MENINGITIS

I. Etiology
  1. Most common organisms in infants: Hemophilus influenzae type B, Meningococcus, Pneumococcus
  2. Aseptic (non-bacterial) meningitis--most common causative organism is Enterovirus (can check for it with PCR of CSF or PCR of stool)
  3. Mycobacterium tuberculosis (rare as a cause of meningitis in U.S.)

II. Clinical findings and diagnostic considerations

  1. Fever
  2. Altered mental status
  3. Meningismus, i.e. stiff neck (absent in 50% of children <18mos old with meningitis)
  4. Kernig's & Brudzinski's sign were both highly insensitive but highly (95%) specific in a study of 297 adults presenting to an ER with suspected meningitis (Clin. Inf. Dis. 35:46, 2002--JW)
  5. In a retrospective study in 167 children hospitalized for meningitis (21 of which turned out to have bacterial meningitis), evaluating sensitivity/specificity of a number of measures for predicting eventual dx of bacterial meningitis, highest sensitivity was for CRP > 20mg/L (91%) and procalcitonin > 0.5ng/mL (89%), or CSF protein > 0.5g/L (86%). (CRP > 20mg/L OR procalcitonin > 0.5ng/mL) was 100% sensitivity for predicting bacterial meningitis (J. Peds. 149:72, 2006--JW)
  6. CSF for WBC count
    1. Analysis of CSF for WBCs is a common diagnostic modality for meningitis (typical cutoff is 10,000,000/L).
    2. Bands-Neither absolute count of "band" cells in CSF nor the proportion of CSF WBC's that were independently predictive of an eventual meningitis diagnosis in a retrospective study of 1,116 pediatric patients with suspected meningitis.  (Pediatrics 123:e967, 2009-JW)
  7. CSF culture
  8. CSF PCR for enterovirus--Negative predictive value 97% in clinical trials.
  9. Clinical decision rules for diagnosis of meningitis
    1. The Bacterial Meningitis Score
      1. Classifies risk of bacterial meningitis as "very low" for pts with CSF pleocytosis (CSF WBC >= 10 cells/microliter) if none of the following are present: positive CSF gram stain, CSF absolute neutrophil count >= 1000 cells/microliter, CSF protein >= 80 mg/dL, peripheral blood absolute neutrophil count >= 10,000 cells/microliter, or history of seizure before or at time or presentation.
      2. In a retrospective cohort study in 3,295 children 29d-19yo presenting to an ED with CSF pleocytosis who had not received antibiotic treatment before LP, a Bacterial Meningitis Score of "very low" risk had sensitivity of 98.3 and negative predictive value of 99.9% for eventual diagnosis of bacterial meningitis.  Both false negatives were in pts < 2mos old (JAMA 297:52, 2007--abst)

III. Treatment

  1. Antibiotics
  2. Corticosteroids
    1. In a meta-analysis of 11 randomized controlled trials of dexamethasone in children with meningitis , for H. influenzae meningitis, dexamethasone was associated with a significantly reduced risk for sensorineural hearing loss. Ditto for pneumococcal meningitis but only if begun before or with 1st dose of antibiotics. For all organisms combined, there was a nonsig. reduction in neurol. deficits other than hearing loss w/dexamethasone.  2d seemed as effective as 4d course. No sig. incr. in adverse f/x with dex, though GI bleeds increased w/4d as opposed to 2d therapy (3% vs. 0.8%). (JAMA 278:925, 1997-JW)
    2. In a randomized trial of 301 adults with suspected bacterial meningitis randomized to dexamethasone 10mg IV Q6h x 4d, starting 15-20min before first dose of abx, vs. placebo, 8wk incidence of "unfavorable outcomes" was sig. lower in dex recipients (15% vs. 25%); total mortality was also lower (7% vs. 15%).  In the subgroup who turned out to have meningococcal meningitis, there was no benefit with dex.  Note that steroids may limit penetration into CSF of vancomycin if that is being used for tx (NEJM 347:1549, 2002--JW)
  3. Glycerol (unclear mechanism; purported to work by scavenging free radicals)
    1. In study in 640 children > 2mo with bacterial meningitis randomized to glycerol 1.5g/kg PO QID x 2d, dexamethasone 0.15mg/kg IV Q6h x 2d, both, or double-placebo; all also received IV ceftriaxone; glycerol group c/w placebo group had sig. lower incidence of in-hospital death (RR 0.58); there was no sig. diff. between dexamethasone and placebo groups (reported in FP News 2/15/05)