Impact of 2g of intrapartum azithromycin on the carriage and antibiotic resistance of Escherichia coli and Klebsiella pneumoniae

Targeted interventions are urgently needed to reduce neonatal morbidity and mortality in sub-Saharan Africa (SSA). Intrapartum azithromycin has been explored to prevent maternal and neonatal sepsis, major contributors to this burden, but raises concerns about antibiotic resistance. At the start of this PhD, data on its impact on Escherichia coli and Klebsiella pneumoniae, leading causes of gram-negative neonatal sepsis were lacking. This thesis investigates the effect of intrapartum azithromycin on the carriage and antibiotic resistance of these bacteria. The analysis conducted includes data from two randomized, double-blind, placebo-controlled trials (PregnAnZI) conducted in West Africa, where pregnant women were randomized (1:1) to receive either oral azithromycin (2g) or placebo. PregnAnZI-1, a proof-of-concept trial with 829 participants in The Gambia, evaluated the effect of intrapartum azithromycin on major gram-positive bacteria causing neonatal sepsis. PregnAnZI-2, a larger trial with approximately 12,000 participants in The Gambia and Burkina Faso, assessed its impact on neonatal sepsis and mortality. In PregnAnZI-1, breast milk (BM), vaginal swabs (VS), and nasopharyngeal swabs (NPS) were collected from all mothers, and NPS from all newborns within 28 days of the intervention. In PregnAnZI-2, rectovaginal swabs (RVS), BM, and NPS were collected from 500 mothers, while rectal swabs (RS) and NPS were obtained from their infants within four months. Microbiological methods were used to identify isolates and perform antibiotic susceptibility testing for E. coli and K. pneumoniae in both PregnAnZI-1 and PregnAnZI-2. A subset of E. coli from maternal RVS and infant RS in PregnAnZI-2 was randomly selected for genomic analysis. Whole-genome sequencing (WGS) was performed to assess the impact of the intervention on antibiotic resistance genes and the genomic diversity of E. coli. In addition, discriminant analysis of principal components (DAPC) was applied to evaluate changes in genomic diversity associated with the intervention. In PregnAnZI-1, microbiological analysis showed an increased prevalence of azithromycin-resistant E. coli in maternal VS and azithromycin-resistant K. pneumoniae in BM. However, there were no significant differences in bacterial carriage or antibiotic resistance among newborns. In PregnAnZI-2, the intervention primarily affected E. coli and K. pneumoniae in infants’ RS. E. coli carriage was lower in the azithromycin arm than in the placebo at days six and 28 post-intervention, but this was accompanied by a higher prevalence of azithromycin-resistant E. coli at these time points. The intervention also increased K. pneumoniae carriage at days six and 28 and azithromycin-resistant K. pneumoniae carriage at day 28. Notably, by the fourth month, the effect of the intervention on bacterial carriage and antibiotic resistance had waned, and there was no effect on maternal samples at any time point. Genomic analysis showed that intrapartum azithromycin increased the carriage of macrolide resistance genes at day six post-intervention. The predominant resistance mechanism was macrolide inactivation via the plasmid-borne gene, mph(A). DAPC found minimal differences in E. coli diversity both pre- and post-intervention, as well as between the azithromycin and placebo arms. However, specific sequence types (ST’s), serotypes, and a phylogroup differed between arms post-intervention. ST349, serotype O153:H9, and phylogroup F were significantly more prevalent in the azithromycin arm, whereas ST101 and serotype O15:H18 were significantly lower in the azithromycin arm. These ST’s, serotypes, and phylogroup have been associated with antimicrobial resistance, virulence, and plasmid dissemination. The increased carriage of K. pneumoniae, a multidrug-resistant pathogen, raises concerns about potential community transmission. Similarly, the higher prevalence of azithromycin-resistant E. coli suggests the intervention may contribute to the spread of antibiotic resistance, reinforced by the increased detection of plasmid-mediated macrolide resistance genes in the azithromycin arm. Furthermore, the intervention influenced the proliferation of specific ST’s, serotypes, and a phylogroup associated with antimicrobial resistance (AMR), and may potentially promote the establishment of resistant clones, an area requiring further investigation. This PhD research highlights the impact of intrapartum azithromycin on E. coli and K. pneumoniae and its findings should be considered alongside the overall benefits of intrapartum azithromycin. This is crucial in assessing its suitability for routine implementation in high-burden settings. Additionally, ongoing surveillance should be an integral component of this intervention to monitor its long-term effects on E. coli and K. pneumoniae.