To determine LDL phenotypes using lipids, lipoproteins, apoproteins, and sdLDL through association rule mining

Abstract

Objective: The atherogenic lipoprotein phenotype is closely associated with the risk assessment of Coronary Artery Disease (CAD) and the monitoring of treatment processes. Particularly, high levels of small dense low-density lipoprotein (sdLDL) and low levels of large buoy-ant low-density lipoprotein (lbLDL) are critical in determining Pattern B. This study aims to determine the lipid phenotype using the Association Rule Mining (ARM) method, based on concentrations of lipids, lipoproteins, apoproteins, and sdLDL.Materials and Methods: This retrospective case-control study utilized analytical research methods. Numerical variables were expressed as mean, standard deviation, median, and min-max values. Statistically significant differences were observed between the low-density lipoprotein (LDL) size categories in terms of triglycerides (TG), LDL, high-density lipoprotein (HDL), apolipoprotein B (ApoB), apolipoprotein E (ApoE), sdLDL, and lbLDL distributions. ARM was employed to detect the lipoprotein phenotype.Results: Statistically significant differences were found between the LDL size categories in distributions of TG, LDL, HDL, ApoB, ApoE, sdLDL, and lbLDL (p(TG)<0.001, p(LDL)=0.03, p(HDL)<0.001, p(ApoB)=0.016, p(ApoE)=0.004, p(sdLDL)<0.001, and p(lbLDL)<0.001). The ARM method revealed that the probability of phenotype B is 100% for sdLDL values in the range of 15.5-109 and lbLDL values in the range of 0-31.5.Conclusion: This study introduces a contemporary approach for detecting lipoprotein phenotypes using ARM, further substantiating the strong correlation between atherogenic phenotypes and sdLDL.