Abstract
Disclosure: M. Mitsunami: None. E.R. Bertone-Johnson: None. A. Eliassen: None. J.E. Chavarro: None. Introduction: Isoflavones, a type of phytoestrogens found mainly in soy and legumes, are structurally similar to estrogens and are consumed at a relatively high rate by Asian populations with an average intake of approximately 15-40 mg/d for isoflavone. They have been reported to alleviate menopausal symptoms, such as hot flashes, suggesting that isoflavones may be biologically active in a low-estrogen environment. Previous studies have reported inconsistent associations between the intake of soy and isoflavones and age at natural menopause. Therefore, our aim was to test the hypothesis that higher isoflavone intake is related to higher age at natural menopause. Methods: We followed 85,810 premenopausal women (mean [SD], 36.5 [4.7] year) who completed a previously validated food frequency questionnaire, had not been diagnosed with cancer, and did not undergo any menopausal hormone therapy as of 1991. Diet was assessed every 4 years. Isoflavone intake was estimated from these reports. Participants self-reported their menopausal status and the type of menopause (natural, surgical, and radiation or chemotherapy) every 2 years. We used proportional hazard regression models with age in months to evaluate the relation of isoflavone intake with age at natural menopause where natural menopause was the study outcome and menopause due to surgery, radiation, or chemotherapy were censoring events. We used restricted cubic splines to assess the possibility of non-linear relations between isoflavone intake and age at menopause using the U.S. average isoflavone intake (0.8mg/d) as a reference while adjusting for potential confounders, including age at menarche, BMI, race/ethnicity, smoking status, physical activity, parity, breastfeeding, time to regular menstrual cycle, energy intake, and two dietary factors identified by principal component analysis. Results: Over 1,079,608 person-years, we observed 37,438 cases of natural menopause. Soy products contributed approximately half of isoflavone intake. There was a non-linear association between isoflavone intake and age at natural menopause (p<0.001). Women with isoflavone intakes below the U.S. national average were more likely to experience earlier onset menopause. Conversely, in women with intakes above the national average, the risk of menopause decreased with increasing isoflavone intakes, albeit plateauing at intakes of approximately 3-4mg/d (90-95(th) percentile of observed distribution). There was a suggestion of a positive association between isoflavone intake and earlier onset menopause for intakes above the 99(th) percentile of the observed distribution (∼14 mg/d). Conclusion: This study suggests a non-linear relation between isoflavone intake and age at natural menopause with a potential later onset of menopause at intake levels higher than typical in the US diet but lower than those typically observed in East Asia. Presentation: 6/2/2024