P1228 The EISER Study: Identifying Microbial Factors Associated with Subclinical Gut Inflammation in Psoriatic Arthritis Patients Who Develop Inflammatory Bowel Disease

Abstract

Abstract Background Nearly 8% of patients with spondyloarthritis (SpA) manifest symptoms that are compatible with active inflammatory bowel disease (IBD). It is well established that microbial dysbiosis is associated with different immune-mediated diseases. We therefore hypothesized that the gut microbiome could be a modulator of intestinal inflammation and factor in the pathogenesis of IBD in SpA patients. Methods As part of the EISER study, we selected a subset of 193 subjects diagnosed with psoriatic arthritis (PsA), a form of SpA, who had no prior diagnosis of IBD. The presence of subclinical symptoms compatible with IBD was evaluated at time of enrollment. For patients with fCAL levels of 80 µg/g or higher, colonoscopy was performed and histological analysis of biopsies collected during the procedure were used to diagnose IBD. Capsule endoscopy was performed on those patients with a negative colonoscopy result. Stool samples were collected from all participants, and microbial DNA was extracted for shotgun metagenomic sequencing using the Illumina HiSeq platform. Sequenced data was processed using MetaPhlAn 4 to estimate taxonomic composition, MMEDS for overall microbiome analysis, and R for additional statistical analyses. Results Twenty-five out of 193 subjects in our cohort (12.95%) had symptoms that were compatible with subclinical IBD (sIBD) and five patients (2.6% of total) were confirmed to have clinical IBD (cIBD). The use of nonsteroidal anti-inflammatory drugs (NSAIDs) resulted in significant differences in overall alpha (Shannon entropy, p=0.03) and beta microbial diversity (PERMANOVA, p=0.043), while the use of proton-pump inhibitors (PPI) resulted in significant differences in beta diversity (PERMANOVA, p= 0.036). Patients with cIBD had significantly higher abundance of taxa including Prevotella copri and Prevotella stercorea. P. copri, P. stercorea and fCAL were all significantly elevated in cIBD patients (Wilcoxon’s p=0.0278, 0.0007, 0.0161), although with a larger effect size in P. copri and P. stercorea (Cohen’s D: P. copri = 1.074; P. stercorea = 1.269; fCAL= 0.629). fCAL was also significantly elevated in patients who used PPIs (p= 4.456e-11), while P. copri and P. stercorea abundances were not impacted (Wilcoxon’s p: P. copri = 0.2394; P. stercorea = 0.0643). Conclusion Microbiome composition was impacted by the use of NSAIDs and PPI. We identified taxa differentially enriched in cIBD patients, including Prevotella copri and P. stercorea. Importantly, while fCAL was significantly higher in cIBD patients, its discriminatory power was not as strong as that of P. copri and P. stercorea, suggesting that these taxa could be potentially used as biomarkers of co-morbid IBD in patients with PsA.