Reviewed: Beavers DP, Beavers KM, Miller M, Stamey J, Messina MJ. Exposure to isoflavone-containing soy products and endothelial function: a Bayesian meta-analysis of randomized controlled trials. Nutr Metab Cardiovasc Dis. August 14, 2010; [epub ahead of print]. doi:10.1016/j.numecd.2010.05.007.

Impaired function of the endothelial cells lining the blood vessels and heart is an early marker of cardiovascular disease. Flow mediated dilation (FMD) of the brachial artery in the upper arm is a non-invasive method for evaluating endothelial physiology. Estrogen treatment improves endothelial function in postmenopausal women, and some studies have shown that soy-based phytoestrogens have vasodilating effects in men and women. Previous meta-analyses of randomized controlled trials (RCTs) testing the effects of soy products on endothelial function were limited by small sample sizes or inclusion of females only. The purpose of this research was to conduct a Bayesian* meta-analysis of RCTs to estimate the effects of isoflavone-containing soy products on endothelial function, as measured by FMD.

The researchers conducted a literature search of the PubMed database through August 21, 2009, to identify eligible RCTs. Search terms included soy, isoflavone, phytoestrogen, endothelial function, and flow mediated vasodilation. The researchers also searched the references cited in the articles and reviewed other sources to identify additional trials: trials that lasted at least 4 weeks; trials that tested dietary ingestion of soy foods, soy protein, or soy isoflavones; and trials that reported FMD as an outcome were included in the analysis. The primary outcome was the mean change in FMD attributed to soy product intervention.

Of the 42 trials identified by the search for possible inclusion in the analysis, 25 were excluded because of short duration, wrong product type, or unsuitable outcome measures. The remaining 17 trials included 1,281 participants from 6 countries. Some participants were healthy, whereas others had high blood cholesterol levels or other symptoms of cardiovascular disease. The average age of participants was 60 years, and 83% of participants were female. Duration of the intervention ranged from 4 weeks to 52 weeks. Six trials tested isoflavone-containing soy products, 7 isolated isoflavones, and 4 soy-based foods or foods fortified with isoflavones. Daily consumption of soy protein ranged from 25 to 40 grams and daily consumption of isoflavones from 33 mg to 120 mg.

The mean increase (improvement) in FMD following intervention with all tested soy products combined was 1.15%. The mean increase in FMD was 0.72% with soy protein intervention and 1.98% with soy isoflavone intervention. Only the soy isoflavone intervention represented a statistically significant improvement in FMD.

The authors discuss possible biological mechanisms for the effects of soy isoflavones on FMD. Impaired release of nitric oxide (NO) leads to endothelial dysfunction by constricting blood vessels and impeding flow. Estrogen appears to improve FMD by increasing the release of NO in endothelial cells, and soy isoflavones may act the same way. Laboratory and animal studies suggest that soy isoflavones increase NO production by increasing the level of nitric oxide synthase, the enzyme that produces NO. Other mechanisms may also be involved, such as reducing inflammation and inhibiting oxidation.

This review concludes that consumption of soy isoflavones modestly but significantly improves endothelial function as measured by FMD, and increasing consumption of soy isoflavones may have beneficial effects on vascular health. In this meta-analysis, soy protein and other soy foods containing isoflavones were associated with a smaller improvement in FMD than isolated soy isoflavones but a greater improvement than controls or placebos.

The authors discuss some limitations in interpreting the findings. The majority of participants evaluated in this meta-analysis were older women. Findings of this meta-analysis may not apply to other subgroups, such as younger people, men, and people with diabetes or other chronic health conditions. Also, FMD is only one method for measuring endothelial function, and other assessment methods may not yield the same results.

Finally, the authors state that additional, larger, randomized, controlled trials are warranted. They recommend that future trials directly compare soy protein with isolated isoflavones to better understand the effects of each on endothelial function. Future trials should also evaluate the dose-response relationship between soy isoflavones and changes in FMD.

—Heather S. Oliff, PhD

*Bayesian statistical analysis differs from classical meta-analyses in that the Bayesian analysis incorporates prior beliefs, does not use P values to measure statistical significance, and it allows calculation of probability (odds ratio).