A175: Dysbiosis of the Enteric Microbiota in Pediatric Spondyloarthritis

Abstract

Background/Purpose:Spondyloarthritis (SpA) affects 1% of the population in the United States. Studies have shown altered flora (dysbiosis) in a variety of inflammatory conditions, including inflammatory bowel disease. We hypothesized that we would identify dysbiosis in pediatric SpA.Methods:Patients were children with enthesitisrelated arthritis. Controls were either healthy children from the community or subjects seen in our clinic and diagnosed with non‐inflammatory causes of joint pain. Stool specimens were collected at home and shipped overnight via commercial carrier. Purified DNA underwent PCR amplification using primers designed to the conserved region flanking the variable IV region from the 16S rDNA gene. The PCR products were run on the miSeq. Analysis was performed with the Quantitative Insight into Microbial Ecology (QIIME) suite. Pairwise comparisons of individual taxa were performed with the Mann‐Whitney U‐test, and multiple‐group comparisons were performed with the non‐parametric Kruskal‐Wallis test. Nominal data was evaluated with the Chi‐Squared or Fisher's exact test.Results:28 children with SpA age (median, range) 14, 7–18 and 13 controls age 13, 5–17 were included. 3 of the subjects had IBD. 21 patients and no controls were on immunosuppressive therapy. The SpA patients had significantly less faecalibacterium (11% vs 4.8%, p = 0.006) and a non‐significant increase in bacteroides (19% vs 11%, p = 0.216). Additionally, akkermansia was present in low quantities in most patients, but was relatively abundant (>2%) in 0/13 controls compared to 8/28 patients (p = 0.040.) Principal coordinates analysis (PCoA; Figure ) demonstrated that while most of the patients clustered with the controls, 10 patients formed their own cluster. Phylogenetic tree analysis (Figure ) likewise demonstrated a small cluster of 10 subjects (9 patients and 1 control) who formed their own cluster separate from other subjects, 8 of whom corresponded to the indicated cluster from the PCoA. These 8 subjects demonstrated significantly increased bacteroides (p < 0.001) and decreased faecalibacterium (p = 0.022) compared to the remainder of the subjects. In contrast, none the patients in this cluster, compared to 8/20 who clustered with the controls, had elevated levels of akkermansia (p = 0.005, Chi‐Squared). Output of Principal Coordinates Analysis (PCoA) of 16S sequencing data. Spondyloarthritis patients are shown in blue, controls in red.imageOutput of unweighted Pair Group Method with Arithmetic Mean (UPGMA) analysis of 16S sequencing data. Controls are indicated with asterisks.imageConclusion:We demonstrate decreased faecalibacterium among patients with SpA, consistent with a suspected anti‐inflammatory effect of f. prausnitzii. We also introduce akkermansia as a possible pathogenetic species in a subset of SpA patients.