The levels of all analytes were lower in Asian than Caucasian women. Overweight women had significantly higher levels of CRP, IL-6, and leptin and lower levels of adiponectin than normal weight women. We did not observe a significant effect of soy foods on the four markers, but leptin increased in the control and not in the intervention group (p = 0.20 for group-time effect); this difference was significant for Asian (p = 0.01) and obese women (p = 0.005). During this 2-year intervention, soy foods did not modify serum levels of CRP, IL-6, leptin, and adiponectin in premenopausal women although leptin levels remained stable among women in the intervention group who were obese or of Asian ancestry. Further studies with diverse markers of inflammation are necessary to clarify the specific effect of soy on immune responses. The results of a 2-year soy intervention at our center and reports from other trials indicate that the preventive effects of soy on breast cancer risk, if they exist, are not mediated by their effects on the major circulating sex hormones. On the other hand, results from animal studies and human trials support the hypothesis that soy or isoflavones reduce chronic inflammation, a possible risk factor for breast cancer. Circulating inflammatory markers, such as cell-adhesion molecules and C-reactive protein (CRP), were decreased in some studies. Genistein appears to down regulate the inflammatory response through its tyrosine kinase inhibitory effects. The choice of markers was driven by the concept of capturing different aspects of the immune response through a combination of cytokine levels as it appears that they predict disease better than single markers. IL-6 is one of the major components responsible for the acute phase protein synthesis by the liver, particularly CRP, a sensitive, non-specific indicator of inflammation. IL-2 is a growth factor for antigen-stimulated T lymphocytes, where it increases cytokine synthesis and B cell proliferation. Leptin, a possible link between nutritional status and the immune function, is a marker of obesity, participates in pro-inflammatory responses, and is an important growth factor for breast cancer. Adiponectin is the most abundant protein in adipocytes with a strong anti-inflammatory function in addition to its anti-atherogenic and insulin-sensitizing properties. The participants were recruited by sending out 10,022 invitations to women who had received a normal screening mammogram. Of these, 975 (9.73%) interested women replied and 352 eligible women aged 35–46 years were identified during a telephone-screening interview. Women were excluded from this study due to use of oral contraceptives or other sex hormones, diagnosis of cancer, hysterectomy, and no intact ovary or no regular menstrual periods. After a run-in period, 220 women were randomized to a soy diet or to the control group and 189 subjects completed 2 years of intervention. The Institutional Review Boards of the University of Hawaii approved the study protocol; participants signed informed consent and gave written permission to use frozen samples for future analyses. This trial was designed to provide two servings of soy per day containing approximately 25 mg aglycone equivalents of isoflavones per serving. A choice of tofu, soy milk, roasted soy nuts, soy bars, and soy protein powder was offered to the participants. The same brands of soy foods were used throughout the intervention and the isoflavone content of food items was monitored by high-pressure liquid chromatography (HPLC) with photo diode array detection. Women in the control group were instructed to maintain their regular diet. Body weight was recorded at baseline and at each study visit. Adherence to the study protocol was high in both arms of this study. Subjects in the intervention group showed an increase in self-reported intake of isoflavones from 4.4 mg/day at baseline to 57.3 mg/day as assessed by unannounced 24-hour recalls, which was confirmed by an increase in urinary isoflavone excretion from 8.0 to 59.8 nmol/mg creatinine. Controls continued their usual soy intake as confirmed by 24-hour recalls and urinary isoflavone excretion. This study makes use of existing serum samples that were collected during the 2-year trial period. We collected blood samples at baseline as well as 3, 6, 12 and 24 months after randomization timed to occur 4–6 days after ovulation as determined by an ovulation kit. Blood was allowed to clot for 30 minutes and centrifuged at 3000 rpm for 15 minutes; aliquoted serum was frozen at -80°C. Not all women had serum samples available for all time points.
