ApoB: The Lipid Number That Matters

Every atherogenic lipoprotein particle — LDL, VLDL, IDL, remnants, Lp(a) — carries exactly one molecule of apolipoprotein B. One particle, one apoB.

So a measurement of apoB is a count of the particles that cause atherosclerosis. A measurement of LDL cholesterol is an estimate of how much cholesterol those particles happen to be carrying.

Those are different quantities, and for a meaningful fraction of people they diverge enough to change a clinical decision.

First, the settled part

There is no serious remaining debate about whether LDL causes atherosclerotic cardiovascular disease.

A European Atherosclerosis Society Consensus Panel synthesized over 200 studies, more than 2 million participants, over 20 million person-years of follow-up and more than 150,000 cardiovascular events — combining genetic studies, prospective cohorts, Mendelian randomization analyses and randomized trials (Ference et al., 2017).

Their conclusion was unambiguous: LDL causes ASCVD, with consistent, dose-dependent relationships between exposure and disease. Any mechanism of lowering plasma LDL particle concentration should reduce event risk in proportion to the absolute reduction and the cumulative duration of exposure — provided the achieved LDL-C reduction is concordant with the reduction in particle number, and there are no competing off-target effects.

That proviso is where this article lives.

The same panel acknowledged that each LDL particle contains a single apoB molecule, making apoB a direct measure of particle number. Under most conditions LDL-C and particle concentration correlate highly. But in metabolic syndrome, diabetes and hypertriglyceridaemia, direct measurement of particle number or apoB may more accurately reflect particle concentration or its effect on risk.

What happens when you put apoB in the model

The decisive test is competitive: include apoB and the conventional lipid measures in the same model and see which survives.

Marston and colleagues did this across the UK Biobank and two large trials, FOURIER and IMPROVE-IT — 389,529 individuals in primary prevention not on lipid-lowering therapy, and 40,430 patients with established atherosclerosis on statins (Marston et al., 2022).

In primary prevention, after full lipid adjustment:

• apoB, per 1 SD: HR 1.27 (95% CI 1.15–1.40), P < .001
• Non-HDL cholesterol, per 1 SD: HR 1.05 (95% CI 0.95–1.15) — no longer significant
• Triglycerides, per 1 SD: HR 1.00 (95% CI 0.96–1.04), P = .71 — no association at all

Adjusted for apoB, triglycerides carried literally no signal.

This addresses a long-running question about whether triglyceride-rich lipoproteins pose risk through their triglyceride content. On this evidence, they pose risk through being particles. What matters is how many atherogenic particles are in circulation, not what is inside them.

Count the particles. The cargo is a detail.

How often does it actually matter?

The obvious objection: apoB and LDL-C are highly correlated, so who cares?

A 2024 analysis in the European Heart Journal answered this with unusual precision. Across 293,876 participants with 11-year follow-up and 19,982 cardiovascular events, the correlation between apoB and both LDL-C and non-HDL-C was r = 0.96 (Sniderman et al., 2024).

Very high. And still insufficient.

At an LDL-C of 130 mg/dL, capturing about 95% of observations required an apoB range spanning 85.8 to 108.8 mg/dL — a range exceeding one standard deviation. Two people with identical LDL-C can sit at meaningfully different particle counts and therefore meaningfully different risk.

The competitive test again: residual apoB remained significant after accounting for LDL-C (HR 1.06, 95% CI 1.04–1.07) or non-HDL-C (HR 1.04, 95% CI 1.03–1.06). The residuals of those cholesterol measures became non-significant when apoB was included.

The authors’ conclusion is direct: the imprecision in individual clinical decision-making that occurs when apoB is not measured is sufficient to be clinically unacceptable. They also note the cost objection is weak — measuring apoB would add roughly 1% in US healthcare terms, and could eliminate unnecessary standard lipid panels at follow-up.

Who is most likely to be discordant

The pattern that emerges from the consensus statement and the discordance analysis: discordance is most likely when triglycerides are elevated, in metabolic syndrome, and in type 2 diabetes. In these states, particles tend to be smaller and cholesterol-depleted — so you need more of them to carry the same cholesterol, and LDL-C understates the count.

This is the clinically dangerous direction. Someone with metabolic syndrome and a reassuring LDL-C may have a particle count that is not reassuring at all.

Two honest limits on all of this. The Marston and Sniderman analyses are observational, however large. And no randomized trial has yet compared an apoB-guided treatment strategy against an LDL-C-guided one for hard outcomes. The case for apoB rests on it being the more accurate measure of the causal quantity, not on a trial showing better outcomes from using it.

What to do with this

• Ask for apoB on your next lipid panel. It is a standard, inexpensive, widely available assay. Most panels omit it out of habit, not cost.
• Pay particular attention if you have elevated triglycerides, metabolic syndrome or type 2 diabetes. This is where LDL-C is most likely to understate your particle count, and where the consensus panel explicitly recommends direct measurement.
• Stop treating triglycerides as an independent lipid target. Adjusted for apoB, the hazard ratio was 1.00. High triglycerides remain a signal worth investigating — for metabolic dysfunction and pancreatitis risk — but not as a separate atherogenic driver.
• Interpret apoB against your overall risk, not a universal cutoff. Appropriate targets differ substantially between primary prevention in a low-risk person and secondary prevention after an event. That conversation belongs with a clinician.
• Remember cumulative exposure. The consensus statement is explicit that risk reduction is proportional to both the absolute reduction and its duration. A moderate apoB at 35 for thirty years is a different exposure from the same value at 65.
• Do not use a normal LDL-C to close the question. With a correlation of 0.96, an LDL-C of 130 mg/dL is compatible with an apoB spanning more than a standard deviation. That is the entire argument, in one number.

The reason to care about a lipid particle count is not sophistication for its own sake. It is that atherosclerosis is the single largest driver of the years we lose to disease, the causal agent is precisely identified, and there is a cheap test that measures it more accurately than the one most people are given. That is an unusually clean opportunity.