Taking the Trauma Out of Traumatic Spinal Taps

Abstract & Commentary

Source: Mazor SS, et al. Interpretation of traumatic lumbar punctures: Who can go home? Pediatrics 2003;111:525-528.

When the pediatric lumbar puncture (LP) is traumatic, confusion arises as to whether cerebrospinal fluid (CSF) white blood cells (WBCs) are from infection or from contaminating blood. To better identify patients with and without CSF pathogens, Mazor and colleagues studied 57 children ages 1 month to 13 years who had traumatic LPs (> 500 red blood cells [RBCs]/mm3) during evaluation for suspected meningitis. All were seen from 1990-1999 at Children’s Memorial Hospital in urban Chicago. Forty-five cases had negative CSF cultures, while 12 (21%) had bacterial pathogens: Neisseria meningitidis (four patients), Haemophilus influenzae type B (three patients), Streptococcus pneumoniae (two patients), Group B streptococcus (two patients), and nontypable H. influenzae (one patient). Children were excluded if they received antibiotics in the prior 72 hours, or had a recent neurosurgical procedure.

For each LP, the Observed:Predicted (O:P) ratio for CSF WBCs was determined by dividing observed (O) by predicted (P) WBCs in the CSF. Predicted CSF WBC was calculated by the following formula: (P) = CSF RBC ´ (blood WBC/blood RBC). Simple ratios of CSF WBCs to RBCs also were determined for each LP.

O:P ratios in children without meningitis were quite low, with a median of 0.064 (range 0.0000054-1.09) as compared to O:P ratios in children with culture-positive CSF (median 1.26; range 0.045-4.72) (p < 0.001). All but one of the 45 meningitis-free children had O:P less than 1.0, while 7 of 12 cases with bacterial pathogens had O:P greater than 1.0, and 5 of 12 had ratios greater than 2.0. In patients without meningitis, the CSF WBC: RBC ratio was far smaller (median 0.001, range 0-4.46) than in cases with positive CSF cultures, in whom median CSF WBC:RBC ratio was 1.98 (range 0.04-24.45) (p < 0.001). For the exclusion of culture-positive meningitis (i.e., proof of no meningitis), the sensitivity, specificity, and positive predictive value of an O:P ratio less than 0.01 were 91.1%, 100%, and 100%, respectively. For a CSF WBC:RBC ratio of less than 1:100, sensitivity, specificity, and positive predictive value (in excluding meningitis) were 84.4%, 100%, and 100%, respectively. The authors conclude that an O:P ratio less than 0.01 and a WBC:RBCratio of less than 1:100 are highly specific for the absence of meningitis in ill children with a traumatic LP.

Commentary by Michael Felz, MD

How practical is this? The less cumbersome value tocalculate is the CSF WBC:RBC ratio. For example, in a febrile child with 10 CSF WBCs, concomitant findings of greater than 1000 CSF RBCs would yield a WBC:RBC ratio of less than 1:100 and speak against meningitis, based on the Mazor study. The calculation of the O:P ratio is more complex but still useful in that peripheral blood WBC and RBC counts are the correction factors for the observed number of WBCs in the traumatic LP.

I found this study applicable in April 2003, when an unprecedented outbreak of enterovirus (Echovirus type 9) meningitis occurred in more than 25 pediatric patients in our urban area. Two examples are illustrative. A 10-year-old girl had two days of 102°F fever, headache, and stiff neck. CSF WBCs were 20/mm3 with RBCs 345, for a WBC:RBC ratio of 0.07. The O:P ratio was 67. Both ratios suggested the presence of meningitis based on the Mazor criteria, although this case involved a viral, not bacterial, pathogen. The second patient was a 9-year-old girl with one day of headache, vomiting, malaise, fever of 101°F, and nuchal rigidity. CSF WBCs were 570/mm3 with RBCs 8, for a ratio of 0.70. O:P ratio was quite high at 57,000. Here again, the ratios reliably predicted the presence of CSF infection, although the etiology was viral. Both children recovered rapidly with conservative management.

The greatest strengh of the Mazor study is in guiding clinicians who must reconcile the appearance of a sick child with laboratory values from a traumatic LP. When the O:P and WBC:RBC ratios are less than 1:100, bacterial meningitis (and perhaps viral—although this was not addressed in this study) is highly unlikely.

Dr. Felz, Associate Professor, Department of Family Medicine, Medical College of Georgia, Augusta, GA, is on the editorial board of Emergency Medicine Alert.