Streptococcus pneumoniae carriage and antimicrobial resistance among Kenyan children discharged from hospital

Abstract

Streptococcus pneumoniae is a leading cause of morbidity and mortality among children under five in sub-Saharan Africa, particularly in the vulnerable period following hospital discharge. This dissertation investigates the dynamics of pneumococcal carriage and antimicrobial resistance (AMR) in this high-risk population through a series of analyses nested within the Toto Bora trial, a randomized, placebo-controlled study of azithromycin in 1,398 Kenyan children discharged from hospital and followed for six months. This collection of nested studies aims to inform strategies for preventing AMR and improving child health by evaluating the impact of azithromycin treatment, identifying risk factors for carriage and resistance, and assessing the clinical consequences of pneumococcal carriage.The first study evaluates the effect of a 5-day course of azithromycin versus placebo administered at hospital discharge on S. pneumoniae carriage and AMR. No significant differences were observed in carriage prevalence or azithromycin resistance at 3- or 6-months post-discharge between treatment arms. This may be due in part to high inpatient antibiotic use in this population, reducing any further impact of azithromycin. Notably, resistance to multiple antimicrobial classes was already prevalent at discharge, with nearly 35% of isolates classified as multidrug-resistant. This study underscores the importance of considering inpatient antibiotic use when evaluating the downstream effects of post-discharge antimicrobial interventions. The second study explores clinical and environmental risk factors for pneumococcal carriage and AMR using longitudinal data from nasopharyngeal swabs collected at discharge and follow-up. Pneumococcal carriage at hospital discharge was inversely associated with inpatient antibiotic use and longer hospital stays whereas carriage at 3 and 6 months was linked to household-level exposures such as unimproved water sources, untreated drinking water, and shared sanitation facilities. Resistance to doxycycline was more common among children with multiple young siblings, while receipt of the recommended doses of pneumococcal conjugate vaccine (PCV) was associated with reduced resistance to doxycycline and clindamycin. These findings highlight the role of water, sanitation, and hygiene (WASH) interventions in shaping post-discharge colonization and support the role of PCVs in AMR mitigation strategies. The third study investigates whether S. pneumoniae carriage at discharge is associated with increased risk of rehospitalization or death. Notably, carriage was not linked to adverse outcomes and was associated with a significantly lower risk of all-cause mortality, even after adjusting for key confounders. This inverse association may reflect protective mucosal immunity conferred by colonization with less virulent strains. Given the persistently high post-discharge mortality observed among children without S. pneumoniae carriage, this group may warrant particular attention in future interventions targeting vulnerable populations. Azithromycin-resistant pneumococcal carriage was associated with poorer clinical outcomes among children treated with azithromycin, underscoring the importance of appropriate antimicrobial therapy. Together, this dissertation provides a comprehensive view of pneumococcal dynamics in the period following discharge from hospital. The findings suggest that short-course azithromycin does not meaningfully alter carriage or antimicrobial resistance patterns in settings with high inpatient antibiotic use, and that household environmental conditions play a critical role in shaping post-discharge colonization. Importantly, pneumococcal carriage at discharge was not associated with increased morbidity or mortality. These insights have implications for antimicrobial policy, vaccine strategy, and post-discharge care in low-resource settings, emphasizing the need for integrated approaches that address both clinical and environmental drivers of AMR and child health.