Purpose: We examined the associations of soy product intake with all-cause, cardiovascular disease (CVD), and cancer mortality and mediations through CVD risk factors based on the Guangzhou Biobank Cohort Study (GBCS), and conducted updated meta-analyses.

Methods: A total of 29,825 participants aged 50 + years were included. Causes of death were identified through record linkage. Soy product intake was assessed by food frequency questionnaire. Cox proportional hazards regression was used to analyze the associations between soy product intake and mortality, yielding hazard ratios (HRs) and 95% confidence intervals (CIs). Mediation analyses with CVD risk factors as mediators, and updated meta-analyses were conducted.

Results: During 454,689 person-years of follow-up, 6899 deaths occurred, including 2694 CVD and 2236 cancer. Participants who consumed soy product of 1-6 portions/week, versus no consumption, had significantly lower risks of all-cause and CVD mortality (adjusted HR (95% CI) 0.91 (0.86, 0.97) and 0.87 (0.79, 0.96), respectively). In participants who consumed soy product of ≥ 7 portions/week, the association of higher intake with lower CVD mortality was modestly mediated by total cholesterol (4.2%, 95% CI 1.0-16.6%). Updated meta-analyses showed that the highest level of soy product intake, versus the lowest, was associated with lower risks of all-cause and CVD mortality (pooled HR (95% CI) 0.92 (0.88, 0.96) and 0.92 (0.87, 0.98), respectively).

Conclusion: Moderate and high soy product intake were associated with lower risks of all-cause and CVD mortality. Our findings provide support for current dietary guidelines recommending moderate soy product intake, and contribute additional evidence regarding the potential protective effects of high soy product intake.