BackgroundObservational epidemiological studies have reported the associations of high body mass index (BMI) with elevated serum uric acid (UA) level and increased risk of postmenopausal breast cancer. However, whether UA is causally induced by BMI and functioned in the BMI-breast cancer relationship remains unclear. MethodsTo elucidate the causality direction between BMI and serum UA, the bidirectional Mendelian randomization (MR) analyses were performed by using summarized data from the largest Asian genome-wide association studies (GWAS) of BMI and UA carried out in over 150,000 Japanese populations. Then, a total of 19,518 postmenopausal women from the Dongfeng-Tongji (DFTJ) cohort (with a mean 8.2-year follow-up) were included and analyzed on the associations of BMI and serum UA with incidence risk of postmenopausal breast cancer by using multivariable Cox proportional hazard regression models. Mediation analysis was further conducted among DFTJ cohort to assess the intermediate role of serum UA in the BMI-breast cancer association. ResultsIn the bidirectional MR analyses, we observed that genetically determined BMI was causally associated with elevated serum UA [beta(95% CI) = 0.225(0.111, 0.339), p < 0.001], but not vice versa. In the DFTJ cohort, each standard deviation (SD) increment in BMI (3.5 kg/m(2)) and UA (75.4 mu mol/l) was associated with a separate 24% and 22% increased risk of postmenopausal breast cancer [HR(95% CI) = 1.24(1.07, 1.44) and 1.22(1.05, 1.42), respectively]. More importantly, serum UA could mediate 16.9% of the association between BMI and incident postmenopausal breast cancer. ConclusionsThe current findings revealed a causal effect of BMI on increasing serum UA and highlighted the mediating role of UA in the BMI-breast cancer relationship. Controlling the serum level of UA among overweight postmenopausal women may help to decrease their incident risk of breast cancer.