I get asked frequently about the correlation between the use of antibiotics and the onset of hemolytic uremic syndrome (HUS) – that is, does the use of antibiotics with an E. coli O157:H7 infection cause a child who would not otherwise develop HUS to in fact develop it?

First, everyone knows that I am not a doctor. That being said, in nearly 18 years of representing families who have suffered through HUS, I have seen cases where children received no antibiotics and suffered HUS and I have seen the reverse.

In 2000 when Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI., published “The risk of hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections” in the N Engl J Med 2000;342:1930-6, physicians (and lawyers) took notice. In pertinent part the article found:

In up to 15% of North American children infected with Escherichia coli O157:H7, hemolytic-uremic syndrome (HUS) develops because of systemic absorption of Shiga toxins produced by the organism. Although antibiotics have been shown in vitro to enhance release of these toxins from injured bacteria their effect on the development of HUS in people infected with E. coli O157:H7 is unknown.

In children with stool cultures positive for E. coli O157:H7, is antibiotic therapy associated with an increased risk of HUS, independent of the severity of the initial diarrheal illness?

The study’s criteria for HUS were met in 10 (14%) of the 71 subjects. HUS developed in 5 (56%) of 9 children who received antibiotic therapy and in 5 (8%) of 62 children who did not (p <% 0.001). Children who received antibiotic therapy were comparable to those who did not with respect to age, sex and the baseline clinical and laboratory features of their illness.

Multivariate analysis showed that the risk of HUS was associated significantly with what were considered to be 2 surrogate markers of disease severity: initial peripheral white blood cell count (p = 0.02) and time elapsed from the day of symptom onset to the day on which stool cultures were obtained (p = 0.008). Risk was directly proportional to the white blood cell count and inversely proportional to the interval between onset of illness and stool cultures. It was inferred from the latter finding that patients with more severe illness were evaluated earlier than those with less severe illness. The relative risk of HUS among the children who were given antibiotic therapy, when adjusted for these 2 variables, was 17.3 (95% confidence interval 2.2-137, p = 0.007).

Although this study is subject to the selection and confounding biases inherent in observational research, it offers compelling evidence of a link between antibiotic therapy and the development of HUS in diarrheal illness caused by E. coli O157:H7. The strength of the association and the biologically plausible effect of antibiotics on the amount of Shiga toxin available for absorption from the intestine support the inference of causality.

Although HUS develops in patients infected with E. coli O157:H7 with or without antibiotic treatment, it occurs much more frequently when antibiotics are given. The findings of this study strongly suggest that these drugs should be withheld in children with acute diarrheal illness until stool cultures confirm growth of an organism for which antibiotic therapy is indicated (e.g., Campylobacter pylori).

In JAMA in 2002 (Aug 28;288(8):996-1001) questions were raised by Safdar N, Said A, Gangnon RE, Maki DG. In part the comment found:

The use of antibiotics for treatment of Escherichia coli O157:H7 infection has become controversial since a recent small study found that it may increase the risk of hemolytic uremic syndrome (HUS). However, other larger studies have reported a protective effect or no association.

To determine whether antibiotic therapy for E coli O157:H7 enteritis increases the risk of HUS.

PubMed and MEDLINE computer searches were performed for studies published from January 1983 to February 2001 using the key words hemolytic uremic syndrome, risk factor, antibiotics, and Escherichia coli O157:H7. Reference lists of relevant publications were reviewed, and 12 experts in the field were contacted to identify additional reports. No language restrictions were applied to the search.

Studies were included if they reported a series of patients with documented E coli O157:H7 enteritis, some of whom developed HUS; had clear definitions of HUS; and had adequate data delineating the relationship between antibiotic therapy and the occurrence of HUS. Nine of the 26 identified studies fulfilled these criteria.

Two authors (N.S. and A.S.) independently reviewed each report identified by the searches and recorded predetermined information relevant to the inclusion criteria. A pooled odds ratio was calculated using a fixed-effects model, with assessment of heterogeneity among the studies.

The pooled odds ratio was 1.15 (95% confidence interval, 0.79-1.68), indicating that there does not appear to be an increased risk of HUS with antibiotic treatment of E coli O157:H7 enteritis. Incomplete reporting of data in individual studies precluded adjustment for severity of illness.

Our meta-analysis did not show a higher risk of HUS associated with antibiotic administration. A randomized trial of adequate power, with multiple distinct strains of E coli O157:H7 represented, is needed to conclusively determine whether antibiotic treatment of E coli O157:H7 enteritis increases the risk of HUS.

So, good readers, what is the answer to the question – “In E. coli O157:H7 cases, is antibiotic therapy associated with an increased risk of HUS, or not?

  • Mary McGonigle-Martin

    Bill, thank you for posting this. Children can get HUS without being given antibiotics. Leaders in the raw milk movement like to lead people to beleive that a child will only get HUS from E.coli 0157:H7 if the child is given antibiotics. This is not true.

  • cheryl berenson RN, MS-MPH student OHSU

    Personally, i would go with the 2002 meta-analysis results until a larger and more rigorous study is completed…as Safdar et al (2002)reported “A randomized trial of adequate power, with multiple distinct strains of E coli O157:H7 represented, is needed to conclusively determine whether antibiotic treatment of E coli O157:H7 enteritis increases the risk of HUS.”
    science rules!!!

  • Elizabeth Armstrong

    Speaking from first hand experience, children do NOT need to be given antibiotics to get HUS.

  • From an email from a Doctor who follows HUS:
    Bill, that is a nice piece you wrote.  The main problem with the original Tarr study was that is was too small to allow stratification according to type of antibiotic, let alone:
    1.  doses/kg given
    2.  when during the diarrheal prodrome the antibiotics were given, and for how long., plus the unanswerable variables such as:
    1.  amount of Stx absorbed per unit of time
    2.  amount of LPS absorbed
    3.  variability in Stx receptors (density) among individuals
    In short, the number of variables is extensive, and the the difficulty in stratifying patients is mind boggling

  • Bruce Clark

    Bill, a more recent study using pigs suggests that it may be the type of antibiotic that is key. The study found in particular that the DNA inhibitor ciprofloxicin was associated with much worse medical outcomes due to increased Stx 2 production. Antibiotics that inhibit protein synthesis such as azithromycin and gentamicin had no effect on Stx production. Cipro has previously been implicated as potentially increasing the risk of HUS and as one of the more commonly prescribed antibiotics for empiric treatment of infectious diarrhea it should probably be avoided by clinicians when there is any risk of shiga toxin producing E. coli infection. See, Zhang et al, Gnotobiotic Piglet Infection Model for Evaluating
    the Safe Use of Antibiotics against Escherichia coli O157:H7 Infection. J Infect Dis. 2009 Feb 15;199(4):486-93.

  • What Bruce said is the utmost truth. Originally when Cipro was first put on the market, it had a warning label that it could cause kidney complications. But then later on the company that makes it removed the warning and its been gone since. What they really need to do is replace the word antibiotics with “CIPRO” as not all antibiotics damage the kidneys when used against e. coli. Then do a study using just cipro and no other antibiotic.

  • Leslie Adams

    My daughter developed HUS and was not given any antibiotics. This was 10 years ago. It was explained to us when we asked why she was not given antibiotics that this would definitely lead to HUS and there was a chance that she would not develop HUS if not given antibiotics. I have never fully understood why she was one who developed the HUS, but the doctors managed her case so beautifully that she is a completely healthy 22 year old. She was 11 yrs. old when she contracted E Coli, was admitted to the hospital 2 days after symptoms appeared and her kidneys failed to 2% functioning by the 4th day. She was started on dialysis and was on dialysis for 8 weeks. She was in the ICU for a week and half and suffered a seizure at one point. She received incredible care from Ochner Hospital in NO. But I remain confused about the antibiotic issue. Would receiving them have prevented the HUS or made it worse?

  • Great to see that – 8 weeks on dialysis – that is quite the ride for all of you. Regarding antibiotic use. There are several articles on the topic and they go both ways – see.www.about-hus.com