Introduction Current BSG guidelines advise routine endoscopic surveillance for patients with Barrett’s oesophagus to enable early detection of oesophageal adenocarcinoma. However, evidence for improved outcomes from surveillance is weak and remains the subject of debate. To address these concerns, BOSS aims to compare the benefits of 2-yearly surveillance endoscopy against endoscopy on an ‘at need’ basis only. As a large, multicentre trial with 10 year follow-up, this ambitious project faced a number of challenges to recruitment, particularly the acceptability of randomising to a control arm with no routine endoscopy, especially in sites where surveillance had been established.

Methods Multi-centre randomised controlled trial (ISRCTN54190466). Inclusion criteria: patients over 18 with endoscopic and histologically proven Barrett’s oesophagus >1 cm. Exclusion criteria: patients unable to consent, unfit for endoscopy, high-grade dysplasia or cancer, or participation in AspECT trial. Target sample size: 3400. Recruitment: patients identified at local centres with new diagnosis, or existing diagnosis of Barrett’s and endoscopy within last 2 years. Follow-up will be for 10 years. Intervention arm will receive 2-yearly surveillance endoscopy, control arm will receive endoscopy on an ‘at need’ basis if symptomatic. Primary outcome: all cause mortality.

Results Recruitment began March 2009, target of 3400 was reached ahead of schedule in October 2011. 3469 consented to be randomised (58.4% of eligible screened patients): 1739 in 2-yearly surveillance, 1730 in ‘at need’ arm. 127 hospitals were open to recruitment, though 3 withdrew from the study, recruitment was closed in one centre by the trial team, and 11 did not recruit any patients. 3780 screened patients did not enter: 1309 ineligible, 2471 declined. Most common reasons for patients to decline were: preference for surveillance, preference not to have endoscopies and a family history of cancer. As of January 2014, 111 patients were no longer in their originally randomised treatment groups: 55 due to patient/clinician decision, 56 for other reasons. For 107 of those 111 patients, data continues to be collected.

Conclusion The successful recruitment to BOSS strongly supports the acceptability of randomisation to the control arm, and patients’ willingness to participate in a long follow-up RCT. The large number of centres and enthusiastic principal investigators have shown continued success, and great promise for BOSS to answer key questions on Barrett’s surveillance.

Reference Fitzgerald RC, et al. Gut 2014:63:7–42

Disclosure of Interest None Declared.