Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

The risk of the hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections

Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI.

N Engl J Med. 2000;342(26):1930-6. [abstract]

 

Reviewed by Mehrengise Cooper BSc (Hons), MB BS, MRCP(Paediatrics), Children's Hospital, Boston MA

Review posted February 23, 2001

I. What is being studied?:

The study objective:

To assess if administration of antibiotics to children infected with Escheria coli 0157:H7 increased the risk of developing the hemolytic-uremic syndrome (HUS).

The study design:

An observational prospective cohort study.

The patients investigated:

71 children under the age of ten years who were infected with Escheria coli 0157:H7 causing diarrhea were studied. They were identified on the basis of positive stool cultures collected within 7 days of developing the diarrheal illness. Following this caregivers completed standardized questionnaires providing demographic information and clinical information regarding the course of the child's illness including any drug prescription Ð particularly looking for any antibiotic administration as specific risk factor.

The children underwent daily blood tests looking at renal function and blood counts until the HUS developed and resolved or until the HUS did not develop Ð it was felt that if HUS had not developed 14 after the onset of the illness, it was then unlikely to do so.

HUS was defined as the triad of

• Hemolytic anemia Ð hematocrit < 30%, with red cell destruction on blood film
• Thrombocytopenia Ð platelet count < 150 000/cubic mm
• Renal insufficiency Ð serum creatinine > upper limit of normal for age

II. Are the results of the study valid?

Primary questions:

1.Were there clearly identified comparison groups that were similar with respect to important determinants of outcome, other than the one of interest?

Yes. The study group comprised those children with positive stool cultures for E. coli 0157:H7 who received antibiotics; and they were compared with those who did not receive antibiotics (within the first 7 days of illness, defined as the onset of diarrhea).

Differences between those treated, and those not treated, with antibiotics were analyzed using independent-sample t-tests for continuous variables, and Fisher's exact or chi-square tests for categorical variables with use of univariate logistic regression for linear trend.

As this was a cohort study and thus not possible to match patients, in addition to the above, multivariate logistic-regression analysis was used to examine the risk of HUS after initial adjustment for initial white cell count and day of illness where stool culture was obtained. Association between exposure to medication during the phase of diarrhea and later development of HUS were estimated by odds ratio and then reported as relative risk. Other factors which included demographic variables, specific symptoms, duration of diarrhea before laboratory investigations and antibiotic therapy (if given), and initial laboratory values were studied to see if a particular variable affected any association between exposure to medication and HUS development.

2. Were the exposures and outcomes measured in the same way in the groups being compared?

Yes. In this study, there were two exposures which were assessed - these were firstly, those children who were infected with E.coli 0157:H7 and were determined by positive stool cultures. The second exposure was to antibiotic administration. These data and other demographic information were picked up from the standardized questionnaires used. Medication use was confirmed by the health care provider and/or review of the child's medical record.

The outcome variable was if HUS developed. The children were all hospitalized (personal communication, Dr Tarr), and blood tests were reviewed daily. If HUS had not developed within 14 days of first becoming unwell, then it would be unlikely to do so, and this was set as the end of the period of follow up.

HUS was defined as stated above with the triad of microangiopathic hemolytic anemia, thrombocytopenia, and acute renal insufficiency.

3. Was follow-up sufficiently long and complete?

Yes. The period of observation was for 14 days. This was based upon the likelihood of developing HUS and period of risk was considered to be not more than 14 days from the onset of diarrhea. All children with HUS were followed to hospital discharge.

Secondary questions:

4. Is the temporal relationship correct?

Yes. HUS was predated by knowledge of if antibiotics had been administered in children known to have positive stool culture for E.coli 0157:H7.

5. Is there a dose-response gradient?

No. The investigators only looked at if antibiotics had been prescribed, with no comment on duration or dose.

9 (13%) received antibiotics, and 10 (14%) developed HUS. Out of the 10 who developed HUS, 5 had received antibiotics.

III. What are the results?

1. How strong is the association between exposure and outcome?

10 (14%) children developed HUS out of 71. The factors of significance in the group developing HUS were related to

• Children given antibiotics, against no antibiotic administration
• Raised initial white cell count
• The day of the initial stool culture

Raised initial white cell count and day of initial stool culture were both found to be independently associated with development of HUS using logistic-regression analysis.

Looking at these factors together with multivariate analysis and adjustments for the day on which initial stool cultures were obtained and initial white cell count, those treated with antibiotics were at higher risk of developing HUS when compared with children who were not treated with antibiotics; the adjusted relative risk was 17.3. In the group who received antibiotics within the first three days of diarrhea, relative risk was 15.0, and after adjusting for day of collecting stool and initial white cell count in the multivariate analysis the adjusted relative risk was 32.3.

When specific antibiotics were assessed, after adjustment for initial white cell count, the relative risk of HUS was 17.7 for trimethoprim-sulfamethoxazole, and 13.4 for beta-lactam antibiotics.

2. How precise is the estimate of the risk?

Although the total number of children studied was of an adequate size, those who received antibiotics were small in number as were those developing HUS. The 95% confidence limits (CI) were large in assessing all relative risks. Looking at those treated with antibiotics with a higher risk of developing HUS the CI were 2.2 - 137; and in those receiving antibiotics within 3 days after onset of illness relative risk CI were 1.3 - 174; those where adjustments were made for initial white count and day of obtaining stool culture CI 1.4 - 737. Looking at the antibiotics used CI with increased risk of developing HUS were 1.2 - 261 in those treated with trimethoprim-sulfamethoxazole, and 1.9 - 96 in those treated with beta-lactam antibiotics. Although the confidence intervals are wide, they do not cross unity.

IV. What are the implications for my practice?

1. Are the results applicable to my practice?

Yes. Severe gastrointestinal infection can lead to dehydration with the sequelae including hypotension, and poor urine output, and/or sepsis. In this study two patients required ICU stays, both suffering with fluid overload and the subsequent pulmonary effects (Personal communication, Dr Tarr). When presented with such an unwell child on the intensive care unit, the clinical scenario is severe enough to warrant commencement of antibiotics, if only for 48 hour sepsis rule-out. As the intensivist it would therefore be essential to know if the administration of an antibiotic would be more risky to the patient than no antibiotic if the organism is known to be E.coli 0157:H7.

2. What is the magnitude of the risk?

The hemolytic-uremic syndrome developed in 5 of the 9 children given antibiotics (56%) when compared with 5 of the 62 children (8%) who were not prescribed antibiotics. In a group where all other characteristics are similar this gives an absolute risk increase of 48%. In those who developed oligoanuric HUS (urine output of < 0.5 ml/kg/h for at least 48 hours during the illness), 22% received antibiotics whilst 3% did not; therefore the absolute risk increase for this outcome is 19%. Therefore the reciprocal calculation shows that for 5 children treated with antibiotics who have E.coli 0157:H7 diarrhea 1 child will develop HUS (NNH - number needed to harm).

3. Should I attempt to stop the exposure?

As a prospective cohort study, Wong et al have clearly shown that those children infected with E. coli 0157:H7 went on to develop HUS more often if they received antibiotics. With this evidence presented, I would be keen to stop this exposure, but from an ICU standpoint every case will be judged separately.

Additional Comments

A group from Montreal conducted a randomized controlled trial looking at the administration of trimethoprim-sulfamethoxazole in children with E.coli 0157:H7 and found no significant effect on progression to HUS, but the numbers affected were small making it difficult to make firm conclusions (1). Other recent case reports also have suggested that treatment of E.coli 0157:H7 prediarrheal symptoms with antibiotics have no benefit (2). This area needs further exploration.

References

1. Proulx F, Turgeon JP, Delage G, Lafleur L, Chicoine L. Randomized, controlled trial of antibiotic therapy for Escherichia coli O157:H7 enteritis. J Pediatr. 1992;121(2):299-303.[abstract]
2. Igarashi T, Inatomi J, Wake A, Takamizawa M, Katayama H, Iwata T. Failure of pre-diarrheal antibiotics to prevent hemolytic uremic syndrome in serologically proven Escherichia coli O157:H7 gastrointestinal infection. J Pediatr. 1999;135(6):768-9. [abstract]

Comments

Back to the EB Journal Club Index