Comparing Low-density Lipoprotein Cholesterol Population Estimates Using Different Predictive Equations: National...

Since 1972, low-density lipoprotein cholesterol (LDL-C) has been calculated by the Friedewald equation, which estimates very low-density lipoprotein cholesterol as triglycerides divided by 5 and is accurate only for triglycerides <400 mg/dL. The Martin equation, published in 2013 (for triglycerides <400 mg/dL), replaced 5 with a factor varying over an array of non-high-density lipoprotein cholesterol and triglyceride levels. This array was extended in 2021 for triglycerides 400-<800 mg/dL. In 2020, the Sampson equation, accurate for triglycerides <800 mg/dL, was developed using multiple least squares regression. This report compares LDL-C as calculated by the Friedewald, Martin, and Sampson equations in a nationally representative sample of adults with triglycerides <400 mg/dL across the distribution of clinical cut points for LDL-C (<70 mg/dL, 70-<100 mg/dL, 100-<160 mg/dL, 160-<190 mg/dL, and ≥190 mg/dL) to assess the impact of equation choice on national estimates.

Using data on 4,461 adults in the 2015-2018 National Health and Nutrition Examination Survey, classification agreement into the LDL-C categories used for clinical management across the three equations was assessed using kappa statistics for men and women overall and by demographic subgroups. A sensitivity analysis assessed classification agreement between the Martin and Sampson equations for adults with triglycerides <800 mg/dL.

During 2015-2018, 9.8%-10.0% of adults age 20 and older had LDL-C levels <70 mg/dL (Friedewald: 10.0%, Martin: 9.8%, Sampson: 9.8%). Less than 3% had LDL-C >190 mg/dL (Friedewald: 2.3%, Martin: 2.4%, Sampson: 2.6%). Very good agreement between the equations was seen in all subgroups (kappa >0.8).

The three equations for LDL-C produce similar U.S. population-level percent distributions for adults age 20 and older across LDL-C categories.