Evaluating low-density lipoprotein cholesterol (LDL-C) equations in young adulthood for long-term risk of atherosclerotic cardiovascular disease: Superior performance of the Martin-Hopkins equation
Y Meng, F Wu, M Juonala, D R Jacobs, L A Bazzano, S R Daniels, T Dwyer, M Kahonen, A R Sinaiko, J Steinberger, E M Urbina, A J Venn, J S A Viikari, O T Raitakari, C G MagnussenAbstract
Background
Accurate low-density-lipoprotein-cholesterol (LDL-C) assessment in young adulthood is critical for atherosclerotic cardiovascular disease (ASCVD) prevention. The optimal LDL-C estimation equation for evaluating long-term ASCVD risks remains uncertain.
Purpose
This study aimed to compare LDL-C levels estimated by Friedewald, Martin-Hopkins, and Sampson equations in young adulthood for evaluating ASCVD risks.
Methods
In seven population-based cohorts from the US, Finland, and Australia within the International Childhood Cardiovascular Cohort Consortium, LDL-C levels were estimated using the Friedewald, Martin-Hopkins, Sampson equations. Discordance was defined as differing LDL-C categorizations (normal <130 mg/dL versus elevated ≥130 mg/dL) between equations. ASCVD events were adjudicated from medical records. Associations between discordance and incident ASCVD were estimated using Fine-Gray regressions. These analyses were replicated in an independent mid-adult cohort from the UK Biobank using harmonized definitions and analytic approaches.
Results
Over a mean follow-up of 22.7 years, 376 fatal and nonfatal ASCVD events among 10550 participants. Although elevated LDL-C estimated by all equations was associated with ASCVD, the Martin-Hopkins equation exhibited the strongest associations [Hazards ratio (95% confidence interval): Martin-Hopkins, 1.69 (1.35-2.12) Sampson, 1.52 (1.22-1.90); Friedewald, 1.47 (1.17-1.84), P <0.05 for harzards ratio comparisons)] and improved risk reclassification (net reclassification improvement: 0.124-0.444). Total discordance ranged from 2.3% to 2.9%. Compared with concordant normal LDL-C group, normal LDL-C by Friedewald but elevated by Martin-Hopkins or Sampson had higher ASCVD risks, with no excess risk in the reverse discordance. Normal Sampson LDL-C but elevated Martin-Hopkins LDL-C had higher risks of ASCVD [Hazards ratio (95% confidence interval): 3.94 (1.80-8.64)], whereas the reverse discordance showed no increased risk [0.86 (0.39-1.91)]. Findings were replicated in an independent mid-adult cohort (the UK Biobank study).
Conclusions
The Martin-Hopkins equation provides more accurate LDL-C–based risk classification, identifying high-risk young adults missed by other equations and minimizing misclassification of those at normal risk, supporting its use for optimizing early ASCVD prevention.