Performance of novel low-density lipoprotein-cholesterol calculation methods in predicting clinical and subclinical atherosclerotic cardiovascular disease risk: The Multi-Ethnic Study of Atherosclerosis

BACKGROUND AND AIMS: This study examined the performance of two novel low-density lipoprotein-cholesterol (LDL-C) calculations, LDL Martin and LDL Sampson, on predicting atherosclerotic cardiovascular diseases (ASCVD) risk compared to traditional LDL Friedewald according to the 2018 American Heart Association/American College of Cardiology (AHA/ACC) primary prevention guidelines.

METHODS: A total of 6701 randomly recruited Multi-Ethnic Study of Atherosclerosis (MESA) participants free of ASCVD at baseline were followed for ASCVD during a median of 13.9 years and for subclinical ASCVD-coronary artery calcium (CAC) during a median of 12.5 years. Prevalence of borderline high triglyceride (≥1.7 mmol/L) was 15.2% and was at 13.5% for high triglyceride (≥2.3 mmol/L).

RESULTS: Applying the criteria of LDL-C<1.8 mmol/L in 40-75 year olds without diabetes mellitus to be exempt from risk discussion, LDL Martin and LDL Sampson classified less individuals in this category than LDL Friedewald (p < 0.001), both had 20 individuals with ASCVD, versus 22 by LDL Friedewald. Positive CAC in the discussion-exempt group were over 38% higher (p < 0.001) when classified by LDL Friedewald than by LDL Martin or LDL Sampson. Individuals with LDL-C≥4.9 mmol/L are recommended to high-intensity statin therapy by the AHA/ACC guidelines. The LDL Friedewald≥4.9 mmol/L group had 20 ASCVD events, versus 21 in LDL Martin and 22 in LDL Sampson group.

CONCLUSIONS: In a multi-ethnic USA population, LDL Martin and LDL Sampson did not over- or under-estimate ASCVD risk compared to LDL Friedewald in primary prevention according to AHA/ACC guidelines, while LDL Friedewald under-estimated subclinical ASCVD risk in the low-risk population. These findings support the replacement of LDL Friedewald by LDL Martin or LDL Sampson for lipid screen in the general population.