373 - Prevalence of bacteremia in febrile children with cancer: Is a peripheral culture necessary?

Poster Number: 373
Publication Number: 373.206

Mohamed Radwan, The Children's Hospital at Monmouth Medical Center, Long Branch, NJ, United States; Ola El Kebbi, American University of Beirut, Beirut, Beyrouth, Lebanon; Adonis Wazir, American University of Beirut Medical Center, London, England, United Kingdom; Dima A. Hamideh, american university of beirut, Hamra, Beyrouth, Lebanon; Sarah Abdul Nabi, American University of Beirut Medical Center, Beirut, Beyrouth, Lebanon; Maha Makki, American University of Beirut, Beirut, Beyrouth, Lebanon; Osama K. Ibrahim, Hasbro Children's Hospital at Rhode Island Hospital, providence, RI, United States; Hashem Nassereddine, American University of Beirut, Dearborn, MI, United States; Rasha D. Sawaya, American University of Beirut, Dublin, Dublin, Ireland

Background: Fever in children with cancer may be the only sign of a severe bacterial infection, including blood stream infections (BSI) that have a reported mortality up to 6%.  In addition, most patients have a central venous catheter (CVC), increasing the likelihood of a Central Line Associated Bloodstream Infections (CLABSI). Distinguishing CLABSI from peripheral BSI can be difficult and therefore drawing a peripheral blood culture (pBC) in addition to a central one remains controversial.

Objective: We aimed to report the prevalence of BSI in febrile children with cancer at our institution, the predictors of BSI and the utility of a pBC in addition to the CVC BC.

Design/Methods: This was a single center, retrospective cohort study of febrile children with cancer aged 1 month - 18 years presenting to the outpatient setting, from 01/01/2014 - 31/12/2018. Fever was defined as a single oral temperature ≥38.3°C or ≥38°C for longer than one hour or ≥38°C on 2 separate occasions over 12 hours. We only included patients presenting within 24 hours of the fever, and who had a pBC and/or CVC BC drawn at presentation. We excluded patients in remission or who had a transplant. To decrease repetitive bias, we only analyzed one randomly selected febrile visit per patient.
We compared the group with at least one positive blood culture to the group without. A p-value < 0.05 indicated statistical significance.

Results: We identified 664 patients, including 343 unique visits. 55% were male, with a mean age of 7 years ± 5. Fever was the sole symptom in 39% and 45% had severe neutropenia. 54% were admitted. See Table 1.
5.2% (18) had a positive BC. Of these, 2 were only peripheral, 12 only central, 3 both and one not specified. See Figure 1.
Negative predictors of BSI included chemotherapy phase (aOR=0.1 CI:[0.01 – 0.9]), and being older than 3 years (aOR=0.3 CI: [0.1-0.8]). On the other hand, patients having fever at presentation to the ED were more likely to have positive blood culture (aOR=1.9 CI:[1.0-3.6]).Conclusion(s): The prevalence of BSI in children with cancer at our institution was only 5.2%, with younger age, fever pattern and chemotherapy phase being risk factors. Further studies are needed to understand reasons behind this low incidence. In addition, as per our data, the value of a pBC did not outweigh the painful, recurrent, needle prick for these children. Perhaps a subset of patients who may benefit from a pBC can be determined in the future.
Table 1: Characteristics of included patients
Figure 1: Description of cases with only a peripheral blood culture positive