Comparison and Agreement of LDL-C Formulas in Cardiovascular Risk Stratification among Malaysia

Abstract

Accurate estimation of low-density lipoprotein cholesterol (LDL-C) is critical for cardiovascular risk stratification and guiding
treatment. Because direct LDL-C assays are not routinely available, calculated LDL-C remains standard in clinical practice. The
Friedewald equation is the most widely used method. Newer formulas, including Sampson, Martin–Hopkins, and Anandaraja,
have been developed to improve accuracy, but their adoption is limited. Discrepancies among these formulas can result in
patient misclassification, potentially leading to under-treatment or persistent cardiovascular risk. However, their comparative
performance has not been thoroughly evaluated in the Malaysian population. This study evaluates the agreement among four
LDL-C calculation formulas and quantifies patient reclassification relative to the Friedewald equation, addressing a gap in
evidence for Malaysia. The analysis included approximately 5665 records from the nationwide Malaysian Health and Wellbeing
Assessment (MyHEBAT). Analyses were performed using SPSS. Data normality was assessed, and parametric tests were
applied as appropriate. LDL-C values were categorized according to the National Cholesterol Education Program Adult
Treatment Panel III (NCEP-ATP III) guidelines. Patient reclassification across LDL-C categories was evaluated using multinomial
logistic regression. The comparison of means indicated that Friedewald differed significantly from Martin–Hopkins (mean
difference [MD] = -0.083, p = 0.003), Sampson (MD = -0.074, p = 0.009), and Anandaraja (MD = -0.093, p = 0.001), with no
significant differences among the three newer equations. Greater discordance was observed among adults with elevated
triglycerides. The Friedewald equation classified a larger proportion of individuals in the very low LDL-C category (<2.6
mmol/L), whereas the Anandaraja formula shifted more patients into the moderate range (3.35–4.04 mmol/L). The Martin–
Hopkins and Sampson formulas produced distributions similar to Friedewald, but notable differences were present in the midto-
high LDL-C categories. The study is limited by reliance on calculated LDL-C values without direct measurement as a
reference standard. Although the large sample size supports generalizability, findings are based on a single population cohort
and may not represent other ethnic or regional groups. Residual confounding factors, including dietary intake, medication use,
and comorbidities, were not assessed. The LDL-C formulas demonstrated variable agreement, resulting in patient
reclassification across risk categories. Friedewald classified more individuals as very low, Anandaraja shifted more toward
moderate, and Martin–Hopkins and Sampson aligned more closely with Friedewald but differed in mid-to-high ranges. These
discrepancies may influence treatment decisions. Future research should validate formulas against direct LDL-C measurement,
assess performance in high-risk subgroups, and consider formula-specific cut-offs to enhance cardiovascular risk assessment.