Background The aim of fistula surgery is to eradicate the disease while preserving anal sphincter function. The efficacy of the Surgisis^® anal fistula plug (Cook Medical, Bloomington, IN, USA) in the treatment of trans-sphincteric fistula-in-ano has been variably reported. Objectives To undertake a randomised comparison of the safety and efficacy of the Surgisis anal fistula plug in comparison with surgeon’s preference for the treatment of trans-sphincteric anal fistulas. Design A randomised, unblinded, parallel-arm, prospective, multicentre clinical trial. Setting Hospitals in the UK NHS involving colorectal surgeons accredited by the Association of Coloproctology of Great Britain and Ireland. Participants Adult patients suffering from trans-sphincteric fistula-in-ano of cryptoglandular origin. Interventions Patients were randomised on a 1 : 1 basis to either the fistula plug or the surgeon’s preference [e.g. fistulotomy, cutting seton, advancement flap or ligation of intersphincteric fistula tract (LIFT) procedure]. Main outcome measures The primary outcome measure was quality of life as measured by the Faecal Incontinence Quality of Life (FIQoL) questionnaire at 12-month follow-up. Secondary outcome measures included clinical and radiological fistula healing rates, faecal incontinence rates, complications rates, reintervention rates and cost-effectiveness. Results Between May 2011 and March 2016, 304 participants were recruited (152 fistula plug vs. 152 surgeon’s preference). No difference in FIQoL score between the two trial groups was seen at the 6-week, 6-month or 12-month follow-up. Clinical evidence of fistula healing was reported in 66 of 122 (54%) participants in the fistula plug group and in 66 of 119 (55%) participants in the surgeon’s preference group at 12 months. Magnetic resonance imaging (MRI) showed fistula healing in 54 of 110 (49%) participants in the fistula plug group and in 63 of 112 (56%) participants in the surgeon’s preference group. Variation in 12-month clinical healing rates was observed: 55%, 64%, 75%, 53% and 42% for fistula plug, cutting seton, fistulotomy, advancement flap and LIFT procedure, respectively. Faecal incontinence rates were low at baseline, with small improvement in both groups post treatment. Complications and reinterventions were frequent. The mean total costs were £2738 [standard deviation (SD) £1151] in the fistula plug group and £2308 (SD £1228) in the surgeon’s preference group. The average total quality-adjusted life-years (QALYs) gain was much smaller in the fistula plug group (0.829, SD 0.174) than in the surgeon’s preference group (0.790, SD 0.212). Using multiple imputation and probabilistic sensitivity analysis, and adjusting for differences in baseline EuroQol-5 Dimensions, three-level version utility, there was a 35–45% chance that the fistula plug was as cost-effective as surgeon’s preference over a range of thresholds of willingness to pay for a single QALY of £20,000–30,000. Limitations Limitations include a smaller sample size than originally calculated, a lack of blinding that perhaps biased patient-reported outcomes and a lower compliance rate with MRI at 12-month follow-up. Conclusions The Surgisis anal fistula plug is associated with similar FIQoL score to surgeon’s preference at 12-month follow-up. The higher costs and highly uncertain and small gains in QALYs associated with the fistula plug mean that this technology is unlikely to be considered a cost-effective use of resources in the UK NHS. Future work Further in-depth analysis should consider the clinical and MRI characteristics of fistula-in-ano in an attempt to identify predictors of fistula response to treatment. Trial registration Current Controlled Trials ISRCTN78352529. Funding This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 23, No. 21. See the NIHR Journals Library website for further project information.