Background: To evaluate the rates of lumbar puncture (LP) in infants with culture-proven sepsis. Study design: We prospectively enrolled 400 infants with early- or late-onset sepsis due to Group B streptococcus (GBS) or Eschericha coli, diagnosed within 90 days of life. Rates of LP and potential variables associated with LP performance were evaluated. Moreover, cerebrospinal fluid (CSF) characteristics and results of the molecular analysis were investigated. Results: LP was performed in 228/400 (57.0%) infants; 123/228 LPs (53.9%) were performed after antibiotic initiation, hampering the ability to identify the pathogen in the CSF culture. However, polymerase chain reaction increased the probability of positive results of CSF analysis compared to microbiological culture (28/79, 35.4% vs. 14/79, 17.7%, p = 0.001). Severe clinical presentation and GBS infection were associated with higher LP rates. The rate of meningitis was 28.5% (65/228). Conclusions: Rates of LP are low in culture-proven neonatal sepsis and antibiotics are frequently given before LP is carried out. Thus meningitis may be underestimated, and the chances of giving an effective therapy to the newborn are reduced. LP should be performed before the start of antibiotics when there is a clinical suspicion of infection.
Keywords: Escherichia coli; group B streptococcus; lumbar puncture; meningitis; newborn; prematurity; sepsis.