Urine bioassay optimisation to assess the association between antimicrobial exposure, pneumococcal carriage and antimicrobial resistance among hospitalised children in Malawi

Understanding antibiotic exposure in low-income settings is crucial when designing and evaluating field research on bacterial carriage and antimicrobial resistance. We evaluated the concordance between reported pre-hospitalisation antibiotic use and antimicrobial activity in urine from children presented with acute respiratory infection (ARI) at a Malawian referral hospital. We further investigated the association between antibiotic exposure and antibiotic resistance prevalence among nasopharyngeal pneumococci. Between November 2016 and July 2019, 102 children aged 1–4 years with ARI were recruited at Queen Elizabeth Central Hospital in Blantyre, Malawi. Nasopharyngeal swab and urine were collected prior to hospital administered antibiotics. History of antibiotic exposure was obtained by questioning the caregiver and reviewing patient-retained medical records. An optimised urine bioassay was used to detect antimicrobial activity in urine. Pneumococci isolated from nasopharyngeal swabs were tested for antibiotic susceptibility by disc diffusion and E-tests. Antimicrobial activity from urine specimens were detected in 64% (65/102) of children. Among 29 children with no reported or no recorded pre-admission antibiotic treatment, antimicrobial activity was detected in 11 cases (38%). Among the 73 children with reported antibiotic treatment before admission, 54 (74%) had antimicrobial activity in urine. Pneumococcal carriage prevalence was 63% in children with pre-admission antibiotics based on parent reports and 62% in those without reported pre-treatment. Based on results from the bioassay the pneumococcal prevalence was 57% in children with antimicrobial activity in urine and 73% in those without activity (p = 0.11). No difference was found in the pneumococcal antibiotic susceptibility in children with or without urine antimicrobial activity. This urine bioassay could be a useful and non-invasive tool for objective assessment of pre-clinic antibiotic treatment, as questionnaires can underestimate antibiotic exposure. This could impact surveillance strategies of bacterial carriage prevalences. A short detection period of some antibiotics in urine may limit the use of the bioassay. We found no signs of selection of resistant pneumococcal strains in children with antimicrobial activity in urine.