The Martin-Hopkins equation, which assesses low-density lipoprotein (LDL) cholesterol levels in blood samples, is used in laboratories in the United States and other countries to guide efforts to lower the risk of cardiovascular disease. Now, a simplified machine learning version of this equation has been shown to match the accuracy of the original in a study of millions of adult and child blood samples in the United States and is now widely accessible. The findings and code published today are: JAMA Cardiology.
We have optimized the LDL cholesterol calculation to make this formula accessible and easy for all laboratories to implement. Our goal is to help clinicians and patients prevent heart attacks and strokes and make better decisions about starting lifesaving treatments. ”
Seth Martin, MD, MHS, Senior Study Author and Director of the Progressive Dyslipidemia Program and Digital Health Lab, Johns Hopkins Center for Cardiovascular Disease Prevention
Accurate assessment of LDL cholesterol is more important than ever, as today’s guidelines recommend lowering and treating LDL cholesterol levels to reduce cardiovascular risk. However, using some formulas to underestimate LDL cholesterol can lead to missed treatment opportunities, a problem Martin-Hopkins formulas can help solve, Martin explains. Its main strength is that it provides the most accurate results for people with low LDL cholesterol, high triglycerides, and high cardiovascular risk.
“A low-cholesterol and high-triglyceride lipid profile is the ultimate stress test for LDL cholesterol calculations,” says Martin. He explains that based on various formulas, a difference of 5, 10, or 20 mg/dL can change eligibility for treatment with treatments such as PCSK9 inhibitors, which have been shown to significantly lower LDL cholesterol levels. “It is cutting-edge examples of this kind that will benefit most from more accurate results,” he added.
Because implementing the 2013 equations may require additional steps for some labs, the researchers created a streamlined code to make it easy for all labs to use.
They first created and tested their machine learning formula using blood samples from 4.9 million U.S. children and adults. These samples are representative of the US population, have a median LDL cholesterol level of 114 mg/dL, and were obtained from a very large lipid database. The researchers then compared the results of the machine learning formula to those of the original Martin-Hopkins equation and examined its accuracy with reference to other common equations used in labs: the Sampson-NIH equation and the Friedewald equation. To assess accuracy, the researchers compared their calculated results to LDL cholesterol levels assessed by ultracentrifugation, a typical tool used in research settings.
Overall, the machine learning version of the Martin-Hopkins equation was found to be similar to the original equation, with a minimum difference of 0.5 mg/dL. Both Martin-Hopkins equations correctly classified 90% of the samples into the correct treatment category, the Sampson-NIH equation correctly classified 86%, the modified Sampson-NIH equation 85%, and the Friedewald equation 83%. Most importantly, researchers found that the Martin-Hopkins equation was the most accurate for classifying high-risk patients with low LDL cholesterol levels, Martin said.
When evaluating people with triglycerides between 200 mg/dL and 399 mg/dL and LDL cholesterol levels below 70 mg/dL, the Martin-Hopkins machine learning equation correctly classified 84% of high-risk samples, the original Hopkins-Martin equation classified 83%, the modified Sampson-NIH equation classified 72%, and the Sampson-NIH The equations have been classified. 61%, 40% classified by Friedewald equation.
As part of this study, more than 3.2 million samples from the original lipid database were used to train machine learning models. An additional 1.6 million samples were used to test the model. We validated these findings in comparison to ultracentrifugation-based measurements using two additional datasets, including a reference laboratory dataset and a patient clinical trial dataset using PCSK9 inhibitors.
Martin explains that further testing the calculations in populations other than those used to develop the model to ensure its accuracy and reliability supports the ability of the calculations to be generalized and widely used in clinical implementation.
The strength of this machine learning equation, the authors note, is that labs can use the code across a variety of systems. Labs can replace the current triglyceride portion used in the Friedewald equation (the first LDL cholesterol equation created and introduced in the 1970s) with a machine-learned version of the Martin-Hopkins equation.
“This updated equation is not only highly accurate, but also transparent and easy to employ in laboratories,” says study author Mark Marzinke, Ph.D., medical director who oversees the test at the Johns Hopkins Hospital Core Laboratory and professor of pathology and medicine at Johns Hopkins University. “We wanted to avoid creating ‘black box’ equations that are opaque or invisible to most users.”
The authors explain that this open-access calculation can improve implementation of the 2026 National Dyslipidemia Guidelines, which recommend preferential use of the Martin-Hopkins calculation for the assessment of LDL cholesterol. This gives patients and their treating clinicians greater confidence in choosing treatments and achieving guideline-recommended LDL cholesterol targets to ensure optimal cardiovascular protection. Depending on your level of cardiovascular risk, guidelines recommend LDL cholesterol goals for levels below 100, 70, and 55 mg/dL.
Other authors include Jihwan Park; Leon Huang, Lori Sokol, Alagaraju Muthukumar Sabina Murphy, Mark Sabatine, Rachana Gurudu And Jeff Meusen.
The Martin-Hopkins equation has no patent or intellectual property restrictions.
The very large database of lipids used to train and test the equations was funded by the David & June Tron Foundation. The FOURIER trial, which provided blood samples to verify the findings, was funded by Amgen.
Martin has received research support from Amgen and Merck. In addition to this work, he has received consulting fees from Amgen, Arrowhead, Chroma, Heartflow, Kanika, Merck, NewAmsterdam, Novartis, Regeneron, and Verve Therapeutics. Marzinke reports receiving grants from the NIH and ViiV Healthcare.
sauce:
Reference magazines:
Park, J. others. (2026) Development and validation of a simplified Martin/Hopkins low-density lipoprotein cholesterol equation using machine learning. JAMA Cardiology. DOI: 10.1001/jamacardio.2026.2314. https://jamanetwork.com/journals/jamacardiology/article-abstract/2851049
