Severe COVID-19 Is Associated With an Altered Upper Respiratory Tract Microbiome

Abstract

Background: The upper respiratory tract (URT) is the portal of entry of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and SARS-CoV-2 likely interacts with the URT microbiome. However, understanding of the associations between the URT microbiome and the severity of coronavirus disease 2019 (COVID-19) is still limited.

Objective: Our primary objective was to identify URT microbiome signature/s that consistently changed over a spectrum of COVID-19 severity.

Methods: Using data from 103 adult participants from two cities in the United States, we compared the bacterial load and the URT microbiome between five groups: 20 asymptomatic SARS-CoV-2-negative participants, 27 participants with mild COVID-19, 28 participants with moderate COVID-19, 15 hospitalized patients with severe COVID-19, and 13 hospitalized patients in the ICU with very severe COVID-19.

Results: URT bacterial load, bacterial richness, and within-group microbiome composition dissimilarity consistently increased as COVID-19 severity increased, while the relative abundance of an amplicon sequence variant (ASV), Corynebacterium_unclassified.ASV0002, consistently decreased as COVID-19 severity increased.

Conclusions: We observed that the URT microbiome composition significantly changed as COVID-19 severity increased. The URT microbiome could potentially predict which patients may be more likely to progress to severe disease or be modified to decrease severity. However, further research in additional longitudinal cohorts is needed to better understand how the microbiome affects COVID-19 severity.

Keywords: COVID-19; SARS-CoV-2; microbiome; mild; moderate; severe COVID-19 outcomes; upper respiratory tract.

Figure 1

(A) Stacked bar charts of the relative abundance of the 12 most abundant amplicon sequence variants (ASVs) are shown for each study participant. The most abundant ASV, Staphylococcus_unclassified.ASV0001, was abundant both in uninfected controls and participants with the full range of COVID-19 severities. The second most abundant ASV, Corynebacterium_unclassified.ASV0002, was highly abundant in uninfected control participants and those with mild-to-moderate COVID-19, but was of low abundance in those with severe or very severe COVID-19. (B) A principal coordinate analysis (PCoA) plot of the Bray–Curtis dissimilarities over the first two axes is shown. Dots represent individual data points and diamonds show the centroids. The 90% confidence data ellipses are shown for each of the COVID-19 severity groups. Overall, microbial community composition was significantly dissimilar among the severity groups (P < 0.001).

Figure 2

Log-transformed bacterial load is shown for each of the COVID-19 severity groups. Each box represents the median and interquartile range, and the mean is shown by the white diamond. Individual points are shown as open circles. Bacterial load of the uninfected control participants was similar to that of those with mild-to-moderate COVID-19. Among those with COVID-19, there was a trend toward increasing bacterial load as disease severity increased.

Figure 3

Bacterial richness and alpha- and beta-diversity results are plotted for each of the COVID-19 severity groups. Each box represents the median and interquartile range, and the mean is shown by the white diamond. Individual points are shown as open circles. In the first three facets, bacterial richness and alpha-diversity are shown for each of the patient groupings. Richness and alpha-diversity (Shannon and Simpson indices) were lowest in the uninfected control participants and generally increased as COVID-19 severity increased. Alpha-diversity (Shannon and Simpson indices) decreased in patients in the ICU with very severe COVID-19. The last facet shows within-group pairwise Bray–Curtis dissimilarities plotted on the y-axis for each of the COVID-19 groups. Larger values indicate that the URT microbial community between the two samples was more dissimilar, while smaller values indicate the opposite. Uninfected control participants had the most similar URT microbiome to each other, while the URT microbiome within each group became more dissimilar as COVID-19 severity increased. The URT microbiomes among those with very severe COVID-19 were very dissimilar to each other.

Figure 4

Relative abundance of the one ASV that was identified as significantly differentially abundant by Kruskal–Wallis testing between the COVID-19 groups and that also had a median which consistently changed as COVID-19 severity increased. Each box represents the median and interquartile range, and the mean is shown by the white diamond. Individual points are shown as open circles. Corynebacterium_unclassified.ASV0002 abundance decreased as disease severity increased.