Fellow: Alison Carroll, MD. Vanderbilt University Medical Center, Monroe Carell Jr. Children’s Hospital at Vanderbilt
Article: Lyons, TW, Garro, AC, Cruz AT, et al. Performance of the Modified Boston and Philadelphia Criteria for Invasive Bacterial Infections. Pediatrics. 2020;145: 1-9. https://doi.org/10.1542/peds.2019-3538.
Summary: The ability of the Boston and Philadelphia criteria to identify infants at low risk for a serious bacterial infection (i.e., urinary tract infection, bacteremia or bacterial meningitis) has not been recently reevaluated. This multicenter retrospective cohort study of infants age 29 to 60 days from 23 tertiary care emergency departments in the US and Canada sought to measure the diagnostic accuracy of the Boston and Philadelphia criteria to identify infants at low risk for invasive bacterial infection (IBI), defined as bacterial meningitis or bacteremia.
A total of 10,928 infants 29 to 60 days old with blood and CSF cultures were included. The Boston criteria was applied to 8344 infants and the Philadelphia criteria to 8131 infants. Overall, 264 (2.4%) of the infants had an invasive bacterial infection. Of the 264 infants with an invasive bacterial infection, 193 had bacteremia and 71 had bacterial meningitis. The Boston criteria identified 133 of the 212 infants with IBI (sensitivity 62.7%; specificity 59.2%) and the Philadelphia criteria identified 157 of the 219 infants with IBI (sensitivity 71.7%; specificity 46.1%). In other words, the Boston criteria misclassified 37% of infants as low risk and the Philadelphia criteria misclassified 28% of infants as low risk. Among the 79 infants misclassified as low risk by the Boston Criteria 17 had bacterial meningitis. Among the 62 infants misclassified as low risk by the Philadelphia criteria 13 had bacterial meningitis.
Key Strengths: This study included a large multinational cohort of infants, increasing the generalizability of the results. The rates of serious bacterial infections found in this study were similar to the rates found in the original derivation studies for the Boston and Philadelphia criteria. The study highlights the changing microbiology of invasive infections in this age group; the top three pathogens identified were GBS, E. coli, and S. aureus.
Limitations: Since the Boston and Philadelphia criteria both require CSF results for application, this study only included infants who already had CSF values. Clinicians who already obtained CSF may have had a higher clinical concern for meningitis, leading the Boston and Philadelphia criteria to be applied to a different (and potentially sicker) patient population from which they were originally derived, creating the potential for selection bias. In addition, the authors were not able to determine infants’ clinical appearance, comorbidities, or indication for undergoing an evaluation. The authors attempted to mitigate this limitation by excluding infants admitted to the ICU as a proxy for critical illness.
Main Takeaway: The Boston and Philadelphia criteria misclassified about 1/3 of infants with an invasive bacterial infection as low risk including a substantial number with bacterial meningitis. This highlights the limitations of these criteria and the need to incorporate new evidence into our clinical practice guidelines for the management of these patients.
Impact on Practice: Reliance on these criteria may result in infants receiving more invasive testing (i.e., lumbar punctures) and hospitalizations, while also potentially missing up to 1/3 of infants with a serious bacterial infection. The incorporation of newer biomarkers such as procalcitonin, C-reactive protein, and multiplex bacterial polymerase chain reaction panels may be useful for developing new risk stratification tools that can rapidly and accurately identify febrile infants in order to reduce unnecessary lumbar punctures, hospitalizations, and exposure to antibiotics.
Article: Lyons, TW, Garro, AC, Cruz AT, et al. Performance of the Modified Boston and Philadelphia Criteria for Invasive Bacterial Infections. Pediatrics. 2020;145: 1-9. https://doi.org/10.1542/peds.2019-3538.
Summary: The ability of the Boston and Philadelphia criteria to identify infants at low risk for a serious bacterial infection (i.e., urinary tract infection, bacteremia or bacterial meningitis) has not been recently reevaluated. This multicenter retrospective cohort study of infants age 29 to 60 days from 23 tertiary care emergency departments in the US and Canada sought to measure the diagnostic accuracy of the Boston and Philadelphia criteria to identify infants at low risk for invasive bacterial infection (IBI), defined as bacterial meningitis or bacteremia.
A total of 10,928 infants 29 to 60 days old with blood and CSF cultures were included. The Boston criteria was applied to 8344 infants and the Philadelphia criteria to 8131 infants. Overall, 264 (2.4%) of the infants had an invasive bacterial infection. Of the 264 infants with an invasive bacterial infection, 193 had bacteremia and 71 had bacterial meningitis. The Boston criteria identified 133 of the 212 infants with IBI (sensitivity 62.7%; specificity 59.2%) and the Philadelphia criteria identified 157 of the 219 infants with IBI (sensitivity 71.7%; specificity 46.1%). In other words, the Boston criteria misclassified 37% of infants as low risk and the Philadelphia criteria misclassified 28% of infants as low risk. Among the 79 infants misclassified as low risk by the Boston Criteria 17 had bacterial meningitis. Among the 62 infants misclassified as low risk by the Philadelphia criteria 13 had bacterial meningitis.
Key Strengths: This study included a large multinational cohort of infants, increasing the generalizability of the results. The rates of serious bacterial infections found in this study were similar to the rates found in the original derivation studies for the Boston and Philadelphia criteria. The study highlights the changing microbiology of invasive infections in this age group; the top three pathogens identified were GBS, E. coli, and S. aureus.
Limitations: Since the Boston and Philadelphia criteria both require CSF results for application, this study only included infants who already had CSF values. Clinicians who already obtained CSF may have had a higher clinical concern for meningitis, leading the Boston and Philadelphia criteria to be applied to a different (and potentially sicker) patient population from which they were originally derived, creating the potential for selection bias. In addition, the authors were not able to determine infants’ clinical appearance, comorbidities, or indication for undergoing an evaluation. The authors attempted to mitigate this limitation by excluding infants admitted to the ICU as a proxy for critical illness.
Main Takeaway: The Boston and Philadelphia criteria misclassified about 1/3 of infants with an invasive bacterial infection as low risk including a substantial number with bacterial meningitis. This highlights the limitations of these criteria and the need to incorporate new evidence into our clinical practice guidelines for the management of these patients.
Impact on Practice: Reliance on these criteria may result in infants receiving more invasive testing (i.e., lumbar punctures) and hospitalizations, while also potentially missing up to 1/3 of infants with a serious bacterial infection. The incorporation of newer biomarkers such as procalcitonin, C-reactive protein, and multiplex bacterial polymerase chain reaction panels may be useful for developing new risk stratification tools that can rapidly and accurately identify febrile infants in order to reduce unnecessary lumbar punctures, hospitalizations, and exposure to antibiotics.