Adam J. Esbenshade

Background Pediatric oncology patients with prolonged (≥96 hours) febrile neutropenia (absolute neutrophil count < 500/μL) often undergo an evaluation for invasive fungal disease (IFD) and other infections. Current literature suggests that beta‐D‐glucan (BDG), galactomannan, bronchoalveolar lavage (BAL), and computed tomography (CT) scans (sinus, chest, and abdomen/pelvis) may help determine a diagnosis in this population. Methods In a retrospective cohort study of all cancer/stem cell transplant patients (diagnosed 2005‐2019) from one pediatric hospital, all episodes with prolonged febrile neutropenia or IFD evaluations (defined as sending a fungal biomarker or performing a CT scan to assess for infection) were identified. Results In total, 503 episodes met inclusion criteria and 64% underwent IFD evaluations. In total, 36.4% of episodes documented an infection after initiation of prolonged febrile evaluation, most commonly Clostridioides difficile colitis (6.4%) followed by a true bacterial bloodstream infection (BSI) (5.2%), proven/probable IFD (4.8%), and positive respiratory pathogen panel (3.6%). There was no difference in sinus CTs showing sinusitis (74% vs 63%, p = 0.46), whereas 32% of abdomen/pelvis CTs led to a non‐IFD diagnosis, and 25% of chest CTs showed possible pneumonia. On chest CT, the positive predictive value (PPV) for IFD was 19% for nodules and 14% for tree and bud lesions. BDG had a PPV of 25% for IFD and GM 50%. BAL diagnosed IFD once and pneumocystis jirovecii pneumonia twice. Conclusions Chest CTs and abdomen/pelvis CTs provide clinically relevant information during the prolonged febrile neutropenia evaluation, whereas BDG, galactomannan, BAL, and sinus CTs have less certain utility.

Background Infections cause significant treatment‐related morbidity during pediatric acute lymphoblastic leukemia/lymphoma (ALL/LLy) therapy. Fevers during periods without severe neutropenia are common, but etiologies are not well‐described. This study sought to describe the bloodstream infection (BSI) and non‐BSI risk in children undergoing therapy for ALL/LLy. Methods Demographic and clinical data were abstracted for febrile episodes without severe neutropenia at two children's hospitals. Treatment courses were stratified by intensity. Multivariate logistic regression evaluated characteristics associated with infection. Results There were 1591 febrile episodes experienced by 524 patients. Of these, 536 (34%) episodes had ≥1 infection; BSI occurred in 30 (1.9%) episodes. No BSIs occurred in episodes with a recent procedural sedation or cytarabine exposure. Presence of hypotension, chills/rigors, higher temperature, and infant phenotype were independently associated with BSI ( p < .05). Of the 572 non‐BSIs, the most common was upper respiratory infection (URI) ( n = 381, 67%). Compared to episodes without infection, URI symptoms, higher temperature, absolute neutrophil count 500−999/μl, and evaluation during a low‐intensity treatment course were more likely to be associated with a non‐BSI ( p < .05) and inpatient status was less likely to be associated with a non‐BSI ( p < .05). Conclusions The BSI rate in pediatric patients with ALL/LLy and fever without severe neutropenia is low, but one‐third of the time, patients have a non‐BSI. Future research should test if the need for empiric antibiotics can be tailored based on the associations identified in this study.