Table 1.
Correlations between host-associated microbial features and SARS-CoV-2 features and outcomes
| Study summary | Sample characteristics |
Samples and microbial analysis | Key findings | Refs | |||
|---|---|---|---|---|---|---|---|
| Groups | Sample size | Female/male | Reported age in years | ||||
| Examined sputum and airway microbial features of patients with COVID-19 | Patients with COVID-19; patients with non-COVID-19 pneumonia | 62; 125 | 22/40; 55/70 | 45 (25–82)a; 48 (15–88)a | Metatranscriptome sequencing of sputum or nasopharyngeal swabs | Airway microbiome in patients with COVID-19 showed reduced α-diversity compared with patients with non-COVID-19 pneumonia | [38] |
| Examined lung microbial features of patients with COVID-19 | Patients with COVID-19; uninfected controls | 19; 23 | NA | NA | Metatranscriptome sequencing of bronchoalveolar lavage fluid samples | Pathogens (e.g., pneumonia-causing Klebsiella oxytoca), immunomodulatory taxa (e.g., lactic acid bacteria and Faecalibacterium prausnitzii), and tobacco mosaic virus were enriched in the COVID-19 group, suggesting microbiota dysbiosis. Analysis of microbial α- and β-diversity also showed large differences between patients and controls | [39] |
| Examined lung microbial features in deceased patients with COVID-19 | Deceased patients with COVID-19 | 20 | 6/14 | 66 (60.75–77.0)b | 16S rRNA sequencing used to profile lung tissue acquired via postmortem needle core biopsies immediately after death | Significant enrichment of pathogenic microbes in lungs. Most prevalent bacterial genera were Acinetobacter (80.70% of total sequences), Chryseobacterium (2.68%), Burkholderia (2.00%), Brevundimonas (1.18%), Sphingobium (0.93%), and Enterobacteriaceae (0.68%), together comprising 92.32% of total sequences and regularly detected in all subjects. Biopsies also revealed that most patients had mixed bacterial and fungal infections | [40] |
| Examined lung microbial features of patients with COVID-19 | Patients with COVID-19; patients with non-COVID-19 pneumonia; uninfected controls | 8; 25; 20 | 3/5; NA; NA | 49.0 (±7.9)a; NA; NA | Metatranscriptome sequencing of bronchoalveolar lavage fluid samples | Relative to healthy controls, sequenced bronchoalveolar lavage fluid samples from patients with COVID-19 were similar to those with community-acquired pneumonia, and were either dominated by pathogens or displayed elevated levels of oral and upper respiratory commensal bacteria | [41] |
| Examined gut microbial features of hospitalised patients with COVID-19 or H1N1 | Patients with COVID-19; patients with H1N1; uninfected controls | 30; 24; 30 | 13/17; 9/15; 13/17 | 55.0 (48.0–62.0)b; 48.5 (33.3–66.8)b; 53.5 (43.8–60.3)b | 16S rRNA (V3-V4) gene sequencing of faecal samples | COVID-19 was associated with significantly reduced bacterial α-diversity, higher relative abundance of opportunistic pathogens (e.g., Streptococcus, Rothia, Veillonella, and Actinomyces). Patients with H1N1 displayed lower diversity and different overall microbial composition compared with patients with COVID-19, with seven microbial biomarkers distinguishing the two cohorts | [42] |
| Examined gut microbial features of COVID-19 patients over the course of hospitalisation | Patients with COVID-19; patients with non-COVID-19 pneumonia; uninfected controls | 15; 6; 15 | 8/7; 2/4; 6/9 | 55 (44–67.5)b; 50 (44–65)b; 48 (45–48)b | Shotgun metagenomic sequencing of faecal samples | COVID-19 was associated with significantly higher relative abundance of opportunistic pathogens (including Clostridium hathewayi, Actinomyces viscosus, and Bacteroides nordii) and lower relative abundance of commensal symbionts (including Eubacterium, Faecalibacterium prausnitzii, Roseburia, and Lachnospiraceae). Changes in abundance of of Bacteroides spp. over the course of hospitalisation were inversely correlated with SARS-CoV-2 load in faecal samples. Several of these bacteria (e.g., Bacteroides massiliensis, Bacteroides dorei, and Bacteroides thetaiotaomicron) have been linked with downregulation of ACE2 in the murine gut, and were inversely correlated with SARS-CoV-2 load in human faecal samples | [43] |
| Examined whether COVID-19 alters nasal, throat and gut microbial features in children | Children with COVID-19; uninfected controls | 9; 14 | NA; NA | 7-139 months; age-matched | 16S rRNA (V4) gene sequencing of faecal samples, nasal swabs, and throat swabs | Relative to healthy controls, sustained enrichment of Pseudomonas veronii in both respiratory and gut microbiomes observed in children with COVID-19, with relative abundance of this taxon exceeding 20% in most of the children with COVID-19 | [179] |
| Examined whether a co-receptor of SARS-CoV-2, heparan sulfate, was sensitive to modification via microbes | NA | NA | NA | NA | No direct measure of microbial composition | In human cell cultures, heparan sulfate was highly sensitive to modification by Bacteroides species. Specifically, Bacteroides ovatus and Bacteroides thetaiotaomicron could catabolise cell-surface heparan sulfate and block SARS-CoV-2 from binding to human lung cells. Across two large human microbiome data sets, two of the established risk factors for COVID-19, older age and male sex, were associated with significant reductions in microbes capable of modifying heparan sulfate | [49] |
| Examined whether gut microbial features could be predictive of blood proteomic biomarkers of severe COVID-19 disease | Patients with severe COVID-19; patients with non-severe COVID-19; uninfected controls | 13; 18; 990 | NA; NA; 668/322 | NA; NA; 58.79 (±5.6)a | Quantitative proteomics analysis of serum samples; 16S rRNA (V3-V4) gene sequencing of fecal samples | Identified set of serum proteomic inflammatory markers capable of predicting progression to severe COVID-19, and established that these serum markers could in turn be accurately predicted by gut microbiota composition. These inflammatory markers were positively associated with the genera Ruminococcus, Blautia, and Lactobacillus, and negatively associated with the genera Bacteroides and Streptococcus and the order Clostridiales | [47] |
| Investigated whether gut microbial features are linked to disease severity in patients with COVID-19 | Patients with COVID-19; uninfected controls | 100; 78 | 47/53; 45/33 | 36.4 (±18.7)a; 45.5 (±13.3)a | Shotgun metagenomic sequencing of faecal samples | Gut microbiota composition covaried with disease severity and dysfunctional immune responses in patients with COVID-19. Microbial composition in recovered patients remained altered compared with individuals not infected with SARS-CoV-2 | [46] |
| Analysed oropharyngeal microbial features in patients with mild, moderate, and severe COVID-19, in patients with non-SARS-CoV-2 infections, and in healthy controls | Patients with mild COVID-19; patients with moderate COVID-19; patients with severe COVID-19; patients with upper respiratory tract infections; uninfected controls | 36; 27; 66; 112; 74 | 24/12; 10/17; 14/46; 75/37; 48/26 | 50 (36.75–55.5)b; 57 (46.25–72.75)b; 64 (53.25–72.75)b; 46 (31.0–57.0)b; 36 (29.75–53.25)b | Shotgun metagenomic sequencing of oropharyngeal swabs | Significantly diminished oropharyngeal microbial diversity and high dysbiosis in hospitalised patients with severe COVID-19, which was further linked to a loss of microbial genes and metabolic pathways. Random forest machine learning revealed that oropharyngeal microbiota abundances of Haemophilus or Streptococcus spp. were most important microbial features for segregating clinical outcomes in patients hospitalised with COVID-19 | [180] |
| Investigated associations between oropharyngeal microbial features and fatality in patients with COVID-19 | Patients with COVID-19; uninfected controls | 192; 95 | 78/114; 57/38 | 58 (49–68)b; 47 (33–61)b | Metatranscriptome sequencing of longitudinal oropharyngeal swabs. Swabs were obtained on days 1, 5, 10, 14, 21, and 28 after admission when patient's condition allowed | Abundance of Streptococcus on admission, particularly that of Streptococcus parasanguinis, was identified as a strong predictor of fatality (39/192 cases of COVID-19 were fatal) | [181] |
| Investigated changes in gut microbial features from acute COVID-19 through postconvalescence | Patients with COVID-19; uninfected controls | 30; 30 | 11/19; 11/19 | 53.5 (39.75–59)b; 53.5 (45.25–58)b | 16S rRNA (V3-V4) gene sequencing of longitudinal faecal samples, acquired during acute phase of infection (from onset of illness to host clearing of virus), convalescence (from host clearing of virus to 2 weeks post discharge from hospital), and postconvalescence (6 months post discharge from hospital) | Significant differences in gut microbial communities between patients with COVID-19 and uninfected controls. Microbial richness was lower among patients with COVID-19 than healthy controls at all three time points. Trend toward increasing richness following infection, but richness remained substantially lower even 6 months after host had cleared the virus | [45] |
| Compared throat and gut microbiomes of patients with COVID-19 with samples from uninfected individuals obtained for an earlier study; human observations then extended to vaccinated and unvaccinated murine models | Patients with COVID-19; uninfected controls | 13; 5 | 7/6; NA | 48 (15–85)b; NA | Metagenomic and metatranscriptomic sequencing of throat swabs and faecal samples from patients with COVID-19, and intestinal and faecal samples from infected vaccinated or infected unvaccinated mice | Gut bacterial diversity lower in patients with COVID-19 relative to uninfected controls. Relative to mild and moderate cases of COVID-19, severe COVID-19 was associated with increase in abundance of opportunistic pathogens (e.g., Corynebacterium, Enterococcus, Campylobacter, Citrobacter, Enterobacter) and a loss of butyrate-producing gut bacteria (e.g., Eubacterium and species such as Faecalibacterium prausnitzii). Relative abundances of both Akkermansia muciniphila and Odoribacter higher in both patients with COVID-19 and infected unvaccinated mice related to infected vaccinated mice, with A. muciniphila also being more transcriptionally active in infected unvaccinated mice, suggesting these taxa are associated with disease processes. As a result of antibiotic treatment in several patients, some bacterial changes were difficult to attribute to effects of SARS-CoV-2 versus concomitant antibiotic exposure | [44] |
Abbreviations: NA, not available
a
Age presented as mean (±S.D.).
b
Age presented as median (interquartile range).