Azithromycin Treatment Response as a Probe to Attribute Bacterial Aetiologies of Diarrhoea using Molecular Diagnostics: A Reanalysis of the AntiBiotics for Children with severe Diarrhoea (ABCD) Trial

Background Multi-pathogen molecular diagnostics enable assignment of diarrhoea aetiology, but defining thresholds of pathogen quantity to accurately attribute aetiology is challenging in high-burden settings where coinfections are common. The Antibiotics for Children with severe Diarrhoea (ABCD) trial provides an opportunity to leverage the azithromycin treatment response to inform which diarrhoea episodes are bacterial. Methods We analysed data from ABCD, which randomized children with watery diarrhoea to azithromycin or placebo. We quantified heterogeneity in the azithromycin treatment response by the quantity of enteric pathogens detected by qPCR as a tool for understanding aetiology. Results The heterogeneity in azithromycin treatment response was most prominent for Shigella . The risk ratio for diarrhoea on day 3 post enrolment for azithromycin compared to placebo was 13% (95% CI:3, 23) lower per log10 increase in Shigella quantity. The protective effect of azithromycin on diarrhoea at day 3 also became stronger as pathogen quantities increased for Vibrio cholerae , ST-ETEC, and tEPEC. No association between pathogen quantity and azithromycin response was observed for Campylobacter , LT-ETEC or EAEC. The associations were consistent for the outcome of 90-day hospitalisation or death. Conclusions The relationships between response to azithromycin treatment and bacterial pathogen quantities observed for Shigella , Vibrio cholerae , ST-ETEC and tEPEC confirm prior evidence that these pathogens are the likely cause of diarrhoea when detected at high quantities. The lack of a similar response pattern for Campylobacter , LT-ETEC or EAEC is consistent with the limited association between pathogen quantity and diarrhoea symptoms previously observed in large studies of diarrhoea aetiology. Key message (3 succinct bullet points, each a single sentence) We investigated whether heterogeneity in treatment response observed in the ABCD trial, where children with diarrhoea were randomised to receive azithromycin or placebo, could be used to inform aetiological attribution of diarrhoea to bacterial enteric pathogens. The protective effect of azithromycin on diarrhoea at day 3 and hospitalisation or death at day 90 became stronger as pathogen quantities increased for Shigella, Vibrio cholerae, ST-ETEC and tEPEC but not for Campylobacter , LT-ETEC or EAEC. The relationships between Shigella, Vibrio cholerae , ST-ETEC and tEPEC quantity and response to antibiotic treatment confirm prior evidence that these pathogens are the likely cause of diarrhoea when detected at high quantities and could be used to inform which diarrhoea cases should be treated with antibiotics.