Understanding Carriage and Transmission of Streptococcus pneumoniae Through a Community Cohort Study in South Africa,

Abstract

Pneumococcal carriage is shaped by host demographics, human immunodeficiency virus (HIV) status, viral co-infections, and microbial community dynamics, yet the longitudinal interplay of these factors remains underexplored in high HIV prevalence settings. Despite the substantial reductions in disease achieved through widespread pneumococcal conjugate vaccine (PCV) use, colonization and invasive disease continue. Gaps remain in understanding how colonization evolves over time, particularly during respiratory viral infections such as influenza and respiratory syncytial virus (RSV). This study investigates how demographic factors, HIV status, and viral co-infections influence pneumococcal colonization and density over time, and examines nasopharyngeal microbiome changes before, during, and after influenza infections to better understand microbial interactions that may modulate colonization. We analyzed data from a prospective household observational cohort study of influenza, RSV and other respiratory pathogens community burden and transmission dynamics in South Africa (the PHIRST study) of 1684 individuals across 327 randomly selected households conducted in two communities in South Africa from 2016 – 2018. Nasopharyngeal (NP) swabs were collected twice-weekly. Pneumococcal detection and quantification, RSV and influenza virus detection were performed using real-time polymerase chain reaction (qPCR) and nasopharyngeal microbiome were profiled via 16S ribosomal ribonucleic acid (rRNA) gene sequencing (V4 region). Colonization episodes were defined using longitudinal qPCR data and analyzed with Markov models to estimate acquisition and clearance rates, adjusting for age, HIV status, and misclassification. Pneumococcal density was summarized as the median across episodes, with associations assessed via multivariable logistic regression. In viral co- infection analyses, changes in pneumococcal density before, during, and after RSV or influenza infection were evaluated using mixed-effects and linear regression models. For microbiome analyses, samples were selected from before, during, and after influenza episodes. Alpha and beta diversity metrics were compared using mixed-effects and regression models. Changes in dominant Operational Taxonomic Units (OTUs) were analyzed using additive log-ratio transformation and regression adjusted for host factors. Predictor-responder analysis was used to identify directional interactions between ix OTUs. Analyses were conducted using Stata and R/RStudio, with significance set at P<0.05. Pneumococcal colonization was detected in 98% (1,655/1,684) of participants. Younger children (<5 years of age) had higher odds of colonization (adjusted odds ratio (aOR) 14.1; 95% confidence interval (CI) 1.8–111.3) and longer episode durations (median 35.5 days, interquartile range (IQR) 17–88) than individuals aged ≥5 years (median 5.5 days, IQR 4–12). People living with HIV (PLWHIV) also had an increased odds of colonization (aOR 10.1; 95% CI 1.3–77.1). An increase in pneumococcal density was observed during RSV infection compared to the period before (log mean before vs. during: 9.3 vs. 10.2 copies/ml; P<0.01). Influenza infection was not associated with a significant mean increase in pneumococcal density (9.6 vs. 9.9 copies/ml; P=0.2), though high influenza viral load (cycle threshold (Ct) <25) was associated with increased pneumococcal density (coefficient 2.8; 95% CI 1.4–4.2). During RSV infection, underweight individuals were less likely to have an increase in pneumococcal density compared to individuals with a normal body mass index (coefficient -1.8; 95% CI -3.5 to - 0.1; P=0.04). Influenza infection was associated with shifts in nasopharyngeal microbiome composition, particularly post-infection, with greater changes in younger individuals, females, and PLWHIV. Alpha diversity remained largely stable however older individuals showed a decrease before an influenza infection compared to after (coefficient=-1.03, P=0.01). Our results showed that beta diversity changed across all three periods. Streptococcus 5 was more abundant during influenza episodes and showed a temporal patter of medium-high-low relative abundance across the periods. Furthermore, we found that Streptococcus 5 abundance before infection predicted increases in both Streptococcus 5 and Moraxella 1 during infection, while Moraxella 1 increases during infection preceded declines in Streptococcus 5 and Dolosigranulum pigrum 4 after infection. Although PCVs have reduced disease burden, children and PLWHIV remain key reservoirs for pneumococcal transmission. Our findings corroborate previous studies demonstrating that viral co-infections are associated with increased pneumococcal density and alterations in the nasopharyngeal microbiome, supporting the concept of a synergistic effect that may promote progression to invasive disease. These results show that respiratory viral infections, particularly RSV, are associated with increases in x pneumococcal colonization density. This suggests that vaccination against respiratory viruses, such as RSV, could play a role in reducing bacterial disease risk. Furthermore, observed microbiome shifts during viral infections may help explain changes in pneumococcal colonization patterns, offering insights into host-microbe interactions relevant to disease development.