How accurate is LDL cholesterol?

Watch TV and you'd get the impression that the world revolves around LDL cholesterol: Commercials for Lipitor, Zetia, Vytorin, etc., all drugs to reduce cholesterol (total and LDL). Your doctor looks first and often only at LDL cholesterol.

If there's so much attention paid to LDL, how accurate is it? 100%? 90%? 80%?

Well, it varies widely. Occasionally, it's truly accurate, but most of the time it's miserably inaccurate . Every single day, I see people with LDL cholesterols that underestimates true (measured) LDL by 40%, 50%, and even over 100%. In other words, LDL cholesterol might be 120 mg/dl by the conventional method, but the genuine measured value might be 160 mg/dl, or even 240 mg/dl. It can be that far off--and it's not rare.

The converse can occasionally be true, though rarely in my experience: that conventional LDL overestimates true LDL. I saw someone in the office today like this, with a conventional LDL of 142 mg/dl but a true measured LDL of 115 mg/dl. I may see one or two more people like this the rest of this year.

Why is LDL so inaccurate? Several reasons:

--LDL in most labs is calculated, not measured. The "Friedewald calculation" derives LDL by substracting HDL and triglycerides (divided by 5) from total cholesterol. The higher triglycerides are, especially above 150 mg/dl, the more inaccurate the calculation becomes. As HDL drops below 50 mg/dl, this also introduces greater and greater inaccuracy.

--LDL particles vary in size. A more accurate representation and measure of LDL's dangers are therefore found in measures of LDL particle number , rather than a weight-based measure or calculation. LDL particle number can be measure as just that, LDL particle number (NMR), or as apoprotein B, the protein in LDL that occurs one apoB per LDL.

I liken conventionally calculated LDL cholesterol to a broken speedometer. You simply won't have an accurate measure of how fast you're going, though you may have a ballpark sense. But try telling that to the state patrol.

Or, as a cardiologist colleague said to me in a similar conversation about LDL: "Well, it's better than nothing!"

The lesson: If you're interested in plaque control, and control or reduction of heart scan score, you need a measured LDL, preferably LDL particle number by NMR or an apoprotein B. Another option is "direct" LDL.