To assess the efficacy and tolerability of phytoestrogens for treatment of menopausal symptoms.
We searched the Cochrane Library and MEDLINE from 1966 to March 2004, using a detailed list of terms related to phytoestrogens and menopausal symptoms and also hand-searched abstracts from relevant meetings.
Randomized trials were eligible if they involved symptomatic perimenopausal or postmenopausal women, compared phytoestrogen with placebo or control, reported hot flush frequency or menopausal symptom scores, and were at least 4 weeks in duration.
Data were extracted onto standardized forms using a prospectively developed protocol. Twenty-five trials involving 2,348 participants met criteria. At baseline, the mean age was 53.1 years, mean duration of menopause was 4.3 years, and mean daily hot flush frequency was 7.1. Mean study duration was 17 weeks. Trials were grouped into categories according to type of phytoestrogen: soy foods, beverages, or powders (n = 11); soy extracts (n = 9); and red clover extracts (n = 5). Of the 8 soy food trials reporting hot flush frequency outcomes, 7 were negative. Five trials of soy foods provided information to calculate effect sizes; these were in the small-to-medium range, favoring placebo in 3 trials and soy in 2. Of the 5 soy extract trials reporting hot flush frequency, 3 (including the 2 largest trials) were negative. Effect sizes were calculated for 2 soy extract trials: one favored placebo with small effect size and the other favored soy with moderate effect size. Red clover trials showed no improvement in hot flush frequency (weighted mean difference −0.60, 95% confidence interval −1.71 to 0.51). Adverse effects were primarily gastrointestinal and taste intolerance in the soy food and beverage trials.
The available evidence suggests that phytoestrogens available as soy foods, soy extracts, and red clover extracts do not improve hot flushes or other menopausal symptoms.
Phytoestrogens, available as soy foods, soy extracts, and red clover extracts, do not improve hot flushes or other menopausal symptoms compared with placebo.
From the *Department of Medicine, University of Minnesota, Minneapolis; and †Center for Chronic Disease Outcomes Research, Veterans Affairs Medical Center, Minneapolis, Minnesota.
Address reprint requests to: Timothy J. Wilt, MD, MPH, Veterans Affairs Medical Center, Section of General Internal Medicine (111–0), One Veterans Drive, Minneapolis, MN 55417; e-mail: Tim.Wilt@med.va.gov.
Received May 6, 2004. Received in revised form July 2, 2004. Accepted July 8, 2004.