Steroids in Pediatric Meningitis: New Large Study Fails to Show Benefit
(Editor’s note: This Journal Club recap is brought to you by Seth, and covers the 10/1 journal club that Raakhee presented. The paper in question is Mongelluzzo J, et al. Corticosteroids and mortality in children with bacterial meningitis. JAMA 2008 May 7;299(17):2048-55. PMID: 1846066)
Corticosteroids have been shown to decrease mortality in the treatment of adults with bacterial meningitis — particularly pneumococcal meningitis — but the data on children are less clear. Although steroid therapy reduces hearing loss in Hib meningitis, vaccination has made Hib (and pneumococcus) much less prevalent in developed nations, and it is unclear whether steroids lower mortality in children with meningitis. This is the question a group of CHoP researchers set out to settle. However, much like the earlier literature, this paper failed to show any benefit from corticosteroid therapy.
The theoretical mechanism for steroid therapy is plausible enough — antibiotics lead to bacteriolysis, which leads to inflammation and cerebral edema. Corticosteroids control this response, but could also lead to decreased CSF penetration of antibiotics; GIB and other direct adverse effects could also result. Finally, there is the fear that steroids would mask a secondary fever if antibiotics failed.
The researchers’ methodology had its strengths and weaknesses. They used a retrospective cohort (children <18 years with bacterial meningitis), as a prospective double-blinded RCT would be difficult due to the relatively low prevalence of meningitis, and the risk associated with randomizing very sick children to treatment groups. The cohort was obtained from the Pediatric Health Information System (PHIS), a network of 27 tertiary care children’s hospitals in 18 states and DC, providing a strong multicenter patient and provider base.
However, study participants were identified through ICD-9 primary discharge code of meningitis, which can be problematic and was likely too narrow. It is conceivable that a large subset of meningitis patients are primarily coded as “fever,” “sepsis,” etc. Also, does this exclude patients who died during admission, a presumably important subset?
Furthermore, the while the usage of corticosteroids in adults has been shown to work, it is administered either 20 minutes prior to antibiotics, or with the first dose. The AAP’s current recommendation for children is similar, if the provider decides to give steroids.
However, in this paper, patients were given steroids at any point within the first 24 hours of hospitalization. As the authors used a fairly advanced statistical analysis (propensity scores), which required both a PhD and a very dense “Methods” section. The first analysis showed no difference when adjusted for propensity scores, except that the sickest patients were much more likely to have been given steroids, and the least sick were much less likely. Resultantly, they repeated the analysis, excluding the sickest and healthiest quintiles. The end result was again an insignificant difference, both in hospital length of stay and in mortality.
The main outcome measures: mortality rates of 6% (15/248) for the corticosteroid group, and 4% (102/2532); a relative risk of 1.5, but the confidence interval “crosses unity” (0.89-2.54). Very nice graphs also demonstrate that mortality and LOS follow essentially the same curves. Therefore, either there is no true difference in outcome, or the sample size (117 deaths among 2780 patients) was too small to demonstrate a difference.
Posted
on Tuesday, October 14th, 2008 at 12:12 am by Nick. Filed under
Meningitis, Infectious Disease, Journal Club, Blog.
You may post a comment.
An excellent review of an interesting article. One point to consider: while steroids did not show an improvement in “mortality” in children, they have been shown to reduce “morbidity”.
In a 2007 Cochrane review (reference below) the authors concluded that steroid use in meningitic children in “high income” countries reduces hearing loss and short term neurologic sequelae. No benefit or harm was found for children in low income countries.
Now, part of this prevention of hearing loss is related to H.Flu meningitis which is presumably decreasing due to immunization. But we don’t have hard data on the new frequency of this disease, and surely it does still exist to some degree. The Cochrane authors recommend use of steroids in meningitic children due to the possible benefits and lack of harm (no study has shown worse outcomes in these children, although there is “theoretical” risk).
None of the studies involve children under one month of age, so there is no evidence in this group.
So the jury is still out. I think it is still worth considering steroids in bacterial meningitic children, especially unimmunized, but can’t say 100% that it will truly help, and remember, it must be given early to be useful. I think this holds even more true in your septic patients that you might consider giving steroids to support their sepsis resuscitation.
The Cochrane review discusses both adults and children and it definately worth a read if you want to understand the evidence on the topic.
van de Beek D, de Gans J, McIntyre P, Prasad K. Corticosteroids for acute bacterial meningitis. Cochrane Database of Systematic Reviews2007, Issue 1. Art. No.: CD004405. DOI: 10.1002/14651858.CD004405.pub2
Comment by Chris Strother on October 21st, 2008 at 10:02 am