BackgroundThe omega-3 polyunsaturated fatty acid eicosapentaenoic acid (EPA) and aspirin both have proof of concept for colorectal cancer (CRC) chemoprevention, aligned with an excellent safety profile.ObjectivesThe objectives were to determine whether or not EPA prevents colorectal adenomas, either alone or in combination with aspirin, and to assess the safety/tolerability of EPA, in the free fatty acid (FFA) form or as the triglyceride (TG), and aspirin.DesignThis was a randomised, blinded, placebo-controlled, 2 × 2 factorial trial.SettingThe NHS Bowel Cancer Screening Programme (BCSP).ParticipantsPatients (aged 55–73 years) identified as ‘high risk’ (i.e. those who have five or more colorectal adenomas of < 10 mm in size or three or more colorectal adenomas if one is ≥ 10 mm in size) at screening colonoscopy.InterventionsThe interventions were capsules containing 2000 mg of 99% EPA–FFA or 2780 mg of 90% EPA–TG (equivalent to 2000 mg of FFA) taken daily, or identical placebo capsules; and 300 mg of aspirin taken daily, or an identical placebo, enteric-coated tablet. Both were taken for ≈1 year until surveillance colonoscopy. All participants and staff were unaware of treatment allocation.Main outcome measuresThe primary outcome was the number of participants with one or more colorectal adenomas [adenoma detection rate (ADRa)] at surveillance colonoscopy. Outcomes were analysed for all participants with observable follow-up data by an ‘at-the-margins’ approach, adjusted for BCSP site and by the need for repeat baseline endoscopy. Secondary outcome measures – these included the number of colorectal adenomas per patient [mean adenomas per patient (MAP)], ‘advanced’ ADRa and colorectal adenoma location (right/left) and type (conventional/serrated).ResultsBetween November 2011 and June 2016, 709 participants were randomised, with 707 providing data (80% male, mean age 65 years). The four treatment groups (EPA + aspirin,n = 177; EPA,n = 179; aspirin,n = 177; placebo,n = 176) were well matched for baseline characteristics. Tissue EPA levels and tolerability were similar for FFA and TG users. There was no evidence of any difference in ADRa between EPA users (62%) and non-users (61%) [risk difference –0.9%, 95% confidence interval (CI) –8.8% to 6.9%] or for aspirin users (61%) versus non-users (62%) (risk difference –0.6%, 95% CI –8.5% to 7.2%). There was no evidence of an interaction between EPA and aspirin for ADRa. There was no evidence of any effect on advanced ADRa of either EPA (risk difference –0.6%, 95% CI –4.4% to 3.1%) or aspirin (risk difference –0.3%, 95% CI –4.1% to 3.5%). Aspirin use was associated with a reduction in MAP [incidence rate ratio (IRR) 0.78, 95% CI 0.68 to 0.90), with preventative efficacy against conventional (IRR 0.82, 95% CI 0.71 to 0.94), serrated (IRR 0.46, 95% CI 0.25 to 0.87) and right-sided (IRR 0.73, 95% CI 0.61 to 0.88) lesions, but not left-sided (IRR 0.85, 95% CI 0.69 to 1.06) adenomas. There was evidence of chemopreventive efficacy of EPA on conventional (IRR 0.86, 95% CI 0.74 to 0.99) and left-sided (IRR 0.75, 95% CI 0.60 to 0.94) adenomas, but not on total MAP (IRR 0.91, 95% CI 0.79 to 1.05) or serrated (IRR 1.44, 95% CI 0.79 to 2.60) or right-sided (IRR 1.02, 95% CI 0.85 to 1.22) adenomas. EPA and aspirin treatment were well tolerated, with excess mild/moderate gastrointestinal (GI) adverse events (AEs) in the EPA alone group. There were six GI bleeding AEs.ConclusionEPA and aspirin treatment were not associated with a reduction in ADRa. However, both agents displayed evidence of chemopreventive efficacy, based on adenoma number reduction, which was specific to adenoma type and location, and is compatible with known anti-CRC activity of aspirin.LimitationsLimitations of the trial included the failure to recruit to the target sample size of 853, and an unexpected switch of EPA formulation mid-trial.Future workA future objective should be to understand the mechanism(s) of action of EPA and aspirin using the trial biobank. Established trial infrastructure will enable future trials in the BCSP.Trial registrationCurrent Controlled Trials ISRCTN05926847.FundingThis project was funded by the Efficacy and Mechanism Evaluation (EME) programme, a MRC and NIHR partnership.