Temporal changes in Streptococcus pneumoniae colonization in children following routine childhood immunization with pneumococcal conjugate vaccine in South Africa

Abstract

Background: Streptococcus pneumoniae remains a leading cause of morbidity and mortality in children <5 years old in sub-Saharan Africa. Pneumococcal conjugate vaccines (PCVs) reduce pneumococcal-associated disease by reducing vaccine-serotype (VT) circulation. Monitoring serotype-specific nasopharyngeal colonization can serve as a proxy to evaluate the impact of vaccination against VT disease. Sensitive tools for detecting pneumococcal serotypes and other bacterial species are required. Current real‑time Polymerase Chain Reaction (PCR) methods do not fully distinguish 13-valent PCV (PCV13) serotypes from non-vaccine serotypes (NVT). The 7-valent-PCV (PCV7) was introduced in the South African immunization program in 2009 (replaced by PCV13 since 2011) using a 2+1 schedule (at 6, 14, and 40 weeks of age). In this study, the impact of routine PCV immunisation on pneumococcus, and other bacterial colonisation among rural (Agincourt, Mpumalanga) and urban (Soweto, Gauteng) dwelling South African children <5 years old, was evaluated 8-9 years after PCV introduction using a comprehensive, real-time PCR. Methods: Designed and previously published assay-sets were combined into a 96-assay reaction set within a nanofluidic real-time PCR platform. The reaction‑set was optimized using reference isolates and synthetic calibrators for analytical performance. To validate the reaction set, diagnostic performance was evaluated through blind analysis of 1 973 archived nasopharyngeal swab samples (NPS) previously serotyped with the referent standard culture‑based Quellung method. The reaction set was applied to NPS collected from children across two study periods, during the early PCV introduction period and after 8-9 years of routine immunization. In Period-1, 630 NPS collected in May to October 2009 were available from Agincourt and 1135 NPS collected in May 2010 to February 2011 were available from Soweto. In Period-2, NPS were collected from 568 children in July 2017 to February 2018 in Agincourt, and 571 children in June to December 2018 in prospective colonization surveys. 1vii Results: The real-time PCR reaction set was analytically sensitive (limit of detection <102 gene equivalents), efficient (90–110%), and had low variation between replicates (r2 > 0.98) for relative quantification. The diagnostic sensitivity and specificity of the reaction set was >80% and >95% respectively for all assay-sets that targeted serotypes previously detected by Quellung at a prevalence >1%. Comparing Period-2 to Period-1, in Agincourt and Soweto, there was a lower overall (76.9% vs. 83.2%; adjusted Odds Ratio [aOR]: 0.65, 95% confidence interval [CI]: 0.48-0.87 and 49.4% vs. 68.1%; aOR: 0.66; 95% CI: 0.54-0.88, respectively) and PCV13-VT colonisation prevalence (14.3% vs. 51.0%; aOR: 0.16, 95%CI: 0.12-0.21 and 18.6% vs. 40.9%; aOR: 0.41; 95% CI: 0.3- 0.56). In Period-2, residual colonization by 19F (5.3% and 8.1%) remained high in both settings compared with Period-1 (10.3%, aOR: 0.52, 95%CI: 0.33-0.82 and 6.6%; 75/1135; aOR: 2.0; 95% CI: 1.09-3.56, respectively). Non-vaccine-serotype (NVT) colonisation was higher in Period-2 in Agincourt (63.2%) than Period-1 (35.6%, aOR: 3.12, 95%CI: 2.45-3.97), while there was no difference in NVT in Soweto between the two periods (37.8% vs. 42.4%; aOR: 0.96, 95% CI: 0.72-1.26. In Agincourt, there was a higher prevalence in Period-2 compared with Period-1 of Acinetobacter baumannii (36.8% vs 1.1%, aOR: 50.11, 95%CI: 23.14-108.50) and Klebsiella pneumoniae (13.2% vs 0.6%, aOR: 22.16, 95%CI: 8.03-61.11). Conclusions: The high‑throughput nanofluidic real‑time PCR method simultaneously detects 15 bacterial species and 92 pneumococcal serotypes, distinguishing to 57 single serotypes and 35 serotypes within 16 subgroups, using 73 serotyping, and 23 bacterial detection assay-sets in 96 samples (including controls), within a single qPCR run. There was an 80% and 59% reduction in colonization by PCV13-VT serotypes in Agincourt and Soweto, 8-9 years after introduction of routine immunisation with PCV, however, there was high residual prevalence PCV13-VT (14.3-18.6%) despite >90% of enrolled children >40 weeks being immunized with PCV in both settings, mainly due to serotype 19F.