To what extent can our intestinal microbiota predict the risk of developing diseases? In the case of multifactorial diseases, bacterial taxa alone do not provide much information, according to a study conducted at the Hospital das Clínicas of the Faculty of Medicine of the University of São Paulo (HCFMUSP), Brazil, and published this year in Gut Microbes. The situation changes, however, when other variables, mainly anthropometric and dietary factors, are considered together with this element.
Danielle Fonseca, a doctoral student in gastroenterology at the Faculty of Medicine of USP, carried out the study. The work was supervised by Dan Linetzky Waitzberg, MD, PhD, and Gabriel da Rocha Fernandes, MD, both of USP.
The researchers evaluated 202 adults, of whom 50 were healthy. The other 152 participants were routinely followed at HCFMUSP for one of the following conditions: Type 2 diabetes, type 1 diabetes, inflammatory bowel disease (Crohn’s disease and ulcerative colitis), plaque psoriasis, rheumatoid arthritis, and systemic lupus erythematosus (SLE).
The composition of the participants’ intestinal microbiota was analyzed through 16S RNA gene sequencing from stool samples. In addition, all individuals included in the research answered questions related to lifestyle habits, bowel patterns, medication use, and dietary habits.
Based on these data, the authors developed predictive models and evaluated the effectiveness of each. Models integrating phenotypic variables with intestinal microbiota taxa had a greater ability to distinguish healthy subjects from those with any disease.
In general, no microbial taxa had significant predictive capacity, according to author Fonseca, who is a specialist in clinical nutrition and pediatrics and now holds a doctorate in gastroenterology. “We found that taxa did not contribute more than 2% of the predictions, which was expected, given the multifactorial nature of the diseases evaluated,” she told the Medscape Portuguese edition.
Integrated Models
When only microbial variables were considered, the predictive model performed poorly for rheumatoid arthritis (area under the curve [AUC], 54.19) and for SLE (AUC, 49.08) and well for type 1 diabetes (AUC, 78.91) and type 2 diabetes (AUC, 72.65). But when phenotypic variables were integrated into the models, the prediction potential increased in the following cases: Rheumatoid arthritis (AUC, 88.03), SLE (AUC, 98.4), type 1 diabetes (AUC, 86.19), and type 2 diabetes (AUC, 96.96).
For type 1 diabetes, the improvement in predictive capacity was mainly associated with the incorporation of information about nutrient consumption, especially folate, cholesterol, zinc, magnesium, and protein. These data were also particularly important for improving the performance of prediction models for rheumatoid arthritis and SLE.
For type 2 diabetes, Fonseca noted that BMI contributed the most to the model’s predictive power.
Beyond Microbial Composition
The intestinal microbiota is currently being associated with all diseases, especially obesity, type 2 diabetes, and inflammatory bowel diseases, said Fonseca. With the reduction in sequencing costs, studying these relationships has become stronger.
Despite the prominence of this research, she added, the patterns of association between microbiota and diseases have not been reproducible.
“The discovery that bacteria alone do not play an important role in determining these diseases confirmed our hypothesis of the nonreproducibility of previous results. The “markers” pointed out by several studies are artifacts of generalist statistical analyses, and even if there is a difference in abundances, they do not have predictive power,” said Fonseca.
For the researcher, it is still necessary to evaluate a lot of data before developing an artificial intelligence system to predict the risk of developing diseases. Particularly in the case of intestinal microbiota, each experiment that identifies a “marker” actually identifies a “candidate” that needs to be tested, measured, and evaluated before being incorporated as a clinical tool, she explained.
Fonseca also emphasized the importance of focusing on patient care. “Our goal is to care, guide, and adapt the individual’s diet, not just to know the microorganisms that inhabit his body,” she said.
This story was translated from the Medscape Portuguese edition using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.
>>> Read full article>>>
Copyright for syndicated content belongs to the linked Source : Medscape – https://www.medscape.com/viewarticle/can-gut-microbiota-predict-multifactorial-disease-risk-2024a10004kq