Background: Antibiotics are commonly given for the treatment of childhood diarrhoea, but are not indicated in most cases. Antibiotics modify the gastrointestinal microbiota, which may have unanticipated effects on the risk of subsequent diarrhoea.
Methods: In a prospective observational cohort study, we assessed the effect of caregiver-reported antibiotic treatment for diarrhoea on the timing of a child's next episode among 434 children followed from birth to 3 years of age in Vellore, India. We estimated median time differences and time ratios from inverse probability of exposure-weighted Kaplan-Meier curves for the time to next diarrhoea episode, comparing children who did and did not receive antibiotics for the previous episode.
Results: Study children had more than five diarrhoea episodes on average in the first 3 years of life, and more than a quarter of all episodes were treated with antibiotics. Children who received antibiotics for their first diarrhoea episode had their second episode on average 8 weeks earlier (median time difference: -8, 95% confidence interval: -10, -3) than children who did not receive antibiotics. The effects of antibiotics on subsequent diarrhoea were greatest at earlier episodes and younger ages, and cefixime had a slightly larger effect compared with cotrimoxazole.
Conclusions: Antibiotic treatment of diarrhoea was associated with reduced time to a subsequent diarrhoea episode, especially among younger infants. Whereas rational use of antibiotics has been advocated to reduce antimicrobial resistance in populations, we show that overuse of antibiotics may also have a direct adverse effect on individual patients.
Keywords: Diarrhoea; India; Kaplan-Meier survival curves; antibiotics; inverse probability weighting; time differences.
© The Author 2015; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association.