A-214 Evaluate LDL-C using NIH equation 2, Friedewald equation, and direct homogeneous measurement

Abstract

Abstract Background It is essential to accurately assess Low-density lipoprotein cholesterol (LDL-C) levels, a key marker for cardiovascular disease often used for treatment recommendations. The Friedewald equation has been used for many years to calculate plasma LDL-C levels, but it has limitations. Therefore, before implementing the NIH equation, we compared LDL-C results from NIH Equation 2 to the Friedewald formula and then to Atellica direct homogeneous assay LDL-C results. Methods A retrospective study of 1200 randomly selected patients out of 53,258 data points pulled from the laboratory information (LIS) system between November 2021 and October 2023. The study evaluated triglycerides, LDL-C, total cholesterol, and HDL-C measurements. The LDL-C calculated using NIH equation 2 has been set in the LIS as a non-reportable test component since November 1, 2021. The results obtained from NIH equation 2 and Friedewald equation were compared. LDL-C levels were measured using an Atellica direct homogeneous assay for an additional 75 specimens. Deming regression analysis compared measured and calculated LDL-C using both formulae. Results The LDL-C calculated using the Friedewald equation has a mean of 106.1 mg/dL and SD 40.1 mg/dL, while the NIH equation has a mean of 108.5 mg/dL and SD 39.9 mg/dL. The average difference between both equations is 2.5 mg/dL, and there is a significant difference between both calculations using paired t-tests (p-value<0.0001). We first compared the calculated LDL-C derived from the Friedewald formula to the NIH equation 2. The regression analysis equation for the Friedewald formula was y = -2.7 + 1.001 X (r = 0.99). When the Triglyceride values <150 mg/dL (N=879), the average difference is 1.3, and the regression equation is -0.45+1.01X(r=0.99). However, when the Triglyceride value >= 150 mg/dL (N=321), the average bias is 5.4 mg/dL, and the regression equation is -11.2+1.05X(r=0.99). The Friedewald formula was compared to the measured LDL-C in 75 specimens; the regression analysis for the Friedewald formula was y = -19.6 + 1.06 X (r = 0.91, p-value <0.0001). The 75 measured LDL-C results were also compared with the NIH formula. The regression analysis of NIH LDL-C with measured LDL-C was y = -5.6 + 0.981 x (r = 0.90, p-value = 0.0002). Both formulae showed a linear relationship against measured LDL-C. The NIH formula outperformed the Friedewald formula; bias difference -7.8(-6.8%) compared to the Friedwald equation bias -12.9(-12.0%). If the NIH equation had been used during this period, an additional 1,116 patients would have had LDL-C results. This is because the Friedewald equation should not be used if the triglyceride value is above 400 mg/dL. Out of the additional patients, 513 (46%) had a high LDL (>100 mg/dL). Conclusions The NIH formula performed better than the Friedewald formula, with a less negative bias when both calculations were compared to direct homogenous measurement. The average difference between the two formulas showed the lowest value when triglyceride < 150 mg/dL. Moving to NIH equation 2 will benefit our population as we had 2% of our patients in the past two years have triglyceride >400 mg/dL.