The study will be conducted in 4 university hospitals in Spain. The reference site will be Institut Català d’Oncologia L’Hospitalet (ICO Hospitalet) – Hospital Universitari Bellvitge. The other three participating sites will be Hospital Clinic de Barcelona (Barcelona), Hospital Germans Trias i Pujol (Barcelona) and Clínica Universidad de Navarra (Pamplona).
Study population and screening of cases for recruitment
This study will include adult patients with haematological malignancies admitted to the haematology wards who are undergoing chemotherapy or haematopoietic stem cell transplant (HSCT) and experience FN; participants will receive empirical antibiotic therapy with one of the BLA studied.
Adult patients (age ≥ 18 years) undergoing chemotherapy or HSCT who present with FN (defined as axillary temperature ≥ 38.0ºC and < 500 neutrophils/mm3 or < 1000 expected to drop within 24-48 hours), and who receive empirical antibiotic therapy with cefepime, piperacillin-tazobactam or meropenem, in monotherapy or in combination with another antibiotic.
Allergy to the study drugs.
Patient receiving systemic antibiotic treatment (except for prophylaxis) at the time of FN onset.
Absence of fever.
Patients with epilepsy.
Severe renal impairment (defined as creatinine clearance rate by MDRD/CKD-EPI < 30 mL/min).
Previously enrolled patients in whom the time between inclusion and the current episode is less than 5 weeks.
Previously enrolled patients without current resolution of the first episode.
Patients admitted in the haematology wards of the participating centres, receiving chemotherapy or HSCT will be followed up daily by the attending physicians. The study investigators will explain the study to the patients that meet the inclusion criteria, and will ask the to sign the informed consent. On FN onset, patients will be randomised to receive EI or II (1:1). The BLA will be chosen according to the clinical criteria of the attending physician. Once randomised, the first BLA dose will be administered within 30 minutes in all patients. The second dose will be administered according to the randomisation group. In the control group (II) the BLA will be administered within 30 minutes at the usual doses recommended for neutropenic patients (cefepime 2 g/8 h, piperacillin-tazobactam 4 g/6 h and meropenem 1 g/8 h), adjusted according to renal function (Table 1). In the study group, the BLA will be administered at the same doses but by EI (the infusion time will be equal to half the time of the dosing interval). Episodes of FN will be classified as “microbiologically documented infection” (with or without bacteraemia), “clinically documented infection” (without microbiological isolation), “fever of unknown origin” or “non-infectious fever”.
Patients will remain in the study until completing a total of 5 days of antibiotic treatment, or until one of the following events occur: discontinuation of BLA based on clinical criteria, severe adverse event (SAE) or death. If any microorganism is isolated, in vitro susceptibility tests will be performed and antibiotic treatment will be targeted accordingly. If the BLA needs to be switched to another BLA also included in the study due to a lack of in vitro activity, the patient will remain in the study. Measurements and visits will start along with the initiation of the second BLA. If the targeted therapy is an antibiotic not included in the study, the patient will be withdrawn from the study. All patients will receive follow-up for 30 days after initiation of the antibiotic treatment.
The determination of plasma BLA concentrations will not influence the dosage, indication or duration of the BLA administered during the study in any case.
The participant timeline during the study is described in Table 2.
- Withdrawal of consent
- Non-compliance with inclusion or exclusion criteria
- Serious adverse reaction
- Protocol violation
In accordance with the current revision of the Declaration of Helsinki and the applicable regulations, a patient has the right to withdraw from the study at any time and for any reason, without this adversely affecting the medical care provided by the patient’s physician. The withdrawal of full consent from a study means that the patient does not wish to continue participating in the study. Any patient withdrawing their consent will be removed from the study treatment and/or observations immediately after the date the patient requests it.
The primary study endpoint is clinical efficacy, measured by the time to defervescence and no need for antibiotic change. To assess this endpoint the rate of patients with defervescence with no change to the antibiotic treatment within the first 5 days of therapy will be calculated. Defervescence will be considered an axillary temperature < 37.5 ºC for 24 hours, assessed at three consecutive time points, with no new documented fever. For this purpose, axillary temperature will be assessed three times daily by the attending nurse. Time (in hours) to defervescence will also be analysed. In addition, prescribed antibiotics will be reviewed every day and recorded in the Electronic Case Report Form (E-CRF) by the participating investigators at each site to estimate the number of patients who will not require an antibiotic change.
The primary endpoint will be analysed by an investigator who will be blind to the treatment group of each patient.
Secondary endpoints will be assessed as follows:
- PK targets: Number of patients in whom the free antibiotic concentration remains above the MIC of the suspected or isolated microorganism for 50%, 75% and 100% of the dosing interval. The same target will also be assessed, taking into account the number of times per patient. Actual MIC values will be used to assess the PK target in those patients with an isolated pathogen. Otherwise, a surrogate MIC will be inferred from the EUCAST database as the highest MIC in the susceptible range for Pseudomonas aeruginosa 26. For this purpose, the nurse in charge of each patient will take blood samples to determine plasma BLA concentrations at 50% and 100% of the dosing interval at visits 1, 2, and 3, and at 75% of the dosing interval at visit 2. The samples will be frozen and subsequently analysed in the biochemistry laboratory at the reference site.
- Bacteraemia clearance: Time in days until bacteraemia clearance. In patients with bacteraemia, daily blood cultures will be performed by the nurse in charge of each patient until bacteraemia clearance.
- Inflammatory biomarkers: Time in days to normalise or decrease > 50% the peak value of CRP, and number of patients who achieve this objective. Blood samples for CRP determination will be drawn at visits 0, 1, 2, 3 and 5, and will be analysed at each site.
- Overall (30-day) case-fatality rate: All patients included in the study will receive follow-up for 30 days after the end of the intervention by the participating investigators. Death by any cause will be recorded in the E-CRF.
- Population PK analysis to characterise cefepime, piperacillin-tazobactam and meropenem PKs in haematological patients with FN: a population PK model will be developed and validated, using concentration-time data obtained from the blood samples of patients at the reference site (ICO Hospitalet). For each antibiotic, a simultaneous analysis of all concentration-time data will be performed by the population approach using the non-linear mixed effect models implemented in NONMEM 7.4 (Icon development). The model will allow us to identify the clinical covariables that modify the PK parameters. Monte-Carlo simulations will be performed to determine the suitability of different dosing regimens assessed by the probability of target attainment. The validation cohort will include data obtained from patients at the other sites.
- Adverse events (AEs): All AEs considered not related to the haematological disease or its treatment and all serious adverse events (SAEs) will be recorded in the E-CRF by the participating investigators. AEs will be identified and classified according to the severity and potential relationship with the BLA and assessed during the intervention (until visit 5). Afterwards, any AEs related to BLA administration and any SAEs until visit 6 (30 days from the beginning of the intervention) will be recorded.
Data collection and management
All data will be collected by the clinical investigators at each participating site and entered in the E-CRF. The collected data will be age, sex, height, weight, type of underlying disease, other comorbidities, date of hospital admission and discharge, HSCT, type and date of HSCT, immunosuppressive therapy, concomitant medication (particularly other systemic antibiotics, prophylactic antimicrobials, fluids, antipyretics and granulocyte-colony stimulating factor), duration of neutropenia, vital signs, duration of fever, clinical examination, blood tests and microbiological results.
Blood cultures will be drawn at the onset of fever. In patients with bacteraemia, daily blood cultures will be taken until clearance. Other cultures will be collected according to the usual practice and clinical criteria.
Cultures will be processed in the microbiology laboratories at each participating site according to the usual techniques (BACTEC Becton Dickinson, BioMerieux, etc.). The microorganisms isolated will be identified according to the usual methods (conventional method, MALDI-TOF, etc.). Antibiotic susceptibility will be studied by the disc diffusion method and/or by microdilution method. In addition, the exact MIC of the antibiotics administered will be determined by a quantitative method, using E-test and/or adequate microdilution plates.
Sample collection for PK/PD analysis
Plasma BLA concentrations will be determined seven times. At visit 1 (which should be performed between 12 and 36 hours from the start of the study antibiotic treatment), the concentrations will be analysed pre-dose (through concentrations or Cmin) and then mid-way through the dosing interval (C50). Subsequently, the Cmin, C50 and the concentration at the 75% of the dosing interval (C75) will be analysed at visit 2 (if not possible, this will be done at visit 3 or 4). Last, Cmin and C50 will be analysed at visit 3 (or if not possible, at visit 4 or 5, consecutive to the visit 2 samples).
At the reference site, intensive sampling will be carried out at visit 2. The schedule according to drug, method of administration and frequency is explained in Table 3a and 3b.
A 5-mL blood sample will be centrifuged at 2000 g for 10 minutes at 4 °C, and the supernatant will be aliquoted in 1.5-mL-microcentrifuge tubes and and stored at –80 °C until analysis. Samples will be sent in dry ice to Bellvitge University Hospital for bioanalysis.
Total plasma BLA concentrations will be analysed using a previously validated method based on ultra-high-performance liquid chromatography coupled to tandem mass spectrometry 27. Protein binding of 30%, 19% and 2% will be applied for piperacillin, cefepime and meropenem, respectively, to the concentrations determined to estimate the unbound fraction 28,29.
Enrolled patients who sign the informed consent will be randomised to receive BLA by either EI or II (1:1). A centralised electronic computer randomisation schedule will be developed by the Biostatistics Department of the reference site. To ensure that an equal number of participants is assigned to each treatment, patients will be randomised in blocks. The randomisation will be performed in computed-generated variable blocks ranging from four to six, and the investigator will be blind to the size of each block. The code numbers for eligible patients will be assigned in ascending sequential order. The allocation list will be stored at the Biostatistics Department of the reference site. At each participating hospital, patients who provide written informed consent and meet the study criteria will be randomised by investigators, who will obtain the assigned treatment and code number from the computer-assisted web site.
The sample size calculation is based on determining whether the administration of BLA by EI (study group) will be clinically more effective than the administration of BLA by II (control group), for the treatment of FN in haematological patients. Previous data suggest that the clinical efficacy rate in the control group is expected to be 0.45 20,21. A relevant clinical efficacy rate in the study group is expected to be 0.70. Each therapy group will required 75 participants (total of 150) to detect statistically significant differences in clinical efficacy. An alpha risk of 0.05 and a beta risk of 0.2 in a bilateral contrast are assumed, and the sample size allows up to 20% losses to follow-up.
An interim analysis will be carried out once 50% of the cases of each arm are recruited. This analysis will evaluate if the estimated sample size is adequate. The results will be issued by an independent committee of experts not participating in this study.
For the primary endpoint, an intention-to-treat analysis will be carried out. The number of patients in both therapy groups with defervescence and with no change in the antibiotic treatment will be compared by a chi-square test.
For the analysis of secondary endpoints, the between-group difference in the percentage of patients reaching the PK target will be compared. The secondary endpoints are: a) days until bacteraemia clearance, b) time until normalisation or 50% decrease of the initial value of the CRP, and c) days until death by any cause within 30 days, and the incidence of these events will be estimated in each group. Survival analysis methods will be used to compare them using the log-rank test. In addition, the rate of therapeutic failure will be compared.
A population PK model will be developed and validated for each BLA studied in haematological patients with FN, using the concentration-time data obtained from the blood samples. The model will allow us to identify the clinical covariates that modify the PK parameters. Monte-Carlo simulations will be performed to determine the suitability of different dosing regimens assessed by the probability of target attainment.
The AEs and SAEs of the study will be described, and their rates will be compared for each study group. The severity, relationship to treatment and resolution will be described.
An analysis by protocol will also be carried out. All study variables will be presented for each group, using descriptive statistics according to the type of variable. The main analysis will be replicated in the following subgroups: patients with “microbiologically documented infection” (with and without bacteraemia), “clinically documented infection”, “fever of unknown origin” or “non-infectious fever”.
Whenever possible, the estimators of a 95% confidence interval will be included. Statistical significance will be set at a probability level < 0.05. The statistical package used to process the data and carry out the analyses will be R, version 3.6.1 or higher for Windows.
In compliance with the standards of Good Clinical Practice, the sponsor will monitor the study tasks, following the approved monitoring plan. Among others, the monitoring tasks will include assessment of the correct application of the inclusion and exclusion criteria, assessment of the quality of the data collected, development and reporting of any AEs, and maintenance of patient confidentiality.
The need for a Data Monitoring Committee was waived by the Ethic Committee because of the low impact on the safety of the study patients.