Impact of improved low-density lipoprotein cholesterol assessment on guideline classification in the modern treatment era—Results from a racially diverse Brazilian cross-sectional study

Abstract

The Martin/Hopkins low-density lipoprotein cholesterol equation (LDL-CN) was previously demonstrated as more accurate than Friedewald LDL-C estimation (LDL-CF) in a North American database not able to take race into account. We hypothesized that LDL-CN would be more accurate than LDL-CF and correlate better with LDL particle number (LDL-P) in a racially diverse Brazilian cohort. We performed a cross-sectional analysis of 4897 participants in the Brazilian Longitudinal Study of Adult Health, assessing LDL-CF and LDL-CN accuracy via overlap with ultracentrifugation-based measurement among clinical guideline LDL-C categories as well as mg/dL and percent error differences. We analyzed by triglyceride categories and correlated LDL-C estimation with LDL-P. LDL-CN demonstrated improved accuracy at 70 to <100 and <70mg/dL (P<.001), with large errors ≥20mg/dL about 9 times more frequent in LDL-CF at LDL-C <70mg/dL, mainly due to underestimation. Among individuals with LDL-C <70mg/dL and triglycerides ≥150mg/dL, 65% vs 100% of ultracentrifugation-based low-density lipoprotein cholesterol calculation fell within appropriate categories of estimated LDL-CF and LDL-CN, respectively (P<.001). Similar results were observed when analyzed for age, sex, and race. Participants at LDL-C <70 and 70 to <100mg/dL with discordantly elevated LDL-CN vs LDL-CF had a 58.5% and 41.5% higher LDL-P than those with concordance (P<.0001), respectively. In a diverse Brazilian cohort, LDL-CN was more accurate than LDL-CF at low LDL-C and high triglycerides. LDL-CN may avoid underestimation of LDL-C and better reflect atherogenic lipid burden in low particle size, high particle count states.