MAPT 2017.

Methods	Multidomain Alzheimer Preventive Trial (MAPT) 4 arms RCT, parallel, (n‐3 ± multidomain intervention vs placebo ± multidomain intervention), 36 months Summary risk of bias: low
Participants	Population: people aged at least 70 years without dementia but with memory complaint, IADL limitation or slow gait speed N: 840 intervention (arms 1 and 3), 840 control (arms 2 and 4) randomised. Numbers analysed differ by outcome. Level of risk for CVD: low Men: 37.2% intervention, 34.5% control. (combined groups) Mean age in years (SD): 75.6 (4.7) and 74.4 (4.4) intervention, 75.1 (4.3) and 75 (4.1) control Age range: not reported Smokers: not reported Hypertension: not reported Medications taken by at least 50% of those in the control group: not reported Medications taken by 20%‐49% of those in the control group: not reported Medications taken by some, but less than 20% of the control group: not reported Location: France and Monaco Ethnicity: not reported
Interventions	Type: supplement (capsule) Comparison: EPA + DHA vs paraffin oil (non‐fat) Intervention Arm 1: omega‐3 (V0137 CA 800 mg/d DHA; 225 mg/d EPA in soft caps). Dose for arms 1 and 3: 1.025 g/d EPA + DHA Arm 3: omega 3 (V0137 CA 800 mg/d DHA; 225 mg/d EPA in soft caps) plus multi‐domain intervention (nutrition, physical exercise, cognitive stimulation, social activities) Control: Arm 2: placebo capsules containing flavoured paraffin oil. All capsules were supplied by Pierre Fabre Médicament (Castres, France) Arm 4: placebo capsules plus multi‐domain intervention (nutrition, physical exercise, cognitive stimulation, social activities) Compliance: adherence to study interventions was assessed every 6 months. For supplementation, adherence was assessed by counting the number of capsules returned by participants (or based on treatment dates if the number of capsules was missing). Furthermore, biological samples were obtained at baseline and after 12 months to assess concentrations of DHA and EPA in red blood cell membranes. Duration of intervention: 36 months
Outcomes	Main study outcome: change in cognitive function ) Dropouts: 200 intervention, 194 control Available outcomes: mortality, CVD events, haemorrhagic stroke, adverse events, functional capacity, other cognitive functions, safety and tolerability Response to contact: no
Notes	Study funding: Gérontopôle of Toulouse, the French Ministry of Health (PHRC 2008, 2009), the Pierre Fabre Research Institute (manufacturer of the polyunsaturated fatty acid supplement), Exhonit Therapeutics, and Avid Radiopharmaceuticals
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned (1:1:1:1) to the combined intervention (i.e. the multidomain intervention plus polyunsaturated fatty acids), the multidomain intervention plus placebo, polyunsaturated fatty acids only, or placebo only. A computer‐generated randomisation procedure (done by ClinInfo, a subcontractor) was used with block sizes of 8 and stratification by centre.
Allocation concealment (selection bias)	Low risk	A clinical research assistant, who was not involved in the assessment of participants, used a centralised interactive voice response system to identify which group to allocate the participant to, and which lot number to administer.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All participants and study staff were blinded to polyunsaturated fatty acid or placebo assignment – both sets of capsules looked and tasted identical. In view of the nature of the multidomain intervention, the study was unblinded for this component, but the independent neuropsychologists who were trained to assess cognitive outcomes were blinded to group assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	All participants and study staff were blinded to polyunsaturated fatty acid or placebo assignment—both sets of capsules looked and tasted identical. In view of the nature of the multidomain intervention, the study was unblinded for this component, but the independent neuropsychologists who were trained to assess cognitive outcomes were blinded to group assignment. Data analysts were not blinded to group assignment, but two data managers, one statistician (CC) and two physicians (SA and BV) did a blinded data review.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1680 participants were enrolled and randomly allocated, the modified intention‐to‐treat population (N = 1525), i.e. 155 excluded (9% over 3 years)
Selective reporting (reporting bias)	Low risk	Protocol registered ClinicalTrials.gov (NCT00672685) – outcomes match report. Because of advances in the field since our trial was designed in 2007, we decided to modify the primary outcome from one cognitive test to a composite cognitive score, which is now thought to be a better endpoint. This protocol amendment was submitted to the local ethical committee on 2 February 2015 and was subsequently approved
Attention	Low risk	Both groups assessed at baseline, 6, 12, 24, 36 months. Groups 1 and 2 only differed by content of capsules.
Compliance	Unclear risk	Adherence to study interventions was assessed every 6 months, by counting the number of capsules returned (or based on treatment dates if the number of capsules was missing). Biological samples were obtained at baseline and after 12 months to assess concentrations of DHA and EPA in red blood cell membranes, but outcomes not reported.
Other bias	Low risk	None noted