Genetic modifiers of triglyceride-mediated mortality in coronary heart disease: a large cohort study of polygenic risk and circulating triglyceride level interaction

Background The relationship between triglyceride (TG) levels and adverse events in coronary heart disease (CHD) patients is debated, and the modifying role of genetic background (TG-PRS) is unclear. Aim This study investigated the non-linear relationship between TG and all-cause mortality across TG-PRS strata to inform personalized risk management. Methods 4304 CHD patients undergoing PCI were enrolled. Restricted cubic spline (RCS) Cox regression analyzed the non-linear TG-mortality association in the overall population, adjusted for confounders (age, sex, BMI, hypertension, diabetes, prior PCI/CABG, COPD, PAD, CVD, hs-CRP, ACS). The population was stratified by TG-PRS tertiles (low, medium, high), with RCS repeated per stratum. Further stratified Cox regression was conducted within each PRS stratum, categorizing participants into low- and high-TG subgroups based on stratum-specific inflection points to examine differences in TG's continuous effect on mortality. Results Among 4304 CHD patients (mean age 57.45 ± 10.14 years, 83% male), 496 (11.5%) died after 10-year follow-up. In the overall population, TG showed an L-shaped association with mortality (P=0.0016) before adjustment, shifting to a non-significant U-shaped pattern after multivariable adjustment (P=0.1965). A significant TG and TG-PRS interaction was identified (P=0.031). Stratified RCS analyses revealed: 1. High TG-PRS: marginally significant U-shape (overall P=0.0548, non-linearity P=0.0386, inflection 1.67 mmol/L; red in Figure2); 2. Medium TG-PRS: significant U-shape (overall P=0.0191, non-linearity P=0.0112, inflection 1.44 mmol/L; orange in Figure2); 3. Low TG-PRS: no significant association (overall P=0.5504, non-linearity P=0.4773; blue in Figure2). Further interaction analyses confirmed: In high TG-PRS (P-interaction=0.0107), each 1 mmol/L TG increment reduced mortality by 48.5% in low-TG subgroups (TG<1.67 mmol/L; HR=0.515, P=0.0133), but not in high-TG subgroups (TG≥1.67 mmol/L; HR=1.057, P=0.5147). A similar pattern was seen in medium TG-PRS (P-interaction=0.0061): significant inverse association in low-TG (TG<1.44 mmol/L; HR=0.461, P=0.0258), but not high-TG (TG≥1.44 mmol/L; HR=1.275, P=0.0644). No significant interaction was found in low TG-PRS (P-interaction=0.2981). Conclusion TG is non-linearly associated with all-cause mortality, modified by TG-PRS. A U-shaped TG-mortality relationship was observed in medium and high TG-PRS groups. TG's protective effect in high-genetic-risk patients was confined to low-TG subgroups, highlighting that TG's clinical relevance depends on genetic background and baseline TG status. The observed association between lower TG and poorer prognosis may reflect reverse causation or disease-related factors (e.g., malnutrition, inflammation). Our findings support stratified, personalized TG management over a universal cutoff, especially for high-genetic-risk individuals.