Source: Smith KE, Wilker PR, Reiter PL, et al. Antibiotic treatment of Escherichia coli O157 infection and the risk of hemolytic uremic syndrome, Minnesota. Pediatr Infect Dis J. 2012;31(1):37-41; doi:10.1097/INF.0b013e31823096a8. See AAP Grand Rounds commentary by Dr. Robert Tolan, Jr. (subscription required).
Question: Among Minnesota residents <20 years old with Escherichia coli O157 infection, does antibiotic treatment increase the risk of progression to hemolytic uremic syndrome?
Question type: Intervention
Study design: Age-matched, case-control comparison
A primary care practitioner called me recently to ask about a 3 year old child with bloody diarrhea and a stool culture positive for Shiga toxin, the presumptive cause of hemolytic uremic syndrome (HUS). He wanted to know risks and benefits of antibiotic treatment in this situation. This study from the Minnesota Department of Health adds to our understanding of management issues, but it's by no means the last word on the subject.
This is not a typical case-control study; both groups consisted of children with E. coli O157 infection, and the comparison groups were actually those who received antibiotics versus those who did not. So, the authors (or perhaps the journal editors) have designated it a "case-case" comparison study. They found that HUS patients were more likely to have received bactericidal antibiotics within the first three days of diarrhea.
In the hierarchy of study design, which is simply a rating that points to the likelihood of the study results subsequently being shown to be wrong, case-control studies rank below randomized controlled trials. Prior cohort or case-control studies have found conflicting results for antibiotic treatment association with HUS.
At the moment, we must deal with less than optimal data on which to base a decision to use antibiotic treatment for a child with diarrhea caused by a Shiga toxin-producing organism. The tough decision, as mentioned by Dr. Tolan in his commentary, is that initially these cases resemble shigellosis, where early antibiotic treatment might be beneficial and at least could limit duration of contagion.
My inclination now is to withhold therapy for most cases, because of this study and others that tend to support the hypothesis that antibiotic treatment could result in increased toxin release and therefore contribute to HUS development. However, I won't be surprised if I'm proven wrong in the future.