377 - The utility of a blood culture in the evaluation of the febrile limping child
Saturday, April 23, 2022
3:30 PM – 6:00 PM US MT
Poster Number: 377 Publication Number: 377.206
Rachelle El Helou, Boston Children's Hospital, Boston, MA, United States; Amir Kimia, Boston Children's Hospital, Boston, MA, United States; Marvin B. Harper, Harvard Medical School, Boston, MA, United States; Natasha Joglekar, Boston Children's Hospital, Sandwich, MA, United States; Assaf Landschaft, Boston Children's Hospital, Bergisch Gladbach, Nordrhein-Westfalen, Germany
Clinical Fellow Boston Children's Hospital Boston Children's Hospital Boston, Massachusetts, United States
Background: Blood cultures (BC) are often obtained in the initial evaluation of children with atraumatic limp and fever. Little is known about the prevalence of bacteremia in this cohort. Routine testing comes with the risk of contaminants that are initially difficult to differentiate from pathogens in osteoarticular infections (OAI).
Objective: To describe the crude prevalence of bacteremia in children presenting to the emergency department (ED) with acute atraumatic limp and fever; and to identify predictors of bacteremia.
Design/Methods: A retrospective cross-sectional review of febrile patients age 1-18 years with acute atraumatic limp presenting to the ED of a tertiary children’s hospital from 2010-2019. Fever was defined as a reported or recorded temperature ≥38.0°C. We excluded patients who had orthopedic comorbidity, were immunocompromised or pretreated with antibiotics. We identified cases using a natural language processing (NLP) model trained and validated on 3698 records then applied to all 2010-2019 notes to identify records for manual chart review. We abstracted clinical, laboratory and radiographic features. We calculated prevalence with Bayesian credible intervals. Predictors of bacteremia were identified using univariate analyses by Fisher’s exact and Mann-Whitney U test followed by a multivariate logistical regression.
Results: The trained NLP model (sensitivity 96.2%, specificity 85.9%) was applied to 478,070 encounters identifying 17,981 for manual review. 689 met inclusion criteria. Median age was 5.3 years [IQR 2.7, 8.8] and 39.5% were female. BC was sent on 75.9% of patients. Patients were more likely to have a BC if they were non-weight bearing (OR 3.6), had focality on exam (OR 2.5) or fever ≥ 39oC (OR 1.5). The overall crude prevalence of bacteremia was 13.7% (95%CI 10.9-17.0%; 70/510), and 11.3% (95%CI 8.3-14.8%) in patients not referred from other sites. The prevalence of contaminants was 1.6% (95%CI 0.7-3.1%). Of 70 patients with bacteremia 72.9% had osteomyelitis, 5.7% had septic arthritis, and 4.3% had combined OAI. The most common pathogens were MSSA 71.6% and MRSA 15.7%. Lyme testing was performed on 61.2% of patients and was positive in 25.1% (95%CI 20.8-29.9%) of those tested. One patient had concomitant Lyme, bacteremia and osteomyelitis. Table 1 presents our patient characteristics and Table 2 provides a multivariate analysis of bacteremia predictors.Conclusion(s): The prevalence of bacteremia in febrile children presenting to the ED with acute atraumatic limp is high and supports the value of obtaining blood cultures during the initial evaluation even in a Lyme endemic region. Table 1: Patient characteristics Table 2: Multivariable logistical regression for Bacteremia predictors