Abstract
Background.
Febrile neutropenia (FN) is common in children, adolescents, and young adults (CAYA) undergoing hematopoietic cell transplant (HCT), particularly following allogeneic haploidentical HCT (haplo-HCT). Optimal empiric antibiotic duration for FN is not well established in this population.
Methods.
A retrospective quasi-experimental study was conducted among CAYA undergoing haplo-HCT between 2010 and 2022 who experienced FN. In 2017 a guideline was implemented for discontinuation of empiric antibiotics within 72 hours of FN in patients without identified infection regardless of persistence of fever. We compared patients who underwent haplo-HCT preimplementation (Pre-I) versus postimplementation (Post-I) of this guideline. The primary outcome was days of therapy (DOT). Secondary outcomes included incidence of bloodstream infections (BSI), intensive care unit (ICU) admission, and mortality from 72 hours of FN onset until neutrophil engraftment. A desirability of outcome ranking (DOOR) with a modified response adjusted for duration of antibiotic risk (RADAR) approach using DOT was applied to comprehensively assess outcomes.
Results.
Patients in the Post-I group (n = 82) had fewer DOT compared to Pre-I (n = 90) with a median difference of 8 days (P < .01). There were no significant differences in BSI incidence, ICU admission, or mortality between groups. The DOOR–modified RADAR analysis demonstrated that Post-I was associated with a 70% (95% confidence interval, 61%–78%) probability of experiencing a more favorable clinical outcome compared to Pre-I (P < .01).
Conclusions.
Early antibiotic discontinuation appears safe and feasible in CAYA haplo-HCT recipients with FN, significantly reducing antibiotic exposure without increasing adverse outcomes. Large prospective multicenter trials are needed to confirm these findings.
Keywords: early antibiotic discontinuation strategy; children, adolescents, and young adults; CAYA; hematopoietic cell transplant from haploidentical donor; haplo-HCT; desirability of outcome ranking; DOOR; febrile neutropenia; FN
Febrile neutropenia (FN) occurs frequently in children, adolescents, and young adults (CAYA) after hematopoietic cell transplant (HCT) [1, 2]. Traditionally, clinicians continue empiric broad-spectrum antibiotics in patients with FN until neutrophil recovery, even without an identified infection. This approach is based on previous studies that highlighted the overwhelming risk of severe infections during neutropenia, particularly bloodstream infections (BSIs) [3–7].
In the over 2 decades since these studies were published [3–7], evolving HCT practices [8] have influenced the epidemiology of post-HCT infections [9, 10], while advancements in the antibacterial armamentarium have impacted management of these infections [11]. In recent literature, fever is reported in 80%–90% of haploidentical HCT (haplo-HCT) recipients [12–14], but BSI occurs in <30% of patients [9, 10], compared to a BSI frequency of 41%–65% in earlier studies [3–7]. Alternative causes for FN in haplo-HCT include alloreactivity from mismatch, cell-infusion fever, cytokine release syndrome, and engraftment syndrome [12, 14, 15]. In addition, infections by multidrug-resistant organisms and Clostridioides difficile have increased worldwide [16–18], negatively impacting HCT recipients who experience poor outcomes from these infections [19, 20]. These epidemiologic changes suggest the need for a paradigm shift in antibiotic therapy post-HCT.
The impact of early discontinuation of empiric antibiotics in high-risk patients with FN is recognized as a knowledge gap in recent guidelines [1]. Studies among adult HCT recipients and one in pediatric allogeneic HCT recipients showed that reducing days of antibiotic exposure did not increase adverse outcomes [21–28]. However, most studies were in adults, had varying antibiotic durations, and used prophylaxis as a step-down strategy.
Given this knowledge gap, we evaluated the impact of an institutional guideline for early antibiotic discontinuation for FN in CAYA undergoing haplo-HCT.
METHODS
Study Design and Population
This retrospective cohort study with quasi-experimental design investigated the impact of an early antibiotic discontinuation guideline implemented in 2017. CAYA <24 years old who underwent haplo-HCT at St Jude Children’s Research Hospital between January 2010 and October 2022 were included. Only the first haplo-HCT was included. Patients with 1 episode of FN without identified infection who remained stable within 72 hours of FN were included (Figure 1). Patients without FN before engraftment and those with infection or clinical instability within 72 hours of FN onset were excluded.
Figure 1.
Flow diagram of patient selection. From the total haplo-HCT, we excluded 48 patients without FN before engraftment as well as 29 repeated HCTs. Afterward, we excluded 51 patients with infection and 23 with clinical instability within 72 hours of FN onset, resulting in a total of 184 patients: 96 in the Pre-I group and 88 in the Post-I group. Finally, for outcome analysis, we excluded 12 patients who achieved neutrophil engraftment before 72 hours of onset of FN. With this, we included 172 patients: 90 in the Pre-I group and 82 in the Post-I group. Abbreviations: BSI, bloodstream infection; FN, febrile neutropenia; haplo-HCT, haploidentical hematopoietic cell transplantation; Post-I, postimplementation; Pre-I, preimplementation; SSTI, skin and soft tissue infection; UTI, urinary tract infection.
Data from electronic medical records were entered into a Research Electronic Data Capture (REDCap) database, including demographics, transplant characteristics, laboratory results, imaging findings, antibiotic administration and indications, and clinical events. Each FN episode was reviewed to assess clinical presentation, infection source, and outcomes.
Study Definitions
FN was defined as oral/axillary temperature ≥38.3°C or ≥38.0°C persisting for 1 hour during neutropenia (absolute neutrophil count [ANC] <500 cells/μL). Only the first FN episode after HCT cell infusion was included.
Neutrophil engraftment was defined as the first of 3 consecutive days with an ANC ≥500 cells/μL.
Patients were considered to have an identified infection if documented through microbiological results, imaging (ie, radiologic diagnosis of pneumonia), or localized clinical signs of infection (ie, signs of cellulitis, phlegmon, etc). Non-BSI infections during the initial evaluation of FN were collected to exclude patients with an identified infection within the first 72 hours of FN onset.
Patients were considered clinically stable if vital signs were within age-based normal ranges except for fever-related tachycardia and tachypnea, and if they did not require hemodynamic or respiratory support. Sepsis and septic shock were defined according to International Phoenix Pediatric Sepsis Criteria [29]. Clinical stability was assessed within 72 hours of onset of FN and adjudicated by the researchers who reviewed each episode of FN (S. C.-R., G. M.).
Intensive care unit (ICU) admission for any indication was recorded. Infection-related ICU admission was defined as associated with a suspected or confirmed infection.
Days of therapy (DOT) was used as a standardized metric of antimicrobial exposure. One DOT was assigned for each antibiotic administered for a therapeutic indication on a given calendar day. When multiple antibiotics were administered on the same day, each antibiotic contributed to 1 DOT. DOT including prophylaxis was calculated using the same approach and included all antibiotics administered for treatment and prophylaxis.
BSI was defined by at least 1 positive blood culture with a pathogenic organism as per National Healthcare Safety Network definitions [30]. For infections with common commensal organisms, the same organism had to be identified from 2 or more separate specimens. The institutional practice throughout the study was to collect a peripheral blood culture by venipuncture and cultures from all lumens of central lines at FN onset.
Empiric Antibacterial Therapy for FN
Empiric antibiotic therapy for FN was determined by institutional guidelines, which changed during the study period. From January 2010 to 2017, guidelines recommended initiation of empiric therapy at FN onset with continuation until neutrophil engraftment and resolution of fever. In February 2017, an updated guideline was implemented recommending antibiotic discontinuation within 72 hours in clinically stable haplo-HCT recipients without infection, irrespective of persistence of fever. This change was based on review of institutional data and implemented following consensus between HCT and infectious diseases specialists. For this study, patients who underwent HCT before February 2017 were included in the preimplementation (Pre-I) group, while patients with an HCT after the guideline change were included in the postimplementation (Post-I) group. Empiric regimens and antibacterial prophylaxis during neutropenia are summarized in Supplementary Figure 1.
Outcomes
All outcomes were assessed from 72 hours after starting antibiotics for FN until neutrophil recovery, comparing Pre-I and Post-I (Figure 2). As both groups received antibiotics for 72 hours following FN onset, the period from 72 hours to engraftment was chosen to assess the impact of early antibiotic discontinuation.
Figure 2.
Observation period for incidence of outcomes including days of therapy, BSI, ICU admission, ICU admission related to infection, sepsis, septic shock, and mortality. Blue lines represent empiric antibiotics, which were given until neutrophil engraftment in the preimplementation group and for 72 hours in the postimplementation group in stable patients without an identified infection. The pink-shaded area represents the observation period for outcomes from 72 hours after onset of FN to neutrophil engraftment. Abbreviations: BSI, bloodstream infection; DOT, days of therapy; FN, febrile neutropenia; HCT, hematopoietic cell transplant; ICU, intensive care unit. Created with BioRender.com.
The primary outcome was DOT. Antibiotic duration metrics are the preferred outcome when evaluating implementation of shorter antibiotic duration guidelines while accounting for the absence of adverse events that would require treatment prolongation [25, 27, 28]. DOT describes overall antibiotic exposure by accounting for both duration and number of antibiotics administered per day [31, 32] and is widely used in pediatric stewardship research [33]. DOT was assessed both including and excluding use of prophylactic antibiotics.
Secondary outcomes were:
Rate of adverse clinical outcomes defined as incidence of BSI, ICU admission, infection-related ICU admission, sepsis, septic shock, and mortality. These were evaluated independently and ranked by severity with desirability of outcome ranking (DOOR) methodology [34] as illustrated in Figure 3.
Fidelity (rate of provider adherence) and sustainability (practice adherence over time) to the Post-I guideline.
Figure 3.
DOOR approach and combination with DOT. Patients were first assigned a rank according to their clinical outcomes within 72 hours after onset of febrile neutropenia until neutrophil engraftment. Within each DOOR rank, patients were further ranked according to their DOT, with those having less DOT ranking lower, as this was deemed a more favorable outcome. Abbreviations: BSI, bloodstream infection; DOOR, desirability of outcome ranking; DOT, days of therapy; ICU, intensive care unit. Created with BioRender.com.
Statistical Analysis
Each patient contributed 1 haplo-HCT and 1 FN episode. Patients’ characteristics were summarized using descriptive statistics and were compared between the Pre-I and Post-I groups using χ2 or Fisher exact test for categorical variables, and Student t-test or Wilcoxon test for continuous variables. FN days and DOT between groups were compared with Wilcoxon test. Adverse clinical outcomes (sepsis, septic shock, ICU admission, death) were compared using χ2 tests. We performed negative binomial regression to adjust DOT and logistic or Firth regression for adverse outcomes. Clinically relevant variables were included in the model. Statistical significance was P < .05. R version 4.4.0 and GraphPad Prism version 10.2.3 were used for analyses.
DOOR Analysis
Each patient was ranked based on clinical outcomes from 72 hours post–antibiotic initiation until neutrophil engraftment. Those with severe outcomes were ranked higher, while patients without adverse complications ranked lowest, as shown in Figure 3. A DOOR–modified RADAR (response adjusted for duration of antibiotic risk) approach was applied [35], integrating the DOOR outcome ranking with the DOT in the same period. Within each DOOR category, patients were additionally ranked according to their DOT, with those with fewer DOT ranked lower, as it is considered a more favorable outcome (Figure 3). Pre-I and Post-I patients were compared using the DOOR win-probability test. Proportional odds regression was performed to adjust DOOR-DOT by clinically relevant covariates.
RESULTS
Characteristics of the Study Population
From 2010 to 2022, 335 haplo-HCTs were performed in 306 patients; 29 received multiple haplo-HCTs. Of all patients, 258 (84.3%) experienced FN episodes. Among these, 184 (71.3%) had FN episodes while clinically stable and without identified infection within 72 hours of onset: 96 in the Pre-I group and 88 in Post-I group (Figure 1). No differences in age, sex, or race existed between groups. The conditioning regimen intensity differed, with higher frequency of reduced-intensity conditioning in Post-I (95.5% vs 80.2%, P < .01). Time from HCT to FN between groups was similar (3 vs 2 days, P = .7). Neutropenia duration was shorter in Post-I (11 vs 12 days, P < .01), with fewer fever days (7 vs 8 days, P = .05). Median ANC at 72 hours after FN onset was comparable between groups (20 vs 0 cells/μL, P = .62) (Table 1).
Table 1.
Baseline Characteristics of Clinically Stable Haploidentical Hematopoietic Cell Transplant Recipients With Febrile Neutropenia and No Infection Within 72 Hours of Onset of Febrile Neutropenia
| Characteristic | All Subjects | Pre-I | Post-I | P Value |
|---|---|---|---|---|
|
| ||||
| No. | 184 | 96 | 88 | |
| Age, y, median (IQR) | 10 (5–14) | 10 (5–15) | 10 (5–14) | .96 |
| Sex | ||||
| Female | 75 (40.8) | 35 (36.5) | 40 (45.4) | .28 |
| Male | 109 (59.2) | 61 (63.5) | 48 (54.6) | |
| Race | ||||
| Asian | 9 (4.9) | 5 (5.2) | 4 (4.5) | .84 |
| Black | 34 (18.5) | 16 (16.7) | 18 (20.5) | |
| White | 128 (69.6) | 67 (69.8) | 61 (69.3) | |
| Other | 13 (7.1) | 8 (8.3) | 5 (5.7) | |
| Ethnicity | ||||
| Hispanic | 50 (27.2) | 19 (20) | 31 (35.6) | .03 |
| Non-Hispanic | 132 (71.7) | 76 (80) | 56 (64.4) | |
| Underlying diagnosis | ||||
| Acute myeloid leukemia | 73 (39.7) | 37(38.5) | 36 (40.9) | |
| B-cell acute lymphocytic leukemia | 51 (27.7) | 23 (24) | 28 (31.8) | |
| T-cell acute lymphocytic leukemia | 12 (6.5) | 6 (6.2) | 6 (6.8) | |
| Other acute leukemiaa | 16 (8.7) | 11 (11.5) | 5 (5.7) | |
| Other hematologic malignancyb | 16 (8.7) | 12 (12.5) | 4 (4.5) | |
| Hematologic nonmalignant diseasec | 11 (6) | 2 (2.1) | 9 (10.2) | |
| Solid tumord | 5 (2.7) | 5 (5.2) | 0 | |
| Conditioning regimen | ||||
| Myeloablative | 19 (10.3) | 16 (16.7) | 3 (3.4) | <.01 |
| Nonmyeloablative | 4 (2.2) | 3 (3.1) | 1 (1.1) | |
| Reduced intensity | 161 (87.5) | 77 (80.2) | 84 (95.5) | |
| Days of neutropenia, median (IQR) | 11 (10–14) | 12 (11–16) | 11 (10–12) | <.01 |
| Days of FN, median (IQR) | 7 (3–9) | 8 (3–10) | 7 (3–8) | .05 |
| Days to antibiotic discontinuation, median (IQR) | 5.5 (3–9) | 9 (6–13.75) | 3 (3–5) | <.01 |
| ANC 72 h after onset of fever, cells/μL, median (IQR) | 20 (0–100) | 0 (0–100) | 20 (10–70) | .62 |
Data are presented as No. (%) unless otherwise indicated.
Abbreviations: ANC, absolute neutrophil count; FN, febrile neutropenia; IQR, interquartile range; Post-I, postimplementation; Pre-I, preimplementation.
Other acute leukemia included: acute megakaryocytic leukemia (6), acute myelomonocytic leukemia (3), acute promyelocytic leukemia (1), mixed phenotype acute leukemia (6).
Other hematologic malignancy included: chronic myeloid leukemia (1), diffuse large B-cell lymphoma (1), Hodgkin lymphoma (3), myelodysplastic syndrome (7), myeloid sarcoma (1), other unclassified leukemia or lymphoma (3).
Hematologic nonmalignant disease included: aplastic anemia (4), hemophagocytic lymphohistiocytosis (1), myeloid aplasia (1), sickle cell disease (5).
Solid tumor included: embryonal rhabdomyosarcoma (1), Ewing sarcoma (2), neuroblastoma (1), soft tissue sarcoma (1).
Guideline Fidelity and Sustainability
Adherence to the guideline was achieved in 60 of 88 patients (68.2%) in the Post-I group and was sustained over time (Supplementary Figure 2). No significant differences were observed in baseline characteristics between patients for whom the guideline was applied and those for whom it was not (Supplementary Table 1).
Clinical Outcomes
Clinical outcomes were evaluated in 90 of 96 Pre-I patients and 82 of 88 Post-I patients who had not achieved engraftment 72 hours after FN onset.
The Post-I group had fewer DOT compared to the Pre-I group (median difference, 8 days [95% confidence interval {CI}, 6–11 days]; P < .01). When prophylactic antibiotics were included, DOT differed by 14 days between groups (95% CI, 10–17; P < .01) (Figure 4A and 4B). Post-I was significantly associated with fewer DOT after adjusting for age, sex, race, underlying diagnosis, and conditioning regimen (Table 2). Days to antibiotic discontinuation were fewer in Post-I compared to Pre-I (3 vs 9 days, P < .01). This difference remained significant when comparing Post-I patients without guideline adherence to the Pre-I group (7 vs 9 days, P < .01).
Figure 4.
A, Comparison of DOT from 72 hours after onset of FN until neutrophil engraftment between the Pre-I and Post-I groups. Antibiotics given for prophylaxis were excluded from this analysis. B, Comparison of DOT including prophylaxis from 72 hours after onset of FN until neutrophil engraftment between the Pre-I and Post-I groups. **** indicates P1<0.001. Abbreviations: CI, confidence interval; DOT, days of therapy; FN, febrile neutropenia; HL, Hodges–Lehmann statistic; IQR, interquartile range; Post-I, postimplementation; Pre-I, preimplementation.
Table 2.
Days of Therapy Negative Binomial Regression Model
| Variable | Rate Ratio | (95% CI) | P Value |
|---|---|---|---|
|
| |||
| Implementation group | |||
| Pre-I (Ref) | 1 | … | |
| Post-I | 0.34 | (.26–.44) | <.01 |
| Age | 1.0 | (.98–1.03) | .77 |
| Sex | |||
| Male (Ref) | 1 | … | |
| Female | 1.0 | (.98–1.0) | .77 |
| Race | |||
| White (Ref) | 1 | … | |
| Asian | 0.93 | (.50–1.85) | .80 |
| Black | 1.21 | (.87–1.69) | .27 |
| Other | 0.86 | (.53–1.44) | .54 |
| Underlying diagnosis | |||
| ALL (Ref) | 1 | … | |
| AML | 0.87 | (.64–1.16) | .33 |
| Othera | 0.81 | (.58–1.13) | .21 |
| Conditioning regimen | |||
| Myeloablative (Ref) | 1 | … | |
| Reduced intensity | 1.0 | (.64–1.53) | .98 |
| Nonmyeloablative | 1.10 | (.41–3.14) | .85 |
Statistically significant P-values (P < 0.05) are shown in bold.
Abbreviations: ALL, acute lymphocytic leukemia; AML, acute myeloid leukemia; CI, confidence interval; Post-I, postimplementation; Pre-I, preimplementation.
Other underlying diagnoses include: acute megakaryocytic leukemia (6), acute myelomonocytic leukemia (3), acute promyelocytic leukemia (1), mixed phenotype acute leukemia (6), chronic myeloid leukemia (1), diffuse large B-cell lymphoma (1), Hodgkin lymphoma (3), myelodysplastic syndrome (6), myeloid sarcoma (1), other unclassified leukemia or lymphoma (3), aplastic anemia (4), hemophagocytic lymphohistiocytosis (1), myeloid aplasia (1), sickle cell disease (5), embryonal rhabdomyosarcoma (1), Ewing sarcoma (1), neuroblastoma (1).
A BSI was diagnosed in 11 (13.4%) patients in the Post-I group and 6 (6.7%) in the Pre-I group 72 hours after FN onset and before engraftment (P = .2). BSI episodes are described in Supplementary Table 2. There were no statistically significant differences in ICU admission, infection-related ICU admission, sepsis, or septic shock between groups (Table 3). No significant differences in BSI and ICU admission risk were observed between groups after adjusting for age, underlying diagnosis, and conditioning regimen (Supplementary Tables 3 and 4). Among the Post-I group, outcomes were not significantly different between patients with and without guideline adherence, except for fewer DOT in patients for whom the guideline was applied (Supplementary Table 1). No deaths occurred in either group during this period.
Table 3.
Adverse Clinical Outcomes From 72 Hours After Febrile Neutropenia Until Neutrophil Engraftment
| Outcome | Pre-I (n = 90) | Post-I (n = 82) | P Value |
|---|---|---|---|
|
| |||
| BSI after 72 h of initial episode of FN | 6 (6.7) | 11 (13.4) | .2 |
| ICU admission | 8 (8.9) | 10 (12.2) | .62 |
| Infection-related ICU admission | 1 (1.1) | 5 (6.1)a | .10 |
| Sepsis | 1 (1.1) | 2 (2.4) | .61 |
| All-cause mortality | 0 | 0 | |
Abbreviations: BSI, bloodstream infection; FN, febrile neutropenia; ICU, intensive care unit; Post-I, postimplementation; Pre-I, preimplementation.
In the Post-I group, ICU admissions related to infection included: hypoxic respiratory failure (2), hypovolemic shock secondary to diarrhea (1), septic shock (2).
DOOR Results
The probability that a patient in the Post-I group had a more desirable (better-ranked) outcome than a Pre-I patient was 47% (95% CI, 41%–52%), indicating comparable outcomes and no statistically significant difference in overall desirability between groups (P = .26) (Figure 5). After adjustment for DOT, the probability that a Post-I patient had a more desirable outcome than a Pre-I patient increased to 70% (95% CI, 61%–78% P < .01). This association remained significant after adjusting age, underlying diagnosis, and conditioning regimen in the proportional odds model (Supplementary Table 5).
Figure 5.
DOOR distribution and comparison between the Pre-I and Post-I groups. The DOOR probability represents the probability that 1 patient in the Post-I group experiences a more favorable outcome than a patient in the Pre-I group. Abbreviations: CI, confidence interval; DOOR, desirability of outcome ranking; ICU, intensive care unit; Post-I, postimplementation; Pre-I, preimplementation.
DISCUSSION
While current guidelines recommend early antibiotic discontinuation for FN in patients at low risk of infectious complications [1, 2], management of FN in high-risk patients, including allogeneic HCT recipients, is not well established. This study evaluated the impact of an early discontinuation guideline on antibiotic exposure and clinical outcomes. We demonstrated that implementation of this guideline reduced antibiotic exposure with favorable safety among CAYA with FN after haplo-HCT. To our knowledge, this is the first pediatric study evaluating early discontinuation at 72 hours irrespective of persistence of fever, and the largest in the contemporary era of posttransplant supportive care.
Our findings align with a prior study investigating early antibiotic discontinuation in pediatric HCT recipients [28]. In that study of children with FN without infection or hemodynamically instability, 17 patients who received empiric antibiotics for 7 days (postintervention) were compared to 24 patients who continued antibiotics until engraftment (preintervention). Similar to our study, there was a significant reduction in antibiotic exposure in this pediatric population [28]. However, that study required patients to be afebrile for 24 hours before antibiotic discontinuation. Adult HCT studies have used stricter criteria, requiring 48–72 hours of apyrexia before antibiotics discontinuation [21, 23, 26, 27, 36]. Only Ram et al, in a randomized controlled trial of 110 adult HCT and cellular therapy recipients, discontinued antibiotics <72 hours regardless of the persistence of fever; this approach resulted in an overall reduction of antibiotic exposure [25]. Ours is the first pediatric study to report outcomes of early antibiotic discontinuation irrespective of ongoing fever.
Our guideline recommended discontinuation of antibiotics in febrile but clinically stable patients, based on institutional data which demonstrated that fever due to noninfectious causes was common in our haplo-HCT population. In our study, there were fewer febrile days before engraftment in Post-I patients than in Pre-I patients. This finding supports previous reports of FN after haplo-HCT [12–15] in which fever was frequently attributable to noninfectious causes, suggesting that the continuation of antibiotic treatment solely due to persistent fever may not be necessary.
Multiple antibiotic exposure metrics have been utilized to measure impact and adherence to early antibiotic discontinuation practices, including calendar days of antibiotic use [23, 26–28], antibiotic-free days [36], fraction of antibiotic-free neutropenia days [25], number of saving days of antibiotic exposure [22], and DOT [21]. In the context of a new guideline that modified both the duration and number of antibiotics used empirically for FN, DOT was selected as it most accurately reflects overall antibiotic exposure by accounting for both changes [31]. Petteys et al used DOT in a retrospective study of 107 adult HCT recipients [21], which compared patients in whom antibiotics were discontinued before engraftment with those who continued until engraftment. In line with our results, they reported a median of 8 fewer DOT in the allogeneic HCT recipients [21].
Unlike adult studies that include prophylaxis in antibiotic exposure estimates [25], we reported DOT with and without prophylaxis to isolate the effect of the empiric antibacterial treatment guideline from changes in prophylaxis practices. Including prophylaxis increased DOT in the Pre-I group, reflecting its contribution to antibacterial exposure during HCT.
BSI events were infrequent in both groups. In the Post-I group there were 11 BSIs compared to 6 BSIs in the Pre-I group. No significant difference in BSI rates was observed between groups, consistent with prior studies that did not demonstrate a significant increase in BSI following early discontinuation [22, 24–26, 28, 36]. However, this study was not powered to detect smaller but potentially clinically meaningful differences, underscoring the need for larger studies.
In our cohort, two-thirds of the BSI in the Post-I group were caused by cefepime-resistant organisms, which is first-line empiric therapy for FN at our institution. This suggests that continued empiric antibiotic therapy would not have prevented these infections and that early discontinuation was not associated with their occurrence. In the Pre-I group, all BSI organisms were cefepime-resistant, as infections occurred during empiric therapy (Supplementary Table 2). These findings highlight the threat of developing antimicrobial resistance in HCT and the need to develop strategies to reduce antimicrobial exposure in immunocompromised hosts.
Early antibiotic cessation was not associated with an increased risk for preengraftment ICU admission, sepsis, septic shock, or death, even after adjusting for potential confounders, similar to the outcomes reported by others [21–28, 36].
To our knowledge, this study is the first to apply the DOOR approach to evaluate safety and efficacy of antibiotic practices in the HCT population [35]. This approach has 2 advantages. First, it allows for the evaluation of multiple outcomes with varying degrees of severity. Additionally, integrating the DOOR rank with DOT, using a concept analogous to RADAR [35], accounts for both antibiotic exposure and severity of clinical outcomes. When applying DOOR alone, we found no significant difference in outcome desirability between the Post-I and Pre-I groups. Applying DOOR–modified RADAR ordering patients within each rank by DOT, the Post-I group had a significantly higher probability of favorable outcomes compared to Pre-I. This emphasizes that early discontinuation reduced antibiotic exposure without increasing risk of adverse events.
Implementing changes in antibiotic guidelines is challenging, requiring multidisciplinary collaboration. Fidelity is generally categorized as high (80%), moderate (51%–79%), and poor (<50%) [37]. We found moderate fidelity (68%) to the guideline, which remained consistent following the first year of implementation (Supplementary Figure 2). This was possible due to the collaboration between the HCT and the infectious diseases services, which has been integrated with the HCT service since 2013. Ongoing institutional efforts are underway to address potential barriers to guideline adherence.
We compared baseline characteristics and outcomes between Post-I patients with and without guideline adherence, finding no significant differences, except DOT (Supplementary Table 1). Comparing Pre-I and Post-I patients showed a 2-day median reduction in antibiotic use, indicating that antibiotics were still stopped before engraftment.
Our study has limitations. This is a retrospective study over a period of 12 years when multiple practice changes occurred, including HCT protocols, conditioning regimens, and graft-versus host disease prophylaxis. Antibacterial prophylaxis was discontinued 2 months after implementation of this guideline and could have impacted our results by increasing the incidence of BSI. Second, given the infrequent occurrence of adverse outcomes, this study is likely underpowered to assess differences in these outcomes. Last, this is a single-center study conducted at an institution with integrated infectious diseases and HCT services, and guideline implementation was evaluated specifically in haplo-HCT; therefore, findings may not be generalizable to other HCT populations.
In conclusion, this study provides evidence of early antibiotic discontinuation as a safe strategy to reduce antibiotic exposure in pediatric patients with FN following haplo-HCT. Our findings demonstrate a significant reduction in antibiotic exposure without differences in adverse outcomes. Multicenter clinical trials are needed to confirm these findings in larger patient populations. Nevertheless, this study represents a step toward optimizing antibiotic stewardship in pediatric HCT recipients, potentially leading to improved outcomes and reduced antimicrobial resistance.
Supplementary Material
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.
Acknowledgments.
We gratefully acknowledge the support of the Pediatric Infectious Diseases Society for the Antimicrobial Stewardship Fellowship Award, which provided mentorship to carry out this study. We thank Dr Winston Gu of St Jude Children’s Research Hospital for medical writing assistance in editing this manuscript. Authors acknowledge the use of BioRender for the design of figures.
Financial support.
This work was supported in part by the American Lebanese Syrian Associated Charities (ALSAC). No additional external funding was received for the study.
Footnotes
Potential conflicts of interest. S. C.-R. is supported by a RAPID Scholars Award from the Academic Pediatrics Association and the Pediatric Infectious Diseases Society. D. R. H. is supported by Merck. R. L. W. is supported by grant 1K12DK111028 from the National Institute of Diabetes and Digestive and Kidney Diseases. G. M. receives funds for sponsored research from Astellas Inc and SymBio Pharmaceutical and grant funding from the National Institutes of Health. All other authors report no potential 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.
Artificial intelligence disclaimer. Authors acknowledge the use of Paperpal, which was used for refining language. The authors remain responsible for the content of the manuscript.
Data availability.
The data underlying this article will be shared on reasonable request to the corresponding author, subject to institutional review board approval and data-sharing agreements.
References
- 1.Lehrnbecher T, Robinson PD, Ammann RA, et al. Guideline for the management of fever and neutropenia in pediatric patients with cancer and hematopoietic cell transplantation recipients: 2023 update. J Clin Oncol 2023; 41:1774–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Averbuch D, Orasch C, Cordonnier C, et al. European guidelines for empirical antibacterial therapy for febrile neutropenic patients in the era of growing resistance: summary of the 2011 4th European Conference on Infections in Leukemia. Haematologica 2013; 98:1826–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pizzo PA, Ladisch S, Ribichaud K. Treatment of gram-positive septicemia in cancer patients. Cancer 1980; 45:206–7. [DOI] [PubMed] [Google Scholar]
- 4.Pizzo PA, Robichaud KJ, Wesley R, Commers JR. Fever in the pediatric and young adult patient with cancer. A prospective study of 1001 episodes. Medicine (Baltimore) 1982; 61:153–65. [DOI] [PubMed] [Google Scholar]
- 5.Busca A, Saroglia EM, Giacchino M, et al. Analysis of early infectious complications in pediatric patients undergoing bone marrow transplantation. Support Care Cancer 1999; 7:253–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bacterial Engelhard D. and fungal infections in children undergoing bone marrow transplantation. Bone Marrow Transplant 1998; 21(Suppl 2):S78–80. [PubMed] [Google Scholar]
- 7.Altuntaş F, Yıldız O, Eser B, et al. Microbiologically documented infections following peripheral blood stem cell transplantation: single center experience. Turk J Haematol 2005; 22:133–45. [PubMed] [Google Scholar]
- 8.Nagler A, Ruggeri A. Haploidentical stem cell transplantation (HaploSCT) for patients with acute leukemia—an update on behalf of the ALWP of the EBMT. Bone Marrow Transplant 2019; 54(Suppl 2):713–8. [DOI] [PubMed] [Google Scholar]
- 9.Yan C-H, Wang Y, Mo X-D, et al. Incidence, risk factors, microbiology and outcomes of pre-engraftment bloodstream infection after haploidentical hematopoietic stem cell transplantation and comparison with HLA-identical sibling transplantation. Clin Infect Dis 2018; 67(Suppl 2):S162–73. [DOI] [PubMed] [Google Scholar]
- 10.Mikulska M, Raiola AM, Galaverna F, et al. Pre-engraftment bloodstream infections after allogeneic hematopoietic cell transplantation: impact of T cell-replete transplantation from a haploidentical donor. Biol Blood Marrow Transplant 2018; 24:109–18. [DOI] [PubMed] [Google Scholar]
- 11.Pizzo PA. Management of patients with fever and neutropenia through the arc of time: a narrative review. Ann Intern Med 2019; 170:389–97. [DOI] [PubMed] [Google Scholar]
- 12.Wang L, Dai B, Gao W, et al. Clinical significance of haplo-fever and cytokine profiling after graft infusion in allogeneic stem cell transplantation from haplo-identical donors. Front Med (Lausanne) 2022; 9:820591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Arango M, Combariza JF. Fever after peripheral blood stem cell infusion in haploidentical transplantation with post-transplant cyclophosphamide. Hematol Oncol Stem Cell Ther 2017; 10:79–84. [DOI] [PubMed] [Google Scholar]
- 14.Imus PH, Blackford AL, Bettinotti M, et al. Severe cytokine release syndrome after haploidentical peripheral blood stem cell transplantation. Biol Blood Marrow Transplant 2019; 25:2431–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Otoukesh S, Elmariah H, Yang D, et al. Cytokine release syndrome following peripheral blood stem cell haploidentical hematopoietic cell transplantation with post-transplantation cyclophosphamide. Transplant Cell Ther 2022; 28:111.e1–8. [DOI] [PubMed] [Google Scholar]
- 16.Centers for Disease Control and Prevention (CDC). Antibiotic resistance threats in the United States. Atlanta, GA: CDC, 2019. [Google Scholar]
- 17.Averbuch D, Tridello G, Hoek J, et al. Antimicrobial resistance in gram-negative rods causing bacteremia in hematopoietic stem cell transplant recipients: intercontinental prospective study of the Infectious Diseases Working Party of the European Bone Marrow Transplantation Group. Clin Infect Dis 2017; 65:1819–28. [DOI] [PubMed] [Google Scholar]
- 18.Austin K, Sweet M, Likar E, et al. Prospective assessment of Clostridioides (formerly Clostridium) difficile colonization and acquisition in hematopoietic stem cell transplant patients. Transpl Infect Dis 2020; 22:e13438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ramanathan M, Kim S, He N, et al. The incidence and impact of Clostridioides difficile infection on transplant outcomes in acute leukemia and MDS after allogeneic hematopoietic cell transplant—a CIBMTR study. Bone Marrow Transplant 2023; 58:360–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Palla P, Konstantinidis I, Boussiou Z, et al. Bloodstream infections due to multidrug-resistant gram-negative organisms in hematopoietic stem cell transplant recipients: a multicenter case-control study in a high-prevalence area. Transplant Cell Ther 2025; 31:596.e1–12. [DOI] [PubMed] [Google Scholar]
- 21.Petteys MM, Kachur E, Pillinger KE, He J, Copelan EA, Shahid Z. Antimicrobial de-escalation in adult hematopoietic cell transplantation recipients with febrile neutropenia of unknown origin. J Oncol Pharm Pract 2020; 26:632–40. [DOI] [PubMed] [Google Scholar]
- 22.Gustinetti G, Raiola AM, Varaldo R, et al. De-escalation and discontinuation of empirical antibiotic treatment in a cohort of allogeneic hematopoietic stem cell transplantation recipients during the pre-engraftment period. Biol Blood Marrow Transplant 2018; 24:1721–6. [DOI] [PubMed] [Google Scholar]
- 23.Paret R, Le Bourgeois A, Guillerm G, et al. Safety and risk of febrile recurrence after early antibiotic discontinuation in high-risk neutropenic patients with haematological malignancies: a multicentre observational study. J Antimicrob Chemother 2022; 77:2546–56. [DOI] [PubMed] [Google Scholar]
- 24.Snyder M, Pasikhova Y, Baluch A. Early antimicrobial de-escalation and stewardship in adult hematopoietic stem cell transplantation recipients. Retrospective review. Open Forum Infect Dis 2017; 4:ofx226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ram R, Amit O, Adler A, et al. Early antibiotic deescalation and discontinuation in patients with febrile neutropenia after cellular therapy: a single-center prospective unblinded randomized trial. Transplant Cell Ther 2023; 29:708.e1–8. [DOI] [PubMed] [Google Scholar]
- 26.Le Clech L, Talarmin J-P, Couturier M-A, et al. Early discontinuation of empirical antibacterial therapy in febrile neutropenia: the ANTIBIOSTOP study. Infect Dis (Lond) 2018; 50:539–49. [DOI] [PubMed] [Google Scholar]
- 27.Rearigh L, Stohs E, Freifeld A, Zimmer A. De-escalation of empiric broad spectrum antibiotics in hematopoietic stem cell transplant recipients with febrile neutropenia. Ann Hematol 2020; 99:1917–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Patel PA, Teherani MF, Xiang Y, et al. Short-course empiric antibiotics in children undergoing allogeneic hematopoietic cell transplantation. Transplant Cell Ther 2023; 29:778.e1–6. [DOI] [PubMed] [Google Scholar]
- 29.Schlapbach LJ, Watson RS, Sorce LR, et al. International consensus criteria for pediatric sepsis and septic shock. JAMA 2024; 331:665–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.National Healthcare Safety Network (NHSN). 2023 CDC/NHSN Patient Safety Component Manual definitions. Atlanta, GA: Centers for Disease Control and Prevention, 2023. [Google Scholar]
- 31.Barlam TF, Cosgrove SE, Abbo LM, et al. Implementing an antibiotic stewardship program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. Clin Infect Dis 2016; 62:e51–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Foolad F, Huang AM, Nguyen CT, et al. A multicentre stewardship initiative to decrease excessive duration of antibiotic therapy for the treatment of community-acquired pneumonia. J Antimicrob Chemother 2018; 73:1402–7. [DOI] [PubMed] [Google Scholar]
- 33.Gerber JS, Jackson MA, Tamma PD, Zaoutis TE. Antibiotic stewardship in pediatrics. Pediatrics 2021; 147:e2020040295. [DOI] [PubMed] [Google Scholar]
- 34.Ong SWX, Petersiel N, Loewenthal MR, Daneman N, Tong SYC, Davis JS. Unlocking the DOOR-how to design, apply, analyse, and interpret desirability of outcome ranking endpoints in infectious diseases clinical trials. Clin Microbiol Infect 2023; 29:1024–30. [DOI] [PubMed] [Google Scholar]
- 35.Evans SR, Rubin D, Follmann D, et al. Desirability of outcome ranking (DOOR) and response adjusted for duration of antibiotic risk (RADAR). Clin Infect Dis 2015; 61:800–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Aguilar-Guisado M, Espigado I, Martín-Peña A, et al. Optimisation of empirical antimicrobial therapy in patients with haematological malignancies and febrile neutropenia (How Long study): an open-label, randomised, controlled phase 4 trial. Lancet Haematol 2017; 4:e573–83. [DOI] [PubMed] [Google Scholar]
- 37.Nurjono M, Shrestha P, Ang IYH, et al. Implementation fidelity of a strategy to integrate service delivery: learnings from a transitional care program for individuals with complex needs in Singapore. BMC Health Serv Res 2019; 19:177. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article will be shared on reasonable request to the corresponding author, subject to institutional review board approval and data-sharing agreements.