Clinical characteristics of febrile young infants with serious bacterial infection in the emergency departments in Korea and validation of Philadelphia criteria and modified Philadelphia criteria: retrospective multicenter study

ABSTRACT

OBJECTIVE: This study examined the characteristics of febrile young infants with a serious bacterial infection (SBI) who visited emergency centers in Korea and validated the Philadelphia criteria and modified Philadelphia criteria to predict the risk of SBI. METHODS: This was a retrospective study conducted on 450 infants aged 31 days to 56 days who visited three emergency centers with fever from September 2014 to August 2017. The characteristics of the SBI patients were analyzed, and the validation of the Philadelphia and modified Philadelphia criteria sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were checked. RESULTS: Of 450 patients, 165 patients (36.7%) had SBI, such as urinary tract infection (33.3%), bacteremia (4.0%), acute osteomyelitis (0.2%), and bacterial meningitis (BM) in two patients (0.4%). The most common pathogen of invasive bacterial infection was Escherichia coli. In the Philadelphia criteria, the sensitivity, specificity, PPV, NPV, and accuracy were 73.9%, 47.0%, 44.7%, 75.7%, and 56.9%, respectively. In the modified Philadelphia criteria that excluded lumbar puncture as a predictor, the sensitivity, specificity, PPV, NPV, and accuracy were 93.3%, 31.9%, 44.3%, 89.2%, and 54.4%, respectively. The most common failed low risk criteria was appearance (43.3%). Two patients with bacterial meningitis were excluded from low risk group by the modified Philadelphia criteria. Although one out of 2 patients met the failed low risk criteria due to their poor condition, this factor is not objective, so BM can be missed. CONCLUSION: Bacterial meningitis was too rare in this study. New criteria are needed to predict SBI. The Philadelphia and modified Philadelphia criteria were not useful for predicting SBI in this study. Other prediction models will be needed to predict SBI in the vaccination era.