Lp(a) and ApoB/ApoA1 ratio independently predict coronary heart disease in type 2 diabetes (Lipids 2026)

Traditional lipid parameters poorly capture coronary heart disease (CHD) risk in type 2 diabetes mellitus (T2DM). This study compared apolipoprotein (ApoA1, ApoB) and lipoprotein(a) [Lp(a)] profiles between T2DM patients with and without CHD and evaluated their combined predictive value. This retrospective study included 140 T2DM patients, 70 with angiographically confirmed CHD and 70 age- and sex-matched without CHD (defined by the absence of cardiac symptoms, normal electrocardiogram, and normal echocardiogram). Fasting serum levels of traditional lipids, ApoA1, ApoB, and Lp(a) were measured, and the ApoB/ApoA1 ratio was calculated. Multivariate logistic regression and ROC curve analysis identified independent predictors and assessed predictive performance. Traditional lipid parameters showed no significant differences between groups (all p > 0.05). CHD patients had significantly lower ApoA1 (1.33 vs. 1.56 g/L, p < 0.001) and higher ApoB (1.40 vs. 1.21 g/L, p < 0.001), ApoB/ApoA1 ratio (1.05 vs. 0.78, p < 0.001), and Lp(a) (20.65 vs. 14.10 mg/dL, p = 0.009). Multivariate analysis identified smoking, ApoB/ApoA1 ratio, and Lp(a) as independent predictors. The ApoB/ApoA1 ratio yielded an AUC of 0.912. A combined model achieved an AUC of 0.976, with bootstrap internal validation confirming robustness (optimism-corrected AUC = 0.959, 95% CI: 0.921-0.982). The ApoB/ApoA1 ratio and Lp(a) were independently associated with CHD in T2DM patients. The combined model demonstrated excellent predictive performance, maintained after internal validation, though external validation is needed.