This article is a collaboration between 鶹ý and:
The LDL cholesterol number, which has been the obsessive focus of physicians and patients for several decades now, is getting a major upgrade. A new and improved method to calculate the LDL cholesterol number is starting to filter into standard laboratory reports.
Until now LDL cholesterol, which is not measured directly by standard blood tests, has been calculated from the Friedewald equation using standard lipid profile data. But the Friedewald equation is a blunt instrument, and it becomes increasingly unreliable at the low LDL levels achieved in patients taking intensive lipid lowering therapy. The new more accurate calculation is based on research performed by Seth Martin and colleagues at Johns Hopkins University.
is the first national laboratory to incorporate the new equation in its lab reports. Martin said that he hoped that Labcorp, the other lab giant, and smaller labs would also convert to his new method. (Martin said that Johns Hopkins has a pending patent on the algorithm.)
"Our algorithm and the Friedewald equation both use the same standard lipid profile data and both divide triglycerides by a number to estimate VLDL cholesterol, but the difference is a personalization versus one-size-fits-all strategy," Martin explained. "Our algorithm selects the division factor out of 180 choices based on the one that best fits the patient's lipid profile, whereas the Friedewald equation divides everyone by 5. Dr. Friedewald et al said in their original paper, 'Simple division of the plasma TG by five does not give a very accurate estimate of VLDL-C'. They tolerated the inaccuracy since patients weren't being treated down to low LDL-C so the estimate was a relatively less important part of the equation. At that time, statins, Zetia [ezetimibe], and PCSK9 inhibitors did not exist. Whereas they had a 400-person dataset, we had over a million, which allowed us to take the personalized approach."
James Stein (University of Wisconsin), said that the "revised conversion factor better reflects the cholesterol component of VLDL. It is empirically derived and is an advance. It will be especially useful in people with very low LDL-C and as triglycerides increase."
The improved LDL number produces a risk estimate that is more concordant with non-HDL cholesterol, . "When our algorithm says that the LDL-C is actually above the guideline cutpoint, non-HDL-C tends to be too," said Martin.
Martin emphasized that the change in LDL calculation does not alter the way people should interpret the LDL number. It is just a more accurate reflection of the amount of LDL. "This is an equivalent reference value for each equation. One interprets LDL-C the same way. The point here is not about recalibrating guideline cutpoints but about getting precise information on the patient to make the best decisions together with them," said Martin.