A-035 Closing the Gap in Lipid Management: Modified Sampson Equation Reduces Missed Treatment Opportunities Compared to Extended Martin at Low LDL-C
Abstract
Low-density lipoprotein cholesterol (LDL-C) is associated with atherosclerotic cardiovascular disease (ASCVD) and is a target for therapy. Calculated LDL-C equations show variable accuracy compared to the gold standard beta-quantification (BQ), particularly at lower LDL-C concentrations commonly used as treatment intensification thresholds. We validated four LDL-C calculation methods (Friedewald [F], extended Martin [eM], Sampson [S], and the recent modified Sampson [mS]) against the BQ-measured LDL-C, then assessed clinical accuracy implications of the equation choice in two real-world datasets. We compared calculated LDL-C equations against BQ measurements in 1,245 patients receiving PCSK9 inhibitors or statins. Clinical impact was evaluated in two datasets: Site A (164,675 patients with ASCVD and lipid panels from 2019) and Site B (235,708 individuals with lipid panels from 2018). We quantified equation concordance at clinically actionable thresholds (55 and 70 mg/dL) using Cohen's kappa and assessed discordance patterns stratified by triglyceride (TG) levels (<150 vs =150 mg/dL). Against the BQ reference, mS demonstrated superior accuracy (mean absolute difference: mS 5.0, S 6.2, eM 7.2, F 7.8 mg/dL) and highest agreement at treatment thresholds (kappa at 55 mg/dL: mS 0.82, S 0.74, eM 0.70, F 0.72; kappa at 70 mg/dL: mS 0.90, S 0.86, eM 0.84, F 0.82). In low-TG samples, all equations underestimated BQ-LDL-C, with mS showing the fewest false-low classifications (55 mg/dL: 4% vs S 8%, eM 10%, F 9%; 70 mg/dL: 5% vs S 9%, eM 11%, F 11%). Conversely, in high-TG samples, eM significantly overestimated LDL-C whereas mS remained very close to BQ (figure). In the real-world datasets, switching from eM to mS would substantially reduce underclassification in the low-TG subgroup, where eM classified patients below treatment thresholds when mS identified them above (55 mg/dL: 25-33% of cases; 70 mg/dL: 12-20% of cases). In high-TG samples, discordance was minimal (0-5%), due to the pattern of overestimation from eM. Compared with eM, mS showed a significant net reclassification improvement (NRI) at both treatment thresholds. At 70 mg/dL, mS achieved NRI of 3.26% (95% CI, 3.12%-3.40%; p<0.001), with 6,452 upward and 1,083 downward reclassifications; notably, 94.5% of downward reclassifications occurred in high-TG patients due to eM overestimation. At 55 mg/dL, NRI was 2.90% (5,664 upward, 891 downward; 89.8% downward in high-TG patients). In low-TG patients, mS demonstrated a similar trend in performance with NRI of 4.79% at 55 mg/dL (5,432 upward, 0 downward reclassifications) and identified 2.9-fold more patients requiring treatment intensification at 70 mg/dL (54.2% vs 26.8% of upward reclassifications occurring with low-TG; p<0.001), representing an improvement of correction of a systematic LDL-C underestimation with eM. LDL-C calculation method affects treatment classification, particularly at lower concentrations where therapeutic intensification decisions occur. The modified Sampson equation demonstrates superior accuracy and reduces false-low results by approximately 50% compared to the extended Martin equation, even at normal triglyceride levels. Given that underestimation leads to missed treatment opportunities in high-risk populations, these findings have important implications for lipid management guidelines and clinical practice. Adoption of more accurate calculation methods could improve appropriate intensification of lipid-lowering therapy.