65 Optimizing Pediatric LDL Estimation: Comparative Performance of Friedewald, Sampson, and Martin–Hopkins Equations in a Large Pediatric Cohort

Background Accurate estimation of low-density lipoprotein cholesterol (LDL-C) is central to pediatric cardiovascular risk stratification. However, commonly used LDL-C equations were derived in adult populations, and their performance in pediatric cohorts remains incompletely characterized. Methods We conducted a retrospective analysis of 28,198 pediatric lipid panels obtained at the Children’s Hospital of Philadelphia (January 2024-January 2026). Patients age ranged from 23 days old to 83 years old with a mean age of 12.9 years old. Additionally, 50.11%% of the results corresponded to female patients. LDL-C was calculated using the Friedewald, Sampson, modified Sampson, and Martin-Hopkins equations, with Martin-Hopkins evaluated across low (3.1), mid (7.5), and high (11.9) adjustable factor ranges. Within our cohort results ranged from 50 to 1,559 mg/dL for total cholesterol, triglycerides from 16 to 2,033 mg/dL, and HDL cholesterol from 6 to 198 mg/dL. Lipid results were classified according to 2018 AHA pediatric thresholds. Concordance with non-HDL cholesterol categories (acceptable, borderline, abnormal) was assessed using percent agreement and Cohen’s kappa. Pediatric age groups were defined according to expert consensus lipid screening recommendations informed by the integrated cardiovascular health schedule, reflecting guideline-driven testing intervals rather than intrinsic age-related risk differences. Results The Martin-Hopkins equation at the highest adjustable factor (11.9) demonstrated the strongest agreement with non-HDL cholesterol classification (91.87% concordance; Cohen’s kappa = 0.834), outperforming all other equations across age and sex strata. When compared to non-HDL cholesterol classification, overall discordance was 8.12% for the Martin-Hopkins equation compared to12.30% for the modified Sampson equation . Discordance with Martin-Hopkins was primarily driven by downward reclassification (69.1%), predominantly from abnormal to borderline categories, whereas discordance with the modified Sampson equation was dominated by upward reclassification (91.7%), largely reflecting shifts from acceptable to borderline. Conclusions LDL-C estimation methods show clinically meaningful variability in pediatric populations. The Martin-Hopkins equation with a higher adjustable factor aligns more closely with non-HDL cholesterol classification, likely due to its reduced dependence on triglyceride levels compared with Sampson-based equations. In pediatric cohorts with moderate to elevated triglycerides, differences in how equations weight triglycerides can systematically affect LDL-C estimates and subsequent risk classification. These findings support the use of the high-factor Martin-Hopkins approach as a more robust strategy for pediatric lipid assessment and laboratory reporting.