How LDL Cholesterol Is Calculated: Friedewald, NIH and Martin-Hopkins

Why LDL is usually calculated, not measured

On a standard lipid panel, three numbers are measured directly: total cholesterol, HDL cholesterol and triglycerides. The LDL cholesterol that appears on the same report is usually not measured. It is calculated from those three values with an equation, because measuring LDL directly is more expensive and until recently required methods that were impractical for routine screening.

That is worth understanding, because it means the LDL on your report is only as good as the equation behind it and the three measurements that went into it. It also means two laboratories can print different LDL values for what is essentially the same blood sample, without either of them making a mistake.

Calculated LDL

An estimate of LDL cholesterol derived from total cholesterol, HDL cholesterol and triglycerides using an equation, rather than a direct laboratory measurement of LDL particles.

The Friedewald equation

In 1972, Friedewald, Levy and Fredrickson published a method in Clinical Chemistry for estimating LDL without using an ultracentrifuge. Their equation is simple enough to do by hand and has been the default in most laboratories for half a century: LDL equals total cholesterol minus HDL minus triglycerides divided by five, with every value in mg/dL.

The division by five is an estimate of the cholesterol carried in very low-density lipoprotein (VLDL), which is assumed to be a fixed fifth of the triglyceride value. That assumption holds reasonably well when triglycerides are moderate and breaks down as they rise, which is why the equation has firm limits.

• It requires a fasting sample, typically 9 to 12 hours, because eating raises triglycerides and distorts the fixed-ratio assumption.
• It is not valid when triglycerides are 400 mg/dL or above. A well-designed calculator refuses to return a number past that limit instead of producing one that looks plausible.
• It tends to underestimate LDL when triglycerides are raised and when LDL itself is very low, which are exactly the situations where an accurate number can matter most.

NIH Equation 2 (the Sampson equation)

In 2020, Sampson and colleagues published a new equation in JAMA Cardiology, derived and validated against beta-quantification, the laboratory reference method. It is often called NIH Equation 2. It is more complicated than Friedewald, which is the price of its accuracy, and it was designed for patients with normal lipids and with raised triglycerides alike.

Its published form is LDL = TC ÷ 0.948 − HDL ÷ 0.971 − (TG ÷ 8.56 + TG × non-HDL ÷ 2140 − TG² ÷ 16100) − 9.44, with TC the total cholesterol and non-HDL equal to TC − HDL, all in mg/dL. The practical differences from Friedewald are that it remains usable up to 800 mg/dL of triglycerides, and that it is more accurate than the older equation when LDL is low or triglycerides are raised.

The table shows the pattern. At moderate triglycerides the two agree to within a few mg/dL. At 350 mg/dL of triglycerides, the Friedewald result of 70 is 13 mg/dL lower than the NIH figure of 83, and a gap of that size could change which side of a treatment threshold a number falls on. That is why the choice of equation matters most when triglycerides are high.

This does not mean the older equation is wrong in general. For most people with moderate triglycerides, the two methods agree closely enough that the choice is unlikely to change anything. The point of knowing about the difference is to recognize the situations where it matters, so that a surprising result is checked and not simply accepted.