Ultimate magazine theme for WordPress.

Intestine and oral microbiomes predict COVID-19 severity

One question that remains unanswered during the ongoing coronavirus disease 2019 (COVID-19) pandemic was why there is a distinct hit-and-miss pattern in major illnesses. An interesting new study that appears on the medRxiv * preprint server shows a highly relevant risk factor: the condition of the oral and intestinal microbiome.

The spring wave of COVID-19 has filled many hospitals and intensive care units with patients gasping for air. Often this number is checked for approval. The need for efficient and reliable risk biomarkers has never been so great.

Study details

The 4C mortality score was described by the International Consortium on Severe Acute Respiratory Tract and Emerging Infections (ISARIC) and the World Health Organization (WHO) in September 2020 to address this need. This far-reaching risk assessment system includes eight variables including age, gender, other pre-existing illnesses, level of consciousness, oxygen saturation in the peripheral blood and C-reactive protein (CRP).

With this information, however, the accuracy of the prediction was only 79%, with 30% of patients at high risk of mortality being missed. This led to an attempt to predict the risk of death using a different method. The researchers in the current study took advantage of the fact that gut microbiomes in COVID-19 patients have severe disorders known as dysbiosis. 23 families of bacteria were specifically linked to disease severity in patients hospitalized with COVID-19.

Using computational and analytical tools, the scientists set up a robust framework to track the networks of connections between microbiota, clinical features and disease severity. They found that Enterococcus, a type of oral and intestinal bacteria, can reliably predict a fatal outcome in these patients.

This small study included 69 COVID-19 patients with moderate to severe symptoms, i.e. those who needed less than or more than 4 liters of oxygen. Of these, 63 had complete medical records. The baseline clinical features were similar in both groups, severe and moderate. Severe patients had to stay in hospital an average of six days longer than moderately ill patients.

When analyzing the comorbid data, the researchers found that a combination of clinical variables, including the severity of COVID-19, had an accuracy of 89% in predicting a fatal outcome. In fact, the need for 4 liters of oxygen was the main factor in predicting such an outcome. If the severity of the disease was not taken into account, the accuracy dropped to 84%. This finding shows that respiratory symptoms are important in predicting COVID-19 outcomes.

Death from COVID-19 is predicted by the severity of respiratory symptoms and other comorbidities commonly used in patient triage.  (A) Area Under the Curve - Receive Operating Curve (AUC-ROC) for a one-time cross-validation to assess the prediction of accuracy of death with COVID-19.  Red lines correspond to the model including all clinical covariates (CC), black lines correspond to the model including all clinical covariates with the exception of disease severity (CC, no severity).  (B) Covariates selected according to the Random Forest Classification model were classified according to their importance for the classification of death as a disease outcome.  (C) For categorical covariates (Yes = 1,

Death from COVID-19 is predicted by the severity of respiratory symptoms and other comorbidities commonly used in patient triage. (A) Area Under the Curve – Receive Operating Curve (AUC-ROC) for a one-time cross-validation to assess the prediction of accuracy of death with COVID-19. Red lines correspond to the model including all clinical covariates (CC), black lines correspond to the model including all clinical covariates with the exception of disease severity (CC, no severity). (B) Covariates selected according to the Random Forest Classification model were classified according to their importance for the classification of death as a disease outcome. (C) For categorical covariates (Yes = 1, No = 0), the number of patients out of the 63 included in the analyzes within a particular category was colored by outcome (Survived, in blue; Died, in red). (D) For numeric variables, whisker charts (median, box interquartile range, 5th and 9th percentiles for lines) are used, with each solid point corresponding to a single patient. (BH adjusted p-value <0.05)

Stool or oral microbiome predict the severity

It is known that a viral infection of the lungs has a long-term impact on the gut microbiome. The researchers therefore used this knowledge to predict the severity of COVID-19 and relate it to other common measures. They tested the effect of using clinical variables only, intestinal microbiome composition only, oral microbiome composition only, the first two combined, and the first and third combined.

They found that the accuracy of the first model was ~ 76%. Again, the comorbidities that best predicted disease severity were such as high cholesterol, Latino race, coronary artery disease, asthma, obesity, breathing difficulties associated with hypoxia, rapid respiratory rate, number of days in hospital, thrombosis, and male gender.

Using the second and third models with the stool or mouth microbiota as predictors, they found accuracies of 92% and 84%, respectively. This is an improvement in accuracy of 122% and 111%, respectively.

The combined models showed the highest predictive accuracy at 96%, suggesting that the oral or gut microbiota is better able to predict COVID-19 severity. Further analyzing the microbiota, the researchers found a type of indicator that can be cultivated in the clinical laboratory.

Top predictor

The three main types of bacteria used to predict the severity of COVID-19 in the gut microbiome were Bacteroides uniformis, Enterococcus faecalis, and Monoglobus pectinilyticus, while those from the oral microbiome were Porphyromonas endodontalis, Veillonella tobetsuensis, and Bifidobacterium breve.

Enterococcus faecalis bacteria known as Streptococcus faecalis.  These bacteria are rounded or oval cocci that typically form cell chains here.  Image Credit: Shutterstock

Enterococcus faecalis bacteria known as Streptococcus faecalis. These bacteria are rounded or oval cocci that typically form cell chains here. Image Credit: Shutterstock

Directional analysis showed that a decrease in the incidence of Enterococcus faecalis and Porphyromonas endodontalis in the intestines and mouth, respectively, in moderately ill COVID-19 patients, or an increase in the incidence of these pathological species in critically ill patients, were the best predictors of severe COVID-19 .

Predictors of moderate COVID-19 were an increase in the incidence of Bacteroides fragilis, Bacteroides caccae, and Clostridium clostridioforme in the stool or Muribaculum intestinal in the mouth.

They could not find a correlation between the number of bacteria of a species and the antibody titers, although higher anti-RBD IgG levels do correlate with survival. This may mean that microbiota and IgG levels are independent predictors of serious outcomes.

Conclusion

“In this study, we have shown that the composition of the stool or oral microbiome can predict the severity of COVID-19 disease with greater accuracy than traditional clinical assessment methods. In particular, two pathobionts in the oral (Porphyromonas endodontalis) or intestinal (Enterococcus faecalis) microbiota can serve as indicator species to reliably predict the severity of SARS-CoV-2 infections. “

This could lead to better risk stratification of patients, especially since Enterococcus faecalis is easy and inexpensive to cultivate. This could help provide earlier support to patients who are likely to develop a fatal disease. The researchers are calling for this bacterium to be included in clinical risk stratification in healthcare.

The severity of the disease is related to uncontrolled inflammation. This could be due to an intestinal dysbiosis that has been burdened with several chronic inflammatory conditions. This area requires further research, particularly to understand the role of regulatory T cells (Tregs), which are responsible for immune modulation under normal circumstances but can be abnormally expressed in COVID-19.

Such studies could help to find out how “the dysbiosis in SARS-CoV-2 infected patients and in particular the accumulation of the pathobionts observed in this cohort can contribute to the severity of COVID-19 disease by changing the Treg development”.

* Important NOTE

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be considered conclusive, guide clinical practice / health-related behavior, or be treated as established information.

Comments are closed.