I Thought I Reversed My Artery Plaque by 48%. Here’s What Really Happened | Simon Hill | EP#439

Episode Overview

• Host Simon Hill (physiotherapist and MSc in Nutrition Science) unpacks the full story behind his serial Coronary Computed Tomography Angiography (CCTA) scans—taken 16.5 months (504 days) apart at age 38 (December 2024) and age 39 (April 2026) under cardiologist Dr. Matthew Budoff at the Lundquist Institute—where one analysis platform reported a headline-grabbing 48% reduction in his coronary artery plaque.
• Hill explains why he had coronary plaque in the first place despite eating a plant-based diet for a decade, detailing the concepts of lifetime cumulative LDL exposure ("area under the curve") and polygenic cardiovascular risk.
• He breaks down the exact nutrition, exercise, sleep, and supplement protocol he followed during the 16.5-month window between scans, noting that he took zero artery-targeted supplements or lipid-lowering medications during this experiment.
• The episode reveals a massive technical discrepancy: when the exact same raw CT scan data was run through three leading quantification platforms (QAngio CT, Cleerly, and HeartFlow), they produced a 6.5-fold spread in total plaque volume and three completely different conclusions (flat, progression, and regression).
• Finally, Hill explains why he refuses to cherry-pick the 48% regression result and why—guided by the March 2026 ACC/AHA Dyslipidemia Guidelines and his family history—he has now started a prescription injectable PCSK9 inhibitor alongside his lifestyle regimen. (

Key Concepts

• Cumulative Lifetime Exposure ("Area Under the Curve"): Arterial plaque is driven by how high your ApoB/LDL cholesterol is and for how long across your entire life, not just your cholesterol levels today. For the first 30 years of his life, Hill ate a standard Australian omnivorous diet (beef, lamb, veal, chicken, milk, butter) and his LDL sat around 125–130 mg/dL. Switching to a plant-based diet at age 30 dropped his LDL into the 70s mg/dL (with an unmedicated ApoB between 70 and 80 mg/dL; 75 mg/dL in June 2025 with an LDL-C of 85 mg/dL), but changing your diet does not automatically reset plaque accumulated over prior decades. Hill notes he cannot test or prove the counterfactual of whether his plaque would be much worse had he not switched diets at age 30, or if it has already been regressing over the last decade.
• Population Prevalence of Subclinical Atherosclerosis: Having plaque in your 30s and 40s is far more common than most asymptomatic people realize. The prospective REACT Study (published in the NEJM), which evaluated 16,000 adults aged 18 to 70 across Denmark and Spain without known heart disease using 3D vascular ultrasound and CCTA, found that among men in their 30s, more than 1/3 had plaque somewhere (carotid, peripheral, or coronary) and 9% had coronary plaque. Among men in their 40s, 2/3 had plaque somewhere and 31% had coronary plaque; by their 60s, more than 9 in 10 people had atherosclerosis.
• Beyond ApoB and Polygenic Risk: While ApoB is causal in plaque development and is the single best lipid marker, two people with identical ApoB can have vastly different disease burdens due to blood pressure, insulin resistance, inflammation, Lp(a), and genetics. Up to 50% of cardiovascular risk is thought to be genetic, mostly polygenic (hundreds or thousands of small variants nudging risk up or down independent of cholesterol). Despite Hill having optimal biomarkers—Lp(a) <14 nmol/L, blood pressure 120/71 mmHg, resting heart rate 62 bpm, and great insulin sensitivity—his father had a heart attack at age 41 and his grandfather at age 49. However, Hill hedges on whether polygenic genomic profiling (which he is testing with medical geneticist Dr. Robert Green) currently changes clinical decision-making or is just expensive, interesting information.
• LDL/ApoB Thresholds for Plaque Regression: While plaque regression can occur at an ApoB of 70–80 mg/dL, trial data shows it is much more likely and pronounced at lower levels. The GLAGOV trial (JAMA), which added the PCSK9 inhibitor evolocumab or placebo to statin therapy and measured plaque via intravascular ultrasound (IVUS), showed a continuous dose-response relationship down to an LDL of 20 mg/dL with no floor, where 81% of patients in the lowest LDL group experienced regression. Hill flags two important caveats: the dose-response analysis in GLAGOV was post-hoc, and the trial subjects had established, symptomatic coronary disease, meaning he is extrapolating to his own asymptomatic context.
• Higher-Fat vs. Low-Fat Plant-Based Diets for ApoB: Contrary to traditional ultra-low-fat vegan heart-reversal protocols, Hill found that a higher-fat, moderate-carbohydrate, high-fiber plant-based diet got his ApoB the lowest and made him feel best. His protocol includes 1.2 to 1.6 g/kg/day of protein (tofu, tempeh, seitan, plant protein powder, legumes, nuts, seeds), whole-food carbohydrates (legumes, oats, potatoes, rice, sourdough), and predominantly unsaturated fats (olive oil, avocados, nuts, seeds, soy). He also cites meta-analytic data on the Portfolio Dietary Pattern (Chiavaroli et al., 2018) and notes that among animal foods, Mediterranean patterns including fatty fish, yogurt, cheese, and even eggs have better cardiovascular evidence.
• Cardiorespiratory Fitness, Sleep, and Supplement Controls: Hill’s baseline lifestyle protocol includes resistance training 3 to 4 times per week (1 hour/session) and cardio via padel 4 to 5 times per week (or stationary bike during a 6-week recovery from a near-complete CFL ankle ligament rupture). His lab-tested $\text{VO}_2\text{ max}$ was 51 a few years ago (estimated mid-to-high 50s now), with a weekly training distribution of ~16 minutes in Zone 5, ~1 hour 40 minutes in Zone 4, and >4 hours across Zones 2 and 3. He sleeps 8 hours per night 90% of the time (in bed before 10:00 PM, waking around 6:00 AM) using an Eight Sleep temperature-controlled mattress cover (a show sponsor, though Hill notes he is not paid to mention them in the editorial segment). Crucially, during the 16.5-month window, he took zero artery-targeted supplements (no nattokinase, aged garlic extract, Vitamin K2, berberine, or red yeast rice) to avoid confounding the experiment, taking only a daily multivitamin (IM8, a show sponsor where Hill discloses he sits on the Scientific Advisory Board; he hedges that trials on multivitamins and cognition are mostly in older populations, so he "holds this loosely"), creatine, 1.5 g of combined EPA/DHA omega-3s, and a prebiotic powder.
• The 3-Platform CCTA Discrepancy and AI Overestimation: Running the exact same Dec 2024 and Apr 2026 scan files (captured on the same 256-slice CT scanner) through three platforms yielded conflicting results:
• QAngio CT (Medis, human-in-the-loop): Baseline $10.4\text{ mm}^3 \rightarrow$ Follow-up $11.5\text{ mm}^3$ ($+1.1\text{ mm}^3$ / $+10.4\%$ observed; $+0.8\text{ mm}^3\text{/yr}$ annualized; Percent Atheroma Volume [PAV] $0.28\% \rightarrow 0.26\%$). Essentially flat.
• Cleerly (Automated AI-QCT): Baseline $61.3\text{ mm}^3 \rightarrow$ Follow-up $74.4\text{ mm}^3$ ($+13.1\text{ mm}^3$ / $+21.4\%$ observed; $+9.5\text{ mm}^3\text{/yr}$ annualized; PAV $1.8\% \rightarrow 1.8\%$). Indicates progression.
• HeartFlow (Automated AI-QCT): Baseline $116\text{ mm}^3 \rightarrow$ Follow-up $60\text{ mm}^3$ ($-56\text{ mm}^3$ / $-48.3\%$ observed; $-40.6\text{ mm}^3\text{/yr}$ annualized; PAV $4.34\% \rightarrow 1.70\%$). Indicates massive regression.
• Why they disagree: Plaque volume is the space between the inner and outer artery walls across ~60 cm of the coronary tree. When total plaque is very small ($11.5\text{ mm}^3$ is ~1 grain of rice; $74.4\text{ mm}^3$ is ~6 grains of rice), tiny differences in where software draws wall boundaries create massive relative swings. Furthermore, the platforms did not measure his total artery size identically between scans, though a mathematical correction holding artery volume fixed at Scan 1 levels did not change the outcome. Dr. Matthew Budoff told Hill he trusts QAngio most. In the 2026 Miami Heart Study (Ichikawa et al., JACC: Cardiovascular Imaging, co-authored by Budoff) examining 2,301 asymptomatic adults aged 40–65, the authors explicitly warned about "the potential for plaque overestimation by AI-QCT" in people with $<20\text{ mm}^3$ of plaque, noting AI frequently flags soft plaque where human readers see normal arteries (though they note it remains unclear if AI is catching real early disease or overdiagnosing due to technical noise).
• Plaque Stability, Composition, and Reference Ranges: Despite the disagreement on volume, all three platforms agreed Hill has a small absolute plaque burden ($<75\text{ mm}^3$) and none showed rapid progression (defined in the PARADIGM study as PAV climbing $\ge 1.0$ percentage point per year). Furthermore, QAngio showed a favorable compositional shift: Hill's dangerous, softer, lipid-rich (low-attenuation) plaque dropped by ~40%, while stable fibrous plaque and dense calcium increased slightly. Comparing his scans to population cohorts highlights flaws in current reference ranges for a 40-year-old: his QAngio result ($11.5\text{ mm}^3$, or $10.9\text{ mm}^3$ two-scan average) is roughly a quarter of the $42.6\text{ mm}^3$ median for men in their 30s with plaque in the REACT study; his Cleerly result ($74.4\text{ mm}^3$) places him at the 55th percentile in the Miami Heart Study (which skews older, average age 53); and HeartFlow's unpublished internal database places him at the 80th percentile.
• Guideline-Driven Pharmacology and How Lifestyle Simplifies Medication: The March 2026 ACC/AHA Dyslipidemia Guidelines reinstated numerical lipid targets: having a Coronary Artery Calcium (CAC) score $\ge 75\text{th percentile}$ for age and sex warrants an LDL target $<70\text{ mg/dL}$. Because Hill’s CAC score was 4 (all in the Left Anterior Descending artery) on both scans—placing him above the 75th percentile for men in their late 30s—he decided to use medication to drive his ApoB to $\le 40\text{ mg/dL}$ to keep his plaque completely flat over the next 25 years (which would drop him from the 55th to the 25th percentile on the Miami Heart curve by age 65). Because his plant-based diet already brought his LDL from 125–130 mg/dL down to 70–80 mg/dL, he only needs a single drug (monotherapy) rather than combination therapy (e.g., statin + ezetimibe). While Hill emphasizes that statins are safe, cheap, life-saving first-line drugs—citing blinded trials (CTT Collaboration, SAMSON) showing $>90\%$ of reported muscle symptoms are not caused by the statin itself, alongside only a small risk of new-onset diabetes in those already at metabolic risk—he chose a biweekly injectable PCSK9 inhibitor for adherence convenience and to avoid those tiny metabolic margins as a lean, insulin-sensitive individual.

Quotes

• At 4:59 - "Plaque in your arteries is about cumulative exposure. It's not about what your cholesterol is today. It's about the area under the curve. How high, for how long, across your whole life." - Explaining why current optimal lipid biomarkers do not grant immunity from atherosclerosis accumulated earlier in life.
• At 5:33 - "I personally had three decades of high LDL cholesterol exposure before I changed anything. And unfortunately, plaque doesn't just reset when your diet does." - Highlighting why 10 years of a plant-based diet did not erase the arterial impact of the first 30 years of his omnivorous diet.
• At 8:47 - "Each of us carries a unique version of that signature, and it could be making us more or less prone to laying down plaque as we age, independent of our cholesterol. It's one of the big reasons why two people with identical ApoB can end up with very different amounts of disease." - Discussing how polygenic risk scores and non-lipid factors influence individual susceptibility to atherosclerosis.
• At 11:30 - "The more they lowered LDL, the more the plaque regressed. There was no floor. The relationship held all the way down to 20 milligrams per deciliter. In that lowest group, 81% showed regression." - Summarizing the post-hoc findings of the GLAGOV intravascular ultrasound trial on PCSK9 inhibitors and plaque volume.
• At 25:16 - "I could have built a program off the back of that, sold you something that claims to guarantee plaque reversal. But the truth is that three platforms analyzed my scans, and only one of them tells that story." - Explaining his decision to share the conflicting data from all three imaging platforms rather than capitalizing on HeartFlow's 48% regression report.
• At 26:22 - "On my most recent scan, QAngio measured 11 and a half cubic millimeters of plaque in my arteries. Cleerly measured 74. HeartFlow measured 60. So same heart, same images, same day, and a six-fold spread between the smallest number and the largest." - Illustrating the massive discrepancy between human-in-the-loop software (QAngio) and automated AI platforms (Cleerly and HeartFlow) on small plaque volumes.
• At 30:36 - "If you're going to track plaque over time like I did, pick one provider and stick with it, whether that's Cleerly, HeartFlow." - Providing a core rule for anyone using quantitative coronary CT angiography to monitor disease progression.
• At 46:55 - "So the lifestyle work, it didn't become irrelevant when I decided to medicate here. It's actually the reason the medication is simple." - Explaining how lowering his LDL by ~50 mg/dL through diet allowed him to reach aggressive ApoB targets using only a single medication instead of multiple drugs.
• At 55:15 - "A short interval plus a low plaque volume is how measurement noise ends up looking like a result." - Explaining why rescanning after only 16.5 months when total plaque is low produces software artifacts rather than reliable biological takeaways.

Takeaways

• Lock in your imaging variables and wait 2 to 5+ years between CCTA scans: If you track coronary plaque via serial CT angiograms, do not rescan in under 2 years. Follow the December 2025 ACC Scientific Statement on Quantitative Coronary Plaque Analysis by waiting at least 2 to 5 years (Hill is waiting 5 years until age 45 for his next scan) and always use the exact same scanner, acquisition protocol, and software analysis provider so you aren't measuring software boundary noise.
• Evaluate your lifetime cholesterol exposure and family history, not just today's diet: Do not assume a clean diet today makes you immune to plaque if you had elevated LDL in your teens and 20s or have a family history of premature heart disease. Consider baseline imaging (such as a CAC score or CCTA in consultation with your physician) and use the 2026 ACC/AHA guideline threshold (CAC $\ge 75\text{th percentile}$ for age and sex) to determine if you need pharmacological lipid lowering to get LDL $<70\text{ mg/dL}$ (or ApoB $\le 40\text{ mg/dL}$ for aggressive stabilization/regression).
• Optimize your diet for unsaturated fats and soluble fiber to do the heavy lifting: To lower ApoB naturally—or to reduce the number/dose of medications needed to hit aggressive targets—replace saturated fats (butter, red meat) with unsaturated fats (olive oil, avocados, nuts, seeds, soy) and high-fiber whole foods (legumes, oats, tempeh, tofu) while maintaining 1.2 to 1.6 g/kg of protein to support muscle mass.
• Prioritize cardiovascularoutcomes data when selecting lipid-lowering therapies: If you and your doctor decide on a PCSK9 inhibitor (or any lipid-lowering drug), prioritize molecules with completed cardiovascular outcomes trials. For example, Hill switched from Lerochol (lerodalcibep-liga), which only has early lipid-lowering data, to Repatha (evolocumab, injected every 2 weeks) because it has proven cardiovascular event reduction data and was the exact molecule proven to regress plaque in the GLAGOV trial.
• Skip unverified "artery-cleansing" supplement stacks and focus on proven fundamentals: Avoid wasting money on overhyped plaque-reversal supplements like nattokinase, berberine, red yeast rice, or Vitamin K2. Instead, build a foundation of 3–4 weekly resistance sessions, regular Zone 2–5 cardio, 8 hours of quality sleep in a cool dark room, and proven lipid lowering; if adding an adjunct arterial supplement, consider 2,400 mg/day of aged garlic extract, which has some clinical imaging trial support for slowing plaque progression (while recognizing it is not in the same evidence league as prescription lipid-lowering drugs).