Background The NIHR By-Band Study aims to compare the effectiveness of adjustable gastric band (AGB) and Roux-en-Y gastric bypass (RYGB) for severe and complex obesity. By-Band had an internal pilot phase to establish trial recruitment and to review whether expansion into a three group trial including sleeve gastrectomy was appropriate. This paper describes how the By-Band internal pilot informed successful adaptation to By-Band-Sleeve.

Methods The internal pilot phase started in two centres in November 2012. After 24.4 centre months of recruitment, data on the pre-agreed decision criteria were reported to the trial funder to inform progression to the main trial. Criteria focused on rates of referrals for surgery, trial eligibility, recruitment and cross over. Evidence from reported and open RCTs including sleeve gastrectomy (SG) was scrutinised and data on SG practice summarised.

Results In two centres 333 patients were screened, 231 (69%) found eligible and 80 consented to randomisation (35%). Cross over rates were 3.5% with similar numbers crossing over from AGB to RYGB and vice versa. The updated Cochrane review of bariatric surgery identified 22 trials of AGB, RYBG or SG. All trials were at a moderate or unclear risk of bias and external validity poor. There was only one trial comparing SG with AGB. Current open trials were also at risk of bias. UK practice showed that whilst AGB was decreasing in incidence in the public sector it remained the commonest procedure undertaken in the reported private cases. An adapted three-group study was successfully proposed to the NIHR and a contract variation agreed in January 2015 to increase the sample size to 1341 to allow comparisons of all three procedures.

Conclusion The study has now recruited 176 patients and is expanding into 7 more centres (total 12). By-Band-Sleeve will be therefore provide high quality evidence to examine which operation is most effective and cost effective for the treatment of severe and complex obesity.

Disclosure of interest None Declared.