Validación de criterios para bajo riesgo de infección bacteriana grave en recién nacidos febriles

criteria for prediction of low risk for serious bacterial infection (sbi) in febrile infants are imperfect, particularly in infants younger than 1 month. we sought to validate new criteria for evaluation of febrile neonates. to this end, 288 febrile newborns with complete evaluation for sepsis were prospectively classified as low-risk for infection if they looked well; had been previously healthy; had no signs of focal infection; did not have persistent, recurrent, or high fever 39.0c); and leukocyte counts of 520 x 109/liter, erythrocyte sedimentation rate <20 mm/hr; and <10,000 leukocytes/ml of urine. these criteria were compared to rochester criteria and false-positive and false-negative rates were compared by z-tests. sbi was present in 68 (23.6%) and bacteremia in 22 infants (7.6%). only 5 of 68 (one with bacteremia) of these infants were classified as low risk using the new criteria (false-negative rate=7.3%, 95% ci (4.711.7%) whereas 15 of 68 (3 with bacteremia) were classified as low risk using the rochester criteria (false-negative rate=22.0%, 95% ci (17.527.4%); p=0.03 vs. new criteria). the lowest false-negative rate was offset by a higher false-positive rate (65.9% vs. 37.7%). the new criteria identifying low risk for serious bacterial infection in febrile neonates had a significantly lower false-negative rate than that of the classic rochester criteria. these new criteria may facilitate adequate care of infants under 1 month of age who present with a history of fever.