ApoB, not LDL: the number that actually predicts risk

Your LDL cholesterol (LDL-C) comes back at 104 mg/dL, dead center, the number a doctor glances at and files under fine. But that one number can sit on top of an ApoB anywhere from 78 to 118: the same cholesterol, a 40-point spread in the thing that actually lodges in an artery wall, and the standard panel can’t tell the two apart. The whole reason that gap exists is the difference between counting cargo and counting trucks.

For decades, “cholesterol” has meant one number: LDL-C. It’s on every panel, it’s what gets treated, and it’s what people quote when they say their numbers look good. It is also a proxy, and an imperfect one.

Particles, not cargo

Picture LDL as a fleet of delivery trucks moving cholesterol through your blood. LDL-C measures the total cargo, how much cholesterol is sitting in all the trucks combined. It says nothing about how many trucks are on the road.

That distinction matters because the cargo isn’t what hurts you. A plaque starts when a particle crosses the artery wall, lodges in the lining, and triggers the slow inflammatory pile-up that becomes atherosclerosis. The cholesterol is freight. The truck is what gets stuck. And every atherogenic particle carries exactly one ApoB protein on its surface, whether it’s LDL, VLDL, IDL, or Lp(a), one apiece. So ApoB is a direct count of the trucks. Count the trucks.

Here’s why the count beats the cargo. You can have trucks running half-empty: small, dense, cholesterol-poor particles, common in anyone who’s insulin-resistant. Your LDL-C reads reassuringly normal because the total freight is modest, but there are far more trucks on the road than that number implies, and each one is a separate chance to crash into an artery wall. The reverse happens too: large, cholesterol-rich particles can push LDL-C up while the actual particle count stays low. Lipidologists call this discordance, and it’s exactly the case a standard panel waves through.

ApoB vs LDL cholesterol: what the evidence says

This isn’t a fringe position. Large pooled analyses of population cohorts and statin trials have repeatedly found ApoB predicts cardiovascular events at least as well as LDL-C, and usually better; once you know someone’s ApoB, the LDL-C adds little. Mendelian randomization work, which uses inherited genetic variants as a natural experiment, points the same way, and that’s the closest thing we have to evidence of cause rather than correlation.

The honest caveat: for most people, most of the time, LDL-C and ApoB agree well enough that watching the cheaper number is fine. Discordance is the exception. The trouble is that the exception is concentrated in the people who can least afford to be missed, the metabolically unhealthy, the ones whose LDL-C looks calm while the risk underneath climbs. ApoB matters most exactly where the standard panel is least trustworthy.

”My panel was normal, so why bother?”

Fair question, and the most common one I get. If your standard lipids came back clean, the cheap number probably is tracking the expensive one, and a single ApoB draw will mostly confirm what you already suspected. So why pay for it?

Because you can’t know you’re in the agreement camp until you’ve looked once. Discordance is invisible from the LDL-C side: a normal LDL-C is consistent with both a low ApoB and a high one. One measurement tells you which of those you’re carrying. If they agree, you’ve bought peace of mind cheaply. If they don’t, you’ve caught the thing the panel was built to miss. That’s why I’d run it at least once even on a clean panel.

Why one number is never enough

Take an ApoB of 95. On its own, a little above where you’d want it, not alarming. But Depth doesn’t look at it on its own. Lined up against the last three panels, that 95 is a slope: 78, then 86, then 95, climbing about eight points a draw while every individual reading stayed inside the “normal” band. One ApoB tells you how many trucks are on the road today. Three of them tell you the fleet is growing, and this one was growing toward a wall.

Then the second signal. That climb tracked almost exactly with a falling fiber intake: as the food log thinned out on soluble fiber through a busy stretch, the ApoB rose in step. The bloodwork alone said “drifting up.” The bloodwork plus the food log said why, and what to change. Neither source could have told you that alone. Your panel sees the particles; your log sees the input. Reading them together turns a vague upward drift into a specific, fixable cause.

What to do with it

1. Measure ApoB at least once alongside your lipid panel. If it disagrees with your LDL-C, trust ApoB. It’s counting the thing that actually lodges.
2. Track the trend, not the value. Three draws over a year tell you more than any single reading, because the slope is the warning, not the dot.
3. Read it against the rest of your data. ApoB next to your fasting insulin, your hs-CRP, and your nutrition is where the cause shows up. A high particle count in someone insulin-resistant means something different from the same number in someone metabolically clean.
4. Pair it with Lp(a) once. The two answer different questions; Lp(a) is largely genetic, so you only need to know it a single time.

Run the trade the data points to, more soluble fiber and the saturated fat pulled back where it crept in, and a number like 95 starts falling toward 88 by the next draw. The line bends. You’d never know to make the trade if you kept watching LDL-C and calling it fine.

Depth lines ApoB up next to your LDL, your inflammation markers, your metabolic picture, and the months of data behind them, and tells you which number is actually steering, instead of leaving you to watch the wrong one.