Seventeen RCTs (n=1,871) were included in the review. Five studies adequately reported randomisation, 10 gave information about method of treatment, three described allocation concealment, five described baseline comparability, 10 had a dropout rate of 20% or less and two performed intention-to-treat analyses. The number of criteria (out of eight) that were graded A was: one in two studies; two in four studies; three in six studies; four in four studies; and five in one study. Twenty-three studies met the inclusion criteria, but only 17 were included in the analysis as six isoflavone studies were included in several other meta-analyses; these were not discussed in detail
Genistein studies (two studies; one was a subgroup analysis): Both RCTs found a significant improvement in hot flush frequency with 54mg genistein.
Soy extract (11 studies): Three out of eight trials found improved hot flush frequency with isoflavones compared with placebo. Five of eight trials found significant improvement in hot flush severity after soy supplement consumption.
Soy protein powder (four studies): One of four studies reported a reduction in hot flush frequency with isoflavones. The three studies that evaluated hot flush severity found no significant reduction.
Adverse events: No proliferative effects of soy isoflavones on the endometrium, vagina or breast tissue was reported in any of the studies. Gastrointestinal disorders occurred after treatment in several studies, but a significant difference between placebo and soy was found in only one study.
No correlation was found between initial hot flushes and outcome.