A single-center retrospective study has found that whether patients received antibiotics within one hour of developing neutropenic fever after a stem cell transplant made no measurable difference to their clinical outcomes, challenging a guideline standard borrowed from sepsis medicine. Among 224 antibiotic-naïve episodes of neutropenic fever occurring within 30 days of hematopoietic stem cell transplantation at one academic medical center between 2018 and 2022, clinical deterioration — defined as a new vasopressor requirement lasting at least six hours, mechanical ventilation, or death within 90 days — occurred in 17 episodes, or 7.6 percent. Receiving empiric gram-negative antibiotics within one hour of fever onset was not associated with reduced deterioration compared with later administration, after adjustment for baseline severity, transplant type, days from transplant, and comorbidity score.
The study, published in Supportive Care in Cancer, was led by Samuel L. Windham and colleagues and analyzed timestamp-validated data from Washington University’s institutional records. The median time to antibiotics was 1.2 hours, with an interquartile range of 0.8 to 2.2 hours, and 86 of the 224 episodes, or 38.4 percent, received antibiotics within the one-hour window. In the primary categorical analysis, the adjusted odds ratio for deterioration with antibiotics within one hour versus later was 1.18, with a 95 percent confidence interval of 0.37 to 3.78 and a P value of 0.78 — a result statistically indistinguishable from no effect. A continuous per-hour analysis was similarly null, with an adjusted odds ratio of 1.03 per additional hour of delay (95 percent CI 0.85 to 1.24, P = 0.80). Among the covariates tested, only allogeneic transplant status was significantly associated with deterioration, with an adjusted odds ratio of 3.81.
The one-hour standard itself has a complicated lineage. The 2010 IDSA febrile neutropenia guideline emphasized prompt empiric antibiotic therapy but did not specify an hourly threshold for adult inpatients; the explicit one-hour target originates in the 2018 ASCO-IDSA outpatient management update, framed for emergency department and ambulatory presentations. Major adult cancer centers, including MD Anderson Cancer Center, Memorial Sloan Kettering, Cleveland Clinic, and Thomas Jefferson, have since operationalized it as an inpatient standard. The recommendation is derived primarily from sepsis and critical care literature, where delays in antibiotic administration have been associated with increased mortality in patients presenting with septic shock. Whether that threshold applies to neutropenic fever in hospitalized transplant recipients — a population that rarely presents with shock — had not been directly tested before this analysis, according to the authors.
The study’s secondary analyses tested assumptions implicit in the one-hour rule. If faster antibiotics improve survival by treating bacteremia before it progresses, then bacteremia should predict mortality. It did not. Recognized pathogen bacteremia occurred in 59 of 224 episodes (26.3 percent), yet 90-day mortality was similar between bacteremic patients (4 of 59, or 6.8 percent) and non-bacteremic patients (7 of 165, or 4.2 percent). After adjustment, the association between recognized pathogen bacteremia and mortality was not statistically significant (adjusted OR 1.98, 95 percent CI 0.50 to 7.90, P = 0.33). A sensitivity analysis using a broader any-positive culture definition, which captured 69 of 224 episodes (30.8 percent), yielded similar findings, as did analyses in the full cohort of 492 episodes.
Perhaps the most striking finding concerned the deaths themselves. Ten of the 11 deaths in the antibiotic-naïve cohort — 91 percent — occurred without any preceding acute physiologic deterioration from the index neutropenic fever episode, meaning no vasopressor requirement and no mechanical ventilation. Detailed characterization showed that these 10 deaths occurred at a median of 45 days after the index fever, with graft-versus-host disease, disease relapse, or progression of the underlying hematologic malignancy as the dominant clinical context. Only one death was preceded by shock or ventilation. The authors argue this pattern suggests that mortality in this population is driven by underlying disease rather than by the febrile episode itself, in sharp contrast to sepsis cohorts presenting to emergency departments, where death typically follows a trajectory of progressive organ dysfunction.
The study also examined whether pursuing a one-hour target might shift blood culture collection into the post-antibiotic window and reduce diagnostic yield. Among 198 blood culture sets obtained within six hours of fever onset, 91 percent (181 of 198) were collected before antibiotic administration, and 53 culture sets (26.8 percent) grew recognized bacterial pathogens. Positivity was similar in pre-antibiotic cultures (49 of 181, or 27.1 percent) and post-antibiotic cultures (4 of 17, or 23.5 percent), with no statistically significant difference after adjustment (adjusted OR 1.03, 95 percent CI 0.28 to 3.85, P = 0.97). In this setting, at least, the one-hour push did not appear to compromise culture yield.
The microbiology of early post-transplant neutropenic fever may explain why the sepsis-derived timing threshold does not translate. Conditioning regimen toxicity causes mucosal barrier injury throughout the gastrointestinal tract, leading to transient translocation of enteric organisms that can trigger fever and bacteremia without the fulminant course of community-acquired sepsis. The organism distribution in the cohort supported this mechanism: among 69 bacteremic episodes, the predominant pathogens were enteric gram-negative rods such as Escherichia coli (19 isolates) and Klebsiella pneumoniae (15), alongside coagulase-negative staphylococci (16) and Pseudomonas aeruginosa (9). Autologous recipients had predominantly gram-negative bacteremia among non-CoNS isolates (44 of 54, or 81.5 percent) compared with allogeneic recipients (12 of 19, or 63.2 percent), a pattern consistent with conditioning-induced mucosal translocation. Fluoroquinolone prophylaxis was used in only 5 percent of episodes, which the authors suggest contributed to the high bacteremia rate and the predominance of susceptible organisms.
The authors subjected their null finding to extensive robustness testing. With 224 episodes and 17 deterioration events, the study had 80 percent power to detect a categorical odds ratio of 4.28 or a per-hour odds ratio of 1.41; detecting a per-hour odds ratio of 1.20 would have required approximately 795 episodes. The reverse fragility index for the primary categorical comparison was 12, exceeding the 11 deterioration events observed in the greater-than-one-hour group. Sensitivity analyses across finer timing increments, small-sample methods including Firth’s penalized regression and exact tests, an expanded composite outcome incorporating fluid boluses, a cohort-inclusion threshold sweep, and a subset restricted to the 145 episodes reaching a single temperature of at least 38.3 °C all supported the null result. By the 95 percent confidence interval, the upper bound on the per-hour time-to-antibiotics effect was an odds ratio of 1.24.
The authors are candid about the study’s limitations. The cohort was predominantly lower-risk: 156 of 224 episodes (69.6 percent) occurred in autologous transplant recipients, the median age-adjusted Charlson comorbidity score was 1, and 90-day mortality was 4.9 percent. Single-center prophylaxis and antimicrobial practices further limit generalizability to centers with predominantly allogeneic transplantation, haploidentical or mismatched donors, myeloablative conditioning for high-risk disease, or a higher prevalence of multidrug-resistant organisms. The limited number of outcome events constrained statistical power and precluded organism-stratified analyses, unmeasured confounding remains possible, and the observational design cannot establish causation. The authors describe the results as hypothesis-generating and point to planned multicenter validation using national electronic health record data.
The findings carry practical implications for inpatient transplant units, though the authors are careful about their boundaries. Modest variation in antibiotic timing — the cohort median was 1.2 hours — did not meaningfully influence outcomes in this continuously monitored inpatient population, potentially allowing clinicians additional time for clinical assessment, culture collection, and deliberate antibiotic selection. The authors explicitly caution that the results should not be interpreted as supporting intentional delays in antibiotic administration, nor extrapolated to settings with prolonged treatment delays; patients presenting with hemodynamic instability should continue to receive antibiotics with maximum urgency per current best practice. The 2010 IDSA guideline’s endorsement of prompt empiric therapy remains intact, and the 2021 Surviving Sepsis Campaign guidelines themselves retreated from a strict one-hour threshold for sepsis without shock, recommending an extended three-hour window. What this study adds is a direct test, in the specific population of early post-transplant adult neutropenic fever, of a threshold that had previously been imported without dedicated outcome evidence at that exact cutoff.
Subject of Research: Time-to-antibiotics and clinical outcomes in neutropenic fever after hematopoietic stem cell transplantation
Article Title: Association between time-to-antibiotics and clinical outcomes in neutropenic fever after hematopoietic stem cell transplant
Article References: Windham, S. L., Guidry, C. A., Sung, A. D., Kollef, M. H., & Simpson, S. Q. (2026). Association between time-to-antibiotics and clinical outcomes in neutropenic fever after hematopoietic stem cell transplant. Supportive Care in Cancer, 34(11), Article 1162. https://doi.org/10.1007/s00520-026-11141-8
Image Credits: AI Generated
DOI: 10.1007/s00520-026-11141-8
Keywords: neutropenic fever, hematopoietic stem cell transplant, time-to-antibiotics, sepsis guidelines, bacteremia, clinical deterioration, HSCT, antibiotic timing, febrile neutropenia, retrospective cohort study, mortality, transplant complications
News Source: Drew Townsend. (October 28, 2026). One-Hour Antibiotic Rule Questioned for Neutropenic Fever After Stem Cell Transplant. Scienmag.
