Abstract
Background
Low-density lipoprotein cholesterol (LDL-C) is an established risk factor for cardiovascular disease (CVD), and a target for lipid-lowering therapy. LDL-C is typically not measured directly but is estimated using the Friedewald formula, which assumes a fixed factor for the ratio of triglycerides (TG) to very low-density lipoprotein cholesterol (VLDL-C). However this assumption is sometimes not valid. The Martin/Hopkins (M/H) formula estimates LDL-C using an adjustable factor for the TG:VLDL-C ratio and is expected to improve upon Friedewald when predicting measured LDL-C, and apolipoprotein B (ApoB), one molecule of which is associated with each LDL particle.
Purpose
We compared values of LDL-C calculated by the Friedewald and M/H methods with respect to their correlation with non-high density lipoprotein cholesterol (non-HDL-C) and ApoB, and their classification of individuals based upon attainment of the threshold value of 70 mg/dl (1.8 mmol/l) of LDL-C. This cut-point is a treatment target for individuals at high risk of CVD in the 2019 ESC guidelines for lipid modification, and a threshold for initiating statin therapy in the 2019 ACC/AHA guidelines.
Methods
In this analysis we included participants in the National Health and Nutrition Examination Survey (NHANES) from 2005–2016, age ≥18, <80 years who had measurements for total cholesterol (TC), TG and HDL-C. LDL-C was calculated using Friedewald and M/H. We correlated LDL-C (calculated using the two methods) with non-HDL-C and ApoB. We identified individuals with LDL-C <70 mg/dl using both methods. When LDL-C (Friedewald) was <70, but LDL-C (M/H) was >70, we classified these participants as discordant. Statistical analyses were performed in IBM SPSS for Windows v26.
Results
12,243 individuals were included. 51.8% were female, mean (±SD) age was 45.5±17.4, 15.3% were treated with statins, ApoB was available for 2179 participants. Mean lipid concentrations (mg/dl) were: TC: 191.5±41.0, TG: 120.0±67.0, HDL-C: 54.1±15.7, LDL-C (Friedewald): 113.3±35.4; LDL-C (M/H): 114.9±35.2. In the whole population, LDL-C (M/H) was more strongly correlated than LDL-C (Friedewald) with ApoB (r=0.935 v 0.894) and non-HDL-C (r=0.981 v 0.944). In statin-treated participants, LDL-C (M/H) was also more strongly correlated with ApoB (r=0.951 v 0.914) and non-HDL-C (r=0.979 v 0.928). 1139 participants had LDL-C (Friedewald) <70 mg/dl. Of these, 206 individuals (18.1%) were discordant, having LDL-C (M/H) >70 mg/dl. Amongst statin-treated patients, 22.9% were discordant. Only 5.5% of individuals with LDL-C (M/H) <70 mg/dl showed reverse discordance (LDL-C (Friedewald) >70 mg/dl).
Conclusions
The M/H method of calculating LDL-C correlates more strongly with non-HDL-C and ApoB than Friedewald. Importantly the discordant results confirm previous observations that Friedewald underestimates LDL-C at low concentrations. This may result in under-use of lipid-lowering therapies.
Funding Acknowledgement
Type of funding source: None