Comparison of the Friedewald, Sampson, and Martin Formulas with Direct Homogeneous Assay of Low-Density Lipoprotein Cholesterol in South American Patients.

BACKGROUND: Increases in obesity and hypertriglyceridemia worldwide, together with lower LDL cholesterol (LDL-c) values from new lipid-lowering agents, prompted the development of new estimation formulas including Martin (M-LDL-c), extended Martin (ME-LDL-c), and Sampson (S-LDL-c) equations. These require validation in different populations. METHODS: Lipid profiles with direct LDL-c (D-LDL-c) by homogeneous assay were collected from 22 748 patients in Argentina. Friedewald (F-LDL-c), S-LDL-c, M-LDL-c, and ME-LDL-c were calculated. Patients were classified into treatment categories using both D-LDL-c and each formula independently. Datasets were subdivided by triglycerides (TG). Lipid ratio plot (LRP) provided indirect comparison to beta quantification (BQ). Passing-Bablok, Bland-Altman, and clinical concordance (kappa index) analyses assessed formula performance against D-LDL-c. RESULTS: LRP analysis showed S-LDL-c had the closest agreement with BQ surrogate line, followed by D-LDL-c. All formulas showed good correlation with D-LDL-c (r > 0.95) and moderate concordance (kappa > 0.6). M-LDL-c and ME-LDL-c demonstrated the best clinical concordance with D-LDL-c overall. Performance decreased with increasing TG for all formulas. F-LDL-c deteriorated from 200 mg/dL (2.28 mmol/L) TG, while S-LDL-c, M-LDL-c, and ME-LDL-c maintained acceptable performance up to 400 mg/dL (4.56 mmol/L) TG. CONCLUSION: S-LDL-c, M-LDL-c, and ME-LDL-c showed better agreement with D-LDL-c than F-LDL-c in this South American population. LRP analysis suggests S-LDL-c may provide most accurate estimation vs gold standard BQ, particularly in hypertriglyceridemia. These findings support the adoption of newer equations in clinical practice.