Mita 2007.

Methods	RCT, parallel, (EPA capsules vs nil), 2 years Summary risk of bias: moderate to high
Participants	Japanese type 2 diabetics N: intervention. 40, control: 41 (analysed 30, 30) Level of risk for CVD: moderate Men: 53% intervention, 67% control Mean age in years (SD): 59 (11.2) intervention 61.2 (8.4) control Age range: not reported Smokers: 40% intervention, 43% control Hypertension: not reported Medications taken by at least 50% of those in the control group: oral hypoglycemics Medications taken by 20%‐49% of those in the control group: insulin, lipid lowering drugs, antihypertensives Medications taken by some, but less than 20% of the control group: antithrombotics Location: Japan Ethnicity: 100% Japanese
Interventions	Type: supplement (EPA oil capsules) Comparison: EPA vs nil Intervention: 1800 mg/d EPA EPADEL capsules (Mochida Pharmaceutical Co Ltd Japan)‐ 98% pure ethyl‐ester EPA (unclear how many caps). Dose: ˜1.8 g/d EPA Control: no intervention Compliance: checked during 3 month reviews throughout trial and 5 participants were excluded for poor compliance but no details on method or results Length of intervention: mean 2.1 (0.2) years
Outcomes	Main study outcome: progression of diabetic macroangiopathy measured by carotid intima‐media thickness and brachial‐ankle pulse wave velocity Dropouts: 10 intervention, 11 control Available outcomes: BMI, lipids, BP, HbA1c, cancer diagnosis Response to contact: not yet attempted
Notes	Blood pressure data not used as groups are different at baseline Study funding: not stated
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients randomly divided into 2 groups matched for age and gender
Allocation concealment (selection bias)	Unclear risk	No details
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open label
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Assessors of main study outcomes were blinded to the treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Dropout rate (26%) over 2 years. All dropouts explained, however, 5 were excluded for poor compliance but no clear predefined protocol for exclusion
Selective reporting (reporting bias)	Unclear risk	No protocol
Attention	Low risk	All participants had the same contact
Compliance	Unclear risk	Compliance measured but no clear methods or reported results
Other bias	Low risk	None noted