Lipid Variability Predicts Long-Term Major Adverse Cardiovascular Events in MINOCA: A Retrospective Cohort Study

Objective: This study investigates the predictive value of long-term variability in low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and lipoprotein(a) [Lp(a)] for major adverse cardiovascular events (MACEs) in patients with myocardial infarction with non-obstructive coronary arteries (MINOCA). Methods: In this retrospective cohort study, 254 consecutive MINOCA patients (80 with MACEs, 174 without) were analyzed over a median 41-month follow-up. Lipid variability (LV) was quantified using standard deviation (SD), coefficient of variation (CV), and variability independent of the mean (VIM) from ≥ 3 measurements. The proportional hazards assumption was tested using Schoenfeld residuals. Collinearity among variability metrics was assessed via variance inflation factor (VIF), and separate Cox models were constructed for each metric to avoid multicollinearity. The independent association between LV and MACE risk was analyzed via multivariate Cox proportional hazards regression models, adjusted for established clinical covariates including age, sex, BMI, smoking, hypertension, diabetes, baseline lipid levels, and statin use during follow-up. The predictive performance was evaluated with time-dependent receiver operating characteristic (ROC) curves. Results: MACE incidence was 31.5% during follow-up. While all variability metrics were higher in the MACE group ( P < 0.05), multivariate analysis identified only HDL-C SD (HR = 3.751, 95% CI: 2.088– 6.740, P < 0.001), HDL-C VIM (HR = 1.620, 95% CI: 1.055– 2.486, P = 0.027), and Lp(a) VIM (HR = 2.229, 95% CI: 1.334– 3.725, P = 0.002) as independent predictors of MACEs, after adjusting for baseline confounding factors, the results were consistent with it. Time-dependent ROC curve analysis indicated that a composite score offered moderate predictive value at 2.5 and 3.5 years of follow-up (AUC of 0.751 and 0.714, respectively), but its performance declined at 4.5 years (AUC = 0.696). Conclusion: In MINOCA, variability in HDL-C and long-term variability in Lp(a) independently predict MACE risk. Beyond achieving lipid targets, maintaining long-term stability may be clinically important, offering a novel perspective for risk stratification and personalized management.