Greenberg RG, Smith PB, Cotton CM, Moody MA, Clark RH, Benjamin DK. Traumatic lumbar punctures in neonates: test performance of the CSF WBC. Pediatr Infect Disease J 2008;27:1047-51.
Question
Among neonates with suspected meningitis, does the cerebrospinal fluid (CSF) white blood cell count need to be adjusted for a traumatic lumbar puncture?
Design
Cohort study performed between 1997 and 2004.
Setting
One hundred fifty neonatal intensive care units in the United States.
Participants
A total of 6374 neonates ≤30 days who underwent a lumbar puncture.
Intervention
CSF white blood cell counts were adjusted downward for traumatic lumbar punctures (defined as CSF specimens with ≥500 red blood cells/mm3) with several commonly used methods.
Outcomes
Sensitivity, specificity, likelihood ratios, and area under the receiver operating characteristic curve of unadjusted and adjusted CSF white blood cell counts for predicting culture or Gram stain–positive meningitis in neonates with traumatic lumbar punctures.
Main Results
Of the lumbar punctures, 2519 (39.5%) were traumatic. One hundred fourteen (1.8%) were culture- or Gram stain–positive for meningitis; 50 neonates with traumatic lumbar punctures had meningitis. The areas under the receiver operating characteristic curve for white blood cell count unadjusted and adjusted by all methods were similar. Adjusting CSF white blood cell counts to account for increased red blood cells did not improve diagnostic utility and resulted in decreased sensitivity with marginal gain in specificity.
Conclusions
Adjustment of white blood cell counts in the setting of a traumatic lumbar puncture does not aid in the diagnosis of culture- or Gram stain–proven bacterial and fungal meningitis in neonates.
Commentary
As acknowledged by the authors, incomplete data for recent exposure to antibiotics (transplacental or neonatal) precluded their ability to assess the potential importance of that factor. Use of positive culture or Gram stain of CSF as the “gold standard” for bacterial or fungal meningitis limited confounding by false-negative culture results because of prior antibiotic therapy, but the high frequency of pretreatment in many nurseries would make analysis of CSF from such subjects particularly useful. The lack of complete data regarding intracranial hemorrhage similarly precluded evaluation of the contribution of consequent arachnoiditis to elevated CSF white blood cell counts in the absence of positive culture or Gram stain. The conundrum posed by coagulase-negative staphylococci as potential contaminant versus true pathogen is recognized and considered separately. Clinicians should recall that CSF with white blood cell counts below the diagnostic threshold does not preclude central nervous system infection. CSF obtained very early after meningeal invasion may reflect insufficient time for generation of an inflammatory cellular response or development of adequate pathogen burden for positive Gram stain. Additionally, slow transport of CSF to the laboratory for analysis may be accompanied by loss of leukocyte viability and a drop in the apparent number of white blood cells.1 Brain abscess and other parenchymal or vascular infectious foci that are not adjacent to CSF-containing spaces may also produce no evident CSF abnormalities. When to perform lumbar puncture in sepsis evaluation remains controversial, although the high frequency of positive CSF culture with negative blood culture in very low birth weight infants >3 days old suggests liberal use of lumbar puncture in such infants.2 These authors present a convincing argument that when CSF is obtained, correction of white blood cell count for excess red blood cells (≥500/mm3) is unnecessary and potentially misleading. An elevated CSF white blood cell count should prompt serious consideration of selection, dosage, and duration of antibiotic therapy appropriate for meningitis even when culture is unrewarding.
2. 2Stoll BJ, Hansen N, Fanaroff AA, Wright LL, Carlo WA, Ehrenkranz RA, et al.To tap or not to tap: high likelihood of meningitis without sepsis among very low birth weight infants. Pediatrics. 2004;113:1181–1186.
FULL TEXT