Nasopharyngeal microbiome composition by SARS-CoV-2 presence and severity

Abstract

The influence of SARS-CoV-2 on the nasopharyngeal microbiome, or vice-versa, is unclear. Nasopharyngeal swabs from Dutch healthcare workers (N = 257) and hospital outpatients with respiratory symptoms (N = 143), leftover after SARS-CoV-2 testing in 2020-2021, were 16S rRNA amplicon sequenced and tested for respiratory viruses by multiplex PCR panel. The healthcare workers were younger and much healthier than the patients, and experienced less severe viral infections. In the healthcare workers, log₁₀ estimated concentrations (ECs) of Corynebacterium were slightly increased in samples with SARS-CoV-2 versus no virus detected, regardless of symptomatology (adjusted regression coefficient 0.52, p = 0.042) but no other bacterial ECs differed. Corynebacterium and Dolosigranulum ECs were higher in very mild/asymptomatic SARS-CoV-2 episodes compared to very mild/asymptomatic episodes with no viruses detected, but lower in mild compared to very mild/asymptomatic SARS-CoV-2 episodes (-1.07, p = 0.015, and -1.37, p = 0.011, respectively). In the patients, similar but non-significant trends by SARS-CoV-2 severity (fatal, severe, moderate versus mild) were seen for Dolosigranulum, but not for Corynebacterium. In this population, the largest nasopharyngeal microbiome composition differences were seen by the presence and severity of comorbidities. These findings suggest that the Dolosigranulum EC decreases with increasing SARS-CoV-2 severity, but the clinical relevance of this finding is unclear.

Keywords: 16S rRNA gene sequencing; COVID-19; Nasopharyngeal microbiome; Netherlands; SARS-CoV-2.

Fig. 1

Nasopharyngeal microbiome compositions of healthcare workers. Panels: a) Top-20 genera by virus type regardless of symptoms;¹ b) Mean RA by episode severity regardless of virus;² c) Mean RA for negative samples by episode severity;^2,3 d) Mean RA by virus type regardless of symptoms;^2,3 e) Mean RA by SARS-CoV-2 episode severity;^2,3 f) Alpha diversity by virus type regardless of symptoms;^2,5 g) MDS Bray-Curtis by virus type regardless of symptoms.^2,6 Abbreviations and footnotes: MDS = multidimensional scaling; PCoA = principal coordinate analysis; RA = relative abundance; WHO = World Health Organization. ¹In the case of viral co-infection (n = 3), SARS-CoV-2 trumps the other virus. ²Co-infections (n = 3) removed. ³All SARS-CoV-2 positive participants (n = 84) and participants positive for other virus (n = 19) removed. ⁴Other virus (n = 19) removed, and participants with no virus detected and mild (n = 7) or unknown (n = 10) symptoms removed. ⁵Statistical testing with Kruskal Wallis test. ⁶Statistical testing with PERMANOVA, with 9999 permutations and adjustment for multiple testing using the Benjamini–Hochberg false discovery rate correction.

Fig. 2

Nasopharyngeal microbiome compositions of patients. Panels: a) Top-20 genera by virus type regardless of symptoms;¹ b) Mean RA by virus type regardless of symptoms;^2,3 c) Mean RA by SARS-CoV-2 episode severity;^2,3 d) Mean RA by comorbidity score; e) Alpha diversity by virus type regardless of symptoms;^2,3,4 f) Alpha diversity by SARS-CoV-2 episode severity;^2,3,4 g) Alpha diversity by comorbidity score;^2,3,4 h) MDS Bray-Curtis by virus type regardless of symptoms;^2,3,5 i) MDS Bray-Curtis by SARS-CoV-2 episode severity.^2,3,5 Abbreviations and footnotes: MDS = multidimensional scaling; PCoA = principal coordinate analysis; RA = relative abundance; WHO = World Health Organization. ¹In the case of viral co-infection (n = 2), SARS-CoV-2 trumps the other virus. ²Co-infections (n = 2) and participants with cystic fibrosis (n = 5) removed. ³Confirmed antibiotic users (n = 21) removed. ⁴Statistical testing with Kruskal Wallis test. ⁵Statistical testing with PERMANOVA, with 9999 permutations and adjustment for multiple testing using the Benjamini–Hochberg false discovery rate correction.

Fig. 2

Nasopharyngeal microbiome compositions of patients. Panels: a) Top-20 genera by virus type regardless of symptoms;¹ b) Mean RA by virus type regardless of symptoms;^2,3 c) Mean RA by SARS-CoV-2 episode severity;^2,3 d) Mean RA by comorbidity score; e) Alpha diversity by virus type regardless of symptoms;^2,3,4 f) Alpha diversity by SARS-CoV-2 episode severity;^2,3,4 g) Alpha diversity by comorbidity score;^2,3,4 h) MDS Bray-Curtis by virus type regardless of symptoms;^2,3,5 i) MDS Bray-Curtis by SARS-CoV-2 episode severity.^2,3,5 Abbreviations and footnotes: MDS = multidimensional scaling; PCoA = principal coordinate analysis; RA = relative abundance; WHO = World Health Organization. ¹In the case of viral co-infection (n = 2), SARS-CoV-2 trumps the other virus. ²Co-infections (n = 2) and participants with cystic fibrosis (n = 5) removed. ³Confirmed antibiotic users (n = 21) removed. ⁴Statistical testing with Kruskal Wallis test. ⁵Statistical testing with PERMANOVA, with 9999 permutations and adjustment for multiple testing using the Benjamini–Hochberg false discovery rate correction.

Fig. 3

Individual estimated concentrations for Corynebacterium and Dolosigranulum. Panels: a) EC of Corynebacterium by SARS-CoV-2 severity in healthcare workers;¹ b) EC of Dolosigranulum by SARS-CoV-2 in healthcare workers;¹ c) EC of Corynebacterium by SARS-CoV-2 severity in hospital patients;² d) EC of Dolosigranulum by SARS-CoV-2 in hospital patients.² Abbreviations: EC = log₁₀ estimated concentrations; WHO = World Health Organization. ¹Co-infections (n = 3), other virus (n = 19), participants with no virus detected and mild (n = 7) or unknown (n = 10) symptoms, and participants with undetermined total bacterial load (n=10) removed. Black dots indicate means. ²Other viruses (n = 52), co-infections (n = 2), confirmed antibiotic users (n = 4), and participants with undetermined total bacterial load (n=6) removed. Black dots indicate means.