Risk-stratifying Infants 29-90 Days Old with Fever and Symptoms of Viral Infections for Serious Bacterial Infections with or Without Bacterial Meningitis

Background: Neonates <28 days old with fever are at high risk of having serious bacterial infections (SBIs). Many criteria exist to risk-stratify febrile infants 29-90 days old into high and low risk for SBIs. Recent data in the post-vaccination era show that concomitant viral symptoms reduce the risk of co-existing SBIs in febrile infants 29-90 days old even further; none have meningitis, though still at risk of urinary tract infections (UTI). No consensus exists to manage febrile infants 29-90 days old with viral symptoms though recent trends favour less aggressive approaches. A clinical conundrum arises to balance the risks of SBI with invasive diagnostic procedures like lumbar punctures (LP), hospitalisations and unnecessary antibiotic use.

Methods: We compared data from two groups of 29-90 day old febrile infants presenting to CE at two different 6-month time-points. Infants in Group A followed our previous modified Rochester criteria and underwent partial septic work-up (blood and urine tests) for risk-stratification into high or low-risk for bacterial meningitis. Infants in Group B followed an updated version where only urine was screened in vaccinated infants with viral symptoms (Figure 1). High-risk infants for both groups were admitted for LP and empiric intravenous antibiotics. Next-day follow-ups were given to all non-admitted low-risk infants.

Admission rates, LP and blood test rates, readmission rates for missed bacterial meningitis were monitored. Subgroup analysis was performed for febrile infants with viral URTI symptoms.

Results: SBI rates were 4.5% (2/44) in Group A’s low-risk infants compared to 7.1% (5/70) in Group B, p=0.573. Comparing Group A to B, LP rates reduced from 43.5% (37/85) to 33.3% (38/114); p=0.142, admission and blood test rates significantly reduced from 64.7% (55/85) to 47.3% (54/114); p=0.015. There were 2 cases of missed invasive SBIs (both bacteraemia) after initial screening of 44 low-risk infants in Group A (4.5%) compared to 2 (1 bacteraemia, 1 bacterial meningitis) in 70 low-risk infants in Group B (2.9%); p=0.633. Close outpatient follow-ups provided an effective safety net in picking these infants up.

In a subgroup analysis of febrile infants with viral URTI symptoms in Group B, 18% (6/34) underwent LP as they were admitted for treatment of UTIs; none of them had meningitis.

Conclusion: Our updated version of modified Rochester criteria was effective in significantly reducing rates of admissions and blood tests in febrile infants 29-90 days old. Patient safety was not compromised as there was no increase in rates of missed invasive SBIs. A subgroup analysis of febrile infants with viral URTI symptoms and concomitant UTI did not have meningitis. Limitations of the study are small sample size and no strict guideline adherence. Such guidelines are effective in small populations like Singapore where only few patients are lost to follow-up.