Reference
Zhang FF, Haslam DE, Terry MB, et al. Dietary isoflavone intake and all‐cause mortality in breast cancer survivors: The Breast Cancer Family Registry. Cancer. 2017;123(11):2070-2079.
Study Design
Prospective epidemiological study
Study Objectives
To examine association between dietary intake of isoflavones and all-cause mortality in a multiethnic cohort of women diagnosed with breast cancer in the United States, with particular attention to hormone receptor status and prediagnosis vs postdiagnosis intake.
Participants
The study population included 6,235 women (17% Hispanic, 12% black, 11% Asian American, 60% white) recruited from The Breast Cancer Family Registry (BCFR), an international population of families with breast cancer from the United States, Canada, and Australia. Participants were limited to the 5 North American sites to maintain consistency of a single food frequency questionnaire (FFQ).
In addition, selection was limited to those diagnosed with a first primary, invasive breast cancer. Women who died of any cause within 1 year of baseline questionnaire, completed the questionnaire more than 5 years before or after diagnosis, or had a questionable FFQ due to 3 or more deviations from normal caloric intake were excluded from the study. The study population (N=6,235) was further divided for subgroup analysis of prediagnosis (n=4,769) and postdiagnosis (n=1,466) intake.
Outcome Measures
Overall mortality was the only outcome measure. There were no secondary outcome measures such as breast cancer–specific mortality or recurrence.
Key Findings
Overall, there was a 21% reduced mortality when comparing the highest (≥1.5 mg) vs lowest (<0.3 mg) quartiles of daily isoflavone intake (hazard ratio [HR]: 0.79; P trend=0.01). Subgroup analysis found that reduction in mortality was only evident in women who had estrogen receptor (ER)-negative tumors (HR: 0.49; P trend=0.005) and women who were not on any hormone therapy for their breast cancer (HR: 0.68; P trend=0.02). Median follow-up was 113 months (9.4 years).
Practice Implications
This study adds to a growing body of evidence that soy consumption is safe in women with a history of breast cancer, regardless of menopausal status or whether the tumor is ER-positive or ER-negative.
While there was a 21% decrease in all-cause mortality overall, when stratified this was statistically significant only for women with a history of ER-negative tumors (51% decrease) or those not on any hormone therapy for their cancer (32% decrease). Of particular clinical interest is that only those with the highest isoflavone intake postdiagnosis, but not prediagnosis, had an associated decrease in mortality.
Studies assessing outcomes of soy intake in women with a history of breast cancer began to emerge in 2009 with the Life After Cancer Epidemiology (LACE) study,1 a prospective, observational study that followed 1,954 US women and tracked recurrence, not mortality. The LACE study found a trend for reduced risk of recurrence with each increasing quintile of consumption of isoflavones vs no soy intake.
Perhaps most notable from the LACE study was the finding that in postmenopausal tamoxifen users there was a statistically significant 60% reduction in recurrence when comparing the highest (>1,453 µg/d) vs lowest (<7.7 µg/d) quintiles of the isoflavone diadzein (HR: 0.48; 95% confidence interval [CI]: 0.21-0.79; P=0.008). The results of the LACE study were profound; they not only negated expectations of interference with soy in women with ER-positive disease history, they suggested possible benefit.
To date, not a single study of moderate soy consumption has found detrimental effects in any women with a history of breast cancer (premenopausal or postmenopausal, ER-negative or ER-positive).
Later in 2009, JAMA published the Shanghai Breast Cancer Survival Study (SBCSS),2 which involved a cohort of 5,033 women living in Shanghai with a history of breast cancer. It was designed to assess not only recurrence but mortality as well. In the SBCSS there was an inverse relationship between increasing amounts of soy intake and both mortality and recurrence.
Specifically, those in the highest quartile of soy isoflavone intake (>6.27 mg/d) had nearly 30% less mortality (HR: 0.71; 95% CI: 0.54-0.92) and 32% less recurrence (HR: 0.71; 95% CI: 0.54-0.87) than those in the lowest quartile (<2.0 mg/d). Perhaps the most intriguing results were seen in women with a history of ER-positive disease. The HR for recurrence in the highest quartile of soy protein intake was 0.65 (95% CI: 0.36-1.17) for nonusers of tamoxifen and 0.66 (95% CI: 0.40-1.09) for tamoxifen users in the same quartile of soy intake. Thus, high intake was not associated with any benefit in those taking tamoxifen, but it was not associated with any increased risk either.
Furthermore, women with a history of ER-positive disease who were not taking tamoxifen but were in the highest quartile of soy consumption had an associated risk of mortality that was significantly lower (HR: 0.65) than those taking tamoxifen who were in the lowest quartile of soy consumption (HR: 0.90). This was the only study of soy intake in women with a history of breast cancer to find such a significant association between soy and decreased mortality risk. Potential explanations include the high soy consumption in Chinese women compared to American women, and the possible presence of genetic differences in a population that has eaten more soy for many generations.
Since these 2 seminal studies began to shift our thinking about soy intake in women with a history of breast cancer, every study and meta-analysis has corroborated the findings. The After Breast Cancer Pooling Project (ABCPP) reviewed the LACE, SBCSS, and the Women’s Healthy Eating and Living Initiative (WHEL) studies and validated each study’s findings.
As the name implies, the ABCPP also pooled all of the participants into a single cohort of 9,514 women. This pooled analysis confirmed an inverse association between soy intake and breast cancer recurrence and/or mortality.3 Specifically, in the ABCPP, consumption of >10 mg/d of isoflavones was associated with a nonsignificant reduction in all-cause mortality (HR: 0.87; 95% CI: 0.70, 1.10) and breast cancer–specific mortality (HR: 0.83; 95% CI: 0.64, 1.07). Pooled analysis also showed that ingestion of >10 mg/day of isoflavone is associated with a statistically significant 25% decrease in recurrence of disease (HR: 0.75; 95% CI: 0.61, 0.92). In addition to these large cohorts, several smaller studies have been published and the findings have been consistent.4-6
Despite this level of evidence, many clinicians still advise women to avoid soy after a breast cancer diagnosis. Why does this advice persist? Until 2009, rodent studies suggested soy may interfere with tamoxifen and/or be estrogenic on breast cancer cells. In the absence of any human data, extrapolation of these studies was all we had to inform our decisions. It was prudent to avoid soy before the human data emerged. Now, not only do we have enough human outcome data to safely advise soy consumption, we have several reasons to explain why rodent models may have been so misleading.
In 2011, Setchell and colleagues compared the differences between how rodents and humans metabolize isoflavones.7 The researchers measured isoflavone metabolism in 4 of the most common rodent models used in mammary tumor research and in 5 different healthy human populations: women who consumed a single serving of either soy nuts, soy milk, or tempeh; men who drank soy milk twice a day for 3.5 days; postmenopausal women who consumed a single dose (50 mg) genistein (a soy-derived isoflavone); men and women who consumed a single dose (20 mg) of enantiomeric pure S-(−)equol; and infants fed soy infant formula at 6 months and a soy germ isoflavone supplement at age 3 years.
They found that humans metabolize (ie, conjugate) nearly all of the isoflavones, leaving <1.0% unconjugated genistein in circulation. Rodents are comparatively poor metabolizers, with unconjugated genistein reaching levels up to 150 times that of humans. Conjugated and unconjugated genistein are different molecules, with different biological effects. The conclusion: Rodents are poor models for testing how isoflavones affect human biology.
There are other considerations as well. Isoflavones are considered phytoestrogens. While it has long been assumed that phytoestrogens are estrogenic, the truth may be less linear and more complicated. We now know that “phytoestrogenic” compounds act more like the drug class known as selective estrogen receptor modulators (SERMs).8-10 The most well-characterized SERM drug is tamoxifen, which is antiestrogenic on breast tissue and estrogenic on the uterus.
A single molecule can have opposing actions in different tissues because of unique proteins inside cells. These proteins act as either coactivators or corepressors, depending on the tissue. What we call phytoestrogens may in fact be more precisely dubbed “phyto-SERMs.”
Isoflavones also have non-ER–driven effects on cells. Some of the anticancer effects include cell differentiation, inhibition of tyrosine kinases, DNA topoisomerase activity, inhibition of matrix metalloproteinases, and suppression of angiogenesis.11,12 In the current study, there was an inverse association between soy intake and overall mortality in women with a history of ER-negative disease (HR: 0.49; P trend=0.005), consistent with findings from the ABCPP. If isoflavones are responsible for this inverse association, it would be through the many anticancer effects that have nothing to do with estrogen receptors.
There is another level of complexity to consider. Soy beans, like all foods, have a complex makeup of thousands of compounds (phytochemicals).11,13 It is easy to understand how research into a single class of compounds (isoflavones) may not adequately predict outcomes of whole-food consumption. While our tendency toward reductionism satisfies our mechanistic mindset in medicine, it’s essential not to conflate mechanistic details with predictive outcomes.
Note that the study reviewed here is the most ethnically diverse study on this subject published to date. The WHEL and LACE study populations were 82% to 85% non-Hispanic white. In the current study, 60% of the participants were white, but 17% were Hispanic, 12% were black, and 11% were Asian American. While the WHEL and LACE studies did not see an overall reduction in all-cause mortality with high isoflavone (>10 mg/d) consumption, the current study did, and its authors suggest this may be due to the more diverse population.
Admittedly, the “gold standard” of evidence-based medicine—a double-blind, randomized, placebo controlled study to conclusively determine that soy foods are safe for all women with a history of breast cancer—has not been done. However, we do have many large, observational studies with consistent and compelling evidence that soy consumption is at the very least not harmful, and may in fact be helpful, for all women with a history of breast cancer.
This study again demonstrates that soy consumption is associated with decreased mortality in women with a history of ER-negative disease. To date, not a single study of moderate soy consumption has found detrimental effects in any women with a history of breast cancer (premenopausal or postmenopausal, ER-negative or ER-positive). The consistent findings regarding the safety of soy consumption should reassure both patients and practitioners that soybeans can be included in a highly plant-based, nutrient-rich diet that is healthy for everyone. More and more it looks like those with a history of ER-negative disease should actually make a concerted effort to include a serving or two of soy foods in their diets each day.
Related: Read our thorough review of the soy studies mentioned in this A&C for more details.
