ApoB vs LDL vs Non-HDL: Which Cholesterol Number Predicts Heart Risk?

Order a cholesterol panel and the number that gets all the attention is LDL cholesterol. Ask a little more and you start hearing about non-HDL cholesterol and ApoB. The frustrating part is that these three numbers often disagree, and the disagreement is not a rounding error: one can call you fine while another says your risk is high. This guide explains what each one measures, why they diverge, and which to trust when they do.

The quick answer

If you remember one thing, remember the order. Across more than 233,000 people, the three numbers line up in a clear hierarchy of accuracy ^[1]. ApoB is the most accurate predictor of heart disease, non-HDL cholesterol is second, and plain LDL cholesterol, the number your report highlights, is the weakest of the three.

ApoB > non-HDL-C > LDL-C. That order holds in study after study, and the gap matters most for exactly the people who assume they are safe.

What each number actually measures

All three are trying to estimate the same thing: how many cholesterol-carrying particles you have that can lodge in an artery wall and build plaque. They just do it with different precision.

• LDL cholesterol (LDL-C) is the amount of cholesterol carried inside your LDL particles. It is a measure of cargo weight in one type of vehicle, and on a standard panel it is usually calculated rather than measured directly. See the LDL cholesterol guide.
• Non-HDL cholesterol is not a separate test you order; it is a number you calculate from your standard lipid panel by taking your total cholesterol and subtracting your HDL. That one subtraction captures the cholesterol in every plaque-causing particle, not just LDL but also VLDL, IDL, and Lp(a). No extra test, no fasting, and it is already sitting on your last lab report.
• ApoB counts the particles themselves. Every atherogenic particle carries exactly one ApoB protein, so one ApoB measurement is a direct headcount of every vehicle on the road, no matter how much cholesterol each happens to carry.

Why they disagree: the discordance problem

Here is the catch that makes LDL-C unreliable. The amount of cholesterol packed into each LDL particle is not fixed. Some people carry their cholesterol in a few large, cholesterol-rich particles; others carry the same total in many small, cholesterol-poor ones.

Because LDL-C measures cholesterol and ApoB measures particles, the two can point in opposite directions. This is called discordance, and it is common. In the Framingham Offspring Study, when particle count and cholesterol disagreed, cardiovascular risk tracked the particle count, not the cholesterol ^[2]. The physiology behind it is well described: ApoB captures atherogenic risk that any cholesterol-based measure structurally cannot ^[3].

In plain terms, it is the number of particles battering your artery walls that drives plaque, not the total cholesterol they happen to be carrying. When the two disagree, the particle count wins.

Why LDL can be normal but ApoB is high

The most important consequence of all this: you can have a perfectly normal LDL cholesterol and a high particle count at the same time.

It happens most in people with insulin resistance, metabolic syndrome, high triglycerides, or type 2 diabetes. In these states the liver churns out many small, dense LDL particles. Each particle carries only a little cholesterol, so LDL-C reads reassuringly low, but there are so many particles that ApoB, which counts them, runs high, and so does your risk. The big meta-analysis found ApoB's advantage was most pronounced in exactly these groups, precisely where LDL-C is most likely to be misleadingly low ^[1].

So if you have a normal LDL alongside high triglycerides, a large waist, or a raised fasting insulin, your LDL number may be hiding your real risk. ApoB is how you see it.

The evidence: which number wins

This is not a fringe view. The hierarchy comes from large, repeated studies.

• The defining meta-analysis pooled 12 prospective studies, more than 233,000 people and nearly 23,000 cardiovascular events, and found ApoB the most accurate of the three, with non-HDL-C clearly beating LDL-C ^[1].
• A separate meta-analysis of 8 statin trials, over 38,000 patients, went further: once you account for ApoB or non-HDL-C, LDL-C adds no independent predictive value at all. People who reached a good LDL-C but still had high ApoB or non-HDL-C stayed at elevated risk ^[4].
• And this is not just statistical. ApoB reflects the particles that are causally responsible for atherosclerosis, which is why lowering it lowers events ^[5].

What to do about it

You do not have to choose between expensive testing and flying blind. There is a clear, practical ladder.

1. Read your non-HDL cholesterol now, for free. It is on every lipid panel you have ever had. Take your total cholesterol and subtract your HDL. That one subtraction gives you a better risk marker than the LDL number your report puts in bold. Ask for it, or just do the math yourself.
2. Order ApoB to get the real picture. It is an inexpensive, widely available blood test that needs no fasting, and it tells you how many atherogenic particles you actually carry. Unlike Lp(a), ApoB responds to diet and lifestyle, so it is also worth retesting to track your progress. The ApoB guide covers ranges and how to lower it; the heart-risk testing guide shows the cheapest way to order it.
3. If you are building a heart panel, ApoB belongs in it. The Cardiac Panel pairs ApoB with the other markers a standard cholesterol test leaves out, like Lp(a) and inflammation.

The bottom line

LDL cholesterol is not useless, but it is the least reliable of the three numbers, and it fails most often in the people who most need an accurate read. Non-HDL cholesterol is a free upgrade hiding on every lab report. ApoB is the gold standard, and it is cheap. When your cholesterol tests disagree, trust the one that counts particles, because your arteries are counting them too.