The 10-year risk model that drives statin guidelines is fundamentally broken: it predicts low risk in people aged 35–55 even though nearly half of all heart attacks and strokes occur before age 60, because the disease has been building silently for decades.
2
ApoB — the count of atherogenic lipoprotein particles — is a more accurate index of cardiovascular risk than LDL cholesterol, non-HDL cholesterol, or triglycerides; Mendelian randomization confirms it is causal, and it can be measured for two dollars.
3
A coronary calcium score of zero does not mean you are safe: calcification marks advanced, late-stage disease; someone with high apoB and a zero CAC can still have aggressive plaque development — the score's false negative is the real danger.
4
Lp(a) measured once in every patient acts as a particle-per-particle multiplier on top of apoB; when both are elevated it creates a double whammy that dramatically shifts the treatment calculus toward early, aggressive lipid lowering.
Protocols
Concrete recipes — what, when, how much, and why
6 items
Replace LDL-C with apoB as the primary lipid treatment target
WhatOrder apoB on every lipid panel. Use apoB as the target for statin titration and add-on therapy decisions. For most treated patients, apoB is the only number needed to follow.
WhenAt the time of any lipid evaluation, whether initial screening, follow-up on statin therapy, or assessment of a patient with metabolic syndrome or diabetes.
DoseLifelong. ApoB was standardized in 1994 and costs roughly $2 at the lab.
For whomAll patients undergoing cardiovascular risk assessment. Critical in patients with metabolic syndrome, hypertriglyceridemia, or diabetes where LDL-C to apoB discordance is greatest.
WhyApoB counts the number of atherogenic particles; LDL-C measures the cholesterol mass inside them. When particles are small (insulin-resistant phenotype), LDL-C systematically underestimates particle count. Mendelian randomization confirms apoB is causal; LDL-C is not the causal driver — the particle is.
CaveatsIn Type III dyslipidemia, apoB alone is insufficient for diagnosis — triglycerides and total cholesterol are needed alongside apoB to diagnose the condition via the apoB app formula.
A doctor receiving a standard lipid panel has five numbers — but if LDL-C is the only therapeutic target, four of those numbers are effectively doing nothing. Adding apoB means you have one number that summarizes all atherogenic particle exposure and that is actually causally actionable. The guest notes: it ticks him off that apoB is not measured everywhere because if you explain to a patient that they have too many bad cholesterol particles, they get it immediately.
Mechanism
Every atherogenic event begins with an apoB-bearing particle crossing the endothelium and getting trapped in the subintimal space. Fewer apoB particles means fewer trapping events per unit time. This is how statins work: they lower apoB particle number, which reduces the flux of particles into the arterial wall.
i can measure your apo b pretty precisely and i know it's yours — when i follow a patient on statins i only have to get one number right
Also said
“particles more than cholesterol — that hasn't moved the american guidelines but on the evidence side apob is better”— The evidence summary: particle number supersedes cholesterol mass on every trial that has examined the discordance.
Screen Lp(a) once universally; treat it as an apoB modifier, not a standalone trigger
WhatMeasure Lp(a) once in every adult patient at first lipid evaluation. If high, note it as a risk modifier. If both Lp(a) and apoB are elevated, lower apoB more aggressively than you would for apoB elevation alone.
WhenSingle baseline measurement at any age. Lp(a) is approximately 90% genetically determined and does not change meaningfully over a lifetime, so repeat testing is unnecessary.
DoseOnce. Use the result lifelong as a contextual modifier.
For whomUniversal screening. Especially important in patients with a family history of premature cardiovascular disease who have otherwise unremarkable apoB levels.
WhyOn a per-particle basis, Lp(a) is more atherogenic than a standard LDL particle. But Lp(a) absolute contribution to risk depends on how much atherogenic burden apoB is already creating. The interaction is multiplicative.
There are currently no approved Lp(a)-specific therapies. The practical clinical response to high Lp(a) is to lower apoB even further than you otherwise would — compressing the overall atherogenic particle load. The guest uses Lp(a) primarily as a framing tool with patients to explain why their normal lipid panel may still confer elevated risk.
i measure an lp little a in everybody and i measure apob i measure the lp little a once when lp little a is high but apob is normal lp little a may not add that much to risk but when you got two of them it's a double whammy
Use the 30-year causal benefit model to frame early treatment decisions for patients under 55
WhatFor patients aged 30–50 with elevated apoB or non-HDL cholesterol, calculate the 30-year cumulative event risk and the projected benefit of starting treatment now vs at age 45 or 55. Present both numbers to the patient.
WhenAt initial lipid evaluation for any patient under 55 with elevated apoB. Specifically important for patients in the 7.5–15% 10-year risk range who fall just below or above statin thresholds.
DoseOne-time calculation updated as apoB changes with treatment.
For whomPatients under 55 with elevated apoB, especially those skeptical about starting statins because they have low 10-year risk per standard calculators.
WhyA 4% 10-year risk number produces inaction. A 30% lifetime event risk before age 65 produced by the same apoB level is actionable. The same pathophysiology, reframed into the time horizon that actually matters to a 38-year-old.
CaveatsThe benefit of early treatment is concentrated in patients with HIGH non-HDL or apoB. Patients with low apoB gain little from treating at 35 vs 55 — this model does not justify universal statin prescribing from age 35.
Published in JAMA Cardiology and Circulation 2019. The key finding: roughly 20% of the population has sufficiently elevated apoB or non-HDL to justify early treatment. The other 80% lose little by waiting until standard guidelines trigger treatment. The model is thus a precision targeting tool, not a mass-prescribing argument.
i say your chances of having an infarct or a stroke before you're 65 are 30 now that's a number you can deal with that's a number that has meaning and we could also calculate how much the risk can be reduced by starting at age 35 or how much you lose by starting at age 55
Diagnose Type III dyslipidemia using the apoB app when cholesterol and triglycerides are both elevated
WhatWhen a patient presents with elevated total cholesterol AND elevated triglycerides but low apoB, suspect Type III dyslipidemia. Use the apoB app — enter total cholesterol, triglycerides, and apoB to get the dyslipidemia subtype diagnosis.
WhenWhenever a new lipid panel shows elevated total cholesterol plus elevated triglycerides. Also when a patient's LDL-C appears paradoxically normal given elevated total cholesterol.
DoseDiagnostic tool — single application per patient presentation.
For whomAny patient presenting with the combination of elevated cholesterol and elevated triglycerides, typically emerging in the mid-30s to 40s.
WhyType III is more common than FH, highly atherogenic, and almost always missed in the U.S. because apoB is not measured. It requires treatment with statins plus fibrates and responds dramatically well. The diagnosis cannot be made from standard lipid panels alone.
CaveatsType III cannot be characterized by apoB alone — the low apoB alongside high cholesterol and high triglycerides is the diagnostic signature. The formula uses all three values.
Type III patients have VLDL particles that fail to be cleared normally by the liver; they circulate and accumulate cholesterol, becoming very cholesterol-rich. Despite the low particle count, these remnants are highly atherogenic per particle. The guest says: we can tell a 38-year-old they have a big problem and with treatment we can take that big problem away — usually within months on statins plus fibrates.
Mechanism
In normal metabolism VLDL converts to IDL then LDL via sequential triglyceride stripping. In Type III this conversion stalls; large cholesterol-rich VLDL remnants circulate long-term, gaining more cholesterol. Result: high cholesterol, high triglycerides, but relatively few particles — hence low apoB.
type 3 cannot be diagnosed in most patients in the united states because apob is not measured — but it can be diagnosed based on the triglycerides total cholesterol and apob there's a formula that we devised
Use CAC as a tiebreaker for borderline treatment decisions — not to override high apoB with a zero score
WhatOrder a coronary calcium score only in patients who are genuinely on the treatment or no-treatment threshold AND who want additional information to make the decision. Do not use a zero CAC to reassure or defer treatment in a patient with clearly elevated apoB.
WhenPatients aged 40–60 who have intermediate 10-year risk and are uncertain about starting statin therapy. Not indicated for patients already clearly above (high apoB + risk factors) or clearly below (low apoB, no risk factors) the treatment threshold.
DoseSingle CT scan. Radiation approximately 1 mSv.
For whomThe patient who says they are not convinced they need a medication. A positive CAC can overcome that hesitation. The key exclusion: CAC should not be used to deny treatment to a patient with clearly elevated atherogenic exposure.
WhyA positive CAC confirms high short-term risk and motivates treatment. A zero CAC provides perhaps 5–10 years of relative safety — but it does not mean disease is absent in a patient with high apoB, because calcification is a late marker of disease already developing for years.
CaveatsCAC is most often positive in men over 60 and women over 65 — ages where the clinical decision is already obvious without imaging. Its information value is highest in the under-60 population.
In a 70-year-old with a CAC of 50, it is sort of like so what — the disease is expected at that age. In a 45-year-old deciding whether to start a statin, a CAC of 200 is meaningful for treatment motivation. But a zero does not override high apoB because plaque can develop for years without calcification before the first calcium signal appears.
if somebody has a high apob the fact that their coronary calcium is negative doesn't mean they don't have a lot of disease and that the disease isn't developing at a rapid rate
Also said
“coronary calcium is negative you're okay and that's the problem for me from my knowledge and interpretation of literature and the pathological studies coronary calcification is an advanced side disease means advanced disease is present”— Explains mechanistically why zero calcification does not equal clean arteries in a high-apoB patient.
In metabolic syndrome patients, always measure apoB alongside LDL-C — discordance is greatest in this group
WhatIn insulin-resistant patients with elevated triglycerides, low HDL-C, obesity, or type 2 diabetes, always measure apoB alongside LDL-C. Treat to apoB target, not LDL-C target.
WhenAt every lipid evaluation in any patient with the metabolic syndrome phenotype.
For whomPatients with metabolic syndrome (high triglycerides, low HDL-C, elevated fasting glucose, central adiposity) and patients with type 2 diabetes — the groups where discordance between LDL-C and apoB is largest.
WhyAs insulin resistance and metabolic syndrome become more prevalent, the fraction of the population whose apoB is substantially higher than their LDL-C predicts grows. Treating to LDL-C targets while leaving apoB elevated is the probable explanation for the recent rise in cardiovascular events despite widespread statin therapy.
As obesity and insulin resistance become more prevalent, a larger fraction of patients have the high-triglyceride lower-HDL phenotype associated with greater discordance between LDL-C and apoB. Treating their LDL-C while their apoB remains elevated produces a false sense of security. This is the structural explanation for why cardiovascular event rates have risen in the past five years despite statin therapy.
as dyslipidemia is growing in the metabolic context meaning if you have more insulin resistance and more type 2 diabetes we know that those phenotypes are associated with greater discordance between apob and ldlc suggesting that you have a greater and greater portion of the population that is being undiagnosed
What's new
Personal practice updates, fresh positions, predictions
6 items
The 10-year risk model produces prevention that starts at age 55–60 — too late for half of all events
~15 min
Current ACC/AHA guidelines calculate 10-year risk using age and sex as the dominant inputs; cholesterol and blood pressure contribute minimally. As a result, almost everyone under 55 scores low-risk, yet nearly half of all first heart attacks and strokes occur before age 60. The disease starts in the first three decades of life and only precipitates clinical events in the fourth. Guidelines are preventing the complications of disease rather than the disease itself.
Why this matters: The structural flaw is not the concept of risk — it is that the risk window is too short and the dominant determinant is age, not the thing you can actually change: apoB.
Background
Herbert Stary's pathology textbook, which showed atherosclerosis in autopsy sections of 26- and 27-year-olds killed in accidents, is what crystallized for Attia that disease onset is decades before clinical presentation.
The guest (Dr. Alan Sniderman, lipidologist at McGill) published a paper in JAMA Cardiology pointing out exactly this failure. Nordestgaard and colleagues at the European guidelines level documented the same blind spot. The mathematical reason: there are far more people under 60 than over 60, so the absolute number of events before 60 is enormous even when the per-person rate is low. By the time guidelines authorize treatment at age 55–60, the disease is well advanced in the arterial wall and can no longer be reversed — only its progression can be modified.
prevention really starts at 55 to 60 but half almost half of all infarcts and strokes occur before the age of 60
Also said
“a lot of our statin prevention therapy is to prevent the complications of disease not to prevent the disease”— States the core critique of the current paradigm in one sentence.
ApoB causality confirmed by Mendelian randomization — it outperforms LDL-C, non-HDL-C, and triglycerides
~55 min
Mendelian randomization studies using gene variants fixed at conception show that apoB is causally related to cardiovascular events and incorporates all the predictive information contained in LDL cholesterol, triglycerides, and even HDL cholesterol. HDL, despite its epidemiological association, is not causally protective — you cannot manipulate it (as CETP inhibitors showed) and change outcomes. ApoB is.
Why this matters: The causality debate has been running since the 1970s; Mendelian randomization closes it. Two different tools — the CETP-inhibitor trials and the Mendelian analyses — both point to the same conclusion.
Background
Bob Lees in 1971 first demonstrated that apoB particle count had no clear relationship to plasma triglycerides or LDL cholesterol — the discordance principle. Ron Krauss and John Gofman earlier showed particles varied in size, which varies cholesterol per particle.
George Davey Smith, a founder of Mendelian randomization methodology, authored analyses showing apoB incorporates and beats triglycerides and LDL-C. The European 2019 guidelines explicitly cited this evidence to upgrade apoB. Yet U.S. guidelines still anchor on LDL-C. The guest finds it stunning that the evidence has not moved the American guidelines in the same way: when every expert in the room has the same prior, it blocks updating.
the mendelian randomization show hdl is not causal whereas they show apob is those are two very important studies
Also said
“apob incorporates and therefore beats triglycerides and ldl cholesterol”— States the Mendelian randomization conclusion directly.
“there's no question that the number of ldl particles is a more accurate index of risk than the ldl cholesterol”— The guest's direct clinical statement on the hierarchy of metrics.
Zero coronary calcium score is a false negative in patients with high apoB — it reassures when it should not
~1 h 20 min
Coronary calcification is a marker of advanced, late-stage atherosclerosis — not early disease. Someone with high apoB and a zero calcium score can still have rapid, non-calcified plaque development. The clinical error is using a negative CAC to rule out the need for treatment in a patient with a clear causal driver. CAC zero buys perhaps 5–10 years of safety, not a clean bill of health.
Why this matters: This reframes CAC's utility: it is a tiebreaker for uncertain cases, not a gatekeeper for treatment. The corollary that a zero scan means you are fine is the most dangerous misuse.
Background
CAC was developed as an imaging adjunct in an era when the only alternative was LDL-C; it was a huge advance. But it was never designed to override a measurable causal driver like apoB.
The guest's specific concern: by the time a man is 60, all American men are at high risk per guidelines anyway — so the test is most likely to be positive at exactly the age where you no longer need it to make a treatment decision. Its incremental value is in younger patients (under 55) who are on the treatment threshold. But even there, a zero score should not override a high apoB, because plaque can develop without calcification for years before the first calcium signal appears.
if somebody has a high apob the fact that their coronary calcium is negative doesn't mean they don't have a lot of disease and that the disease isn't developing at a rapid rate
Also said
“coronary calcium is negative you're okay and that's the problem for me from my knowledge and interpretation of literature and the pathological studies coronary calcification is an advanced side disease means advanced disease is present”— States explicitly why the zero-score reassurance is clinically misleading.
Type III dyslipidemia: missed diagnosis because apoB is not measured — high cholesterol, high triglycerides, low apoB
~45 min
Type III dyslipidemia (remnant-type) is a highly atherogenic condition characterized by elevated total cholesterol and elevated triglycerides but — paradoxically — low apoB. The VLDL particles fail to convert normally to LDL; they circulate long, become cholesterol-rich, and are highly atherogenic. The diagnosis is invisible if only triglycerides and LDL-C are measured. It is more common than familial hypercholesterolemia and responds well to statins plus fibrates.
Why this matters: This syndrome cannot be diagnosed without apoB — the very test that most physicians in the U.S. do not order. Patients get misclassified as high-cholesterol with high triglycerides when the real diagnosis is a distinct, treatable lipoprotein processing defect.
The guest developed a calculator (apoB app at apobapp.app) that uses total cholesterol, triglycerides, and apoB to diagnose any atherogenic dyslipidemia subtype. Type III presents around age 35–40 and can be treated to near-normal within months. The guest frames not measuring apoB as an ethical failure: without it, you cannot characterize a lipid phenotype at all.
type 3 cannot be diagnosed in most patients in the united states because apob is not measured
Also said
“there is no phenotype without putting an apob in there because they're lipoprotein particles they're disorders of lipoprotein particle metabolism”— The philosophical statement: every dyslipidemia subtype requires apoB to be fully characterized.
The causal benefit model: 30-year risk projection replaces 10-year risk for treatment decisions starting at age 35
~20 min
Sniderman, with collaborators at McGill, Duke, and Harvard, developed the causal benefit model — using apoB or non-HDL cholesterol to project risk over 20–30 years rather than 10. For a 35-year-old with high non-HDL, the 30-year risk of a first infarct or stroke before age 65 may be 30%. That number is actionable. The model also calculates how much benefit is lost by delaying treatment from age 35 to 45 to 55.
Why this matters: A 30% lifetime event risk is a number patients can act on; a 4.1% 10-year risk produces inaction. The model shifts the conversation from 'your current risk' to 'your causal exposure accumulated over time.'
Background
The paper was published in JAMA Cardiology. A companion Circulation paper (2019) quantified the costs of delay of intervention starting at ages 35, 45, and 55.
Key finding from the 2019 Circulation paper: if your non-HDL is LOW, starting treatment at 35 vs 55 gives very little extra benefit. The gain is concentrated entirely in people with HIGH non-HDL or high apoB. That means the model is not a case for mass statin therapy from age 35 — it is a case for identifying the roughly 20% of the population with elevated atherogenic particles and treating them early and aggressively. The rest can wait.
i say your chances of having an infarct or a stroke before you're 65 are 30 now that's a number you can deal with that's a number that has meaning
Also said
“we published a paper in circulation i think it's 2019 looking at what are the costs of delay of intervention starting at age 35 45 and 55”— Names the specific paper that quantifies the early-vs-late treatment benefit calculation.
Lp(a) screened once per patient; it is only additive risk when apoB is also elevated
~48 min
The guest measures Lp(a) once in every patient. On a per-particle basis Lp(a) is more atherogenic than a standard LDL particle. But when Lp(a) is high and apoB is normal, the overall risk elevation is modest. The dangerous phenotype is high Lp(a) plus high apoB — a double whammy that materially changes the treatment threshold.
Why this matters: Reframes Lp(a) not as a standalone risk factor but as a conditional modifier of apoB-driven risk. Pure Lp(a) elevation without apoB co-elevation may be over-weighted in current clinical practice.
The guest uses Lp(a) measurement primarily as an additional frame for the patient conversation about their future risk — not as an independent treatment trigger. The emerging Lp(a)-lowering drugs (PCSK9 inhibitors partially lower it; dedicated RNA therapies are in trials) have not yet produced outcome data, so the clinical management of isolated high Lp(a) remains indirect — namely, drive apoB even lower to offset the additional per-particle risk.
i measure an lp little a in everybody and i measure apob i measure the lp little a once when lp little a is high but apob is normal lp little a may not add that much to risk but when you got two of them it's a double whammy
Recommendations
Products, supplements, and tools mentioned in the episode
Free web-based calculator that takes total cholesterol, triglycerides, and apoB as inputs and returns the lipoprotein disorder diagnosis — including Type III remnant dyslipidemia, familial hypercholesterolemia, and other atherogenic dyslipidemias.
The app was available as a mobile app but as of the episode had stopped working on mobile; it remains functional on the web at apobapp.app. The guest describes it as the practical diagnostic tool he uses to characterize every patient's lipoprotein phenotype. It walks the physician through the diagnostic logic step by step. Attia uses it to look up edge cases. The guest and Attia discuss how a doctor receiving five numbers from a standard lipid panel effectively has only one actionable number — LDL-C — and four that are decorative. The apoB app converts the combination of values into a diagnosis with treatment implications.
apo b app where you could plug in the total cholesterol triglyceride and apob and you get the diagnosis of any of the atherogenic capable dystrophies
Pathology of the Human Atherosclerotic Plaque (Herbert Stary et al.)
Book
The pathology textbook Sniderman gave Attia approximately 8 years before the episode. It showed autopsy sections of coronary arteries from 26- and 27-year-old accident victims with already-established atherosclerosis, proving that disease begins decades before clinical events.
Attia describes buying as many copies as he could find on Amazon after receiving one from Sniderman. The book documents how by the third decade of life, oxidized apoB-bearing particles are already being engulfed by macrophages and thickening the intima, even in the absence of any clinical disease or calcification. It made viscerally real the concept that waiting for 10-year risk to trigger treatment means waiting until the artery architecture is already destroyed — which you cannot restore.
i probably read it eight years ago and it actually made sense with what he said because in addition to going through in great detail the pathological staging of atherosclerosis it had many autopsy sections of coronary arteries of people who had died for other reasons and notably they were quite young
Measure Lp(a) once in every patient at first lipid evaluation
Practice
The guest's personal practice: Lp(a) measured once at baseline as part of every patient's initial lipid workup, used to contextualize the apoB-based risk estimate lifelong.
Lp(a) is 80–90% genetically determined and does not change meaningfully with lifestyle, diet, or most medications. There is no clinical reason to repeat it. The value is permanent context: a patient with Lp(a) at the 95th percentile and apoB at the 70th percentile has a very different risk profile than a patient with low Lp(a) and the same apoB, and that difference should be reflected in how aggressively apoB is lowered as a compensatory response.
i measure an lp little a in everybody and i measure apob i measure the lp little a once
Present patients with 30-year causal benefit calculation, not 10-year risk score
Practice
Presenting patients with their 20–30 year event risk and the calculated benefit lost by delaying treatment from age 35 to 45 or 55, rather than a 10-year risk score.
Based on the 2019 Circulation paper by Sniderman and collaborators. The key insight: for patients with high non-HDL or apoB, the gain from early treatment is large and concentrated. For patients with low levels, there is little benefit to early treatment — so this is not mass statin prescribing, it is precision early targeting of the high-risk 20%.
vs alternatives
The standard 10-year ACC/AHA risk calculator produces a number like 4.1% that creates false reassurance. The causal benefit model reframes this as: in the group you belong to, about 30% will have a heart attack or stroke before age 65. That framing dramatically changes patient engagement with treatment decisions.
i say your chances of having an infarct or a stroke before you're 65 are 30 now that's a number you can deal with that's a number that has meaning
Lines worth pulling out — contrarian, specific, or perfectly phrased
6 items
atherosclerosis it's a disease in the tissue and almost everything that lipid people talk about is in plasma and if we don't understand the natural history of the disease how can we construct a strategy to prevent it
The foundational argument in one sentence — plasma cholesterol measurements divorced from pathology have led the field astray for 50 years.
prevention really starts at 55 to 60 but half almost half of all infarcts and strokes occur before the age of 60
The central paradox of the 10-year risk model stated as a single devastating fact.
particles more than cholesterol — that hasn't moved the american guidelines but on the evidence side apob is better
Summarizes 40 years of debate between particle counting and cholesterol mass measurement in one line.
there is no phenotype without putting an apob in there because they're lipoprotein particles they're disorders of lipoprotein particle metabolism
The absolute statement: characterizing any dyslipidemia without apoB is incomplete — not just suboptimal but definitionally incomplete.
if somebody has a high apob the fact that their coronary calcium is negative doesn't mean they don't have a lot of disease and that the disease isn't developing at a rapid rate
The clearest clinical statement of why CAC zero should not override apoB in the treatment decision.
as dyslipidemia is growing in the metabolic context meaning if you have more insulin resistance and more type 2 diabetes we know that those phenotypes are associated with greater discordance between apob and ldlc suggesting that you have a greater and greater portion of the population that is being undiagnosed
Explains why cardiovascular events have risen despite statin therapy — a structural miss driven by using the wrong metric in the fastest-growing patient population.
Sign in to share feedback
Tell us if this brief hit the mark or missed it — feedback feeds back into the next iteration of the prompt.
Reading is free for everyone. A free account adds the personal layer: save protocols, follow experts, and see how the other experts weigh in on this same topic.
Educational summary of the cited expert source — not medical advice. Open the source recording linked above and consult a qualified physician before acting on any protocol.
