A new equation for estimating low-density lipoprotein cholesterol concentration based on machine learning

Lei Chen,Chen Rong,Peidu Ma,Yiyang Li,Xue Deng,Muxing Hua
DOI: https://doi.org/10.1097/md.0000000000037766
IF: 1.6
2024-04-14
Medicine
Abstract:Low-density lipoprotein cholesterol (LDL-C) is a critical risk factor for atherosclerotic cardiovascular disease and is both an initiating factor and therapeutic target in primary and secondary prevention. [ 1 ] Most laboratories measure LDL-C directly or estimate it using the Friedewald equation. However, according to the 2021 ESC Guidelines, the Friedewald equation is only recommended when triglyceride (TG) 50 mg/dL, if these parameters are exceeded, it is recommended to consider using the Martin or Sampson equation as a supplement or direct determination. However, studies have demonstrated that the Friedewald equation may be less accurate in calculating LDL-C concentration than expected, [ 2 , 3 ] highlighting the need for improved accuracy in the calculation of LDL-C concentration. Currently, the widely used and validated LDL-C calculation equations are derived from samples of European and American populations, with fewer equations derived from Chinese population samples. Additionally, commonly used equations typically utilize TG, total cholesterol (TC), and high-density lipoprotein cholesterol (HDL-C) as variables, with their highest power being 2. Furthermore, it has been shown that age and gender may influence LDL-C concentration levels. [ 4 , 5 ] Therefore, the aim of our study is to evaluate the influence of different variable combinations and highest powers on LDL-C concentration and to explore an equation that can accurately calculate LDL-C concentration. By comparing it with the LDL-C concentration calculated by the Martin/Hopkins and extended Martin/Hopkins, Sampson, and Friedewald equations, we aim to determine the optimal LDL-C calculation equation suitable for the Chinese population.
medicine, general & internal
What problem does this paper attempt to address?