Bacterial nasopharyngeal colonisation in children in South Africa before and during the COVID-19 pandemic: an observational study.

Background: Deployment of non-pharmaceutical interventions such as face masking and physical distancing during the COVID-19 pandemic could have altered the transmission dynamics and carriage of respiratory organisms. We evaluated colonisation with Streptococcus pneumoniae and other upper respiratory tract bacterial colonisers before and during the COVID-19 pandemic.

Methods: We did two cross-sectional surveys in Soweto, South Africa from July 3 to Dec 13, 2018 (pre-COVID-19 period) and from Aug 4, 2021, to March 31, 2022 (COVID-19 period) in healthy children (aged ≤60 months) who had recorded HIV status and had not received antibiotics in the 21 days before enrolment. At enrolment, we collected nasopharyngeal swab samples from child participants. Following nucleic acid extraction, nanofluidic quantitative PCR was used to screen all samples for 92 S pneumoniae serotypes and 14 other bacteria. The primary objective was to compare the prevalence and density of pneumococcal nasopharyngeal colonisation, overall and stratified by 13-valent pneumococcal conjugate vaccine (PCV13) serotypes and non-vaccine serotypes. Secondary study objectives included a comparison of serotype-specific pneumococcal colonisation and density, as well as colonisation by the 14 other bacteria in the COVID-19 versus pre-COVID-19 period. We used an adjusted multiple logistic and linear regression model to compare the colonisation prevalence and density between study periods.

Findings: We analysed nasopharyngeal swabs from 1107 children (n=571 in the pre-COVID-19 period; n=536 in the COVID-19 period). We observed no change in overall pneumococcal colonisation between periods (274 [51%] of 536 in the COVID-19 period vs 282 [49%] of 571 in the pre-COVID-19 period; adjusted odds ratio [aOR] 1·03 [95% CI 0·95-1·12]). The prevalence of PCV13 serotypes was lower in the COVID-19 than in the pre-COVID-19 period (72 [13%] vs 106 [19%]; 0·87 [0·78-0·97]), whereas the prevalence of non-typeable S pneumoniae was higher (34 [6%] vs 63 [12%]; 1·30 [1·12-1·50]). The mean log density for overall pneumococcal colonisation was lower in the COVID-19 period than in the pre-COVID-19 period (3·96 [95% CI 3·85-4·07] vs 4·72 [4·63-4·80] log genome equivalents per mL; p<0·0001). A lower density of non-vaccine serotypes (3·63 [3·51-3·74] vs 4·08 [3·95-4·22] log genome equivalents per mL; p<0·0001) and non-typeable S pneumoniae (3·11 [2·94-3·29] vs 4·41 [4·06-4·75] log genome equivalents per mL; p<0·00001) was also observed in the COVID-19 period. There was no difference in the density of PCV13 serotypes between the periods. The prevalence of colonisation during the COVID-19 versus pre-COVID-19 period was lower for non-typeable Haemophilus influenzae (280 [49%] vs 165 [31%]; aOR 0·77 [95% CI 0·71-0·84]), Moraxella catarrhalis (328 [57%] vs 242 [45%]; 0·85 [0·79-0·92]), and Neisseria lactamica (51 [9%] vs 13 [2%]; 0·64 [0·52-0·78]), but higher for Acinetobacter baumannii (34 [6%] vs 102 [19%]; 1·55 [1·35-1·77]) and Staphylococcus aureus (29 [5%] vs 52 [10%]; 1·28 [1·10-1·50]).

Interpretation: There were variable effects on the colonisation prevalence and density of bacterial organisms during the COVID-19 compared with the pre-COVID-19 period. The lower prevalence of PCV13 serotype together with other respiratory organisms including non-typeable H influenzae and M catarrhalis could have in part contributed to a decrease in all-cause lower respiratory tract infections observed in South Africa during the initial stage of the COVID-19 pandemic. The pathophysiological mechanism for the increase in A baumannii and S aureus colonisation warrants further investigation, as does the clinical relevance of these findings.

Funding: The Bill & Melinda Gates Foundation.