Abstract
Introduction The high-fat diet (HFD) has complex health implications, shaped by the composition of macronutrients and food sources. However, existing dietary assessment tools lack the precision required for effective risk stratification. Objectives This study developed novel HFD scores to accurately characterize dietary patterns and examined their associations with total and cardiovascular disease (CVD) mortality, while also identifying relevant protein biomarkers. Methods Data from the UK Biobank and NHANES were utilized to develop a novel HFD scoring system incorporating both macronutrient ratios and quality indicators. Mortality associations were evaluated using Cox proportional hazards models, Kaplan-Meier analysis, and restricted cubic splines (RCS). Proteomic analysis was conducted to identify plasma proteins associated with mortality. Results The unhealthy HFD score showed a linear correlation with increased total and CVD mortality. In contrast, the healthy HFD score exhibited a U-shaped relationship with CVD mortality. On a 30-point HFD scale, individuals with low total mortality risk in the UK and US should maintain scores below 15, while those with low CVD risk in the UK should aim for scores between 10 and 15. An online HFD risk calculator was developed for practical application. Proteomic analysis revealed 48 proteins linked to total mortality and 153 proteins associated with CVD mortality, with significant enrichment in immune regulation and cardiovascular pathways. Machine learning models identified key predictors, such as EDA2R and NTproBNP, which demonstrated mediation effects of 12.92% and 13.86%, respectively. Conclusion These findings establish a precision nutrition framework that links HFD patterns, proteomic biomarkers, and mortality outcomes, offering actionable insights for clinical and public health intervention.</p>