LONGEVITY LATEST ISSUE 23 COMPANION · 12 AUGUST 2026
LONGEVITY LATEST · DEEP DIVE
Your Test Weighs the Cargo. The Damage Is Done by the Lorries.
One protein per particle, millions of particles, and the reason lifetime exposure beats any number you can measure on a Tuesday morning.
By Christian Thomsen · Companion to Issue 23 · 12 August 2026 · ~7-minute read
You've read the newsletter, so you know where the evidence landed: an A for apoB, an A for the non-HDL number already sitting on your printout, an A for a single lifetime Lp(a), and a Hype Check on the £299 panel that sells you a more elaborate description of something you could have known for £39.
This article is about the biology underneath that, and about why the whole field spent forty years measuring the most convenient thing rather than the causal one.
Why we measure cholesterol at all
Fat doesn't dissolve in blood. To move cholesterol and triglycerides around the body, the liver packages them into lipoproteins — spherical parcels with a fatty core and a protein shell that lets them travel in water. The parcels come in families: VLDL leaving the liver loaded with triglyceride, IDL as it sheds that cargo, LDL as the shrunken remainder, and Lp(a), which is an LDL particle with a second protein bolted on. Every one of those, without exception, carries precisely one molecule of apolipoprotein B.
HDL is the exception that proves the point: it uses a different protein entirely, apolipoprotein A-I, and it does not lodge in artery walls.
Here is the historical accident. In the 1970s, measuring cholesterol was cheap and measuring proteins was not. So the field built its risk models, its trials and eventually its guidelines around the cholesterol carried inside the particles, because that was the number the machines produced. LDL cholesterol was never a claim about mechanism. It was a proxy, and it was a good one, and everyone gradually forgot that it was a proxy at all.
What actually happens in the artery wall
Atherosclerosis begins when an apoB-carrying particle crosses the endothelium — the single-cell lining of the artery — and gets stuck in the layer beneath. Particles below about 70 nanometres cross freely, which covers LDL, IDL and remnant particles. Once inside, apoB binds to proteoglycans in the arterial matrix, and the particle is retained. Not passing through. Retained.
Retained particles oxidise. Oxidised particles recruit monocytes, which become macrophages, which engorge themselves on modified lipid and turn into foam cells. Foam cells die, spill their contents, and the resulting necrotic mush becomes the lipid core of a plaque. Everything downstream of that first retention event is consequence.
Which tells you what the rate-limiting step is. Not how much cholesterol a particle is carrying. How many particles arrive at the wall.
Two particles carrying different amounts of cholesterol do the same amount of damage. The cargo varies. The lorry is what gets stuck.
This is the entire case for apoB in one sentence, and it is why a person with small, cholesterol-depleted particles — the classic pattern in insulin resistance and type 2 diabetes — can post a reassuring LDL cholesterol while carrying an unreassuring number of particles. The cholesterol assay counts the freight. The disease counts the vehicles.
The variable nobody puts on the printout
If retention is cumulative, then risk should depend not on your particle count today but on the integral of that count over your life. Exposure multiplied by years. And that prediction has been tested about as cleanly as anything in cardiology, using a trick of genetics.
Some people inherit variants that leave them with slightly lower LDL cholesterol from birth. Those variants are allocated at conception, independent of diet, income, or anything else that normally confounds an observational study — nature's own randomised trial. Ference pooled nine such variants across 312,321 participants and found that each mmol/L of genetically lower LDL cholesterol came with a 54.5% reduction in coronary heart disease.
A statin started in middle age, dropping the same number by the same amount, delivers roughly a third of that.
Same molecule. Same magnitude of change. Triple the effect — and the only variable that differs is how many decades the lower level was in force. A caveat worth noting: Mendelian randomisation estimates have been argued to be inflated, and the comparison is between a lifetime and a decade rather than between two randomised arms. But the direction has been replicated too many times to wave away, and it converges with what the artery-wall biology predicts.
What this changes about the numbers you buy
Three consequences fall out of the cumulative model, and they aren't the ones people usually draw.
The first is that the age at which you find out matters more than the precision of the finding out. A roughly correct particle count at thirty-five is worth more than an exquisitely precise one at sixty-two, because the integral has been running the whole time either way. This is also the honest argument for measuring Lp(a) young — not because anything can currently lower it, but because it tells you how hard to push everything else, starting from a decade when pushing is cheap.
The second is that discordance is not a curiosity. When apoB and LDL cholesterol disagree, the biology says the particle count is the one tracking the mechanism, and the discordance studies say the same thing empirically — apoB won nine of nine head-to-head comparisons in the 2025 review of 593,354 people. The people most likely to be misled are precisely the people whose metabolic health is already deteriorating.
The third is a caution against the thing this article could easily be used to justify. Nothing here argues for finer resolution. Particle size, subfraction distribution, small dense LDL as a separate quantity — the 2026 guideline recommends against measuring these routinely, and the mechanism explains why. If retention is driven by particle number, then splitting the count into size bins adds detail without adding decisions. Once you know how many lorries there are, their dimensions are a description, not an instruction.
What this means for you
If I were designing a protocol from this evidence, it would be boringly simple, and it would look almost nothing like a health-optimisation dashboard.
Work out your non-HDL cholesterol from the last blood test you already had — total minus HDL, no appointment, no cost. Measure Lp(a) once, before fifty, and never again. Add apoB if you have raised triglycerides, diabetes, metabolic syndrome, or a statin whose job you'd like confirmed. Then spend the remaining effort on the exposure rather than the measurement, because the integral responds to years of lower particle count and is entirely indifferent to how many markers you sampled.
The most expensive item on that list costs about thirty pounds and you buy it once.
Sources and further reading
1. Marston NA, Giugliano RP, Melloni GEM, et al. Association of apolipoprotein B-containing lipoproteins and risk of myocardial infarction in individuals with and without atherosclerosis. JAMA Cardiol. 2022;7(3):250–256. PMID: 34773460 — 389,529 statin-naive UK Biobank participants; adjusted HR 1.38 per SD of apoB; LDL-C and triglycerides lost significance after adjustment for apoB.
2. Sniderman AD, et al. ApoB, LDL-C, and non-HDL-C as markers of cardiovascular risk: a systematic review of discordance analyses. J Clin Lipidol. 2025 — 15 studies, 593,354 participants; apoB outperformed LDL-C in 9 of 9 comparisons.
3. Sniderman AD, Williams K, Contois JH, et al. A meta-analysis of LDL-C, non-HDL-C and apoB as markers of cardiovascular risk. Circ Cardiovasc Qual Outcomes. 2011;4(3):337–345 — 12 reports, 233,455 subjects, 22,950 events; standardised RRR 1.43 apoB, 1.34 non-HDL-C, 1.25 LDL-C.
4. Ference BA, Yoo W, Alesh I, et al. Effect of long-term exposure to lower LDL cholesterol beginning early in life on the risk of coronary heart disease: a Mendelian randomization analysis. J Am Coll Cardiol. 2012;60(25):2631–2639 — 312,321 participants; 54.5% (95% CI 48.8–59.5) lower CHD risk per mmol/L, approximately threefold the effect seen with statins started later in life.
5. 2026 ACC/AHA/Multisociety Guideline on the Management of Dyslipidemia. Circulation. 2026 — restores LDL-C and non-HDL-C goals; recommends Lp(a) measurement at least once in all adults; recommends against routine advanced lipoprotein testing (NMR, ion mobility, gradient gel electrophoresis, density gradient ultracentrifugation).
6. Kronenberg F, Mora S, Stroes ESG, et al. Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. Eur Heart J. 2022;43(39):3925–3946 — lifetime-risk modelling across Lp(a) strata in 415,274 UK Biobank participants of European ancestry.
7. Ho HVT, Sievenpiper JL, Zurbau A, et al. The effect of oat β-glucan on LDL-cholesterol, non-HDL-cholesterol and apoB for CVD risk reduction. Br J Nutr. 2016;116(8):1369–1382 — 58 RCTs, n=3,974; median 3.5 g/day lowered apoB by 0.03 g/L.
8. National Institute for Health and Care Excellence. Cardiovascular disease: risk assessment and reduction, including lipid modification. NICE guideline NG238. 2023 — UK guidance; non-HDL cholesterol as the preferred measure; non-fasting samples acceptable.
© 2026 FrontWave Media Ltd · Longevity Latest