Prevention of bone resorption by intake of phytoestrogens in postmenopausal women: a meta-analysis

Abstract

Phytoestrogens as selective estrogen receptor modulators like compounds may consider as a therapeutic option in osteoporosis. In this regard, the effect of phytoestrogens on bone biomarkers was examined in several trials which their results are controversial. We aimed this meta-analysis to evaluate the net effect of phytoestrogens on bone markers. A thorough search was conducted from 2000 to 2010 in English articles. All randomized clinical trials were reviewed, and finally, 11 eligible randomized clinical trials were selected for meta-analysis. Totally 1,252 postmenopausal women were enrolled in the study by considering the changes of pyridinoline (Pyd), desoxypyridinoline (Dpyd), bone alkaline phosphatase, and osteocalcin concentrations in urine and serum after phytoestrogens consumption. The urine Pyd and Dpyd levels decreased significantly in phytoestrogens consumers. Effect size and effect size for weighted mean difference of urine Pyd levels showed -1.229171 (95% confidence interval (CI) = -1.927639 to -0.530703) and -9.780623 (95% CI = -14.240401 to -5.320845), respectively, a significant results in comparison to control group and significant results for Dpyd -0.520132 (95% CI = -0.871988 to -0.168275) and -0.818582 (95% CI = -1.247758 to -0.389407), respectively. Meta-analysis indicates that phytoestrogens intake can prevent bone resorption, but its benefits on bone formation are not significant. This favorable effect was observed in low doses and in at least 3 weeks of phytoestrogens intake.

Fig. 1

Flow diagram for study selection

Fig. 2

a Individual and pooled effect size for the outcome of “Pyd” in the studies considering phytoestrogens comparing to placebo therapy. b Individual and pooled effect size for weighted mean difference for the outcome of “Pyd” in the studies considering phytoestrogens comparing to placebo therapy. c Publication bias indicators for the outcome of “Pyd” in the studies considering phytoestrogens comparing to placebo therapy (effect size). d Publication bias indicators for the outcome of “Pyd” in the studies considering phytoestrogens comparing to placebo therapy (effect size for weighted mean difference)

Fig. 3

a Individual and pooled effect size for the outcome of “Dpyd” in the studies considering phytoestrogens comparing to placebo therapy. b Individual and pooled effect size for weighted mean difference for the outcome of “Dpyd” in the studies considering phytoestrogens comparing to placebo therapy. c Publication bias indicators for the outcome of “Dpyd” in the studies considering phytoestrogens comparing to placebo therapy (effect size). d Publication bias indicators for the outcome of “Dpyd” in the studies considering phytoestrogens comparing to placebo therapy (effect size for weighted mean difference)

Fig. 4

a Individual and pooled effect size for the outcome of “BALP” in the studies considering phytoestrogens comparing to placebo therapy. b Individual and pooled effect size for weighted mean difference for the outcome of “BALP” in the studies considering phytoestrogens comparing to placebo therapy. c Publication bias indicators for the outcome of “BALP” in the studies considering phytoestrogens comparing to placebo therapy (effect size). d Publication bias indicators for the outcome of “BALP” in the studies considering phytoestrogens comparing to placebo therapy (effect size for weighted mean difference)

Fig. 5

a Individual and pooled effect size for the outcome of “OC” in the studies considering phytoestrogens comparing to placebo therapy. b Individual and pooled effect size for weighted mean difference for the outcome of “OC” in the studies considering phytoestrogens comparing to placebo therapy. c Publication bias indicators for the outcome of “OC” in the studies considering phytoestrogens comparing to placebo therapy (effect size). d Publication bias indicators for the outcome of “OC” in the studies considering phytoestrogens comparing to placebo therapy (effect size for weighted mean difference)