Introduction
Bacteremia occurs in up to 23–29% of febrile neutropenia (FN) episodes. Reference Klastersky, Ameye and Maertens1,Reference Viscoli, Cometta and Kern2 The National Comprehensive Cancer Network, Infectious Diseases Society of America, and American Society for Microbiology recommend obtaining at least two and up to four sets of blood cultures from different anatomic sites at the onset of FN. Reference Freifeld, Bow and Sepkowitz3,Reference Baden, Swaminathan and Almyroudis4 The collection of two sets, instead of one, is associated with an increased sensitivity for detecting bloodstream infections. Reference Tarai, Jain, Das and Budhiraja5 It also allows better discrimination between true pathogens and contaminants, thereby improving the diagnostic yield, which is critical in guiding appropriate antimicrobial therapy in immunocompromised patients. Reference Mermel, Allon and Bouza6
A recent manufacturing interruption of blood culture bottles compatible with the BACTEC system, which began in June 2024, led to the implementation of conservative measures, including an institutional restriction on the number and frequency of blood culture collections. We herein sought to determine the impact of this shortage on the blood culture yield for FN, using real-world data.
Methods
This is a single-center, retrospective, observational study evaluating adult patients with neutropenic fever who had blood cultures collected within the Oklahoma University Medical Center, from April 1, 2024, through November 30, 2024. An episode of FN was defined as a single oral temperature greater than or equal to 101°F (38.3°C) or a temperature greater than or equal to 100.4°F (38°C) sustained over a 1-hour period and an absolute neutrophil count (ANC) < 0.5*103/uL. Reference Freifeld, Bow and Sepkowitz3 Febrile episodes during the study period that were separated by at least 7 days were considered as distinct episodes. Blood cultures collected within 7 days after the index blood culture was obtained were considered follow-up blood cultures. Episodes occurring between April 1, 2024, and July 31, 2024, were labeled to have occurred during the “Liberal period.” During that period, there were no restrictions on blood culture orders by number or frequency. Episodes occurring between August 1, 2024, and November 30, 2024, were labeled to have occurred during the “Restriction period.” In response to the blood culture shortage that occurred during that period, the hospital system implemented changes to conserve the limited supply, which came into effect on August 1, 2024. Hard stops on blood culture ordering in the electronic medical record were placed to limit ordering to one blood culture set, including an aerobic and anaerobic bottle, every 48 hours per patient. No restrictions were placed on the volume of collected blood per bottle. Exceptions were allowed per clinical scenario as assessed and approved by on-call Infectious Diseases specialists. Basic demographic and clinical data were collected. The primary outcome evaluated was the prevalence of bacteremia in patients with FN between the Liberal period and the Restriction period. The secondary outcome was the diagnostic yield of blood cultures depending on the number of blood culture sets obtained, using a previously described definition. Reference Tarai, Jain, Das and Budhiraja5 Episodes resulting in a single positive set blood culture for commensal microbes (coagulase-negative Staphylococcus, Gemella spp., Corynebacterium spp., Rothia spp.) were considered contaminants. Chi-square and Mann–Whitney U tests were used to compare groups as appropriate. Statistical analysis was performed on IBM SPSS version 29.0.0.0.
Results
A total of 236 episodes of FN were identified during the study period (134 episodes during the Liberal period, and 102 episodes during the Restricted period), involving a total of 173 patients (Table S1). The median temperature, ANC, and presence of central venous catheter were comparable between the two groups. There was no significant difference in the rate of detected bacteremia on the initial blood culture between the Liberal and Restricted periods (30.6% vs 24.5%, P = 0.3). Follow-up blood cultures were more often obtained during the Liberal period (57.5% vs 52.0%) with similar rates of positivity (15.6% vs 9.4%, P = 0.2). Only 6.5% (Liberal period) and 7.5% (Restriction period) of follow-up blood cultures detected a different organism than the initial culture (Table 1). Higher ANC was associated with significantly lower rates of microbial growth on univariate and multivariate analysis (odds ratio 0.004 for every 1k/µL increase in ANC, 95% confidence interval: 0–0.09, P < 0.001). No difference in microbial growth was noted when comparing blood cultures obtained during Restriction and Liberal periods (Table S2).
Table 1. Characteristics of Neutropenic Fever Episodes and Blood Culture Results
When stratified by number of initial blood culture sets obtained, the diagnostic yield of blood culture for bacteremia increased with the number of blood culture sets obtained without reaching statistical significance (19.6%, 25.8%, 30.4%, 46.6% for 1, 2, 3, 4 sets of blood cultures, respectively, P = 0.12), including when the yield of one set was compared to the yield of two sets (P = 0.29) (Table 2).
Table 2. Diagnostic Yields of Initial Blood Cultures Stratified by Number of Sets
Discussion
Our study found no significant difference in the rate of microbial growth on initial blood cultures for patients with FN between the Liberal and Restricted periods. We also noted a trend of increasing diagnostic yield of blood cultures with more sets of blood at the onset of a febrile neutropenic episode, although not statistically significant. Anaerobes, coagulase-negative Staphylococcus, and Escherichia coli appear to have a lower growth rate on initial cultures during the Restriction period.
Interestingly, these results are not concordant with the literature evaluating cohorts of non-neutropenic patients, where higher positivity rates were observed with paired-set blood cultures compared with a single set. Reference Elantamilan, Lyngdoh and Khyriem7 Their study also noted a higher rate of Gram-negative growth and a lower rate of Gram-positive growth on cultures, when compared with our study. When studying the need for peripheral sticks in 241 pediatric patients with FN and bacteremia, the authors found that 15% of the positive blood cultures originated from the central line, while 24% came from the peripheral site alone, suggesting a risk of missed diagnoses if only a single culture were obtained. Reference Sarquis, Ibáñez and De La Maza8
Repeat blood culture yield was low in both of our cohorts. Despite that, fewer repeat blood culture sets with longer intervals were used in the Restricted period. Serody et al. showed that among hematopoietic transplant recipients with FN, the prevalence of positive blood culture after day one of fever was low at 4.6%, and most repeat cultures demonstrated the same organism as the initial cultures. Reference Serody, Berrey and Albritton9 Such findings have been reproduced by other studies. Reference Sheu, Molina Garcia and Shrivastava10
Based on our study findings and evidence from the literature, our institute is continuing the policy limiting repeat blood cultures with intervals less than 48 hours, while resuming two sets of blood cultures after the shortage ended.
Our analysis has some limitations. Being retrospective allows for potential biases. The relatively small number of episodes may have contributed to the limited power of our study to find statistical significance among the compared groups. Additionally, we did not have data on total blood volume collected, which is known to affect the diagnostic yield of blood cultures.
In conclusion, this blood culture bottle shortage inadvertently created opportunities for diagnostic stewardship in patients with FN, particularly in limiting the use of low-yield follow-up blood cultures. Future studies with a larger sample size and controlled designs with a focus on neutropenic patients are needed to determine the optimal number of blood culture bottles in FN.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2025.10310.
Financial support
This study received no financial support.
Competing interests
J.S. reports research grants paid to his institution from Ansun BioPharma, Community Infusion Solutions, Eurofins Viracor, F2G, Mundipharma, Pulmotect, and Shionogi. All other authors report no conflicts of interest.
Open access
