JELIS 2007.

Methods	Japan Eicosapentaenoic acid Lipid Intervention Study (JELIS) RCT, parallel, 2‐arm (EPA capsule vs nil), 5 years                                    Summary risk of bias: moderate or high
Participants	People with hypercholesterolaemia N: intervention, 9326, control 9319 (analysed intervention 9326, control 9319) Level of risk for CVD: moderate (Patients with hypercholesterolaemia) Men: 32% intervention, 31% control Mean age in years (SD): 61 (8) intervention 61 (9) control Age range: 40‐75 years Smokers: 20% intervention, 18% control                                                Hypertension: 36% intervention, 35% control                                                Medications taken by at least 50% of those in the control group: statins Medications taken by 20%‐49% of those in the control group: calcium channel blockers, other antihypertensives Medications taken by some, but less than 20% of the control group: beta‐blockers, antiplatelet, hypoglycemics, nitrates Location: Japan Ethnicity: Japanese
Interventions	Type: supplement (EPA capsule)    Comparison 1: EPA vs nil             Intervention: 3 × 2 × 300 mg capsules/d EPA ethyl ester (total dose of 1.8 g/d EPA), after meals. Dose: 1.8 g/d EPA Control: nothing (though all in both groups received "appropriate" dietary advice). All patients in both groups were on statins. Compliance: monitored by local physicians and measuring plasma fatty acids concentrations. Study drug regimens, 71% adhered EPA intervention, 73% adhered EPA control, 74% adhered statin Duration of intervention: maximum 5 years, mean 4.7 (1.1) years
Outcomes	Main study outcome: major coronary events Dropouts: 1766 intervention, 1582 control (but all had endpoint evaluation) Available outcomes: major coronary events: sudden cardiac death, fatal or non‐fatal MI, unstable angina, angioplasty or CABG. Also all‐cause mortality, stroke, peripheral artery disease, cancer, lipids, rise in blood sugar, fasting glucose, HbA1c Response to contact: no
Notes	Study funding: Mochida Pharmaceutical Company
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Statistical co‐ordination centre: "permitted block randomisation with a block size of 4"
Allocation concealment (selection bias)	Low risk	Centralised. Statistical coordinating centre (see above)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded as there was no placebo. Quote: "[o]pen label blinded end point"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"Clinical endpoints ... reported by local physicians were checked by members of a regional organizing committee in a blinded fashion. Then an endpoints adjudication committee ... confirmed them once a year without knowledge of the treatment allocation"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Well documented, dropout numbers low
Selective reporting (reporting bias)	Unclear risk	NCT00231738 registered October 2005, recruitment November 1996 to November 1999, main results published 2007. Rationale and design paper published in 2003 (reported baseline characteristics, so before completed follow‐up, but after data collection began). All reported outcomes appear to have been published.
Attention	Low risk	Slight, as no placebo provided to control group, but only capsules to intervention group. Otherwise 2 groups appeared to be treated equally
Compliance	Unclear risk	Monitored by local physicians and measuring plasma fatty acids concentrations. Study drug regimens,71% adhered EPA intervention, 73% adhered EPA control, 74% adhered statin
Other bias	Low risk	No further bias noted