Dysken 2014

Methods	Design: multicentre, randomised, double‐blind, placebo‐controlled trial Duration: 6 months to 4 years
Participants	Diagnosis (including criteria used): possible or probable AD (NINCDS‐ADRDA) Number of participants: 613 total randomised (completers: 140 in vitamin E group, 140 in placebo group, 142 in memantine group, 139 in the vitamin E + memantine group) Age: 53 to 96 years, mean age 78.8 years Gender: 97% men Cognitive status (e.g. MMSE): 21 Ethnicity: 86% white Inclusion criteria: diagnosis of possible or probable AD (NINCDS‐ADRDA); presence of a carer (friend or relative); informed consent; MMSE 12 to 26; administration of maintenance dosage of acetylcholinesterase inhibitors for at least 4 weeks; agreement not to take vitamin E/memantine outside the study. Exclusion criteria: non‐Alzheimer's primary dementia; current major depression, delirium, alcohol or psychoactive substance abuse or dependency; schizophrenia, or delusional disorder as defined by DSM‐IV; presence of any uncontrolled systemic illness that would interfere with participation in the study or life expectancy of < 1 year; pregnant or intention to become pregnant; enrolment in another interventional clinical trial; current prescription with more than one acetylcholinesterase inhibitor; current prescription for warfarin; use of vitamin E supplements in the past 2 weeks; use of memantine in the past 4 weeks or known intolerance; estimated creatinine clearance < 5 mL/minute; use of amantadine in the past 2 weeks.
Interventions	Treatment 1: vitamin E (2000 IU total daily dose divided into 2 doses) Treatment 2: memantine (20 mg/day) Treatment 3: vitamin E + memantine Treatment 4: placebo
Outcomes	ADCS‐ADL (months 0, 6, 12, 18, 24, 30, 36, 42, 48) MMSE (months 0, 6, 12, 18, 24, 30, 36, 42, 48) ADAS‐Cog (months 0, 6, 12, 18, 24, 30, 36, 42, 48) DS (months 0, 6, 12, 18, 24, 30, 36, 42, 48) NPI (months 0, 6, 12, 18, 24, 30, 36, 42, 48) Caregiver Activity Survey (months 0, 6, 12, 18, 24, 30, 36, 42, 48) Adverse events
Notes	On average, carers reported that vitamin E was taken on 65% of the days.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants randomised centrally using "...a random permuted block design of randomly varying sizes between 4 and 12."
Allocation concealment (selection bias)	Low risk	"…participants were randomized centrally into one of the four treatment groups…"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"The patient, caregivers, and all site investigators were blinded to the treatment assignment." Participants received "matching placebos for vitamin E were hard‐gelatin, liquid‐filled capsules containing soybean oil." The authors did not describe whether blinding was successful.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	"The patient, caregivers, and all site investigators were blinded to the treatment assignment." The authors did not describe how blinding was maintained and whether it was successful.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Incomplete outcome data was comparable between the vitamin E group and placebo group (12 participants excluded in each due to a lack of follow‐up data).
Selective reporting (reporting bias)	Low risk	Study protocol was published, which included the primary and secondary outcomes as included in the main paper.
Other bias	Low risk	No other sources of bias noted.