Hidden Sources Of Heart Attacks & How To Prevent Them | Dr. Mark Hyman

Hyman notes that at Function Health, 95% of members show abnormal particle numbers and 89% show abnormal small LDL on fractionation — populations that had all been 'cleared' by standard panels. The full panel also includes liver function, kidney function, uric acid, vitamin D, sex hormones, and a cholesterol absorption/production test (Boston Heart Lab 'cholesterol balance') to distinguish hyper-absorbers from high producers. ApoB alone is a reasonable minimal upgrade: it is cheap, universally available, and tells you the number the standard LDL misses.

Hyman's personal example: one of his patients — an overweight, insulin-resistant woman — dramatically improved all lipid biomarkers on a ketogenic diet (LDL dropped 100 points, triglycerides dropped 200, HDL rose 30). Another patient — lean, fit, cycling 50 miles a day — had the opposite response on the same diet. Genetics determines optimal macronutrient composition. Foods with specific blood-sugar-regulating properties he names: bitter melon, fenugreek, ginger, cinnamon, turmeric.

Aerobic exercise specifically improves HDL subfractions — increasing the large protective HDL2 particles — while HIIT lowers triglycerides and non-HDL cholesterol. No drug safely raises HDL quality. The biggest mortality gains happen in moving from sedentary (under 2,000 steps) to mildly active (4,000–8,000 steps).

The Cell Metabolism RCT cited by Hyman used a controlled feeding design (caloric intake matched between groups), isolating the timing effect from caloric restriction. Results: improved insulin sensitivity, improved beta-cell function, lower blood pressure, and reduced oxidative stress. Frontiers and Current Atherosclerosis Reports also show improvements in triglycerides and LDL with TRE.

Hyman positions omega-3 as part of a four-supplement baseline for cardiac health alongside fiber, vitamin D, and plant sterols. A systematic review using Mendelian randomization found higher dietary omega-6 intake (soybean, corn, canola oils) was associated with higher risk of cardiovascular events — partly because omega-6 oxidizes LDL more readily, generating pro-inflammatory leukotrienes and prostaglandins that drive plaque inflammation.

Hyman describes a 63-year-old patient with horrible cholesterol numbers — high small LDL, high Lp(a), high apoB — who had perfectly clean coronary arteries on CCTA. This spared him statin therapy he did not need and may have caused harm. The AI system classifies plaque type: low-density non-calcified (soft, high rupture risk), non-calcified (possibly scar tissue), and calcified (stable). It generates a 3D vessel map with plaque volume and stenosis quantification by vessel.

Hyman mentions the cholesterol absorption/production test as the way to differentiate treatment: hyper-absorbers get ezetimibe or plant sterols plus extra fiber; high producers get production-targeting agents (statins or berberine). He cites himself as a hyper-absorber who reduced cholesterol dramatically with a cholesterol-blocking strategy rather than statin therapy. Akkermansia muciniphila is specifically mentioned for blood sugar support and gut health, connecting the microbiome to cardiovascular risk via endotoxin (lipopolysaccharide) permeability.

Hyman frames the discrepancy with a lamp-post analogy: 'The light's better here' — doctors check LDL because it's easy to measure and easy to drug, not because it's the best predictor. At Function Health, 95% of members show abnormal lipid particle numbers and 89% show abnormal small LDL on proper fractionation panels — numbers that look perfectly fine on a standard cholesterol test. The European Heart Survey of 4,300 heart attack patients across 110 centers found 31% had full-blown type 2 diabetes and 36% had pre-diabetes on glucose tolerance testing — conditions driving the atherogenic dyslipidemia, not the LDL number.

Hyman describes seeing a patient with a 'perfect' standard cholesterol panel — total cholesterol 148 — who on fractionation had extremely high lipid particle number, high small LDL, and a carbohydrate-heavy diet with excess visceral fat. ApoB would have flagged this immediately. The mechanism: small LDL particles bypass LDL receptor clearance, stay in circulation longer, become oxidized, get taken up by macrophages, and form the foam cells at the base of arterial plaque. Lowering LDL weight without lowering particle count (apoB) leaves the atherogenic substrate intact.

Hyman cites the Minnesota Coronary Experiment — a controlled trial in 9,000 institutionalized subjects comparing corn oil vs. butter — where LDL dropped in the corn oil group but heart attacks increased linearly with every point of LDL reduction. The study was buried for 30 years. The mechanism for insulin resistance driving dyslipidemia: visceral fat acts as an inflammatory factory, releasing cytokines that drive oxidative stress, endothelial dysfunction, and conversion of LDL to small dense atherogenic particles. This is called atherogenic dyslipidemia and has its own ICD code.

The AI phenotyping system stages plaque 0–3 per vessel (right coronary artery, left anterior descending, circumflex) by plaque volume, stenosis degree, and plaque density. Low-density non-calcified soft plaque is classified as high-rupture-risk. Hyman recommends baseline CCTA at age 40, repeat at 50, and more frequently if actively treating risk factors. He describes reversing plaque on serial scans using biomarker-guided lifestyle and drug interventions. No financial relationship between Hyman and Clearly.

HMG-CoA reductase — the enzyme statins inhibit — also synthesizes CoQ10, an essential mitochondrial co-factor. Statin users show mitochondrial damage on muscle biopsy even when asymptomatic and without elevated CPK. A 12-week RCT showed statins blunted every fitness biomarker improvement in a supervised exercise program. Statins also inhibit vitamin K2 synthesis (involved in arterial calcification protection via matrix GLA protein) and selenoprotein synthesis (including glutathione peroxidase, a major antioxidant enzyme). Hyman notes that some patients have a genetic statin-intolerance variant detectable in advance.

Lp(a) is structurally similar to LDL but carries an additional apolipoprotein(a) that mimics plasminogen, interfering with fibrinolysis. This creates prothrombotic risk independent of LDL or apoB levels. Hyman notes a racial disparity: sub-Saharan African and African-American populations have higher average Lp(a) levels. If Lp(a) is elevated in an otherwise excellent metabolic context (clean imaging, low apoB, no insulin resistance), it appears less clinically dangerous. Optimal Lp(a) is under 20 mg/dL; levels between 30–50 carry a 60% higher rate of coronary calcification.