A Modified Sampson–NIH Equation with Improved Accuracy for Estimating Low Levels of Low-Density Lipoprotein-Cholesterol
Maureen Sampson, Rafael Zubirán, Anna Wolska, Jeffrey W. Meeusen, Leslie J. Donato, Allan S. Jaffe, Giorgio Melloni, Robert P. Giugliano, Marc S. Sabatine, Nicholas Marston, Alan T. Remaley
Abstract
BACKGROUND: Cardiovascular guidelines have long recommended low-density lipoprotein-cholesterol (LDL-C) as the primary target for lipid-lowering therapy. Recent guidelines have emphasized the importance of achieving low LDL-C levels; hence, the accurate measurement of low LDL-C is increasingly clinically relevant. METHODS: Using lipid panel test results from the Mayo Clinic (n = 24 590) and the FOURIER clinical trial of evolocumab (n = 9605), the following modified Sampson equation was developed by least-squares regression to match LDL-C (mg/dL) by the β-quantification reference method, by combining terms into non High Density Lipoprotein Cholesterol (nonHDLC = Total Cholesterol - High Density Lipoprotein Cholesterol) and forcing the coefficient to be one. RESULTS: The modified Sampson equation demonstrated significant improvement in its concordance to the reference method compared to other equations (the Lin Concordance Correlation Coefficient 0.992, P < 0.001). By overall kappa analysis, it showed the best agreement to the reference method at the 55 mg/dL cutpoint (1.4 mmol/L, 0.98 [P < 0.001], Sampson-NIH: 0.96, Martin-Hopkins: 0.96, Friedewald: 0.94) and the 70 mg/dL cutpoint (1.8 mmol/L, 0.97 [P < 0.001], Sampson-NIH: 0.94, Martin-Hopkins: 0.95, Friedewald: 0.92). The false classification rate of the modified Sampson equation was also significantly lower compared to the other equations at 55 mg/dL (15%, [P < 0.001], Sampson-NIH: 29%, Martin-Hopkins: 28%, Friedewald: 37%) and 70 mg/dL (18%, [P < 0.001]; Sampson-NIH: 30%, Martin-Hopkins: 28.%, Friedewald: 34%). The new equation increases the percentage of correctly classified patients with low LDL-C by approximately 10% to 20% over the other equations based on its net reclassification index. CONCLUSIONS: The modified Sampson equation shows improved accuracy compared to other equations for low LDL-C. It more accurately identifies high-risk patients, who are not at their LDL-C goals and could benefit from more intensive lipid-lowering therapy.