Background
Historically, antimicrobial treatment for bloodstream infection (BSI) has been delivered strictly by the intravenous (IV) route, although IV therapy can be costly and carries risk of toxicity and other complications. Reference Kane-Gill, Kirisci, Verrico and Rothschild1 An emerging body of literature in pediatric and adult patients demonstrates that enteral (PO) antimicrobials reach sufficient concentrations to treat BSI and are safer than IV-only therapy without compromising effectiveness. Reference Grau, Clarivet, Lotthé, Bommart and Parer2–Reference Gunter, Wingler, Cretella, Wagner, Barber and Stover6 Despite no mortality association in adult oncology patients with BSI, IV-to-PO conversion is uncommon in immunocompromised pediatric patients. Reference Itoh, Hadano, Saito, Myokai, Nakamura and Kurai7,Reference Tossey, El Boghdadly and Reed8 Assessments of safety and effectiveness of PO antimicrobials for BSI in pediatric hematology/oncology (HEMONC) and bone marrow transplant (BMT) patients are lacking.
Conversion to PO antimicrobials was uncommon locally, leading our center to establish guidelines encouraging HEMONC/BMT providers to change from IV-to-PO antimicrobials in their pediatric patients with BSI. This implementation review assesses IV-to-PO adoption among patient cases meeting criteria for enteral antimicrobials. Mortality and infection recurrence are presented as balancing measures.
Methods
Study design and setting
In 2019, our freestanding, pediatric, quaternary care hospital implemented a local guideline recommending transition to PO treatment of bacterial or candidal BSI among HEMONC/BMT patients meeting specific criteria, including central line removal (CLR), negative blood cultures, and enteral administration of other medications and nutrition (see Supplemental Material for protocol). This Antimicrobial Stewardship-developed guideline established CLR as a criterion for conversion as it represents substantial source control and provides an opportune time to change antimicrobial routes, although it may not universally be an IV-to-PO requirement. Recommendations were integrated into HEMONC divisional guidelines with re-education occurring annually during divisional Standard Practice training. Education to infectious diseases providers reinforced consideration for IV-to-PO with CLR. Daily audit and feedback by an Antimicrobial Stewardship pharmacist identified eligible cases and encouraged teams to follow guideline recommendations.
This single-center, retrospective assessment describes this guideline implementation, identifying HEMONC/BMT patients with BSI transitioned to PO antimicrobials after CLR. Patients were identified between October 3, 2020, and July 30, 2023. The start date denotes implementation of a new electronic health record for our institution.
BSI-CLR cases were collected for patients admitted to the HEMONC or BMT services when CLR occurred within 14 days of the initial positive blood culture.
Patients were considered eligible for inclusion in this assessment if they were neutropenic or had syndromes that may decrease absorption, although the latter differed from protocol criteria (Supplementary Materials). Infectious diseases consultation was, similarly, not reviewed for inclusion. Patients treated for Mycobacterium infections were excluded from analysis because these pathogens frequently require prolonged treatment durations and multiple, concurrent antimicrobials.
Data were collected using Tableau (Seattle, WA) and the Epic health record (Verona, WI). Retrospective chart review was conducted to validate data.
Outcomes and definitions
The primary outcome was the proportion of eligible BSI-CLR cases among HEMONC/BMT patients with antimicrobial IV-to-PO conversion. Secondary outcomes include infection with any pathogen, recurrence of incident infection, or patient mortality within 30 days of CLR.
PO conversion was defined as ordering an enteral antimicrobial suitable to treat the patient’s BSI within 14 days of CLR. Proportion of PO therapy was calculated as the number of days of enteral antibiotics relative to all antibiotic days of therapy after CLR. Additional variables such as admitting service, unit of care, pathogenic organisms, absolute neutrophil count, and antimicrobial order details were also collected. Absolute neutrophil count was included from the lab drawn closest to the day of CLR, < 200 cells/mm3 defining neutropenia. The unit of care was defined as the unit in which the patient experienced their first positive blood culture. Polymicrobial BSI cases identified more than one pathogen on blood culture. IV-to-PO conversion of any antimicrobial used to treat a pathogen within a polymicrobial infection categorized the entire case as a PO conversion.
Data analysis
Fisher’s exact test was used to compare categorical variables. All analyses were performed using Stata version 14 software (College Station, TX). The institutional review board approved this assessment.
Results
We examined 76 BSI-CLR cases among 63 HEMONC/BMT patients. Patients with at least 1 BSI case meeting PO eligibility criteria had a median age of 5 years (n = 49 patients, range: 0–20 yrs) compared to 10 years for patients not meeting eligibility (n = 14, range: 0–23 yrs). Other demographic data are available in Table 1.
Demographics of cases based on protocol eligibility for IV-to-PO conversion
PO, enteral; IV, intravenous; HEMONC, hematology-oncology service; BMT, bone marrow transplant service; CLR, central line removal; PICU, pediatric intensive care unit; BSI, bloodstream infection; NS, not significant.
Protocol criteria for IV-to-PO eligibility were met in 57 of 76 cases (75%). Cases not meeting criteria were more likely to occur in the ICU and relative to neutropenia (Table 1).
Conversion to PO antimicrobials occurred during 29 of 57 (50.8%) PO-eligible BSI cases and none of the IV-only episodes (Table 2). The median duration of PO therapy was 6 days (range: 2–22), compared to median total treatment course of 10 days (range: 3–28). Patients changed to enteral antimicrobials received PO agents on 278 of 407 days of therapy after CLR.
PO conversion and patient outcomes for BSI-CLR cases
PO, enteral; BSI, bloodstream infection; CLR, central line removal; IV, intravenous; NS, not significant.
Re-infection within 30 days of CLR occurred in 3 cases, all among patients who remained on IV antimicrobials. One of those reinfections was recurrence of the incident non-aeruginosa Pseudomonas identified prior to CLR.
Ten cases resulted in patient expiration within 30 days of the CLR; mortality was more likely to occur among patients not meeting PO eligibility. Antimicrobials were converted to enteral therapy for one of these cases, treating a Rothia and Granulicatella BSI, although this patient’s death appeared to be related to uncontrolled Aspergillus infection.
Pathogens identified among all BSI cases, as well as enteral agents used for completion of therapy, are available in Supplementary Materials Table 1.
Discussion
After implementation of an IV-to-PO guideline for BSI antimicrobials among HEMONC/BMT patients at our institution, 50.8% of eligible cases were converted to enteral therapy after CLR. All-cause mortality and infection within 30 days remained low after conversion.
Hesitancy toward using definitive oral therapy may persist, despite proven safety and efficacy in immunocompromised patients, due to long-standing beliefs that IV antimicrobials are more effective than enteral for treating BSI. Reference Wald-Dickler, Holtom and Phillips9 Historical, conservative practices, such as reticence to de-escalate empiric febrile neutropenia coverage after identification of a targeted pathogen, may also manifest as hesitancy among providers to prescribe enteral therapy in the setting of neutropenia in pediatric patients. Similar reluctance to use enteral antibiotics was described in an assessment of IV-to-PO therapy in adult oncology patients with gram-negative BSI. Reference Tossey, El Boghdadly and Reed8
The findings of our quality improvement initiative illustrate that these historical approaches to care for immunocompromised patients are modifiable. Antimicrobial Stewardship teams can facilitate changes in practice through development of shared guidelines with HEMONC/BMT teams, education regarding safety and risks of IV-to-PO versus IV-only practices, and routine surveillance to identify opportunities to make guideline-concordant changes. With only half of eligible patients undergoing IV-to-PO conversion, our findings also highlight that optimizing these practices likely requires iterative cycles of optimization to identify barriers and solutions to improve implementation.
Our requirement for CLR was conservative relative to many IV-to-PO protocols, which may have limited consideration of patient eligibility. However, we believe that this approach improved the willingness of providers to consider enteral therapy instead of obtaining peripheral access to continue IV antimicrobials. This study was also limited by its single-center and retrospective perspectives.
Overall, our intervention was adopted in half of eligible cases, suggesting that opportunities persist to improve IV-to-PO conversion in pediatric HEMONC/BMT patients after CLR for BSI. Among patients who successfully converted to PO antimicrobials, no adverse safety signals were detected.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ash.2025.10136.
Acknowledgments
None.
Financial support
No external funding was used for this project.
Competing interests
All authors declare no conflicts of interest.
Open access
