Key clinical point: Nonresponse to methotrexate (MTX) therapy can be predicted based on the gut microbiome of an individual patient newly diagnosed with rheumatoid arthritis (RA).

Major finding: A model developed using machine learning predicted 83.3% of patients who did not respond to MTX and 78% of MTX responders.

Study details: An analysis of DNA from fecal samples obtained from a training cohort of 26 patients with new-onset RA (NORA), a validation cohort of 21 patients with NORA, and a control group of 20 patients with RA.

Disclosures: This study was funded by the National Institutes of Health, the Rheumatology Research Foundation, the Searle Scholars Program, various funds from the Spanish government, the UCSF Breakthrough Program for Rheumatoid Arthritis-related Research, and the Arthritis Foundation Center for Excellence. Four authors report consultancies and memberships on scientific advisory boards with pharmaceutical and biotechnology companies that do not overlap with the current study.