Healthy nutrient-rich dietary patterns and mortality in older Chinese: a 16-year follow-up of Guangzhou Biobank Cohort Study

Using different techniques to derive dietary patterns (DPs) could evaluate of real-world diet behaviors and provide DP recommendations. Therefore, we identified DPs using hybrid methodologies, and examined the associations of DPs with all-cause and cardiovascular disease (CVD) mortality among older Chinese. Using data from the Guangzhou Biobank Cohort Study, dietary intake was assessed using a validated food frequency questionnaire. DPs were derived using hybrid methods including reduced rank regression (RRR) and partial least squares (PLS), focusing on nutrients commonly insufficient in Asian diets. Associations of the DPs with mortality and CVD risk factors were examined using Cox regression and generalized linear models, respectively. Of 19,598 participants with an average follow-up of 15.8 years, 4,966 deaths occurred. Two DPs were derived based on the riboflavin-density, potassium-to-sodium ratio, and vitamin C-density. The DPs derived from both RRR and PLS featured high intakes of green vegetables, yellow/orange fruits, whole grains, and low intakes of refined grains and plant oils, with additional high intakes of fish identified by RRR and milk by PLS. These DPs were associated with lower all-cause and CVD mortality risks. Compared with the lowest quartile, the highest quartiles showed lower risks of all-cause (HRs: 0.89-0.91, all P<0.01) and CVD mortality (HRs: 0.79-0.82, all P <0.01). Moreover, both DPs were associated with favorable cardiometabolic profiles, including lower systolic blood pressure, triglycerides and hsCRP levels, and higher HDL-c levels. These findings suggest that nutrient-rich DPs using hybrid methods may support the development of dietary recommendations to reduce mortality among older Chinese.