacking for febrile neutropenic patients.
We designed a single-center, nonblinded, randomized trial to compare extended infusion (4 hours) and bolus infusion (30 minutes) of piperacillin-tazobactam or ceftazidime in high-risk patients with febrile neutropenia. The primary endpoint was overall response on day 4, defined as the combination of resolution of fever, sterile blood cultures, resolution of clinical signs and symptoms, and no need for a change in the antibiotic regimen. Outcome was adjudicated by investigators blinded to treatment allocation.
Of 123 enrolled patients, 105 had febrile neutropenia and were included in the intention-to-treat analysis: 47 in the extended infusion arm and 58 in the bolus infusion arm. Overall response occurred in 35 (74.4%) patients treated with extended infusion and 32 (55.1%) patients treated with bolus infusion (P = .044). The superiority of extended infusion was greatest for patients with clinically documented infections (overall response, 68.4% [13/19] vs 35.7% [10/28]; P = .039) and specifically for those with pneumonia (80% [4/5] vs 0% [0/8]; P = .007).
Extended infusion of β-lactams is associated with superior treatment outcomes compared with bolus infusion for high-risk patients with febrile neutropenia. The benefit of extended β-lactam infusion may be greatest for patients with pulmonary infections.
NCT02463747.
More than 80% of patients with hematologic malignancies develop fever during at least 1 cycle of chemotherapy [1–3]. Fever may be the only sign of severe infection during neutropenia, and patients with profound and prolonged neutropenia, hemodynamic instability, bacteremia, pulmonary infiltrates, and organ dysfunction are at high risk of complications and death [2]. Early detection of febrile neutropenia and prompt initiation of broad-spectrum β-lactam antibiotics, such as piperacillin-tazobactam and cefepime, are essential to prevent adverse outcomes [2]. However, given the potential of neutropenic patients to deteriorate rapidly and the rising incidence of infections with multidrug-resistant gram-negative organisms [4], strategies to optimize medical treatment are a priority.
The bactericidal activity of β-lactam antibiotics is determined by the duration of time during which the free plasma drug concentration exceeds the bacterial minimal inhibitory concentration (MIC) [5, 6]. Emerging data suggest that standard dosing schemes of β-lactams routinely fail to achieve desired pharmacokinetic/pharmacodynamic (PK/PD) targets [7]. Extended or continuous β-lactam infusion is associated with higher rates of PK/PD target attainment and may improve clinical outcomes and survival rates of critically ill patients [8–12]. Patients with hematological malignancies and neutropenia and critically ill patients share physiological characteristics that alter drug PK, namely, increased volume of distribution (Vd) and drug clearance [13, 14]. However, there are only limited observational studies to support extended β-lactam infusion in patients with febrile neutropenia [14, 15]. Mathematical modeling predicts that standard dosing of piperacillin-tazobactam is unlikely to achieve a PK target of >50% time over MIC in neutropenic patients, whereas the probability of PK target attainment is >90% using extended infusion [13].
We aimed to compare, in a randomized prospective study, clinical outcomes of patients with high-risk febrile neutropenia treated with either intermittent bolus infusion or extended infusion of broad-spectrum β-lactams.
METHODS
Study Design and Patients
This was a single-center, open-label, nonblinded, prospective, randomized trial comparing the efficacy of extended infusion of piperacillin-tazobactam or ceftazidime with standard, intermittent bolus infusion in high-risk patients with febrile neutropenia. The primary objective was to determine whether extended infusion, which optimizes β-lactam PK/PD parameters, produces superior clinical response compared with intermittent bolus infusion in the setting of febrile neutropenia. The study was performed at the Tel Aviv Sourasky Medical Center (TASMC) from July 2015 through May 2017. No modifications to the study protocol were made during the course of the study. The TASMC Institutional Review Board approved the protocol, and all patients provided written informed consent. Interim analysis of the data was performed after the enrollment of 50 patients, and the results were reported to the institutional review board, which approved the continuation of the study. All authors had access to the primary clinical data.
Male and female patients aged ≥18 years who were undergoing hematopoietic cell transplantation (HCT) or receiving induction or consolidation chemotherapy for acute leukemia were eligible to enroll into the study. Patients were excluded if they were scheduled for outpatient follow-up during neutropenia, if they were receiving maintenance chemotherapy for acute lymphoblastic leukemia, and if their calculated creatinine clearance was less than 40 mL/min. Patients infected or colonized with bacteria resistant to study antibiotics within 30 days prior to enrollment were also excluded from this study. Enrollment and randomization were done prior to hospital admission for HCT patients and before starting chemotherapy for patients with acute leukemia. Patients were randomly assigned in advance at a 1:1 ratio using computerized random number generation to receive either bolus infusion or extended infusion of β-lactam in case of a febrile neutropenia event. Allocation was concealed in sequentially numbered sealed opaque envelopes that were opened by a study coordinator who marked the patient’s file with the allocation arm.
Study Therapy
Patients who met criteria for febrile neutropenia were treated empirically with piperacillin-tazobactam (4.5 g intravenously every 8 hours) or ceftazidime (2 g intravenously every 8 hours) in case of a documented history of penicillin allergy. Febrile neutropenia was defined as a single oral temperature of ≥38.3°C or a temperature of ≥38.0°C sustained for >1 hour and an absolute neutrophil count (ANC) <500 cells/mm3, or an ANC that was expected to decrease to <500 cells/mm3 over the next 48 hours. Patients were randomly assigned to receive either intermittent bolus infusion, where each antibiotic dose was administered as a 30-minute infusion, or extended infusion, where each antibiotic dose was administered over 4 hours. Patients in the extended infusion arm received a single loading dose of study medication over 30 minutes, followed 6 hours later by the first extended infusion of study medication. Study drugs were prepared by nursing staff according to local standard practice.
Treatment with vancomycin was initiated in accordance with treatment guidelines [2]. Amikacin was added to the antibiotic regimen of patients from both arms with hemodynamic instability unresponsive to a bolus of intravenous crystalloid solution. Patients who continued to be hemodynamically unstable despite treatment with a broad-spectrum β-lactam and amikacin were switched to a carbapenem (imipenem-cilastatin or meropenem). A fluoroquinolone was added for patients with pulmonary infiltrates on chest imaging. Patients who did not respond to a bolus of intravenous crystalloid solution were started on noradrenaline continuous infusion and were followed in-ward by an intensive care specialist. Antibiotic regimens were adjusted according to microbiological culture results, when available, in consultation with an infectious diseases specialist.
Chemotherapy Regimens and Supportive Care
All patients were hospitalized at a single designated ward in single-bed rooms equipped with high-efficiency particulate air filters and were treated by a single team of 2 hematologists and 3 infectious diseases consultants. Patients with acute myeloid leukemia were given induction chemotherapy based on 3 days of daunorubicin (45–90 mg/m2) and 7 days of continuous cytarabine infusion (100 mg/m2). In these patients, consolidation was based on either high-dose cytarabine (in patients aged <60 years) or intermediate-dose cytarabine (in patients aged ≥60 years). Patients with acute lymphoblastic leukemia were treated according to the German acute lymphoblastic leukemia (GMALL) 2003 protocol. Eligibility for allogeneic or autologous HCT was determined by an expert team based on patient disease status and the HCT comorbidity index (HCT-CI) score. Reduced-intensity conditioning was based on fludarabine (150 mg/m2) with busulfan (6.4 mg/kg intravenously) or melphalan (140 mg/m2). Myeloablative regimens were based on either busulfan (12.8 mg/kg intravenously) with cyclophosphamide (120 mg/kg) or fludarabine (120 mg/m2), or cyclophosphamide (120 mg/kg) and fractionated total-body irradiation (12 Gy). Graft-versus-host disease prophylaxis included methotrexate 15 mg/m2 on day 1 and 10 mg/m2 on days 3 and 6, administered with cyclosporine or with mycophenolate mofetil in cases of severe mucositis [16]. In cases of unrelated donor, low-dose antithymocyte globulin was added to the preparative regimen. Stem-cell source was G-derived peripheral-blood stem cells. Autologous HCT was performed with a preparative regimen based on high-dose melphalan (140–200 mg/m2) in patients with myeloma or the BEAM (BiCNU, Etoposide, Ara-C, Melphalan) protocol in patients with lymphoma.
Supportive care included growth factors in patients undergoing transplantation from day +7, in patients with acute myeloid leukemia from day +14, and in patients with acute lymphoblastic leukemia according to the GMALL-2003 protocol. Prophylaxis with acyclovir was given to all transplanted patients, and prophylaxis with ciprofloxacin and fluconazole was given to all patients, except patients with myeloma undergoing autologous HCT.
Study Outcomes
Study endpoints were assessed by study investigators who were blinded to each participant’s random assignment. The primary study endpoint was overall response on day 4 post-symptom onset, defined as a composite of 4 criteria: resolution of fever for at least 24 hours; microbiological eradication (for microbiologically documented infection): sterile cultures on days 3 and 4; clinical response (for clinically documented infection): resolution of signs and symptoms of infection; and no need for a change in the antibiotic regimen (addition of an aminoglycoside or a fluoroquinolone within 48 hours of initiating treatment was not considered treatment failure). Treatment was considered successful if all criteria were met.
Secondary endpoints were breakthrough bloodstream infection ≥5 days after initiation of treatment, recurrent fever ≥5 days after initiation of treatment, infection with Clostridium difficile, death within 30 days of enrollment into study, duration of hospitalization, acute kidney injury (doubling of serum creatinine level) within 4 days, and use of noradrenaline due to persistent hypotension.
Statistical Analyses
Sample size was calculated for the comparison of dichotomous variables between 2 treatment arms. We planned a trial in which the intervention and control arms were independent and the ratio of patients in the 2 arms was 1:1. We used the 2-sided χ2 test to evaluate the null hypothesis. Historical data at TASMC showed a rate of treatment failure of 0.7. Assuming that the incidence of treatment failure in the intervention arm was 0.4, we calculated a sample size of 42 patients in each arm in order to reject the null hypothesis with a power of 0.8 and a 1-sided alpha of 0.05. Since around 80% of enrolled patients were expected to experience febrile neutropenia, sample size was set at 53 patients for each study arm.
All primary and secondary analyses were prespecified. Outcomes were assessed among all patients who developed febrile neutropenia (intention-to-treat [ITT]analysis), and patients who completed at least 48 hours of a study antibiotic (piperacillin-tazobactam or ceftazidime; per-protocol analysis). For the overall and individual endpoints, the observed proportion of patients with a favorable outcome was determined and 95% confidence intervals (CIs) were calculated using the normal approximation to the binomial distribution. The ITT cohort was specified as the population for which the primary endpoint was determined. Student t and Wilcoxon rank-sum tests for independent samples were used to compare normally and nonnormally distributed continuous variables, respectively, and Fisher’s exact test was used for categorical variables. P values were 2-sided; P values ≤ .5 were considered to represent statistical significance. Statistical analyses were performed in Stata version 13 (StataCorp, College Station, Texas).
RESULTS
Enrollment
A total of 123 patients provided informed consent and underwent randomization (intermittent bolus infusion, n = 63; extended infusion, n = 60; Figure 1). There were no significant differences between the study arms with respect to baseline patient demographics and comorbidities (Table 1). Interim safety analysis performed after enrollment of 50 patients did not reveal safety concerns.
Characteristics of 123 Enrolled Patients
| Characteristic | Bolus Infusion (n = 63) | Extended Infusion (n = 60) | P Value |
|---|---|---|---|
| Age, years; median (IQR) | 60.1 (45.8–66.3) | 60.4 (50.9–67.5) | .74 |
| Sex, female | 28 (44.4) | 23 (38.3) | .58 |
| Weight, kg; median (range) | 75 (50.4–111.7) | 81 (58–108) | .13 |
| Primary diagnosis | .46 | ||
| Acute leukemia/myelodysplastic syndrome | 26 (41) | 18 (30) | … |
| Lymphoma | 17 (27) | 17 (28) | … |
| Multiple myeloma | 18 (29) | 24 (40) | … |
| Other | 2 (3) | 1 (2) | |
| Reason for admission | .29 | ||
| Allogeneic HCT | 21 (33) | 18 (30) | … |
| Autologous HCT | 29 (46) | 36 (60) | … |
| Induction for acute leukemia | 7 (11) | 4 (7) | … |
| Consolidation for acute leukemia | 6 (10) | 2 (3) | … |
| Myeloablative regimena | 12 (57) | 8 (44) | .43 |
| HCT-CI; median (IQR)a | 2 (2–3) | 2 (1–3) | .21 |
| Karnofsky score; median (IQR) | 100% (90–100) | 100% (80–100) | .15 |
| Neutropenia <500/µL, days; median (range) | 7 (1–46) | 7 (1–37) | .32 |
| Neutropenia <100/µL, days; median (range) | 6 (0–36) | 6 (0–30) | .42 |
| Fluoroquinolone prophylaxis | 35 (56) | 34 (57) | .98 |
| Antibiotic use, days; median (range)b | 1 (0–14) | 2 (0–17) | .61 |
| Febrile neutropenia | 58 (92.0) | 47 (78.3) | .041 |
| Hospital days before febrile neutropenia; median (range) | 12 (1–24) | 12 (1–22) | .6 |
| Clinically documented infection | 29 (46.0) | 20 (33.3) | .19 |
| Pneumonia | 8 (12.7) | 6 (10.0) | .7 |
| Colitis | 4 (6.3) | 2 (3.3) | .6 |
| Exit site infection | 5 (7.9) | 4 (6.6) | 1.0 |
| Microbiologically documented infection | 11 (7.4) | 10 (6.6) | 1.0 |
| Bloodstream infection | 10 (15.8) | 7 (11.6) | .6 |
| Estimated glomerular filtration rate (mL/min/1.73 m2),c mean ± standard deviation | 109.4 ± 17.7 | 109.8 ± 21.5 | .9 |
| Primary antibiotic treatment | .6 | ||
| Piperacillin-tazobactam | 52 (89.6) | 44 (93.6) | … |
| Ceftazidime | 4 (6.9) | 3 (6.3) | … |
| Meropenem | 2 (3.4) | 0 (0) | … |
| Vancomycin co-treatment | 22 (39.2) | 19 (42.2) | .8 |
| Characteristic | Bolus Infusion (n = 63) | Extended Infusion (n = 60) | P Value |
|---|---|---|---|
| Age, years; median (IQR) | 60.1 (45.8–66.3) | 60.4 (50.9–67.5) | .74 |
| Sex, female | 28 (44.4) | 23 (38.3) | .58 |
| Weight, kg; median (range) | 75 (50.4–111.7) | 81 (58–108) | .13 |
| Primary diagnosis | .46 | ||
| Acute leukemia/myelodysplastic syndrome | 26 (41) | 18 (30) | … |
| Lymphoma | 17 (27) | 17 (28) | … |
| Multiple myeloma | 18 (29) | 24 (40) | … |
| Other | 2 (3) | 1 (2) | |
| Reason for admission | .29 | ||
| Allogeneic HCT | 21 (33) | 18 (30) | … |
| Autologous HCT | 29 (46) | 36 (60) | … |
| Induction for acute leukemia | 7 (11) | 4 (7) | … |
| Consolidation for acute leukemia | 6 (10) | 2 (3) | … |
| Myeloablative regimena | 12 (57) | 8 (44) | .43 |
| HCT-CI; median (IQR)a | 2 (2–3) | 2 (1–3) | .21 |
| Karnofsky score; median (IQR) | 100% (90–100) | 100% (80–100) | .15 |
| Neutropenia <500/µL, days; median (range) | 7 (1–46) | 7 (1–37) | .32 |
| Neutropenia <100/µL, days; median (range) | 6 (0–36) | 6 (0–30) | .42 |
| Fluoroquinolone prophylaxis | 35 (56) | 34 (57) | .98 |
| Antibiotic use, days; median (range)b | 1 (0–14) | 2 (0–17) | .61 |
| Febrile neutropenia | 58 (92.0) | 47 (78.3) | .041 |
| Hospital days before febrile neutropenia; median (range) | 12 (1–24) | 12 (1–22) | .6 |
| Clinically documented infection | 29 (46.0) | 20 (33.3) | .19 |
| Pneumonia | 8 (12.7) | 6 (10.0) | .7 |
| Colitis | 4 (6.3) | 2 (3.3) | .6 |
| Exit site infection | 5 (7.9) | 4 (6.6) | 1.0 |
| Microbiologically documented infection | 11 (7.4) | 10 (6.6) | 1.0 |
| Bloodstream infection | 10 (15.8) | 7 (11.6) | .6 |
| Estimated glomerular filtration rate (mL/min/1.73 m2),c mean ± standard deviation | 109.4 ± 17.7 | 109.8 ± 21.5 | .9 |
| Primary antibiotic treatment | .6 | ||
| Piperacillin-tazobactam | 52 (89.6) | 44 (93.6) | … |
| Ceftazidime | 4 (6.9) | 3 (6.3) | … |
| Meropenem | 2 (3.4) | 0 (0) | … |
| Vancomycin co-treatment | 22 (39.2) | 19 (42.2) | .8 |
All numbers represent number of patients (percent of total in treatment arm), except where otherwise specified.
Abbreviations: CI, comorbidity index; HCT, hematopoietic cell transplantation; IQR, interquartile range.
aAllogeneic HCT only.
bBefore onset of febrile neutropenia, excluding prophylactic ciprofloxacin.
cCalculated with the Chronic Kidney Disease Epidemiology Collaboration formula.
Characteristics of 123 Enrolled Patients
| Characteristic | Bolus Infusion (n = 63) | Extended Infusion (n = 60) | P Value |
|---|---|---|---|
| Age, years; median (IQR) | 60.1 (45.8–66.3) | 60.4 (50.9–67.5) | .74 |
| Sex, female | 28 (44.4) | 23 (38.3) | .58 |
| Weight, kg; median (range) | 75 (50.4–111.7) | 81 (58–108) | .13 |
| Primary diagnosis | .46 | ||
| Acute leukemia/myelodysplastic syndrome | 26 (41) | 18 (30) | … |
| Lymphoma | 17 (27) | 17 (28) | … |
| Multiple myeloma | 18 (29) | 24 (40) | … |
| Other | 2 (3) | 1 (2) | |
| Reason for admission | .29 | ||
| Allogeneic HCT | 21 (33) | 18 (30) | … |
| Autologous HCT | 29 (46) | 36 (60) | … |
| Induction for acute leukemia | 7 (11) | 4 (7) | … |
| Consolidation for acute leukemia | 6 (10) | 2 (3) | … |
| Myeloablative regimena | 12 (57) | 8 (44) | .43 |
| HCT-CI; median (IQR)a | 2 (2–3) | 2 (1–3) | .21 |
| Karnofsky score; median (IQR) | 100% (90–100) | 100% (80–100) | .15 |
| Neutropenia <500/µL, days; median (range) | 7 (1–46) | 7 (1–37) | .32 |
| Neutropenia <100/µL, days; median (range) | 6 (0–36) | 6 (0–30) | .42 |
| Fluoroquinolone prophylaxis | 35 (56) | 34 (57) | .98 |
| Antibiotic use, days; median (range)b | 1 (0–14) | 2 (0–17) | .61 |
| Febrile neutropenia | 58 (92.0) | 47 (78.3) | .041 |
| Hospital days before febrile neutropenia; median (range) | 12 (1–24) | 12 (1–22) | .6 |
| Clinically documented infection | 29 (46.0) | 20 (33.3) | .19 |
| Pneumonia | 8 (12.7) | 6 (10.0) | .7 |
| Colitis | 4 (6.3) | 2 (3.3) | .6 |
| Exit site infection | 5 (7.9) | 4 (6.6) | 1.0 |
| Microbiologically documented infection | 11 (7.4) | 10 (6.6) | 1.0 |
| Bloodstream infection | 10 (15.8) | 7 (11.6) | .6 |
| Estimated glomerular filtration rate (mL/min/1.73 m2),c mean ± standard deviation | 109.4 ± 17.7 | 109.8 ± 21.5 | .9 |
| Primary antibiotic treatment | .6 | ||
| Piperacillin-tazobactam | 52 (89.6) | 44 (93.6) | … |
| Ceftazidime | 4 (6.9) | 3 (6.3) | … |
| Meropenem | 2 (3.4) | 0 (0) | … |
| Vancomycin co-treatment | 22 (39.2) | 19 (42.2) | .8 |
| Characteristic | Bolus Infusion (n = 63) | Extended Infusion (n = 60) | P Value |
|---|---|---|---|
| Age, years; median (IQR) | 60.1 (45.8–66.3) | 60.4 (50.9–67.5) | .74 |
| Sex, female | 28 (44.4) | 23 (38.3) | .58 |
| Weight, kg; median (range) | 75 (50.4–111.7) | 81 (58–108) | .13 |
| Primary diagnosis | .46 | ||
| Acute leukemia/myelodysplastic syndrome | 26 (41) | 18 (30) | … |
| Lymphoma | 17 (27) | 17 (28) | … |
| Multiple myeloma | 18 (29) | 24 (40) | … |
| Other | 2 (3) | 1 (2) | |
| Reason for admission | .29 | ||
| Allogeneic HCT | 21 (33) | 18 (30) | … |
| Autologous HCT | 29 (46) | 36 (60) | … |
| Induction for acute leukemia | 7 (11) | 4 (7) | … |
| Consolidation for acute leukemia | 6 (10) | 2 (3) | … |
| Myeloablative regimena | 12 (57) | 8 (44) | .43 |
| HCT-CI; median (IQR)a | 2 (2–3) | 2 (1–3) | .21 |
| Karnofsky score; median (IQR) | 100% (90–100) | 100% (80–100) | .15 |
| Neutropenia <500/µL, days; median (range) | 7 (1–46) | 7 (1–37) | .32 |
| Neutropenia <100/µL, days; median (range) | 6 (0–36) | 6 (0–30) | .42 |
| Fluoroquinolone prophylaxis | 35 (56) | 34 (57) | .98 |
| Antibiotic use, days; median (range)b | 1 (0–14) | 2 (0–17) | .61 |
| Febrile neutropenia | 58 (92.0) | 47 (78.3) | .041 |
| Hospital days before febrile neutropenia; median (range) | 12 (1–24) | 12 (1–22) | .6 |
| Clinically documented infection | 29 (46.0) | 20 (33.3) | .19 |
| Pneumonia | 8 (12.7) | 6 (10.0) | .7 |
| Colitis | 4 (6.3) | 2 (3.3) | .6 |
| Exit site infection | 5 (7.9) | 4 (6.6) | 1.0 |
| Microbiologically documented infection | 11 (7.4) | 10 (6.6) | 1.0 |
| Bloodstream infection | 10 (15.8) | 7 (11.6) | .6 |
| Estimated glomerular filtration rate (mL/min/1.73 m2),c mean ± standard deviation | 109.4 ± 17.7 | 109.8 ± 21.5 | .9 |
| Primary antibiotic treatment | .6 | ||
| Piperacillin-tazobactam | 52 (89.6) | 44 (93.6) | … |
| Ceftazidime | 4 (6.9) | 3 (6.3) | … |
| Meropenem | 2 (3.4) | 0 (0) | … |
| Vancomycin co-treatment | 22 (39.2) | 19 (42.2) | .8 |
All numbers represent number of patients (percent of total in treatment arm), except where otherwise specified.
Abbreviations: CI, comorbidity index; HCT, hematopoietic cell transplantation; IQR, interquartile range.
aAllogeneic HCT only.
bBefore onset of febrile neutropenia, excluding prophylactic ciprofloxacin.
cCalculated with the Chronic Kidney Disease Epidemiology Collaboration formula.
Schematic of patient randomization and study populations. Abbreviation: ITT, intention-to-treat.
Schematic of patient randomization and study populations. Abbreviation: ITT, intention-to-treat.
Febrile neutropenia occurred in 105 patients, who were included in the ITT analysis: 58 (92.0%) in the bolus infusion arm and 47 (78.3%) in the extended infusion arm (P = .041). Of these, 96 (91.4%) were treated initially with piperacillin-tazobactam, 7 (6.6%) were treated with ceftazidime, and 2 (1.9%) were treated with meropenem. Forty-one patients (40.5%) received vancomycin as part of their initial antibiotic regimen. Ninety-one patients received at least 48 hours of a study antibiotic regimen and were included in the per-protocol analysis: 48 (83% of the patients with febrile neutropenia) in the bolus infusion arm and 43 (91%) in the extended infusion arm. Fourteen patients were excluded from the per-protocol analysis: 11 due to early antibiotic switch to a carbapenem, 2 who were initially started on a carbapenem, and 1 whose antibiotic treatment was discontinued after <24 hours due to symptom resolution (Figure 1).
Documented Infections
Forty-seven patients (44.7%) had clinically documented infections. Sixteen (34.0%) had bloodstream infection, 13 (27.6%) had opacities on chest imaging consistent with inflammatory infiltrates, 11 (23.4%) had central venous catheter exit site or tunnel infection, and 6 (12.7%) had enterocolitis.
Microbiologically documented infections were observed in 21 patients (20.0%), including 16 patients with bloodstream infection (Table 2). Gram-negative organisms were isolated in 15 (71.4%) patients, gram-positive organisms in 5 patients (23.8%), and Candida spp. in 1 (4.7%) patient. Two of 10 Enterobacteriaceae (20%) were extended-spectrum β-lactamase producers. There were no infections with carbapenem-resistant Enterobacteriaceae, vancomycin-resistant Enterococci, or methicillin-resistant Staphylococcus aureus.
Culture Results of Patients With Febrile Neutropenia
| Organism | Intermittent Bolus | Extended Infusion | ||
|---|---|---|---|---|
| All MDI | BSI | All MDI | BSI | |
| N = 11 | N = 10 | N = 10 | N = 7 | |
| Gram positive | 3 (27) | 3 (30) | 2 (20) | 2 (42) |
| Coagulase negative Staphylococcus | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Viridans group Streptococcus | 1 (9) | 1 (10) | 1 (10) | 1 (14) |
| Granulicatella spp. | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Listeria monocytogenes | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Gram negative | 8 (72) | 7 (70) | 7 (70) | 4 (57) |
| Escherichia coli | 4 (36) | 3 (30) | 4 (40) | 2 (29) |
| Klebsiella pneumoniae | 2 (18) | 2 (20) | 0 (0) | 0 (0) |
| Pseudomonas aeruginosa | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Extended spectrum β-lactamase– producing Enterobacteriaceae | 2 (18) | 1 (10) | 0 (0) | 0 (0) |
| Piperacillin-tazobactam (S) | 4 (66) | 4 (80) | 4 (80) | 2 (66) |
| Ceftazidime (S) | 4 (66) | 4 (80) | 5 (100) | 3 (100) |
| Campylobacter spp. | 1 (9) | 1 (10) | 1 (10) | 0 (0) |
| Fusobacterium nucleatum | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Leptotricha buccalis | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Candida spp. | ||||
| C. albicans | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Organism | Intermittent Bolus | Extended Infusion | ||
|---|---|---|---|---|
| All MDI | BSI | All MDI | BSI | |
| N = 11 | N = 10 | N = 10 | N = 7 | |
| Gram positive | 3 (27) | 3 (30) | 2 (20) | 2 (42) |
| Coagulase negative Staphylococcus | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Viridans group Streptococcus | 1 (9) | 1 (10) | 1 (10) | 1 (14) |
| Granulicatella spp. | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Listeria monocytogenes | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Gram negative | 8 (72) | 7 (70) | 7 (70) | 4 (57) |
| Escherichia coli | 4 (36) | 3 (30) | 4 (40) | 2 (29) |
| Klebsiella pneumoniae | 2 (18) | 2 (20) | 0 (0) | 0 (0) |
| Pseudomonas aeruginosa | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Extended spectrum β-lactamase– producing Enterobacteriaceae | 2 (18) | 1 (10) | 0 (0) | 0 (0) |
| Piperacillin-tazobactam (S) | 4 (66) | 4 (80) | 4 (80) | 2 (66) |
| Ceftazidime (S) | 4 (66) | 4 (80) | 5 (100) | 3 (100) |
| Campylobacter spp. | 1 (9) | 1 (10) | 1 (10) | 0 (0) |
| Fusobacterium nucleatum | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Leptotricha buccalis | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Candida spp. | ||||
| C. albicans | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
All numbers represent number of isolates (percent of total in treatment arm).
Abbreviations: BSI, bloodstream infection; MDI, microbiologically documented infection; S, susceptible to drug.
Culture Results of Patients With Febrile Neutropenia
| Organism | Intermittent Bolus | Extended Infusion | ||
|---|---|---|---|---|
| All MDI | BSI | All MDI | BSI | |
| N = 11 | N = 10 | N = 10 | N = 7 | |
| Gram positive | 3 (27) | 3 (30) | 2 (20) | 2 (42) |
| Coagulase negative Staphylococcus | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Viridans group Streptococcus | 1 (9) | 1 (10) | 1 (10) | 1 (14) |
| Granulicatella spp. | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Listeria monocytogenes | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Gram negative | 8 (72) | 7 (70) | 7 (70) | 4 (57) |
| Escherichia coli | 4 (36) | 3 (30) | 4 (40) | 2 (29) |
| Klebsiella pneumoniae | 2 (18) | 2 (20) | 0 (0) | 0 (0) |
| Pseudomonas aeruginosa | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Extended spectrum β-lactamase– producing Enterobacteriaceae | 2 (18) | 1 (10) | 0 (0) | 0 (0) |
| Piperacillin-tazobactam (S) | 4 (66) | 4 (80) | 4 (80) | 2 (66) |
| Ceftazidime (S) | 4 (66) | 4 (80) | 5 (100) | 3 (100) |
| Campylobacter spp. | 1 (9) | 1 (10) | 1 (10) | 0 (0) |
| Fusobacterium nucleatum | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Leptotricha buccalis | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Candida spp. | ||||
| C. albicans | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Organism | Intermittent Bolus | Extended Infusion | ||
|---|---|---|---|---|
| All MDI | BSI | All MDI | BSI | |
| N = 11 | N = 10 | N = 10 | N = 7 | |
| Gram positive | 3 (27) | 3 (30) | 2 (20) | 2 (42) |
| Coagulase negative Staphylococcus | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Viridans group Streptococcus | 1 (9) | 1 (10) | 1 (10) | 1 (14) |
| Granulicatella spp. | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Listeria monocytogenes | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Gram negative | 8 (72) | 7 (70) | 7 (70) | 4 (57) |
| Escherichia coli | 4 (36) | 3 (30) | 4 (40) | 2 (29) |
| Klebsiella pneumoniae | 2 (18) | 2 (20) | 0 (0) | 0 (0) |
| Pseudomonas aeruginosa | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Extended spectrum β-lactamase– producing Enterobacteriaceae | 2 (18) | 1 (10) | 0 (0) | 0 (0) |
| Piperacillin-tazobactam (S) | 4 (66) | 4 (80) | 4 (80) | 2 (66) |
| Ceftazidime (S) | 4 (66) | 4 (80) | 5 (100) | 3 (100) |
| Campylobacter spp. | 1 (9) | 1 (10) | 1 (10) | 0 (0) |
| Fusobacterium nucleatum | 1 (9) | 1 (10) | 0 (0) | 0 (0) |
| Leptotricha buccalis | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
| Candida spp. | ||||
| C. albicans | 0 (0) | 0 (0) | 1 (10) | 1 (14) |
All numbers represent number of isolates (percent of total in treatment arm).
Abbreviations: BSI, bloodstream infection; MDI, microbiologically documented infection; S, susceptible to drug.
Treatment Outcomes
Primary Endpoint
Overall treatment response on day 4 occurred in 67 (63.8%) patients in the ITT population and 65 (71.4%) patients in the per-protocol population. In the ITT population, the primary endpoint occurred in 32 patients (55.1%) in the bolus infusion arm vs 35 patients (74.4%) in the extended infusion arm (absolute difference, 19.3%; 95% CI, 1.4%–37.1%; number needed to treat [NNT] with extended vs bolus infusion, 5 patients for each clinical response; P = .044). In the per-protocol analysis, the primary endpoint occurred in 30 patients (62.5%) in the bolus infusion arm and 35 patients (81.4%) in the extended infusion arm (absolute difference, 18.8%; 95% CI, 0.9%–36.8%; P = .063; Table 3, Figure 2).
Treatment Outcomes of Patients With Febrile Neutropenia
| Outcome | Intention-to-Treat Population | Per-Protocol Population | ||||
|---|---|---|---|---|---|---|
| Intermittent Bolus (n = 58) | Extended Infusion (n = 47) | P Value | Intermittent Bolus (n = 48) | Extended Infusion (n = 43) | P Value | |
| Overall response | 32 (55.1) | 35 (74.4) | .044 | 30 (62.5) | 35 (81.4) | .063 |
| Clinically documented infection | 10/28 (35.7) | 13/19 (68.4) | .039 | 9 (42.8) | 13 (81.2) | .041 |
| Pneumonia | 0/8 (0) | 4/5 (80) | .007 | 0/7 (0) | 4/4 (100) | .003 |
| Microbiologically documented infection | 2/11 (18.1) | 4/9 (44.4) | .3 | 1/6 (16.6) | 4/7 (57.1) | .26 |
| Bloodstream infection | 1/10 (10) | 3/6 (50) | .1 | 1/6 (16.6) | 3/5 (60) | .24 |
| Components of overall response | ||||||
| Persistent fever | 17 (29.3) | 13 (27.6) | 1.0 | 14 (29.1) | 10 (23.2) | .6 |
| Clinical failure | 11 (18.9) | 3 (6.3) | .08 | 5 (10.4) | 1 (2.3) | .2 |
| Microbiological failure | 2 (3.5) | 0 (0) | .5 | 1 (2.0) | 0 (0) | 1.0 |
| Premature antibiotic change | 21 (36.2) | 10 (21.2) | .13 | 13 (27.0) | 6 (13.9) | .1 |
| Switch to carbapenem | 14 (24.1) | 9 (19.1) | .6 | 7 (14.5) | 5 (11.6) | .7 |
| Aminoglycoside added | 6 (10.3) | 3 (6.3) | .7 | 3 (6.2) | 3 (6.9) | 1.0 |
| Fluoroquinolone added | 2 (3.4) | 3 (6.3) | .6 | 2 (4.1) | 2 (4.6) | 1.0 |
| Secondary outcomes | ||||||
| Fever, days (median, range) | 2 (1–17) | 2 (1–9) | .9 | 2 (1–17) | 2 (2–9) | .9 |
| Use of noradrenaline due to hypotension | 10 (17.2) | 4 (8.5) | .2 | 4 (8.3) | 2 (4.6) | .6 |
| Acute kidney injury | 6 (10.3) | 3 (6.3) | .7 | 4 (8.3) | 3 (6.9) | 1.0 |
| Clostridium difficile infection | 2 (3.4) | 0 (0) | .5 | 2 (4.2) | 0 (0) | .5 |
| Breakthrough bloodstream infection | 7 (12.0) | 2 (4.2) | .18 | 7 (14.5) | 2 (4.6) | .1 |
| Breakthrough fever after day 4 | 10 (17.2) | 6 (12.7) | .5 | 9 (18.7) | 5 (11.6) | .3 |
| Length of stay, days (median, range) | 24 (13–145) | 23 (15–61) | .3 | 23 (15–145) | 22 (15–61) | .3 |
| Death, 30 days | 2 (3.7) | 1 (2.5) | 1.0 | 1 (2.2) | 1 (2.7) | 1.0 |
| Outcome | Intention-to-Treat Population | Per-Protocol Population | ||||
|---|---|---|---|---|---|---|
| Intermittent Bolus (n = 58) | Extended Infusion (n = 47) | P Value | Intermittent Bolus (n = 48) | Extended Infusion (n = 43) | P Value | |
| Overall response | 32 (55.1) | 35 (74.4) | .044 | 30 (62.5) | 35 (81.4) | .063 |
| Clinically documented infection | 10/28 (35.7) | 13/19 (68.4) | .039 | 9 (42.8) | 13 (81.2) | .041 |
| Pneumonia | 0/8 (0) | 4/5 (80) | .007 | 0/7 (0) | 4/4 (100) | .003 |
| Microbiologically documented infection | 2/11 (18.1) | 4/9 (44.4) | .3 | 1/6 (16.6) | 4/7 (57.1) | .26 |
| Bloodstream infection | 1/10 (10) | 3/6 (50) | .1 | 1/6 (16.6) | 3/5 (60) | .24 |
| Components of overall response | ||||||
| Persistent fever | 17 (29.3) | 13 (27.6) | 1.0 | 14 (29.1) | 10 (23.2) | .6 |
| Clinical failure | 11 (18.9) | 3 (6.3) | .08 | 5 (10.4) | 1 (2.3) | .2 |
| Microbiological failure | 2 (3.5) | 0 (0) | .5 | 1 (2.0) | 0 (0) | 1.0 |
| Premature antibiotic change | 21 (36.2) | 10 (21.2) | .13 | 13 (27.0) | 6 (13.9) | .1 |
| Switch to carbapenem | 14 (24.1) | 9 (19.1) | .6 | 7 (14.5) | 5 (11.6) | .7 |
| Aminoglycoside added | 6 (10.3) | 3 (6.3) | .7 | 3 (6.2) | 3 (6.9) | 1.0 |
| Fluoroquinolone added | 2 (3.4) | 3 (6.3) | .6 | 2 (4.1) | 2 (4.6) | 1.0 |
| Secondary outcomes | ||||||
| Fever, days (median, range) | 2 (1–17) | 2 (1–9) | .9 | 2 (1–17) | 2 (2–9) | .9 |
| Use of noradrenaline due to hypotension | 10 (17.2) | 4 (8.5) | .2 | 4 (8.3) | 2 (4.6) | .6 |
| Acute kidney injury | 6 (10.3) | 3 (6.3) | .7 | 4 (8.3) | 3 (6.9) | 1.0 |
| Clostridium difficile infection | 2 (3.4) | 0 (0) | .5 | 2 (4.2) | 0 (0) | .5 |
| Breakthrough bloodstream infection | 7 (12.0) | 2 (4.2) | .18 | 7 (14.5) | 2 (4.6) | .1 |
| Breakthrough fever after day 4 | 10 (17.2) | 6 (12.7) | .5 | 9 (18.7) | 5 (11.6) | .3 |
| Length of stay, days (median, range) | 24 (13–145) | 23 (15–61) | .3 | 23 (15–145) | 22 (15–61) | .3 |
| Death, 30 days | 2 (3.7) | 1 (2.5) | 1.0 | 1 (2.2) | 1 (2.7) | 1.0 |
Treatment Outcomes of Patients With Febrile Neutropenia
| Outcome | Intention-to-Treat Population | Per-Protocol Population | ||||
|---|---|---|---|---|---|---|
| Intermittent Bolus (n = 58) | Extended Infusion (n = 47) | P Value | Intermittent Bolus (n = 48) | Extended Infusion (n = 43) | P Value | |
| Overall response | 32 (55.1) | 35 (74.4) | .044 | 30 (62.5) | 35 (81.4) | .063 |
| Clinically documented infection | 10/28 (35.7) | 13/19 (68.4) | .039 | 9 (42.8) | 13 (81.2) | .041 |
| Pneumonia | 0/8 (0) | 4/5 (80) | .007 | 0/7 (0) | 4/4 (100) | .003 |
| Microbiologically documented infection | 2/11 (18.1) | 4/9 (44.4) | .3 | 1/6 (16.6) | 4/7 (57.1) | .26 |
| Bloodstream infection | 1/10 (10) | 3/6 (50) | .1 | 1/6 (16.6) | 3/5 (60) | .24 |
| Components of overall response | ||||||
| Persistent fever | 17 (29.3) | 13 (27.6) | 1.0 | 14 (29.1) | 10 (23.2) | .6 |
| Clinical failure | 11 (18.9) | 3 (6.3) | .08 | 5 (10.4) | 1 (2.3) | .2 |
| Microbiological failure | 2 (3.5) | 0 (0) | .5 | 1 (2.0) | 0 (0) | 1.0 |
| Premature antibiotic change | 21 (36.2) | 10 (21.2) | .13 | 13 (27.0) | 6 (13.9) | .1 |
| Switch to carbapenem | 14 (24.1) | 9 (19.1) | .6 | 7 (14.5) | 5 (11.6) | .7 |
| Aminoglycoside added | 6 (10.3) | 3 (6.3) | .7 | 3 (6.2) | 3 (6.9) | 1.0 |
| Fluoroquinolone added | 2 (3.4) | 3 (6.3) | .6 | 2 (4.1) | 2 (4.6) | 1.0 |
| Secondary outcomes | ||||||
| Fever, days (median, range) | 2 (1–17) | 2 (1–9) | .9 | 2 (1–17) | 2 (2–9) | .9 |
| Use of noradrenaline due to hypotension | 10 (17.2) | 4 (8.5) | .2 | 4 (8.3) | 2 (4.6) | .6 |
| Acute kidney injury | 6 (10.3) | 3 (6.3) | .7 | 4 (8.3) | 3 (6.9) | 1.0 |
| Clostridium difficile infection | 2 (3.4) | 0 (0) | .5 | 2 (4.2) | 0 (0) | .5 |
| Breakthrough bloodstream infection | 7 (12.0) | 2 (4.2) | .18 | 7 (14.5) | 2 (4.6) | .1 |
| Breakthrough fever after day 4 | 10 (17.2) | 6 (12.7) | .5 | 9 (18.7) | 5 (11.6) | .3 |
| Length of stay, days (median, range) | 24 (13–145) | 23 (15–61) | .3 | 23 (15–145) | 22 (15–61) | .3 |
| Death, 30 days | 2 (3.7) | 1 (2.5) | 1.0 | 1 (2.2) | 1 (2.7) | 1.0 |
| Outcome | Intention-to-Treat Population | Per-Protocol Population | ||||
|---|---|---|---|---|---|---|
| Intermittent Bolus (n = 58) | Extended Infusion (n = 47) | P Value | Intermittent Bolus (n = 48) | Extended Infusion (n = 43) | P Value | |
| Overall response | 32 (55.1) | 35 (74.4) | .044 | 30 (62.5) | 35 (81.4) | .063 |
| Clinically documented infection | 10/28 (35.7) | 13/19 (68.4) | .039 | 9 (42.8) | 13 (81.2) | .041 |
| Pneumonia | 0/8 (0) | 4/5 (80) | .007 | 0/7 (0) | 4/4 (100) | .003 |
| Microbiologically documented infection | 2/11 (18.1) | 4/9 (44.4) | .3 | 1/6 (16.6) | 4/7 (57.1) | .26 |
| Bloodstream infection | 1/10 (10) | 3/6 (50) | .1 | 1/6 (16.6) | 3/5 (60) | .24 |
| Components of overall response | ||||||
| Persistent fever | 17 (29.3) | 13 (27.6) | 1.0 | 14 (29.1) | 10 (23.2) | .6 |
| Clinical failure | 11 (18.9) | 3 (6.3) | .08 | 5 (10.4) | 1 (2.3) | .2 |
| Microbiological failure | 2 (3.5) | 0 (0) | .5 | 1 (2.0) | 0 (0) | 1.0 |
| Premature antibiotic change | 21 (36.2) | 10 (21.2) | .13 | 13 (27.0) | 6 (13.9) | .1 |
| Switch to carbapenem | 14 (24.1) | 9 (19.1) | .6 | 7 (14.5) | 5 (11.6) | .7 |
| Aminoglycoside added | 6 (10.3) | 3 (6.3) | .7 | 3 (6.2) | 3 (6.9) | 1.0 |
| Fluoroquinolone added | 2 (3.4) | 3 (6.3) | .6 | 2 (4.1) | 2 (4.6) | 1.0 |
| Secondary outcomes | ||||||
| Fever, days (median, range) | 2 (1–17) | 2 (1–9) | .9 | 2 (1–17) | 2 (2–9) | .9 |
| Use of noradrenaline due to hypotension | 10 (17.2) | 4 (8.5) | .2 | 4 (8.3) | 2 (4.6) | .6 |
| Acute kidney injury | 6 (10.3) | 3 (6.3) | .7 | 4 (8.3) | 3 (6.9) | 1.0 |
| Clostridium difficile infection | 2 (3.4) | 0 (0) | .5 | 2 (4.2) | 0 (0) | .5 |
| Breakthrough bloodstream infection | 7 (12.0) | 2 (4.2) | .18 | 7 (14.5) | 2 (4.6) | .1 |
| Breakthrough fever after day 4 | 10 (17.2) | 6 (12.7) | .5 | 9 (18.7) | 5 (11.6) | .3 |
| Length of stay, days (median, range) | 24 (13–145) | 23 (15–61) | .3 | 23 (15–145) | 22 (15–61) | .3 |
| Death, 30 days | 2 (3.7) | 1 (2.5) | 1.0 | 1 (2.2) | 1 (2.7) | 1.0 |
Absolute difference between study arms with respect to overall response on day 4 in the intention-to-treat population.
Absolute difference between study arms with respect to overall response on day 4 in the intention-to-treat population.
Patients with microbiologically documented infection, bloodstream infection, clinically documented infection, and pneumonia had significantly lower rates of overall response compared with patients without these characteristics (Supplementary Table 1). Specifically, the overall response rates for each patient group compared with the rest of the study cohort were: microbiologically documented infection, 30% (6/20) vs 71.7% (61/85), P = .0013; bloodstream infection, 25.0% (4/16) vs 70.7% (63/89), P = .0011; clinically documented infection, 48.9% (23/47) vs 75.8% (44/58), P = .0075; and pneumonia, 30.7% (4/13) vs 68.4% (63/92), P = .012. We performed prespecified analyses of the primary endpoint stratified for these covariates.
For patients with clinically determined infections, overall response occurred more frequently in the extended infusion arm vs the bolus infusion arm (ITT: absolute difference 32.7%, 95% CI 5.2%–60.1%, NNT 3, P = .039; per-protocol: absolute difference 38.3%, 95% CI 9.8%–66.9%, NNT 3, P = .041). Specifically, significant differences in overall response between the bolus infusion and extended infusion arms were observed for patients with pulmonary infiltrates (ITT: 0% [0/8] vs 80% [4/5], P = .007; per-protocol: 0% [0/7] vs 100% [4/4], P = .003, respectively; Table 3, Figure 2). No significant differences in overall response were found between study arms for patients with clinical infections other than pneumonia (Table 3, Figure 2).
Stratification for microbiologically documented infection was limited by the small size of this subgroup. Although there was a trend for a greater overall response rate with extended vs bolus infusion in patients with bloodstream infection, the difference was not statistically significant (Table 3, Figure 2).
Among components of the primary endpoint used to classify treatment failure, clinical worsening or failure to improve were nonsignificantly more frequent among patients treated with bolus infusion (Table 3, Figure 2). The number of days with fever and the proportion of patients who were switched to a secondary antibiotic regimen did not differ between study arms (Table 3, Figure 2).
Secondary Endpoints
Rates of use of noradrenaline for hypotension and breakthrough bloodstream infection were nonsignificantly higher in patients treated with bolus infusion vs those receiving extended infusion. Duration of fever, length of hospital stay, and rates of acute kidney injury and Clostridium difficile infection were similar between study arms (Table 3, Figure 2). Three patients died by day 30, 2 in the bolus infusion arm and 1 in the extended infusion arm. Rates of breakthrough fever occurring after treatment day 4 and breakthrough bloodstream infection were also similar between treatment arms.
DISCUSSION
In this prospective, randomized study, we compared the efficacy of extended infusion of broad-spectrum β-lactams with standard bolus infusion in high-risk patients with hematologic malignancies and febrile neutropenia. For the primary study endpoint, extended β-lactam infusion resulted in a higher overall response rate compared with bolus infusion in the ITT population. The superiority of extended β-lactam infusion was noted in the subgroup of patients with clinically documented infection and specifically in those with pneumonia. Similar results were observed in the per-protocol analyses.
Of the individual components comprising the primary study endpoint, clinical response, microbiological response, and premature change of the primary antibiotic regimen were all associated with nonsignificantly higher rates of favorable response in patients treated with extended β-lactam infusion. No difference was observed in the time to resolution of fever. Secondary endpoints of requirement for noradrenaline for hemodynamic support and breakthrough bloodstream infection were less frequent in patients in the extended β-lactam infusion arm, but the difference did not reach statistical significance.
Animal models and human studies have shown that β-lactams have time-dependent antibacterial activity, meaning that the PK/PD index determining their efficacy is the time during which the unbound plasma drug concentration exceeds the bacterial MIC (ƒT > MIC). Maximal bacterial killing and host survival are achieved with ƒT > MIC greater than 40% to 50% of the dosing interval [17, 18]. It is increasingly recognized that standard antibiotic dosing schedules frequently fail to achieve this PK/PD target, particularly in critically ill patients [7] and in patients with cancer [19–21]. In an observational study of critically ill patients, extended (2 to 4 hour) infusion of piperacillin-tazobactam was associated with significantly higher rates of PK/PD target attainment and lower rates of mortality in patients with respiratory infection and those with high sequential organ failure assessment scores [9]. An observational study of patients with febrile neutropenia found that extended infusion of meropenem was independently associated with a higher rate of clinical success on day 5 and fewer additional antibiotics prescribed [7]. However, prospective, randomized trials that compared extended or continuous β-lactam infusion with bolus schedules have yielded conflicting results. Some prospective trials demonstrated higher rates of clinical response in patients treated with extended infusion [8, 10], whereas others found no significant benefit [22, 23]. These discrepancies among study results may be explained by differences in trial design, target population, and treatment endpoints. Specifically, the benefit of extended infusion seems to be greatest for patients with pulmonary infection [8–10], whereas a study of patients with complicated abdominal infection found no advantage to continuous piperacillin-tazobactam infusion [23]. Importantly, there have been no prospective randomized studies to assess the efficacy of extended β-lactam infusion in febrile neutropenic patients.
Certain characteristics of patients with febrile neutropenia should be taken into account when interpreting the results of our study. First, only a fraction of patients with febrile neutropenia have underlying infection [24, 25]. Moreover, resolution and reoccurrence of fever may be affected by multiple noninfectious factors, such as mucositis, recovery from neutropenia, febrile reactions to blood products and cytotoxic drugs, and neoplastic fever [26]. These variables may dilute observed antibacterial effects. Thus, it is not surprising that extended β-lactam infusion increased the response rate for patients with clinically documented infection, whereas the duration of fever was similar between treatment arms. Second, based on insights from animal models [14, 27], the relative effect of PK/PD target attainment on antibiotic treatment outcome is expected to be greater in neutropenic vs nonneutropenic patients. Last, pulmonary infiltrates, bloodstream infection, and breakthrough infection are all associated with a high risk of complications and mortality in patients with febrile neutropenia [24]. Thus, although our study was not powered to detect differences in mortality between treatment arms, the ability of extended β-lactam infusion to improve clinical outcomes in these subgroups of patients is of particular importance. Of note, the observed benefit from PK/PD optimization in patients with pneumonia is consistent with studies showing that standard dosing of piperacillin-tazobactam frequently fails to achieve epithelial lining fluid concentrations that exceed the MIC of many causative organisms [28].
Some limitations of this study should be noted. This was a small, single-center trial. The study was unblinded, but the investigators who reviewed and adjudicated clinical outcomes were unaware of the treatment arm allocation of each patient. Thirteen percent of the ITT population were not included in the per-protocol analysis, mainly due to early (<48 hours) switch to treatment with a carbapenem. Nonetheless, treatment effects were consistent across the per-protocol and ITT analyses. Finally, plasma drug levels were not determined in this study.
In conclusion, this prospective, randomized study showed superior treatment outcomes for high-risk patients with febrile neutropenia treated with extended as compared to intermittent bolus β-lactam infusion. Optimization of β-lactam dosing is a broadly available and cost-effective strategy to improve outcomes of high-risk patients with hematological malignancies and febrile neutropenia. These findings should be further validated in larger, multicenter trials.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Note
Potential conflicts of interest. R. B. has received consulting fees from Merck Ltd. and Pfizer pharmaceuticals. All remaining authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.



Comments