Lewis 2006.

Methods	Randomisation method: separate site prepared randomisation codes Blinding: double Allocation concealment: adequate Design: parallel‐group Number of women screened: 792 Number randomly assigned: 99 Number of dropouts: 12% (12/99): two in soy group (one could not accept Rx, one adverse event), five in flaxseed group (two could not accept Rx, one adverse event, one medical, one personal), five in placebo group (one could not accept Rx, one adverse event, two medical, one protocol violation) Number analysed: 87 Intention‐to‐treat analysis: no Power calculation: not reported Duration: 16 weeks Timing: not stated Location: Toronto and Calgary, Canada Funding: Canadian Institutes of Health Research
Participants	Inclusion criteria: 45 to 60 years old, natural menopause with last menses in previous one to eight years, Menoquol vasomotor score > 3.0 Exclusion criteria: medical or surgical menopause; inflammatory bowel disease; malabsorption syndrome; uncontrolled thyroid disorder; known allergy or intolerance to muffin ingredients or any serious and active medical or social condition likely to affect quality of life during the study; no ET in past three months; no phytoestrogens, steroids or antibiotics in past month Mean age, years: 53 Recruitment method: mailings to family practice, to previous participants of menopause workshops and to gynaecologists and family physicians
Interventions	Flaxseed muffins (50 mg/d lignans) Soy muffins (42 mg/d isoflavones) Placebo muffins (low levels of lignans and no isoflavones) Dose, duration and timing of administration: one muffin per day
Outcomes	Primary: Menoquol vasomotor score (0 to 6) Secondary: number of flushes per day Severity of flushes (0 to 6) Gastrointestinal side effects
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Separate site prepared the randomisation codes
Allocation concealment (selection bias)	Low risk	Separate site ensured concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Stated as double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Stated as double‐blind
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Moderate dropout—not balanced between groups
Selective reporting (reporting bias)	Unclear risk	Inadequate data on side effects