Comparing different assessments of remnant lipoprotein cholesterol: The very large database of lipids

Background: Remnant lipoprotein cholesterol (RLP-C) is a risk factor for atherosclerotic cardiovascular disease, but there is no standard method for measurement. Some studies have used very low-density lipoprotein cholesterol estimated by the Friedewald equation to approximate RLP-C using a basic lipid panel, whereas others have attempted to measure RLP-C with ultracentrifugation. Objective: The aim of the study was to compare RLP-C levels estimated from basic lipid parameters to those measured by ultracentrifugation. Methods: We analyzed 1,350,908 individuals from the Very Large Database of Lipids, comparing one estimate of RLP-C using basic lipid parameters (RLP-C_estimated = non–high-density lipoprotein cholesterol − Friedewald-estimated low-density lipoprotein cholesterol for triglycerides <355 mg/dL [4 mmol/L], or non–high-density lipoprotein − directly measured low-density lipoprotein for triglycerides ≥355 mg/dL) to levels measured by vertical auto profile ultracentrifugation (RLP-C_measured = dense subfraction of very low-density lipoprotein cholesterol + intermediate-density lipoprotein cholesterol). We calculated correlations between RLP-C_estimated and RLP-C_measured along with median within-subject differences between RLP-C_estimated and RLP-C_measured across quintiles of RLP-C_estimated. We also assessed correlations with RLP-C estimated from basic lipid parameters using a novel method of calculating low-density lipoprotein cholesterol with a patient-specific conversion factor (RLP-C_estimated-N). Results: Our cohort was 48% male, and median (interquartile range) age was 59 (49–69) years. Median (interquartile range) RLP-C_estimated and RLP-C_measured were 23 (16.4–33.2) and 24 (19–32) mg/dL, respectively. The correlation between RLP-C_estimated and RLP-C_measured was 0.76. Based on the specified definition of RLP-C_estimated, the correlation between RLP-C_estimated and triglyceride/5 for triglyceride < 355 mg/dL was exactly 1.0. RLP-C_estimated was lower than RLP-C_measured in the first and second quintiles of RLP-C_estimated but greater in the highest quintile. The correlations with RLP-C_estimated-N were 0.98 and 0.81 for RLP-C_estimated and RLP-C_measured, respectively. Conclusions: A previously used estimate of RLP-C using basic lipid parameters correlates weakly with remnants measured by ultracentrifugation. Our findings emphasize the need to standardize definitions and measurements of RLP-C.