The Modern Lipid Playbook Part 2: THE TESTING — Why Your Lipid Panel Isn't Enough
In a Nutshell
Standard lipid panels miss hidden heart disease because LDL-C measures cholesterol cargo, not the number of ApoB-carrying particles that actually cause plaque. ApoB and Lp(a) are now guideline-recommended because they directly count these dangerous particles and identify genetic risk that LDL-C overlooks. Even with perfect bloodwork, only coronary CT angiography with AI plaque analysis can detect existing non-calcified plaque that stress tests routinely miss until it's too late.
These notes were generated by AI and may contain inaccuracies.
A 62-year-old retired engineer named David appeared completely healthy with supposedly perfect cholesterol labs. He exercised, ate well, and saw his doctor every year. His LDL was in the 40s, below even today's aggressive targets, and he received a clean bill of health from his primary care doctor. When additional tests were run that had never been ordered before, his widowmaker artery was already 75 to 99% blocked.
3 weeks ago, 11 medical specialty societies rewrote the guidelines on cholesterol testing for the first time in 8 years. ApoB and Lp(a), the two tests that would have flagged David's heart disease decades earlier, are now officially standard of care, no longer optional, not just advanced. The standard lipid panel you've been getting from your primary care doctor is now, by consensus, not enough.
David's father had bypass surgery, congestive heart failure, and died of his heart disease. David himself had been told for 25 years that he was fine, but came in because something didn't add up.
The speaker's own ApoB was 128, which is especially high. The CAC score was zero, and the lack of plaque was confirmed recently with a CT cardiac angiogram. Treatment is now being received preventatively because this is not something to sit on. The speaker's grandfather died of stroke, grandmother of a heart attack, aunt of aortic dissection, and father is still recovering from a heart transplant. This set of diseases caused by atherosclerosis is completely preventable.
The most accessible place to start is bloodwork. When doctors order cholesterol panels, you typically get four numbers: total cholesterol, LDL-C or LDL cholesterol, HDL-C, and triglycerides. Most of the time, results appear in the green and everything looks fine.
The moment you're told that your labs look good might be one of the most dangerous moments in preventive medicine, because the test that your doctor just used has a fundamental flaw built into it, and most doctors aren't even aware it's there.
Blood carries fat in little vehicles called lipoproteins, which can be thought of as delivery trucks on a highway. LDL trucks carry cholesterol to tissues, while HDL trucks carry it back to the liver. VLDL trucks carry triglycerides, the type of fat the body makes from extra calories.
ApoB is a protein stamped on the outside of every dangerous truck. Counting ApoB is counting the number of trucks. There's also a special lipoprotein called lipoprotein(a) that has an extra sticky protein bolted to it, which is entirely genetic.
LDL-C is the number that everyone has traditionally focused on. In many cases, it's not even measured directly as part of a standard panel - it's calculated. This goes back to 1972, where William Friedewald published a simple equation: take total cholesterol, subtract HDL, subtract triglycerides divided by five, and that supposedly gives LDL-C.
The divided by five is a fixed ratio assuming a triglyceride to VLDL cholesterol ratio. This assumes five triglycerides in every VLDL truck, but that's not true for every person - it's highly variable. Textbooks teach that this equation breaks down when triglycerides are over 400.
The data shows that the equation starts getting unreliable well before 400. Above 150 triglycerides, LDL-C is already underestimated by about 18 mg/dL if using the calculated value. That's enough to flip someone from needing treatment to being told they're fine.
The patients whose results are most likely to be wrong are the patients who need accurate numbers the most. People with metabolic syndrome, diabetes, obesity, and overweight have higher triglycerides and higher risk for heart disease. About a third of American adults have higher triglycerides. The equation is breaking down in the exact population where getting it right matters most.
In 2013, Seth Martin at Johns Hopkins published a better equation called the Martin Hopkins method, derived from 1.35 million lipid profiles. Instead of one fixed ratio, it uses 180 different ratios based on actual triglyceride and non-LDL levels. Quest Diagnostics now uses it for all their panels. It's a meaningful improvement, but still has the same fundamental problem.
LDL-C measures cholesterol content, the cargo inside the particles. It doesn't count the particles themselves. It's actually the particles that get into the artery wall and cause plaque.
David's LDL was in the 40s - 40 mg per deciliter. By every guideline and every standard metric, that is great. His doctor looked at that number and said he looked great for 25 years. Even with a perfect LDL, there's no way to know what's already in the artery wall.
LDL-C is still checked due to inertia. It's what doctors learn in medical school, what the electronic health record defaults to, and what insurance universally covers with no prior authorization. Insurance coverage is needed for expensive medications like PCSK9 inhibitors, so LDL-C is still relied upon.
The bloodstream is a highway with vehicles (LDL particles) and cholesterol inside the LDL, HDL, and VLDL trucks. What causes the accident is not the cargo, but the vehicles themselves.
In scenario one: 10 big trucks each carrying 10 boxes equals 100 boxes and 10 vehicles. In scenario two: 50 motorcycles each with two boxes equals the same 100 boxes, but 50 vehicles instead of 10. That equals five times the amount of traffic and five times the chance that something will crash into the artery wall.
ApoB is the license plate on every single vehicle. Every atherogenic lipoprotein particle carries exactly one ApoB molecule on its surface. Every LDL, every VLDL, every IDL, every Lp(a) gets a tag, but not HDL. When ApoB is measured, every particle in the blood that can theoretically bind to the artery wall and cause damage is counted.
As of the new guidelines, ApoB is finally in the clinical risk algorithm. The AMORIS study on 175,000+ men and women in Sweden following them for over a decade showed that ApoB, and in particular the ApoB to ApoA1 ratio, outperformed every conventional cholesterol measure as a predictor of fatal heart attack. Total cholesterol, LDL-C, and HDL ratios were all outperformed by ApoB alone.
ApoA1 is the counterpart on HDL particles. It's offered in some labs and showed as a strong predictor in research, but is not recommended to be tested separately or routinely because it doesn't add enough value where ApoB alone captures so much risk.
HDL itself is really not as useful as previously thought for risk determination. One in five people, 20%, have a mismatch between their LDL cholesterol and their actual particle count.
The CARDIA study followed young adults for 25 years, and when LDL-C and ApoB disagreed, ApoB predicted who actually got coronary calcium. A 2024 analysis in JAMA Cardiology showed huge variability. Two patients with the same LDL-C of 110 could have ApoB levels of 80 versus 140, which is the difference between low risk and high risk. The standard lipid panel cannot tell which category someone falls into.
Half of heart attacks occur in people who have perfectly normal LDL-C. A study of half a million people showed that multimodal risk prediction combining genetics, ApoB, and hs-CRP would outperform any single biomarker alone.
The European Society of Cardiology recommends ApoB of less than 65 mg per deciliter for high-risk patients. In the longevity world, many push for under 60, some even under 50, based on Mendelian randomization data showing that lifelong low ApoB exposure dramatically reduces atherosclerosis risk. The ApoB test is about 15 bucks and is now guideline recommended. If a doctor won't order it, it can be self-ordered on many consumer platforms online.
David had a high ApoB number due to Lp(a). About 90% of Lp(a) level is genetically determined. You can't eat your way to a high Lp(a) - you're born with it and it just gets worse over time. Lp(a) is essentially an LDL particle with an extra protein, apo(a), stuck to it. That protein makes it extra sticky, extra atherogenic, prothrombotic, and pro-inflammatory all at once.
20 to 25% of the global population has elevated levels. Prevalence is higher in people of African and South Asian ancestry, though seen in Caucasian descent as well. As of the 2026 guidelines, universal one-time screening is now consensus recommendation. Every adult should know their Lp(a).
When David's Lp(a) was ordered, it came back as 217 nanomoles per liter. Optimal is below 75, so he was nearly three times the upper limit. The speaker's own Lp(a) is 13, which was lucky due to genetics.
Lp(a) has been checked multiple times, but it basically doesn't change and only needs to be checked once because there are no therapies that really lower it and no amount of lifestyle will change it. There is some natural variability, but more likely the assay is not standardized. Different labs report in different units (milligram per deciliter versus nanomole per liter) with imperfect conversion. There are also differences picked up day-to-day if checked on different lab panels.
Advanced lipid panels include the NMR panel, LabCorp's lipid profile, Quest's Cardio IQ, or similar. These use nuclear magnetic resonance to count and size LDL particles. They correlate really well with ApoB, with R equals 0.9 to 0.95. However, NMR costs $200 plus for information that is captured by the $15 ApoB test.
The small dense LDL subfractions are intellectually interesting, but no randomized trial has ever used LDL particle size to guide therapy. Every lipidologist consulted doesn't think there's a lot of great value except for research interest. ApoB alone is sufficient for everyone, and money can be saved, especially if not being covered.
Four drug companies spent billions trying to improve or raise HDL numbers. Torcetrapib raised HDL by 72% and the trial was stopped early because patients died more. Three other CETP inhibitors followed the same pattern. The only one that showed benefit, anacetrapib, worked by lowering LDL, not by raising HDL. Merck didn't bother trying to file for approval.
HDL level doesn't tell how well HDL is actually functioning. It's like counting garbage trucks without knowing if they're picking up trash. HDL is better understood as a signal rather than a shield. People with high HDL tend to be healthier because the behaviors that raise HDL (exercise, not smoking, moderate weight) are themselves protective. HDL is like a thermometer, not the fever itself.
High triglyceride levels are a bad metabolic signal indicating insulin resistance, visceral fat, and too many refined carbs. The remnant particles they create are genuinely atherogenic. This has been proven in Mendelian randomization studies.
The PROMINENT trial showed that lowering triglycerides without lowering ApoB does nothing. 10,000 patients on pemafibrate lowered their triglycerides by 26% and yet there was zero cardiovascular benefit. Triglycerides should be checked if genuinely helpful for other reasons, but lowering triglycerides only helps if ApoB is also lowered. It's better to decrease metabolic risk overall.
Most standard labs only check fasting glucose and maybe A1C. By the time fasting glucose is elevated, insulin resistance has likely been present for a decade. The body compensates for years, and the pancreas cranks up insulin production to force glucose into cells. Glucose looks normal on labs, but elevated insulin is quietly driving inflammation, fat storage, VLDL production, and accelerating atherosclerosis.
A meta-analysis of over 65,000 subjects found that each standard deviation increase in HOMA-IR was associated with a 46% increased cardiovascular disease risk. Fasting insulin is one of the best tests to get because it is cheap, widely available, and tells something that fasting glucose and A1C cannot.
hs-CRP (high-sensitivity C-reactive protein) is a blood marker of inflammation elevated by any inflammation, whether infectious, autoimmune, or even vigorous exercise. The JUPITER trial in 2008 on 17,000+ healthy people with normal LDL cholesterol but elevated hs-CRP used rosuvastatin and reduced major cardiovascular events by 44%. The action was on the inflammatory side.
The CANTOS trial in 2017 on over 10,000 patients with elevated CRP after a heart attack used canakinumab, which blocks inflammation without touching cholesterol. CRP dropped by 37% and resulted in 15% fewer cardiovascular events without lowering cholesterol, proving the inflammatory hypothesis.
hs-CRP is profoundly non-specific. Even a cold, dental procedure, or tough workout will cause it to go up. The ACC/AHA recommends averaging two measurements two weeks apart because of this variability. It tells that something is inflamed somewhere, but doesn't necessarily indicate coronary arteries. It's ordered with every lab check because it's quick and easy.
David's hs-CRP was 0.4 mg/L, which is low and clean.
Blood tests measure what's floating in the bloodstream, not what's already embedded in the artery wall. Even if every blood marker is perfect, ApoB is managed, Lp(a) is known, and inflammation is low, it tells nothing about whether there's already plaque in artery walls. A perfect cholesterol panel doesn't mean clean arteries, and that gap is where people die.
To find actual disease, imaging is needed, and the first test from a doctor is not the best one. The first test that everybody gets is the wrong test first. The stress test creates the most dangerous false reassurance in all of medicine.
Stress tests put electrodes on the chest, walking on a treadmill with increasing speed and incline. A stress test only catches blockages from when they're above 70% stenosis. Two-thirds of acute heart attacks come from lesions that are less than 50% narrowed. The dangerous plaques aren't the big, stable ones that block flow - they're smaller, inflamed, lipid-rich ones that rupture without warning.
This has been known for decades. A 1995 study by Falk and an earlier 1988 study by Ambrus demonstrated this. Stress tests are still ordered as screening tools despite this knowledge.
In 1987, a landmark paper in the New England Journal showed that coronary arteries don't just passively get clogged. They remodel, meaning they expand outward to accommodate growing plaque, which is called positive or compensatory remodeling. The blood vessel lumen stays the same size and grows outwards until 40% of the vessel wall is filled with plaque. Stress tests will be completely normal during this process.
Even old school cardiac angiographies with dye will look perfectly normal because they're looking at the hole in the middle, not the plaque in the wall. Exercise ECG sensitivity is 68%, meaning it misses a third of significant disease. In women, sensitivity drops to roughly 60% and false positive rates are two to three times higher due to estrogen effects on ST segments.
Women are both more likely to be falsely reassured and more likely to undergo unnecessary cardiac catheterization from false positives. The WISE study showed that women with ischemia but no obstructive coronary artery disease on angiography has significantly elevated mortality. One in three cardiac deaths in women occurred with angiograms that looked clean.
Women with chest pain, signs of ischemia, and clean angiograms were told nothing is wrong and sent home. Nuclear stress testing adds radiation at 11 to 24 millisieverts, equivalent to about 400 to 800 chest x-rays, but does not solve the fundamental flow limitation problem. The balanced ischemia trap occurs when all three coronary territories are equally diseased, producing normal results on nuclear perfusion scans because there's no relative difference to detect.
Passing a stress test might be the most dangerous phrase in cardiology.
A stress test asks only whether blood flow is restricted at that moment. Most heart attacks come from plaque rupture in the future at unpredictable times, not from current flow restriction. Only imaging that visualizes the artery wall can identify rupture risk.
The ABI measures the blood pressure ratio between ankle and arm. It's cheap, involves no radiation, and detects peripheral disease away from the body's center. However, diabetics may get falsely normal readings due to medial arterial calcification. A normal ABI doesn't provide information about coronary arteries, so it should generally be skipped.
CIMT is an ultrasound measuring carotid artery wall thickness. The 2013 ACC/AHA guidelines removed it from recommended risk assessment due to operator dependency, lack of standardized protocols, and medial hypertrophy from hypertension causing increased CIMT without atherosclerosis. This creates false positives in patients with high blood pressure. Some longevity doctors still use it for being radiation-free, but specialists do not recommend it.
CAC scoring uses a quick low-dose CT scan with about 1 mSv radiation, typically costing around $100 out of pocket. It measures calcified plaque in coronary arteries with no IV or contrast required. The MESA study of 6,800 patients across four ethnic groups showed CAC adds to risk prediction beyond traditional factors. A CAC of zero indicates very low 10-year event risk under 1 to 2%.
Ethnic variation exists: white and South Asian populations have higher CAC prevalence at the same risk level, while Black and Hispanic populations have lower CAC prevalence. Universal CAC thresholds may underestimate risk in Black patients.
A CAC score of zero does not mean absence of atherosclerosis, only absence of calcified atherosclerosis. CAC is completely blind to non-calcified, soft, lipid-rich, inflammatory, vulnerable plaque most likely to rupture and cause heart attacks.
When patients with zero CAC undergo CCTA, 9.3% have atherosclerotic plaque, and of those, 47% have significant stenosis. A 2023 case series described a 31-year-old with severe two-vessel coronary artery disease despite CAC of zero. Young patients preferentially develop non-calcified plaque since calcification is a later-stage process.
CAC scores often increase on statin therapy because statins stabilize plaque by calcifying it. Rising CAC on treatment is generally a good sign. Lean mass hyper-responders on low-carb carnivore diets may have high LDL-C with CAC of zero, but lifetime cumulative exposure to atherogenic particles drives atherosclerosis risk. The question isn't whether arteries look clean today, but whether they'll stay clean tomorrow.
The Cudos CTA trial by Norbert and Feldman is designed to address this question. David's CAC score was 1,522, nearly four times the severe threshold of 400. After 25 years on statins with LDL in the 40s, his arteries were loaded with plaque. His wife had CAC of zero, likely due to family genetics and estrogen protection until menopause.
CCTA provides 3D visualization of coronary arteries and their walls, showing plaque quantity and composition. AI analysis identifies calcified, partially calcified, non-calcified, and low attenuation plaque (lipid-rich, dangerous). The procedure requires fasting, IV placement, possible beta-blocker for heart rate control, and takes 30 to 45 minutes total. Contrast may create a sensation of needing to urinate.
The SCOT-HEART trial of 4,146 patients with stable chest pain showed CCTA group had 41% reduction in coronary heart disease death or non-fatal MI at 5 years. Management changed in 23%, diagnosis changed in 27%. The 10-year follow-up in 2025 showed sustained benefit.
The PROMISE trial of over 10,000 symptomatic patients compared CCTA versus functional testing. CCTA found more disease, led to more appropriate treatment, and resulted in fewer invasive catheterizations showing no obstructed disease.
The MESA study establishes subclinical atherosclerosis present in over 50% of adults over 45 with risk factors. Traditional risk calculators underestimate risk in South Asian and Black populations.
David's CCTA showed 75% stenosis of the left anterior descending artery. HeartFlow and Cleerly AI analysis quantified plaque burden and characterized composition. He experienced unstable angina, underwent emergent catheterization and LAD stenting. Treatment includes Repatha (PCSK9 inhibitor), ezetimibe, colchicine for inflammation, aspirin, and ticagrelor.
Cleerly, founded by James Min from Weill Cornell, quantifies total plaque volume in cubic millimeters by type. Low-attenuation plaque (less than 30 Hounsfield units) corresponds to lipid-rich necrotic core. Their TCT 2025 data showed LAP volume is a stronger predictor of future heart attacks than calcified plaque.
HeartFlow started with FFRCT using computational fluid dynamics to estimate functional significance without invasive catheterization. Their Decide Registry showed plaque analysis changed management in over 50% of patients.
Elucid's PlaqueIQ is the only platform validated against histology. When Elucid identifies lipid-rich plaque, it's proven against ground truth tissue.
Caristo measures inflammation in pericoronary fat. The Caristo CT study of 3,912 patients showed higher pericoronary fat attenuation predicted adverse cardiac events.
Limitations include severe kidney disease (eGFR under 30), true iodine allergy, heart rates over 65 bpm, atrial fibrillation, CAC scores above 1,000 creating artifact, and BMI over 40 degrading image quality. Modern protocols deliver 1 to 5 mSv radiation. Availability varies, with rural areas underserved. Cost ranges from $500 to $2,500. Incidental findings like lung or thyroid nodules may trigger unnecessary workups.
CPT code 75577 for AI-driven coronary plaque quantification from CCTA went live January 1, 2026, with Medicare reimbursement of about $1,021. United Healthcare and Cigna announced coverage. The 2021 guidelines gave CCTA class I recommendation for intermediate-risk patients. SCOT-HEART 2 trial is underway for asymptomatic screening.
Self-pay CCTA at $300 to $600 is increasingly available. Existing CT scans can be submitted to Cleerly or HeartFlow for AI analysis.
Microvascular disease disproportionately affects women. The WISE study showed women with chest pain, ischemia signs, and clean angiograms had worse outcomes. Coronary arterioles weren't dilating properly, measurable by coronary flow reserve on PET but rarely tested.
The 2026 guidelines recommend discussing statins at age 30 because atherosclerosis is cumulative. Twenty extra years of LDL reduction produces dramatically better lifetime benefit. The ASCVD risk calculator uses 10-year risk, potentially missing high lifetime risk in younger patients.
South Asians have higher risk due to higher insulin resistance and elevated Lp(a). Standard calculators underestimate risk. Familial hypercholesterolemia affects 1 in 250 people, with 90% unaware they have it.
Blood work first: apoB, Lp(a), standard lipid panel, fasting insulin, hs-CRP, HbA1c. CAC score if imaging isn't accessible. CCTA with AI plaque analysis when indicated is definitive screening. Stress testing for functional assessment when plaque is already known.
Minimum blood work costs about $100 out of pocket. CAC scoring around $100. CCTA with AI plaque analysis around $1,000 or more. APOE genotyping identifies 2, 3, or 4 alleles, with APOE4 requiring more diligence.
Genetic testing for familial hyperlipidemia may be considered, but the honest truth is that bad lipids will be detected through the other tests discussed first. More specialized tests include carotid imaging, peripheral vascular assessment, coronary flow reserve testing for those worried about microvascular disease, and polygenic risk scores for borderline risk patients. Cardiologists who specialize in preventative work should be consulted for these advanced testing options.
Even in the best medical centers, there remains a significant lag behind current science. Advocating for appropriate testing can be difficult, as demonstrated by an example where a partner with a family history of heart attack was denied a CTCA with AI by a preventative cardiologist who stated, "No. Why would I do that? You seem young, pretty healthy. Your lab numbers are pretty good." This response was considered terrible given the low risk of preventative testing in someone with strong family history.
"I know it's hard to advocate for yourself, but do so. Try to get the testing that you need."
Education about lipid health is valuable to share with others who may believe their cholesterol is fine or who passed a stress test and think they're in the clear. Part three will address fixing the issues identified through proper testing.
Keep Hillary Lin, MD in your library
Save the videos and channels worth coming back to, and find them again in one place.
