Reducing Lipid Panel Error Allowances to Improve the Accuracy of Cardiovascular Risk Stratification.

Abstract

The standard lipid panel forms the backbone of atherosclerotic cardiovascular disease risk assessment. Suboptimal analytical performance, along with biological variability, could lead to erroneous risk assessment and management decisions. The current National Cholesterol Education Program (NCEP) performance recommendations have remained unchanged for almost 3 decades despite improvements in assay technology. We investigated the potential extent of risk misclassification when the current recommendations are met and explored the impact of improving analytical performance goals. We extracted lipid panel data for 8506 individuals from the NHANES database and used these to classify subjects into 4 risk groups as recommended by the 2018 US Multisociety guidelines. Analytical bias and imprecision, at the allowable limits, as well as biological variability, were introduced to the measured values to determine the impact on misclassification. Bias and imprecision were systematically reduced to determine the degree of improvement that may be achieved. Using the current performance recommendations, up to 10% of individuals were misclassified into a different risk group. Improving proportional bias by 1%, and fixing imprecision to 3% across all assays reduced misclassifications by up to 10%. The effect of biological variability can be reduced by taking the average of serial sample measurements. The current NCEP recommendations for analytical performance of lipid panel assays allow for an unacceptable degree of misclassification, leading to possible mismanagement of cardiovascular disease risk. Iteratively reducing allowable error can improve this.