Most public heart-health content is written for the average person with lifestyle-driven risk. It doesn't work well for people whose primary risk driver is genetic — where you can do everything "right" and still build plaque, or have terrible numbers on paper and carry almost no risk. This is a living reference doc, started July 2026, built around two real (anonymised) case studies.
Not medical advice. This is a personal reference framework, not a clinical recommendation — genetics decides which parts of this apply to any given person, and that requires an actual clinician working from your own results.
Cardiovascular risk from cholesterol comes down to three questions, roughly in this order of importance:
How many atherogenic particles do you make, and for how long have you been exposed to them? Mostly genetic — set by LDL receptor function, ApoB production, Lp(a) level, and how the body clears remnant lipoproteins (this is where ApoE genotype comes in). Lifestyle nudges this, but for a lot of people it doesn't move it nearly enough.
Is that exposure actually causing plaque, right now, in your arteries? This is what imaging (CAC, CT angiogram, carotid ultrasound) answers directly, rather than inferring it from a blood number.
What's your inflammatory and metabolic context? Smoking, diabetes, blood pressure, visceral fat, hs-CRP — these determine how "reactive" your arteries are to a given particle burden, and they're the part lifestyle affects most directly.
The mistake in most public discussion is collapsing all three into "cholesterol is good/bad." Genetics sets the dose, imaging tells you the actual effect, and metabolic health is the modifier. Two people with identical LDL can have completely different outcomes depending on #1 and #3 — the exact contrast between the two case studies below.
A quick correction to how these genes are usually described, then the interesting part. There are three common ApoE variants, not four — ε2, ε3, ε4 — which pair up into six possible genotypes. ε3/ε3 is the population "default" (~60% of people); ε3/ε4 (the genotype in Case study A) sits in about a quarter of the population; ε4/ε4 is rare. There was historically a genuinely rare fourth electrophoretic variant labelled "ApoE1," but it's a distinct, uncommon mutation, not a fourth mainstream population type. The three common forms differ at just two amino acid positions, which changes how well the ApoE protein binds the liver's LDL receptor and clears cholesterol-carrying remnant particles out of the blood — that single mechanical difference drives most of the downstream lipid effects.
The evolutionary timeline runs backwards from the numbering. ε4 is the ancestral, oldest form — for roughly the first 96% of modern human history, everyone was ε4/ε4. ε3 arose from a mutation around 220,000 years ago and gradually became dominant; ε2 is the newest, appearing only around 70,000–80,000 years ago. So the ε3/ε4 genotype in Case study A is literally a mix of the original hunter-gatherer-era allele and the newer, lower-risk one.
Why hasn't the "risky" ε4 allele been bred out? A few converging, evidence-backed (though not fully settled) explanations:
The mismatch, tied together: for almost all of human history — short lifespans from injury, infection and childbirth, constant pathogen exposure, unreliable food supply — ε4's package of more available circulating cholesterol, a more reactive immune system, and better fertility was a net win, and its costs mostly arrived at ages almost nobody reached. The modern environment flips the equation: we live long enough, pathogen load is low, and food is constantly abundant, so the ancestral upside barely gets used while the downside gets 60–80 years to compound. It's the same "evolutionary mismatch" logic behind things like the thrifty-genotype model of diabetes risk or the sickle-cell/malaria trade-off — just playing out in lipid and immune biology instead.
This is worth understanding because it explains why "ultra healthy lifestyle, still progressive disease" isn't a contradiction — for a lot of people, diet was never the main lever to begin with.
Most circulating cholesterol isn't coming from your plate. The liver manufactures roughly 75–80% of it; only about 20–25% comes from food. On top of that, the body runs a tight feedback loop: eat more dietary cholesterol, and the intestines absorb proportionally less of it while the liver dials down its own production (via HMG-CoA reductase — the same enzyme statins target) to hold blood levels roughly steady. This compensation is well-documented enough that the 2015–2020 US Dietary Guidelines dropped the old 300mg/day dietary cholesterol cap entirely, because the evidence didn't support it mattering much for most people. The often-cited illustration: an 88-year-old man who ate around 25 eggs a day for years had entirely normal cholesterol, because his body increased bile acid production and cut both absorption and synthesis to compensate. Saturated fat is a somewhat more consistent lever on LDL than dietary cholesterol itself — the two get conflated a lot in public messaging, but they're not the same thing.
The exception that proves the rule: a genetically-influenced minority known as "hyperresponders" whose cholesterol does move meaningfully with diet, particularly with saturated fat. And this loops straight back to genetics — ε4 carriers are more likely to fall into that hyperresponsive group, and some trial data shows ε4 carriers get a proportionally larger LDL/ApoB benefit than ε3/ε3 people when saturated fat is swapped for low-glycaemic-index carbohydrate. So the practical read for the genotype in Case study A: diet is a genuinely bigger lever than it is for the average ε3/ε3 person, worth using deliberately rather than assuming it won't help — but it's still working against a liver that's producing and clearing particles according to a baseline set by genetics, not diet, which is exactly why lifestyle alone wasn't enough to prevent plaque from forming in that case.
| Test | What it tells you | Where it fits |
|---|---|---|
| Standard lipid panel | Cheap, universal, but LDL-C is a calculated estimate and can mislead, especially with high triglycerides | Baseline only — don't stop here with family history |
| ApoB | Direct count of atherogenic particles | The real number to track; more accurate than LDL-C |
| Lp(a) | Genetically fixed, independent risk factor | Check once in a lifetime — a high value raises targets on the rest of the panel |
| ApoE genotype | Determines remnant lipoprotein clearance speed; ε4 carriers clear slower, run higher LDL | Explains why the numbers behave as they do; flags a poorer average statin response |
| Genetic panel + Dutch Lipid Clinic Score | Confirms or scores probability of familial hypercholesterolemia | Worth doing with strong family history + high LDL from a young age |
| CAC score | Directly measures calcified plaque burden | The single best "does this matter for me" test |
| CT coronary angiogram | Sees calcified and non-calcified plaque, plus stenosis severity | Step up from CAC when borderline or genetic risk is high |
| Carotid ultrasound | Independent, radiation-free window onto systemic atherosclerosis | Good corroborating and repeatable tracking test |
| hs-CRP | Inflammatory modifier of risk on top of particle number | Adds context for borderline lipid numbers |
Lifestyle is always the floor, rarely enough alone for genetic cases: saturated-fat-aware, plant-forward diet with regular oily fish or omega-3, resistance and aerobic exercise, weight/visceral fat management, good sleep, not smoking. It meaningfully affects the inflammatory/metabolic layer and triglycerides, modestly affects LDL/ApoB — and does essentially nothing to Lp(a). Doing all of this and still progressing isn't a failure of effort, it's the genetics working as expected.
Pharmacological ladder, roughly in order of use:
Statin — first line, biggest evidence base, but ~3–5% have genuine intolerance (vs. a much larger nocebo-driven dropout rate); a supervised rechallenge is worth doing before writing them off entirely.
Ezetimibe — blocks cholesterol absorption in the gut, modest LDL reduction (~15–20%) alone, well tolerated, additive to statins.
Bempedoic acid — statin-free mechanism, no muscle symptoms, meaningfully effective; increasingly used alone or combined with ezetimibe in statin-intolerant patients.
PCSK9 inhibitors or inclisiran (injectable) — largest LDL reductions after other drugs are maxed out. In Australia, PBS access is gated fairly tightly by LDL threshold, prior therapy, and diagnosis criteria — worth clarifying exactly which criterion is being missed, since private funding is also an option.
On the horizon, not yet proven by outcomes trials: pelacarsen and olpasiran, both Lp(a)-lowering (80–95%+ reductions in phase 2) — the first real test of whether lowering Lp(a) actually prevents events, reading out through 2026.
This cuts both ways, and treating it as one-directional is where a lot of public commentary goes wrong.
Prescribing by LDL number alone, in a primary-prevention, average-genetic-risk person, produces a very poor number-needed-to-treat at the margins — around 400 over five years for the lowest-risk patients now brought into guideline thresholds, up from ~40 in the 1990s. Statins show no measurable event benefit in patients with a genuinely CAC = 0 score, absent other independent risk drivers.
Common in low-carb/keto-adjacent commentary: generalising the LMHR argument to people whose high particle count comes from a completely different mechanism — impaired LDL-receptor clearance, a poor-clearing ApoE4 genotype, or genetically fixed high Lp(a) — none of which have anything to do with diet.
The honest framework, stated plainly: genetics decides who's in which bucket, and the imaging is what actually tells you which bucket you're in — the lipid panel by itself doesn't.
High cholesterol on a standard panel, CAC score of zero, no smoking, no diabetes. On current evidence, this is close to the textbook case where imaging overrides the cholesterol number for near-term decision-making — a candidate for periodic re-scanning rather than immediate statin therapy, provided there's no family history of premature CAD or other independent risk driver. A good side-by-side illustration of why "high cholesterol" alone is an incomplete question.
None of this is a substitute for a cardiologist working from your own results. It's a framework for asking better questions, not a protocol to follow.