Evaluation of febrile infants under 3 months of age: is routine lumbar puncture warranted?

Fever may be the sole clinically evident presentation of serious bacterial infection (SBI) in a very young infant, and therefore lumbar puncture is still widely regarded as a mandatory procedure in the sepsis workup of febrile infants under 2 months of age. In this retrospective study, we evaluated the frequency and the diagnostic value of cerebrospinal fluid examination in 492 infants aged less than 3 months who were hospitalized because of fever during 1988-1994. The patients were categorized as being at "high risk" or "low risk" for SBI according to current clinical and laboratory criteria. Among the 492 infants, 196 (40%) were identified as "high-risk" for SBI, and 296 (60%) were at low risk. Among the overall series of infants, 60 babies (12%) were subsequently proven with bacterial infection. Among the 196 "high-risk" babies, 26% had bacterial infection, compared to only 3% of the 296 infants at low risk (p < 0.0001), denoting a sensitivity of 85% and a specificity of 65% of the clinical classification criteria. Lumbar puncture was done to 186 (46%) infants upon hospital admission; 176 punctures yielded satisfactory samples of cerebrospinal fluid (CSF). Sixteen (3%) patients had abnormal CSF findings: 2 of them had positive bacterial cultures and 14 were compatible with aseptic meningitis. The 2 patients with purulent meningitis were clinically very ill and were immediately recognized as deserving a lumbar puncture. Of the 14 patients with aseptic meningitis, 13 were initially screened as being at high risk for serious infection, and therefore underwent a lumbar puncture. Over the years of this survey, a declining trend for performing lumbar puncture in "low-risk" young febrile infants became evident: during 1988-1992, evaluation of sepsis included a lumbar puncture in 45% of the infants, compared to 27% during the following 2 years (p < 0.0001). Not one instance of purulent meningitis evolved among the infants in whom lumbar puncture was not performed. Our observations suggest that hospitalized young febrile infants may safely be spared a lumbar puncture when they do not meet the proposed criteria for being at high risk, or when their clinical and laboratory picture suggests being at low risk for SBI.