How can laboratories assess LDL-C test accuracy without access to β-quantification?

Low-density lipoprotein cholesterol (LDL-C) remains a cornerstone of cardiovascular risk assessment. Yet most clinical laboratories face a fundamental challenge: the β-quantification reference method for LDL-C is technically complex, resource-intensive, and largely unavailable outside of specialized research laboratories. As a result, most clinical laboratories cannot readily assess the accuracy of their LDL-C methods.  

In routine practice, LDL-C is commonly estimated using equations that depend upon the results of the standard lipid panel¹. Although the Friedewald equation has been used for decades, newer approaches—including the Martin–Hopkins and Sampson-NIH equations—have been developed to improve the accuracy of LDL-C calculation¹. The accuracy of LDL-C equations, however, depends upon the accuracy of their input values, thus depending on how a lipid panel is done, the relative accuracy of the different LDL-C equations can change. Although not as widely utilized, LDL-C can also be directly measured using homogenous assays¹. Importantly, switching between these different methods can significantly alter LDL-C values, especially in samples with high triglyceride (TG) concentrations.  

To address the lack of accessibility of the LDL-C reference method for comparison studies, we recently described the Lipid-ratio plot². It is a simple graphical tool that allows laboratories to indirectly assess LDL-C accuracy without performing β-quantification. It is based on an linear relationship observed between two different lipid ratios, namely LDL-C/non-high-density lipoprotein cholesterol (non-HDL-C) and the square root of TG/non-HDL-C. When plotted with β-quantification test results, it produces a negative linear regression line with a characteristic slope and intercept that can be used as a benchmark. LDL-C values generated by equations or direct assays can be plotted in the same way and compared to β-quantification. Deviations in the slope from β-quantification reflect the presence of proportional bias in the test method, whereas shifts in the intercept indicate constant bias. Because the Lipid-ratio plot solely relies on lipid panel test results, it can be constructed using existing laboratory data. Simulation analyses show that as few as 80 appropriately distributed samples are sufficient to reliably estimate the slope and intercept of any LDL-C test method.  

When applied to LDL-C equations, the Lipid-ratio plot demonstrated that the Sampson-NIH equation most closely aligned with β-quantification. In contrast, the Friedewald equation showed a negative proportional bias with increasing TG, whereas the Martin-Hopkins equation showed a positive proportional bias. Certain direct LDL-C assays showed assay-specific errors, indicating that not all direct methods perform equally well. 

Like any statistical tool, the Lipid-ratio plot has limitations. As an indirect method, it depends on the validity of the established β-quantification reference relationship, but this was determined with two large databases. In addition, because it relies on ratios rather than absolute values, it is better for detecting overall systematic bias rather than evaluating accuracy at a specific LDL-C cutpoint. Despite these limitations, the Lipid-ratio plot offers laboratories a practical and inexpensive quality-assessment tool for evaluating new LDL-C methods. Its use has the potential to improve LDL-C test accuracy and cardiovascular risk assessment. Software to generate the Lipid-ratio plot is freely available at: Item - Sampson Lipid equations with Lipid Ratio Plots - figshare - Figshare

References

1. Wolska, A., et al., New Methods for Calculating LDL-Cholesterol and Related Biomarkers of Atherosclerotic Cardiovascular Disease Risk. Current Atherosclerosis Reports, 2026. 28(1): p. 27. 
2. Gcingca, T., et al., Lipid Ratio Plot: A Simple Graphical Approach for Investigating the Accuracy of LDL Cholesterol Equations or Direct Assays. The Journal of Applied Laboratory Medicine, 2025. 10(5): p. 1154-1167. 

Acknowledgement:
​This research was supported by the Intramural Research Program of the National Institutes of Health (NIH). The contributions of the NIH authors are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. 

To whom correspondence should be addressed:
​Anna Wolska, Lipoprotein Metabolism Laboratory, Translational Vascular Medicine Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, 9000 Rockville Pike, Bldg. 10/Rm. 8N220, Bethesda, MD 20892, Tel: 301-496-3707, Fax: 301-402-1885, e-mail: [email protected]