Identifying characteristics of asymptomatic Barrett’s oesophagus and oesophageal cancer patients to inform novel prevention and early detection opportunities

Data analysis identified that 46% of Barrett’s oesophagus patients were classified without reflux symptoms using the definition where there was no reflux symptoms recorded in medical case note. Similarly, 40% of Barrett’s oesophagus patients were categorised as without reflux symptoms using the definition where no anti-reflux medication was prescribed prior to incident Barrett’s oesophagus diagnosis. In addition, characteristics were compared between the Barrett’s oesophagus patient groups with and without reflux symptoms.

Multivariable logistic regression analysis found that Barrett’s oesophagus patients without reflux are more likely to present with anaemia, gastrointestinal bleeding, weight loss, diabetes and oesophageal stricture, but less likely to present with epigastric pain and Barrett’s seen macroscopically during endoscopy. <br><br>Secondly, a prospective cohort study was conducted on all Barrett’ s oesophagus patients from the NIBR between 2009 and 2021 and matched to prescribing data from the NI Enhanced Prescription Database (n=16,283).

To classify Barrett’s oesophagus patients without reflux, the definition of never prescribed an anti-reflux medication or medication from the BNF Group 1.3 prescribed on the date of Barrett’ s oesophagus diagnosis or after was applied. Results demonstrated that 16% of the cohort were coded as “Barrett’s oesophagus patients without reflux” and the remaining 84% were symptomatic with reflux at the time of diagnosis. Barrett’s oesophagus patients without reflux were statistically more likely to be younger and male.

The risk of progression to oesophageal adenocarcinoma and high grade dysplasia was also reported for both symptom groups, concluding that there was no significant difference in progression with a rate of 0.13% per year for both Barrett’s oesophagus patients with reflux (95% CI: 0.10-0.16) and Barrett’s oesophagus patients without reflux (95% CI: 0.08-0.22), which highlights that screening or surveillance techniques for Barrett’ s oesophagus should not rely on reflux symptom profile for stratifying risk. <br><br>Thirdly, a cohort real-world data analysis was conducted utilising data from the OCCAMS consortium on over 4,000 oesophageal adenocarcinoma patients from the UK diagnosed between 2010 and 2021.

The aim was to detect an accurate proportion of those oesophageal adenocarcinoma patients without reflux symptoms at diagnosis. We found that 2,493 (82%) patients were coded as positive and 562 (18%) patients were coded as negative for reflux symptoms at oesophageal adenocarcinoma diagnosis. Oesophageal adenocarcinoma patients without reflux were more likely to be a non-White ethnicity, non-smokers, not overweight/obese, and not have diabetes than symptomatic patients. <br><br>In addition, to develop the strength of the previous study, a sub-group of oesophageal adenocarcinoma patients from the OCCAMS data that had completed a questionnaire were re-analysed, with the inclusion of self-reported exposure information.

It was found that 14% were coded as oesophageal adenocarcinoma patients without reflux symptoms with the self-reported reflux variable, but 27% were without reflux based on their clinical records reflux variable. Therefore, a key finding consisted of the inconsistencies found between the number of patients that were reflux negative or positive between the self-reported questionnaire and clinical notes. Associations between certain variables and the oesophageal adenocarcinoma patients without reflux symptoms were identified, such as, oesophageal adenocarcinoma patients without reflux were significantly associated with heavy alcohol consumption (≥14 units per week), however, due to the sample size of 346 oesophageal adenocarcinoma patients, these results are not statistically powered. <br><br>Finally, a systematic review with narrative synthesis was undertaken to examine the diagnostic accuracy of sponge or balloon non-endoscopic devices for sampling oesophageal tissue to detect Barrett’ s oesophagus and/or oesophageal cancer, which are currently being promoted for improving early detection, screening at-risk patients and triaging high risk patients for endoscopy.

Thirty-six studies were included, with thirty-one in the main outcome analysis and five secondary outcome studies. The results showed moderate to high reported sensitivities and specificities of sponge and balloon devices such as Cytosponge™ or, EsoCheck ®, with many of the studies not reporting the proportion of people that were offered and/or accepted the test. Additionally, only three sponge device studies targeted a population of patients without reflux symptoms.

This is the first systematic review to investigate the diagnostic accuracy of multiple non-endoscopic techniques for detecting oesophageal adenocarcinoma and/or Barrett’ s oesophagus. <br><br>In conclusion, there is a significant proportion of oesophageal adenocarcinoma and Barrett’ s oesophagus patients that are without reflux symptoms, meaning they will be excluded from current screening trials and diagnostic guidelines.

This highlights a need to consider expanding the eligibility criteria of ongoing screening trials. In addition, the studies identified distinct characteristic more likely in the patient groups without reflux symptoms, however, no clear trend was established, making it difficult to recommend inclusions to a risk-stratification model. For the emerging screening trials utilising non-endoscopic devices, the diagnostic accuracy of devices differ and our results reported heterogeneity between study’s methods of reporting results for each device, identifying a need to standardise the approach for evaluating the effectiveness of these devices. <br><br><i>Thesis is embargoed until 31 July 2027.</i><br>