Abstract

BACKGROUND: Data on changes in pneumococcal serotypes in hospitalised children following the introduction of the pneumococcal conjugate vaccine (PCV) in low-income and middle-income countries are scarce. In 2016, Mongolia introduced the 13-valent PCV (PCV13) into the national immunisation programme. We aimed to describe the trend and impact of PCV13 introduction on pneumococcal carriage in hospitalised children aged 2-59 months with pneumonia in Mongolia over a 6-year period. METHODS: In this active surveillance programme, children aged 2-59 months with pneumonia who met the study case definition (cough or difficulty breathing with either respiratory rate ≥50 beats per min, oxygen saturation <90%, or clinical diagnosis of severe pneumonia) were enrolled between April 1, 2015, and June 30, 2021, from four districts in Ulaanbaatar. We tested nasopharyngeal samples collected at enrolment for pneumococci using lytA real-time quantitative PCR and conducted molecular serotyping and detection of antimicrobial resistance (AMR) genes with DNA microarray. We used log-binomial regression to estimate prevalence ratios (PRs) of pneumococcal carriage, comparing prevalence in the periods before and after the introduction of PCV13 and between vaccinated and unvaccinated children for three outcomes: overall, PCV13 vaccine-type, and non-PCV13 vaccine-type carriage. PRs were adjusted with covariates that were identified by use of a directed acyclic graph, informed by relevant literature. FINDINGS: A total of 17 688 children were enrolled, of whom 17 607 (99·5%) met the study case criteria. 6545 (42·5%) of 15 411 collected nasopharyngeal swabs were tested for pneumococci. In all age groups, a similar prevalence of pneumococcal carriage was shown between the pre-PCV13 period and post-PCV13 period (882 [48·0%] of 1837 vs 2174 [46·2%] of 4708; adjusted PR 0·98 [95% CI 0·92-1·04]; p=0·60). Overall, vaccine-type carriage reduced by 43·6% after the introduction of PCV13 (adjusted PR 0·56 [95% CI 0·51-0·62]; p<0·0001). Younger children (aged 2-23 months) showed a 47·7% reduction in vaccine-type carriage (95% CI 41·2-53·5; adjusted PR 0·52 [95% CI 0·46-0·59]; p<0·0001), whereas children aged 24-59 months had a 29·3% reduction (12·6-42·8; 0·71 [0·57-0·87]; p=0·0014). Prevalence of 6A, 6B, 14, 19F, and 23F decreased following the introduction of PCV13; however, 19F and 6A remained common (5·8% and 2·9%). Non-vaccine-type carriage increased (adjusted PR 1·49 [95% CI 1·32-1·67]), with 15A, NT2, and 15B/C being the most prevalent serotypes. Overall, 1761 (89·3%) of 1978 analysed samples contained at least one AMR gene. The percentage of samples with any AMR gene decreased with vaccine introduction (92·3% in the pre-PCV13 period vs 85·3% in the post-PCV13 period; adjusted odds ratio 0·49 [95% CI 0·34-0·70]), with similar decreases for samples with at least three AMR genes (46·8% vs 27·6%; 0·44 [0·36-0·55]). INTERPRETATION: 6 years after the introduction of PCV13 in Mongolia, the prevalence of vaccine-type carriage and AMR genes showed a reduction among young hospitalised children with pneumonia. Reductions in vaccine-type carriage are likely to result in reductions in pneumococcal pneumonia. FUNDING: GAVI, the Vaccine Alliance.