Abstract 4142639: Assessing low-density lipoprotein cholesterol (LDL-C) equations in childhood for long-term risk of cardiovascular disease events: Superior performance of the Sampson equation

Background: Low-density lipoprotein cholesterol (LDL-C) is the primary lipid factor for diagnosing and managing lipid disorders in children. Although multiple equations are available for estimating LDL-C, the best one for assessing long-term cardiovascular (CV) risk is unclear. Hypothesis: Childhood LDL-C levels estimated from different equations differently associates with adult CV events. Aim: To compare the prospective association of childhood LDL-C levels estimated by the Friedewald, Martin/Hopkins, Sampson, and DeLong equations with adult CV events. Methods: Data were from seven prospective cohorts in the US, Finland, and Australia within the International Childhood Cardiovascular Cohort Consortium. LDL-C levels were estimated using four equations: Friedewald, Martin/Hopkins, Sampson, and DeLong. Discordance was defined as differing LDL-C categorizations (acceptable <110 mg/dL versus elevated ≥110 mg/dL) between equations. CV events were adjudicated from medical records. Associations between discordance and incident CV events were estimated using proportional-hazards regression, adjusted for sex, childhood age, systolic blood pressure, and body mass index. Results: Over a mean follow-up of 35 years, 171 fatal CV events occurred among 22262 participants (49.4% male; 24.2% Black; mean age 11.9 years in childhood) and 376 fatal/nonfatal CV events among 11627 participants. Although LDL-C levels estimated by different equations were similarly associated with CV events, predictive utility varied with discordant classifications (discordance ranging from 1.2-4.7%). Children with acceptable LDL-C by the Friedewald equation but elevated by the Martin/Hopkins, Sampson, or DeLong equations had an increased risk for CV events compared to those with concordant LDL-C (Figure). Similarly, discordance between acceptable Martin/Hopkins LDL-C but elevated LDL-C by the Sampson or DeLong equations indicated an increased risk. Conversely, discordance between elevated DeLong LDL-C but acceptable Sampson LDL-C indicated a decreased risk. Conclusions: The Sampson equation provides the most accurate assessment of long-term CV risk—outperforming the underestimation by Friedewald and Martin/Hopkins and the overestimation by DeLong— suggesting its potential for estimating elevated childhood LDL-C and improving early risk stratification.