Introduction Minimally invasive methods for sampling the oesophageal microbiota may be clinically useful for detection and risk stratification of patients with Barrett’s oesophagus. The Cytosponge is a non-endoscopic device that samples epithelial cells and bacteria along the length of the oesophagus. This study aimed to determine the utility of the Cytosponge for sampling the oesophageal microbiota, and to assess whether the device could detect altered microbiota in Barrett’s oesophagus.

Methods 251 samples from 67 patients and relevant negative controls underwent 16 S rRNA gene sequencing (in duplicate) on the Illumina MiSeq platform. A subset of Cytosponge samples (N = 31) was compared with matched fresh frozen endoscopic biopsies (N = 31), endoscopic brushings (N = 31) and throat swabs (N = 13). The microbial DNA yield for the different oesophageal sample types was analysed using quantitative PCR for overall bacterial abundance. 16 S rRNA gene data was further interrogated from patients representing the Barrett’s progression sequence, including normal squamous oesophagus (N = 20), Barrett’s oesophagus (N = 24), and high grade dysplasia (N = 23).

Results 16 S rRNA gene data showed the majority of genera overlapped between Cytosponge samples and endoscopic brushes and biopsies (84%, cut-off >0.0001% abundance; 100%, cut-off >0.1% abundance), but the Cytosponge samples were enriched for genera known to be prevalent in the stomach and oral cavity such as Campylobacter and Fusobacterium, respectively (q < 0.05, Metastats). The Cytosponge yielded higher quantity of microbial DNA in comparison to endoscopic brushes or endoscopic biopsies (P < 0.001). Furthermore, the Cytosponge detected decreased diversity in patients with high grade dysplasia, in comparison to normal squamous controls (P < 0.05), suggesting that some taxa may have a competitive advantage over others in this disease state. There was also a trend towards decreased diversity in non-dysplastic Barrett’s oesophagus compared to normal squamous controls, which reached significance for the Shannon index and inverse Simpson index (P < 0.05). There was no difference in microbial diversity between non-dysplastic Barrett’s oesophagus and high grade dysplasia.

Conclusion The Cytosponge was capable of sampling the microbiota in the oesophagus and proximal stomach, and detected decreased diversity in patients with high grade dysplasia compared to normal squamous oesophagus.

Disclosure of Interest D. R. Fels Elliott: None Declared, A. Walker: None Declared, M. O’Donovan: None Declared, J. Parkhill: None Declared, R. Fitzgerald Grant/research support from: RCF developed the Cytosponge™ technology with MRC-Technology, which provided devices for research. The Cytosponge™ has recently been licensed to Covidien, and RCF has no direct financial relationship with Covidien.