Comparison of the Friedewald, Martin/Hopkins, and Sampson Equations for Calculating Low-Density Lipoprotein Cholesterol

Abstract Background/Aims: Ultracentrifugation, which is the reference method for low-density lipoprotein-cholesterol measurement, has been replaced by direct assays because it is not practical in standard clinical laboratories. However, due to the high cost and limited added value of direct assays, predictive equations are used. In this study, our aim was to compare the predictive performances of the Friedewald, Martin/Hopkins and Sampson equations to a direct assay when the triglyceride level is below 400 mg/dL Methods: A total of 8,632 lipid profiles analyzed with Abbott Architect c8000 and c400 devices (Abbott Laboratories, Abbott Park, IL,) were included in the study. Those that misclassified patients based on low-density lipoprotein-cholesterol thresholds of 70 mg/dL and 190 mg/dL were identified as clinically relevant errors. Results: The percentage of results meeting desirable bias specifications (±5.4%) was 52.5%, 58.2%, and 56.2% in the Friedewald, Martin/Hopkins, and Sampson equations, respectively. The percentage of results meeting the total allowable error specifications (±11.8%) was 84%, 89.1% and 87.7% in the Friedewald, Martin/Hopkins and Sampson equations, respectively. The Friedewald, Martin/Hopkins and Sampson equations showed negative proportional biases with a median bias of −2.3%, -1.9% and -0.7%, respectively. When triglyceride level was between 200–399 mg/dL, the median percentage bias was -8.2% for the Friedewald equation which exceeded the desirable bias criteria. The rates of clinically relevant errors for the Friedewald, Martin/Hopkins, and Sampson equations were 6%, 4.9%, and 5.5%, respectively. Conclusions: Our comparative analysis of LDL-C estimation equations using the Abbott platform within the Turkish population reveals that while the Martin/Hopkins and Sampson equations exhibit only marginal differences in performance, both demonstrate superior accuracy relative to the Friedewald equation. Although the Martin/Hopkins yields marginally preferable outcomes, Sampson equation remains a clinically viable alternative.