Dear Editor:
Since gut represents the largest immunological organ and its resident microbiota are well-characterized to significantly orchestrate host immune responses, there were a growing interest in interrogating the potential impact of the gut microbiota on host susceptibility to SARS-CoV-2. Various studies have confirmed that alterations in gut microbiota composition are significantly associated with COVID-19 progression and severity of illness.1,2 However, these studies to date have not completely eliminated confounding factors, thereby establishing a definitive causal relationship between COVID-19 and the gut microbiota. Mendelian randomization (MR), as a statistical methodology utilizing genetic variations as instrumental variables, has increasingly adopted as a promising tool for revisiting causal relationships between exposures and outcomes, with the controlling for confounding factors. MR studies can be conducted by using large-scale genetic data from biobanks, which can provide high statistical power and reduce the potential for bias. Mendelian randomization has been leveraged to investigate the causal effects of multiple influential factors on outcomes and severity of COVID-19.3-6 These findings hold tremendous potential to inform public health policies and interventions aimed at mitigating the burden of COVID-19.
This study aims to explore the directional relationship between gut microbiota and the severity of COVID-19 using MR method. The genetic variants correlated with exposure, the gut microbiota composition detected in fecal sample, were sourced from the largest genome-wide meta-analysis to date, conducted by the MiBioGen consortium.7 The GWAS data of COVID-19 patients were obtained from the latest r7 version based on COVID-19 Host Genetics Initiative data (HGI) (https://www.covid19hg.org/results/r7/). According to the HGI standard, the outcomes were stratified into three distinct categories (mild, moderate and severe). In MR analysis, various methods were applied, including inverse-variant weighted, MR-Egger, weighted mode, weighted median, and simple median. Only SNPs with a significant threshold (p <5×10-8) were considered for instrumental variable selection. Pleiotropy and heterogeneity were assessed via MR-Egger, Cochran’s Q test and ”Leave-one-out” sensitivity analysis (Figure 1). All data used in the current study were publicly accessible.
The results of MR analysis revealed a causal relationship between certain genera of gut microbiota and the severity of COVID-19, as shown in Table 1 . Notably, increased composition of Dorea at genetic level was positively correlated with all outcomes, whileBifidobacterium was related to mild and moderate outcomes, and the Rikenellaceae RC9 gut group was linked to moderate and severe outcomes (Supplementary Figure 1). After performing sensitivity analysis, we have demonstrated that the causal pathway was not influenced by any potentially influential SNPs, thereby validating the robustness of our findings. Additionally, MR‒Egger was performed as well, in which no potentially significant horizontal pleiotropy or outliers were detected (Supplementary Figure 2-5). To further summarize the evidence and verify the results of MR, we conducted a meta-analysis incorporating the original data included in our study. The overall effect sizes were quantified and no significant evidence of heterogeneity was found (Supplementary Figure 6).
Since previous studies concerning the relationship between gut microbiota and COVID-19 susceptibility were limited by a lack of causal relationships, small sample sizes, insufficient evidence due to the timing of fecal sample collection and inability to rule out confounding factors, our study represents the first time that MR has been utilized to confirm the causal impact of specific genera of gut microbiota on the severity of COVID-19. Further investigation into the specific bacteria associated with severity of illness among COVID-19 patients is also warranted. For instance, Dorea and Bifidobacterium are crucial producers of short-chain fatty acids (SCFAs), which have been found to have potent immunomodulatory properties, along with the ability in mediating CD8+ T cell immune response.8 Elevated levels of Rikenellaceae RC9 gut group have been shown to modulate TLR signaling pathways, thereby increasing the susceptibility of the gut to inflammation.9 Bacteroides play a role in regulating host immunity and suppressing colonic ACE-2 expression. Furthermore, an inverse correlation has been observed betweenBacteroides levels and the severity of COVID-19 infection.1 An increased abundance of intestinalPrevotella is associated with adverse outcomes in several inflammatory disorders, such as rheumatoid arthritis, ankylosing spondylitis, as well as systemic immune activation in HIV patients.10 Hence, exploring the microbiota that is intricately linked with COVID-19 has the potential to unravel the exact pathophysiological mechanisms that underpin SARS-CoV-2 infection.
In conclusion, our results implied that targeted screening of gut microbiota compositions in fecal samples for susceptible populations could become an effective yet feasible strategy for preventing moderate to severe cases of COVID-19, paving the way for more personalized treatments.