Serum cortisol levels in children with acute bacterial and aseptic meningitis.

OBJECTIVE: To study serum cortisol levels in acute childhood meningitis with respect to the severity of illness and the outcome. DESIGN: Prospective observational study. SETTING: Pediatric services of a tertiary care teaching and referral hospital. SUBJECTS: A total of 30 consecutive children, 2 months to 12 yrs of age, with suspected bacterial meningitis. METHODS: Serum cortisol levels (by enzyme-linked immunosorbent assay) obtained at admission were correlated with clinical characteristics (including Glasgow Coma Scale and Pediatric Risk of Mortality scores) recorded at admission and with neurologic and hearing status 2 months after discharge using SPSS 10.0. RESULTS: Mean +/- sd serum cortisol was 467 +/- 251 ng/dL in patients with bacterial (n = 16) and 319 +/- 159 ng/dL in aseptic meningitis (n = 14, p = .068). Glasgow coma scale score, systolic blood pressure, age, Pediatric Risk of Mortality, and cerebrospinal fluid protein were significant independent predictors of serum cortisol on stepwise multivariate regression analysis (each had an R change of >5%). Patients with neurologic or hearing sequelae had significantly higher median serum cortisol (450 ng/mL, n = 12) than those without sequelae (300 ng/mL, n = 17; p = .043 by Mann-Whitney U test). On multivariate logistic regression analysis, a serum cortisol of >/=420 ng/mL (odds ratio, 0.022; 95% confidence interval, 0.01-0.43) and systolic blood pressure (odds ratio, 1.35; 95% confidence interval, 1.04-1.74) were significant independent predictors of neurologic and hearing sequelae. CONCLUSION: Low serum cortisol is uncommon in acute bacterial meningitis of nonmeningococcal pathogenesis. Very high levels are likely to be associated with sequelae.