Why lowering LDL has not eliminated heart disease

• Heart disease remains a leading cause of death despite decades of increasingly powerful LDL-lowering medications. Coronary bypass surgery is still needed when blockages become too advanced for less invasive procedures, and heart surgeons remain in demand despite predictions that cholesterol drugs would make their work obsolete.
• LDL lowering reduces some cardiovascular events, but the observed reduction is smaller than expected under a model in which LDL dominates atherosclerosis. Some patients take statins for decades, meet cholesterol targets, and still require bypass surgery.
• Lowering LDL earlier, further, and for longer does not by itself resolve why disease persists. Additional metabolic factors, including insulin resistance, are relevant to the broader disease process.
• LDL remains a possible contributor to atherosclerosis. The question is how its substantial reduction relates to the more limited reduction in cardiovascular disease.
• Advanced heart disease is common with aging, while preventing its progression to the point of surgery remains an important objective.

Cardiovascular endpoints, mortality, and trial design

• Cardiovascular event reduction and longer overall survival are distinct outcomes. Improvement in a cardiovascular endpoint need not demonstrate an all-cause mortality benefit, and deaths from other causes remain relevant to patients.
• Lifelong genetically lower LDL, including PCSK9 loss-of-function variants, and Mendelian randomization findings raise questions about overall longevity as well as cardiovascular outcomes. The findings cited in the conversation do not establish longer lifespan through lower LDL alone.
• Early termination of trials after cardiovascular or composite endpoints can leave mortality questions unresolved. In the FOURIER PCSK9-inhibitor trial, the intervention arm had numerically more all-cause deaths, without a statistically significant difference. [1]
• Randomization strengthens clinical evidence, while selection, duration, endpoints, run-in periods, subgroup differences, and statistical adjustments influence interpretation. Statins can also have anti-inflammatory effects beyond LDL reduction.
• Evidence obtained mainly in metabolically unhealthy populations or patients with established cardiovascular disease does not directly establish identical effects in metabolically healthy people with elevated LDL.

Metabolic health and individualized risk

• KETO-CTA examines people with very high LDL after carbohydrate restriction and otherwise favorable metabolic markers, using coronary imaging to investigate the LDL-risk relationship in this distinct physiology. [2]
• Insulin resistance receives less attention than cholesterol in routine cardiovascular care. Smoking is an established cardiovascular risk, and glucose and insulin dysfunction are additional factors warranting direct measurement.
• The cited subgroup analyses of statin and PCSK9 trials associate lower insulin resistance with smaller observed medication benefits. Direct study of metabolically healthy groups can provide information not available from adjustments across mixed populations.
• Decisions for metabolically healthy low-carbohydrate patients with high LDL can incorporate inflammation, glucose control, insulin resistance, lipoprotein characteristics, and actual arterial disease alongside LDL.
• Coronary imaging allows assessment of existing disease and subsequent changes. Monitoring arterial plaque and metabolic health can inform decisions about cholesterol medication in selected patients.

Making prevention practical

• Stay Off My Operating Table focuses on avoiding advanced heart disease and surgery; Stay Off My Kitchen Table develops the associated dietary and practical approach. [3][4]
• Medical interventions and prescriptions can leave patients dependent on healthcare while dietary advice remains ineffective. Prevention involves understanding and improving metabolic health through everyday choices.
• Calorie restriction, exercise instructions, and low-fat diets can be difficult to sustain when they produce persistent hunger. Whole-food, low-carbohydrate eating can improve satiety, energy, and adherence.
• Lifestyle coaching and follow-up can support behavioral change, with successful preventive care eventually reducing the need for ongoing medical supervision.
• Medical care often addresses advanced damage after disease develops, while long-term dependence on that care remains common.

Diabetes and the hospital food contradiction

• Substantial improvement or remission of type 2 diabetes through carbohydrate restriction has been demonstrated in Virta Health's clinical work, while hospital meals commonly contain substantial carbohydrate.
• Postoperative glucose control matters because elevated glucose is associated with worse cardiac-surgery outcomes. Patients may receive intravenous insulin while also consuming carbohydrate-heavy hospital meals.
• Lower carbohydrate intake can reduce glucose exposure and insulin requirements. A trial comparing carbohydrate restriction with standard postoperative glucose management would examine the consequences of these different approaches.
• Richard Bernstein's low-carbohydrate approach to type 1 diabetes uses smaller carbohydrate loads and insulin doses. Ovadia's brother, who has longstanding type 1 diabetes, has also used less insulin after carbohydrate restriction.
• Joseph Kraft's insulin-testing work is associated with a high prevalence of insulin resistance among heart-attack patients that conventional diabetes diagnosis or LDL measurement can miss. Historical dietary priorities focused more on saturated fat and cholesterol than John Yudkin's concerns about sugar.

Why metabolic risk cannot be reduced to LDL

• LDL of 40 mg/dL with HbA1c of 7% does not automatically indicate better metabolic or cardiovascular health than LDL of 130 mg/dL with HbA1c of 5%. Very low LDL does not eliminate risk associated with metabolic syndrome.
• Patients can substantially lower HbA1c with carbohydrate restriction while experiencing higher LDL. Decisions focused solely on the LDL rise may overlook improvements in other metabolic markers.
• David Unwin's low-carbohydrate primary-care data showed little mean LDL change alongside metabolic improvements, while individual LDL responses differed.
• A meta-analysis of 41 randomized trials, involving 1,379 participants over an average 19.4 weeks, linked lower BMI to larger LDL increases, intermediate BMI to little average change, and the highest BMI to LDL decreases. [5]
• The 2025–2030 US dietary guidelines give greater prominence to real foods, meat, and full-fat dairy while retaining a saturated-fat limit. The combined guidance creates tension with older dietary messages centered on cholesterol. [6]

Meat, plants, and dietary individualization

• The EAT-Lancet focus on predominantly plant-based eating does not establish that plant-only whole-food diets are superior to whole-food diets containing meat. Meat provides important nutrients and can contribute to optimal nutrition. [7]
• Vegan diets can support health with careful planning and supplementation, including B12, while requiring continued attention to nutritional adequacy.
• Carnivore eating can function as an elimination diet. Removing multiple foods may reduce symptoms, and gradual reintroduction can help identify individual intolerances.
• Reintroduction of fruit, vegetables, or other carbohydrates does not independently establish that carnivore eating was unsuccessful. Dietary tolerance, activity, glucose regulation, and preferences vary between people.
• Greater insulin resistance can require tighter carbohydrate restriction than the intake tolerated by a lean, highly active individual. Dietary suitability depends on individual outcomes.

Continuous glucose monitoring

• A two-week continuous glucose monitor can identify individual meal responses and reveal glucose patterns not captured by general dietary advice.
• Normal glucose regulation permits temporary post-meal increases followed by a return toward baseline, such as under 100 mg/dL within roughly two to three hours. Persistent elevation is a different pattern.
• Responses to potatoes, strawberries, and other carbohydrate-containing foods differ between individuals. Measurements can help define a sustainable carbohydrate intake.
• Better metabolic flexibility and glucose regulation may have implications for cardiovascular, brain, and cancer-related health beyond changes in LDL.

What cardiac surgery actually involves

• Ovadia wanted to become a surgeon in childhood and completed general and cardiothoracic training despite predictions that cholesterol medications would reduce demand for cardiac surgery.
• Cardiopulmonary bypass permits a stopped heart while blood circulates through equipment that supplies oxygen and removes carbon dioxide. John Gibbon developed the heart-lung machine following his experience with a fatal pulmonary embolism.
• Cardiopulmonary bypass may use continuous, nonpulsatile circulation despite normal pulsatile blood flow. The comparisons mentioned do not establish a clear advantage for pulsatile circulation during surgery.
• Ventricular assist devices can maintain continuous blood flow in severe heart failure, with kidney, stroke, and other complications remaining possible.
• Major cardiac operations involve substantial physiological disruption, intensive care, ventilation, chest tubes, and recovery. Prevention can avoid burdens that even successful surgical care cannot eliminate.

Ovadia's metabolic recovery and clinical approach

• Ovadia previously experienced obesity and prediabetes. Calorie counting and increased exercise initially reduced weight, followed by regain.
• A gluten-free dietary experiment with his wife in 2013 also reduced bread and pasta intake, improved energy, and initiated a sustained dietary change.
• Gary Taubes's lectures and the books The Case Against Sugar and Good Calories, Bad Calories contributed to Ovadia's revised view of obesity and metabolic disease. He subsequently achieved and maintained substantial weight loss. [8][9]
• Carbohydrate restriction and intermittent fasting now coincide with more consistent energy during six- to eight-hour operations than Ovadia experienced when younger and metabolically unhealthy.
• A roughly 30-day carbohydrate-restriction trial, with measurements before and after, can reveal changes in glucose, triglycerides, appetite, and energy. Subjective improvement can support long-term adherence.
• Ovadia's preventive practice, education, books, and podcast developed from a desire to reduce progression to advanced disease and surgery.

Why prevailing medical ideas are difficult to change

• Ovadia's medical training centered on the cholesterol hypothesis, with little exposure to John Yudkin's work on sugar. Training can establish a dominant model without extensive examination of competing explanations.
• Career incentives, funding, and publication practices can favor established research questions. Work on lean mass hyper-responders may be viewed as professionally risky by early-career researchers.
• Clinicians and researchers may have invested substantial professional effort in prevailing assumptions. Revising those assumptions requires reevaluating earlier advice and decisions.
• Scientific and media gatekeeping can give conventional interpretations greater visibility and restrict the circulation of dissenting work without requiring coordinated action.
• Cardiac surgical techniques have changed through repeated technical advances. Cardiovascular prevention can also evolve when observations conflict with established expectations.

Insulin resistance, GLP-1 drugs, and lipid quality

• GLP-1 medications improve cardiometabolic outcomes without LDL reduction being their main mechanism. Their cardiovascular effects are relevant to metabolic dysfunction beyond cholesterol.
• GLP-1 drugs can help some patients initiate difficult lifestyle changes, while durable improvement also depends on diet and behavior.
• Small-dense LDL proportion, oxidation, glycation, and inflammation may have associations with risk that differ from total LDL and ApoB concentrations. A high absolute particle count need not correspond to identical risk in every metabolic context.
• Systemic metabolic dysfunction can alter both the lipid profile and cardiovascular risk, producing an association between the two without resolving every causal contribution.
• The 2017 European Atherosclerosis Society consensus supports a causal role for LDL in atherosclerosis. Its application to metabolically healthy carbohydrate-restricted groups with unusually high LDL is the question examined here. [10]

KETO-CTA: the cohort and its central question

• KETO-CTA includes about 100 metabolically healthy ketogenic participants, many lean, with markedly elevated LDL, low triglycerides, high HDL, low inflammation, and favorable glucose markers.
• Mean age was approximately 55 years, and ketogenic dieting had continued for roughly four to five years. Six participants had LDL at or above 400 mg/dL, with the highest near 595 mg/dL.
• Some participants had high absolute numbers of small-dense LDL particles even when these represented a low proportion of total LDL particles. The cohort allows examination of predictions based on absolute ApoB exposure.
• Longitudinal KETO-CTA used coronary CT angiography about one year apart. In the analyses described, LDL and ApoB were not associated with plaque progression in the manner predicted by a simple lipid-centered model. [11]
• The earlier KETO publication found low baseline plaque levels in the high-LDL ketogenic cohort and included a comparison with Miami Heart participants. [12]
• Small baseline plaque volumes increase the difficulty of distinguishing biological change from measurement variability, making imaging precision and blinding important.

The conflicting coronary image analyses

• The Cleerly quantitative dataset contained unusual findings: progression in all participants, regression in none, and an average increase in plaque volume of approximately 50% in one year.
• Participants with calcium scores of zero had greater measured plaque progression than participants with positive calcium scores in that dataset.
• Feldman's inspection of anonymized raw measurements identified further irregularities, and the original Cleerly analysis was found not to have been fully blinded.
• Blinded HeartFlow measurements and prespecified Medis QAngio measurements aligned more closely with semi-quantitative findings than with the Cleerly measurements.
• Multiple quality-control efforts preceded the decision to exclude the delivered Cleerly dataset. The concern centered on that dataset rather than the general utility of quantitative coronary imaging.
• Medical decisions depend on the reliability of image interpretation. Disagreement between methods creates a need to examine both the underlying images and measurement processes.

Reading plaque measurements correctly

• Coronary CT angiography reconstructs coronary anatomy, semi-quantitative methods grade affected arterial segments, and quantitative methods estimate plaque volume.
• Heart motion, scan timing, changing vessel geometry, and low plaque volumes can produce measurement differences without corresponding biological progression.
• The smallest detectable change at a specified confidence level differs from the smallest numerical change produced by software. Feldman's approximate HeartFlow figure was 60 mm³, leaving smaller changes difficult to distinguish from noise.
• Percent atheroma volume is plaque volume divided by vessel volume. A decrease in vessel volume can increase this ratio despite an absolute reduction in plaque.
• A patient's maximum left anterior descending stenosis changed from 81% to 82%, while the length of severe narrowing contracted from several millimeters to a pinpoint. The maximum percentage alone missed the anatomical improvement.
• Severe stenosis in an asymptomatic patient does not automatically require stenting. Imaging interpretation must account for symptoms, actual anatomy, and disease progression.

Long-term uncertainty, lean mass hyper-responders, and mechanisms

• Under a simple cumulative LDL-exposure model, the cohort's lifelong exposure would predict more disease than was observed at baseline. Children with homozygous familial hypercholesterolemia provide a contrasting high-LDL example.
• Repeat imaging after five or six years could establish whether plaque progression accelerates and whether LDL or ApoB becomes associated with its rate. The existing observation period cannot resolve lifelong risk.
• Carbohydrate restriction and the lean mass hyper-responder phenotype do not eliminate the possibility of heart attacks. Some cardiovascular events can occur as a cohort of 55-year-olds ages, while isolated cases cannot quantify population risk.
• Prospective enrollment using predefined metabolic criteria reduces the selection problems of retrospectively assembled cases. Narrow ranges of normal glucose, blood pressure, and inflammation also limit correlations detectable within the cohort.
• A five-patient Journal of Clinical Lipidology case series included features such as markedly elevated triglycerides and xanthomas. Those characteristics differ from the classic metabolically healthy lean mass hyper-responder pattern. [13]
• The classic lean mass hyper-responder pattern includes LDL at least 200 mg/dL, HDL at least 80 mg/dL, and triglycerides no greater than 70 mg/dL. LDL responses form a continuum and may vary with body composition and weight-loss history. [14]
• Atherosclerosis includes active interactions among lipoproteins, arterial endothelial cells, and transport into the vessel wall. The 2020 European Atherosclerosis Society mechanistic paper addresses processes beyond passive cholesterol accumulation. A fuller model of cardiovascular disease incorporates more than increasingly intensive LDL reduction. [15]

References

1. [00:22] Evolocumab and Clinical Outcomes in Patients with Cardiovascular Disease — https://doi.org/10.1056/NEJMoa1615664
2. [00:27] The impact of carbohydrate restriction-induced elevations in low-density lipoprotein cholesterol on progression of coronary atherosclerosis: the ketogenic diet trial study design — https://doi.org/10.1097/MCA.0000000000001395
3. [00:39] Stay Off My Operating Table — https://ifixhearts.com/books/somot/
4. [00:40] Stay Off My Kitchen Table — https://ifixhearts.com/books/somkt/
5. [01:12] Increased low-density lipoprotein cholesterol on a low-carbohydrate diet in adults with normal but not high body weight: A meta-analysis — https://doi.org/10.1016/j.ajcnut.2024.01.009
6. [01:17] Dietary Guidelines for Americans, 2025–2030 — https://realfood.gov/guidelines
7. [01:21] Food in the Anthropocene: the EAT-Lancet Commission on healthy diets from sustainable food systems — https://doi.org/10.1016/S0140-6736(18)31788-4
8. [02:20] The Case Against Sugar — https://www.penguinrandomhouse.com/books/213737/the-case-against-sugar-by-gary-taubes/
9. [02:21] Good Calories, Bad Calories — https://www.penguinrandomhouse.com/books/176680/good-calories-bad-calories-by-gary-taubes/9781400033461/
10. [03:00] Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies — https://doi.org/10.1093/eurheartj/ehx144
11. [03:02] RETRACTED: Longitudinal Data From the KETO-CTA Study: Plaque Predicts Plaque, ApoB Does Not — https://doi.org/10.1016/j.jacadv.2025.101686
12. [03:32] Carbohydrate Restriction-Induced Elevations in LDL-Cholesterol and Atherosclerosis: The KETO Trial — https://doi.org/10.1016/j.jacadv.2024.101109
13. [03:43] Ketogenic diets, not for everyone — https://doi.org/10.1016/j.jacl.2020.10.005
14. [03:45] Elevated LDL Cholesterol with a Carbohydrate-Restricted Diet: Evidence for a Lean Mass Hyper-Responder Phenotype — https://doi.org/10.1093/cdn/nzab144
15. [03:54] Low-density lipoproteins cause atherosclerotic cardiovascular disease: pathophysiological, genetic, and therapeutic insights — https://doi.org/10.1093/eurheartj/ehz962