LDL-C calculation formulas are proposed to overcome the difficulty of standardizing the dosage but also its cost. However, a performance evaluation is required for each to determine the best formula used for the best estimate of LDL-C. This study was conducted to compare the 3 calculation formulas in comparison to the direct LDL-C test. It is a retrospective and analytical study conducted at the biochemistry laboratory of CHN Dalal Jamm. The study population includes patients with a lipid profile prescription, who met the required pre-analytical conditions. The LDL-c concentration was determined in parallel for each patient with the Friedewald, Martin-Hopkins, and Sampson-NIH formula, as well as the direct test method. The study included 119 patients with a mean age of 54±14 years, with a predominance of female (male ratio of 0.63). Statistically significant correlations and negative biases were observed between the LDL-c calculation methods (Friedewald (r=0.902; bias = -0.255), Martin-Hopkins (r=0.895; bias = -0.239), Sampson-NIH (r=0.901; bias = -2.255)) and the direct test method. LDL values were underestimated by the various calculation formulas, in particular for triglyceride levels < 1.5 g/l and LDL values ≥ 1.89 g/L. Sampson's formula showed better overall agreement at 63.87% and a lower downward reclassification rate at 32.77% compared to direct doses. Sampson's formula seems more accurate in estimating LDL-C in our population compared to Friedewald and Martin's formula.
Keywords: Friedewald, LDL-cholesterol estimation formulas, Martin–Hopkins, Sampson, Senegalese adult population, direct dosing
