Han 2002.

Methods	Design: parallel‐group Number screened for inclusion: not stated Number randomly assigned: 82 Number of dropouts: two (one from each arm, one due to poor response, one due to nausea—not stated which arms they were on) Number analysed: 80 Intention‐to‐treat analysis: no Power calculation: none stated Duration: four months Timing: August 1999 to February 2000 Location: university clinic, Brazil Funding: unclear. Investigators acknowledge the co‐operation of a doctor employed by food supplement manufacturer Eugenbio
Participants	Inclusion criteria: women 45 to 55 years of age; "in menopause" at least 12 months, no hormonal treatment for at least 12 months, intact uterus, FSH > 25 U/L, oestradiol < 20 pg/mL, having hot flushes Exclusion criteria: taking lipid‐lowering drugs, antidiabetic medications, soybean‐derived products or herbal supplements; uncontrolled hypertension, stroke or transient ischaemic attack, cancer diagnosed within past five years, previous myocardial infarction Age, years: mean active arm 48 (SE 1.1), placebo arm 49 (SE 1.3) Recruitment method: not stated
Interventions	Phytoestrogen: isoflavone capsules Formulation: soy protein 50.3 mg and isoflavone 33.3 mg (genistein 23.3 mg, daizein 6.2 mg, glycitein in aglycone form 3.8 mg) per capsule Placebo Dose, duration and timing of administration: one capsule eight hourly (= 100 mg isoflavone daily) for four months
Outcomes	Menopausal symptoms: hot flashes (Kupperman Index) Compliance: examination of prescriptions/pills Endometrial thickness
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerised random number generator
Allocation concealment (selection bias)	Low risk	Numbered coded envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded researchers and participants
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcomes recorded by an independent gynaecologist
Incomplete outcome data (attrition bias) All outcomes	Low risk	Minimal dropout
Selective reporting (reporting bias)	High risk	Adverse events not reported