Alterations in Gut Microbiota of Patients With COVID-19 During Time of Hospitalization

Abstract

Background & aims: Although severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infects gastrointestinal tissues, little is known about the roles of gut commensal microbes in susceptibility to and severity of infection. We investigated changes in fecal microbiomes of patients with SARS-CoV-2 infection during hospitalization and associations with severity and fecal shedding of virus.

Methods: We performed shotgun metagenomic sequencing analyses of fecal samples from 15 patients with Coronavirus Disease 2019 (COVID-19) in Hong Kong, from February 5 through March 17, 2020. Fecal samples were collected 2 or 3 times per week from time of hospitalization until discharge; disease was categorized as mild (no radiographic evidence of pneumonia), moderate (pneumonia was present), severe (respiratory rate ≥30/min, or oxygen saturation ≤93% when breathing ambient air), or critical (respiratory failure requiring mechanical ventilation, shock, or organ failure requiring intensive care). We compared microbiome data with those from 6 subjects with community-acquired pneumonia and 15 healthy individuals (controls). We assessed gut microbiome profiles in association with disease severity and changes in fecal shedding of SARS-CoV-2.

Results: Patients with COVID-19 had significant alterations in fecal microbiomes compared with controls, characterized by enrichment of opportunistic pathogens and depletion of beneficial commensals, at time of hospitalization and at all timepoints during hospitalization. Depleted symbionts and gut dysbiosis persisted even after clearance of SARS-CoV-2 (determined from throat swabs) and resolution of respiratory symptoms. The baseline abundance of Coprobacillus, Clostridium ramosum, and Clostridium hathewayi correlated with COVID-19 severity; there was an inverse correlation between abundance of Faecalibacterium prausnitzii (an anti-inflammatory bacterium) and disease severity. Over the course of hospitalization, Bacteroides dorei, Bacteroides thetaiotaomicron, Bacteroides massiliensis, and Bacteroides ovatus, which downregulate expression of angiotensin-converting enzyme 2 (ACE2) in murine gut, correlated inversely with SARS-CoV-2 load in fecal samples from patients.

Conclusions: In a pilot study of 15 patients with COVID-19, we found persistent alterations in the fecal microbiome during the time of hospitalization, compared with controls. Fecal microbiota alterations were associated with fecal levels of SARS-CoV-2 and COVID-19 severity. Strategies to alter the intestinal microbiota might reduce disease severity.

Keywords: Bacteria; Coronavirus; Fecal Nucleic Acid; Gut Microbiome.

Graphical abstract

Figure 1

Schematic diagram of stool sample collection, SARS-CoV-2 PCR test results and hospitalization duration in patients with COVID-19 (n = 15). “CoV” denotes patient with COVID-19. Stool specimens were serially collected for shotgun metagenomics sequencing and quantitative RT-PCR test for SARS-CoV-2 virus; “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection. “+ve throat swab”: the first positive result for SARS-CoV-2 virus in nasopharyngeal/throat/pooled swabs; “-ve throat swab”: the first negative result for SARS-CoV-2 virus in 2 consecutive negative nasopharyngeal/throat/pooled swab tests, on which patient was then discharged.

Figure 2

Gut microbiome alterations in patients with COVID-19 and longitudinal changes over the disease course. (A) The effect size of subject metadata in gut microbiome composition, as determined by PERMANOVA test. ∗∗P < .01; ∗P < .05. (B) Microbiome community alterations in COVID-19, viewed by NMDS (nonmetric multidimensional scaling) plot based upon Bray-Curtis dissimilarities. The microbiomes were compared among healthy controls (n = 15), COVID-19 (abx−, n = 7), COVID-19 (abx+, n = 8), and pneumonia controls (n = 6). (C) Dissimilarity of the gut microbiome of patients with COVID-19 to that of healthy controls during the disease course. The microbiome dissimilarity was calculated as Bray-Curtis dissimilarity. The gray area denotes the range of Bray-Curtis dissimilarities among gut microbiomes of healthy controls, and the solid black line indicates the median dissimilarity among healthy individuals. “CoV” denotes patient with COVID-19. “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Figure 3

Correlation between gut bacteria and fecal SARS-CoV-2 shedding in patients with COVID-19 over the disease course. (A) Longitudinal changes in fecal viral loads of patients with COVID-19. (B) Bacteria significantly associated with fecal viral load during disease course, as determined by Spearman correlation test.

Figure 4

Schematic summary of the gut microbiome alterations in COVID-19. In healthy individuals, Eubacterium, Faecalibacterium prausnitzii, Roseburia, and Lachnospiraceae taxa are prevalent in their gut microbiome. However, the gut microbiome of patients with COVID-19 is characterized by enrichment of opportunistic pathogens and depletion of commensals in the gut. Such gut dysbiosis persists during the COVID-19 disease course, even after clearance/recovery of SARS-CoV-2 infection. Baseline fecal abundance of the bacteria Coprobacillus, Clostridium ramosum, and Clostridium hathewayi showed significant correlation with COVID-19 severity, whereas an anti-inflammatory bacterium Faecalibacterium prausnitzii showed an inverse correlation. Four Bacteroidetes members, including Bacteroides dorei, Bacteroides thetaiotaomicron, Bacteroides massiliensis, and Bacteroides ovatus, known to downregulate ACE2 expression in the murine gut, showed significant inverse correlation with fecal SARS-CoV-2 viral load in patients with COVID-19.

Supplementary Figure 1

Longitudinal changes of fecal abundance of Eubacterium ventriosum in patients with COVID-19 over the disease course. Bacterial species abundance is expressed as fractional abundance (%). “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Supplementary Figure 1

Longitudinal changes of fecal abundance of Eubacterium ventriosum in patients with COVID-19 over the disease course. Bacterial species abundance is expressed as fractional abundance (%). “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Supplementary Figure 1

Longitudinal changes of fecal abundance of Eubacterium ventriosum in patients with COVID-19 over the disease course. Bacterial species abundance is expressed as fractional abundance (%). “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Supplementary Figure 1

Longitudinal changes of fecal abundance of Eubacterium ventriosum in patients with COVID-19 over the disease course. Bacterial species abundance is expressed as fractional abundance (%). “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Supplementary Figure 1

Longitudinal changes of fecal abundance of Eubacterium ventriosum in patients with COVID-19 over the disease course. Bacterial species abundance is expressed as fractional abundance (%). “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Supplementary Figure 1

Longitudinal changes of fecal abundance of Eubacterium ventriosum in patients with COVID-19 over the disease course. Bacterial species abundance is expressed as fractional abundance (%). “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.

Supplementary Figure 7

Longitudinal changes of the fecal microbiome in patients with COVID-19, at the community level, during the disease course. “CoV” denotes patient with COVID-19. “D0” denotes baseline date when the first stool was collected after hospitalization; the following timepoints starting with “D” represent days since baseline stool collection.