Comparison of Various Low Density Lipoprotein Cholesterol Calculators: Is It Time for the Friedewald Equation to Go in India?

BACKGROUND AND OBJECTIVES

Low-density lipoprotein cholesterol (LDL-c) is the main target in the management of dyslipidemia in primary and secondary prevention of atherosclerotic cardiovascular disease (ASCVD). Our study aims to compare the conventional and most popular Friedewald equation, and the more recent Sampson equation (literally unknown in India but more accurate), to the directly estimated LDL-c by the homogenous method. No such comparative study has been done in India.

METHODS

A multicenter retrospective analysis of the lipid profiles of patients whose LDL-c was derived directly by the homogenous method was included. Indirect LDL-c was estimated using the Friedewald and Sampson equations. The accuracy, error, and level of correlation between direct LDL-c and the two equations were assessed for concordance and discordance based on LDL-c categories (≤50, 51-70, 71-100, 101-130, ≥131 mg/dL). Upward discordance is when the calculated LDL-c was at least one category higher and downward discordance is when the calculated LDL-c was at least one category lower than the corresponding direct LDL-c (DLDL-c). Subgroup analyses were also done for triglycerides (TGLs) <150, 151-199, ≥200 mg/dL. TGL >400 were assessed separately.

RESULTS

A total of 30,810 samples from four centers were studied. The overall mean DLDL-c was 107.86 mg/dL. The mean estimated LDL-c by Friedewald's equation (FLDL-c) and Sampson's equation (SLDL-c) were 101.37 and 104.37 mg/dL, respectively. There was statistical significance (-value < 0.0005) in the mean difference when DLDL-c was compared with combined SLDL-c and FLDL-c as per the Games-Howell multiple comparison study, in which the mean difference was greater with the Friedewald equation (6.495) than with the Sampson equation (3.491). The overall concordance between FLDL-c vs DLDL-c and SLDL-c vs DLDL-c was nearly similar (74 vs 77%, respectively). The overall discordance upward was 5% with FLDL-c and 7% with SLDL-c (-value < 0.0005). The overall discordance downward for FLDL-c and SLDL-c was 22 and 15% (-value < 0.0005). When subgroup analysis was done for LDL-c ≤ 70, the downward discordance was 41% (FLDL-c) and 32% (SLDL-c), giving a false sense of goal LDL-c being achieved more often with FLDL-c. For LDL-c > 70, the corresponding downward discordance was 16 and 11%, respectively. Starting from TGL ≥ 200 mg/dL, there was consistent underestimation of LDL-c by the Friedewald equation. When TGL was > 400, the mean DLDL-c was 109 mg/dL, while for FLDL-c, it was 82 mg/dL and for SLDL-c, it was 98 mg/dL.

CONCLUSION

Sampson's equation improves the accuracy of the indirectly calculated LDL-c compared to Friedewald's equation. The FLDL-c is estimated falsely low at LDL-c ≤ 70 mg/dL, giving a lower sense of LDL-c goal being achieved, and these patients will be undertreated. By switching to Sampson's equation, 9% in the LDL-c ≤ 70 mg/dL and 5% in the LDL-c > 70 mg/dL can be correctly classified. The shift to Sampson's equation requires a simple formula change and can be easily embraced in India-a small change but in the right direction.