EPIC‐2 2008

Methods	EPANOVA in Crohn's Disease, Study 2 (EPIC‐2) RCT, parallel, 2 arms (omega 3 vs MCT), 58 weeks Summary risk of bias: moderate or high
Participants	Adults with a confirmed diagnosis of Crohn's Disease and a Crohn's Disease Activity Index (CDAI) score < 150 who are responding to steroid induction therapy N:  intervention, 189, control 190 (187 intervention, 188 control analysed) Level of risk for CVD: low (people with quiescent Crohn's disease) Men: 48.1% intervention, 41.1% control Mean age in years (SD): 38.5 (13.8) intervention, 40.0 (13.6) years control Age range: > 16 years Smokers: 25.1% intervention, 37.2% control                                    Hypertension: unclear Medications taken by at least 50% of those in the control group: systemic corticosteroids – prednisolone, budesonide (but tapered and discontinued during the study) Medications taken by 20%‐49% of those in the control group: only reported for prior 12 months Medications taken by some, but less than 20% of the control group: only reported for prior 12 months Location: Canada, Europe, Israel, USA Ethnicity: not reported
Interventions	Type: supplement (capsule) Comparison: EPA + DHA vs SFA (medium chain triglycerides of short SFAs) Intervention: 2 × 2 1 g gelatin capsules omega‐3 free fatty acids (Epanova) providing total dose ˜2.2 g/d EPA, 0.8 g/d DHA. Dose: ˜3.0 g/d EPA + DHA                    Control: 2 × 2 1 g capsules medium chain triglyceride oil Compliance: measured by patient interviews and pill counts, 75.4% adhered intervention, 81.4% adhered control Length of intervention: mean 58 weeks
Outcomes	Main study outcome: maintain Crohns symptomatic remission Dropouts: 114 intervention, 112 control Available outcomes: mortality, CV events (nil), cancer diagnoses, adverse events Response to contact: yes (data provided)
Notes	Study funding: Tillotts Pharma, authors had extensive financial disclosures
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by number generator. Used a centralised randomisation procedure via interactive voice recognition system
Allocation concealment (selection bias)	Low risk	Centralised randomisation (see above)
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double blinding stated, identical capsule (slow‐release capsules). Neither investigator nor participant knew the allocation. However no information provided on capsules taste or smell
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Study states double‐blind but does not state that outcome assessors were blinded or provide a mechanism for this
Incomplete outcome data (attrition bias) All outcomes	High risk	Number of dropouts and reasons provided, however 114 of 189 in intervention group and 112 of 190 in control group terminated early.
Selective reporting (reporting bias)	High risk	NCT00074542. First received 2003, study start 2002. Published 2008.  Some outcomes, such as quality of life, stated in trials registry but not in published papers
Attention	Low risk	As investigators were blinded, attention bias was not possible.
Compliance	Unclear risk	Measured by patient interviews and pill counts, 75.4% adhered intervention, 81.4% adhered control
Other bias	Low risk	No further bias noted