Performance of the enhanced Sampson-NIH equation for VLDL-C and LDL-C in a population with familial combined hyperlipidemia

Abstract

IntroductionLow-density cholesterol (LDL-C) has long been estimated by the Friedewald formula (F-LDL-C); however, this method underestimates LDL-C in patients with hypertriglyceridemia (HTG) or low LDL-C levels. The Martin (M-LDL-C) and Sampson (S-LDL-C) formulas partially resolve these limitations. Recently, Sampson et al. developed a new equation (eS-VLDL-C) that includes ApoB. This new equation could be particularly useful in FCHL, which is characterized by the predominance of triglyceride-rich VLDL and a discordance between LDL-C and ApoB. MethodsVery low-density lipoproteins (VLDL-C) was measured in 336 patients with FCHL by sequential ultracentrifugation. LDL-C was estimated by subtracting VLDL-C, estimated by the different equations, from non-HDL cholesterol. Spearman correlations, R2, mean squared error (RMSE), and bias were used to compare the accuracy of the different equations. Concordance of the estimated LDL-C values with LDL-C thresholds and ApoB was also assessed by their kappa coefficients and ROC analysis. ResultsOverall population had a mean age of 47 years, and 61.5% were women. 19.5% had type 2 diabetes, hypertension was present in 20.8%, and only 12.2% were on statin treatment. Both S-LDL-C and eS-LDL-C performed similarly, and better than M-LDL-C and F-LDL-C. In Bland-Altman analysis, eS-LDL-C showed the lowest bias, better performance in HTG, and better concordance with LDL-C treatment goals compared to other formulas (e.g. ρ: 0.87, 95% CI 0.84–0.89). ConclusionsLDL-S and LDL-eS equations estimate the concentration of LDL-C with greater accuracy than other formulas. The LDL-eS has best performance in estimating LDL-C with lower RMSE than other formulas.