Nahas 2007.

Methods	Design: parallel‐group Number randomly assigned: 80 Number dropped out: four (two in each group—reasons given) Number analysed: 76 Intention‐to‐treat analysis: no Power calculation: not reported Duration: 10 months Timing: not stated Location: Climacterium Outpatient Service of the Department of Gynecology in Sao Paolo University Funding: Ativus Farmaceutica Brazil and Fundacao Lucentis de Apoio a Cultura, Ensino, Pesquisa e Extensao
Participants	Inclusion: postmenopausal, 45 years of age or older with good overall health, spontaneous amenorrhoea for at least 12 months, FSH level > 40 mIU/mL, average of five or more vasomotor symptoms per day Exclusion: strict vegetarian; high‐fiber or high‐soy diet; history of breast cancer, endometrial carcinoma, cardiovascular disease, thromboembolic disorders, undiagnosed vaginal bleeding, chronic alcoholism or chronic gastrointestinal diseases. Not on hormonal therapy or phytoestrogens for previous six months Mean age, years: 55.7 Recruitment method: from outpatient menopause clinic at university
Interventions	Soy isoflavones extract (n = 40) 250 mg (Glycine max AT) corresponding to 100 mg/d of isoflavones, administered twice daily in capsules containing 125 mg soy extract plus 50 mg of isoflavones each (50% genistein and 35% daidzein) Placebo (n = 40): two lactose tablets daily Followed up for 10 months with evaluations at baseline and at four, seven and 10 months
Outcomes	Kupperman Menopausal Index Daily diary of hot flushes Symptoms Body mass index Vaginal cytology (maturational value, pH) Transvaginal ultrasonography (endometrial thickness) Adverse events Mammography Lipids, plasma levels of isoflavones
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	'Randomly assigned to one of two groups'. Computer randomised
Allocation concealment (selection bias)	Low risk	Centralised computerised randomisation using software by a statistician who was unaware of the study protocol
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	Low risk	Isoflavone group: Two withdrew because of flatulance and epigastralgia; control group: two withdrew because of depression and family problems and not wishing to continue
Selective reporting (reporting bias)	Low risk	Original protocol was not viewed, but outcomes listed in the methods section of the paper were reported in the results