#19 - Dave Feldman: stress testing the lipid energy model

Episode Summary

Executive Summary: Peter Attia and Dave Feldman debate whether very high LDL-C/LDL-P in lean, low-carb “hyper-responders” is benign or atherogenic. Attia argues the LDL hypothesis remains strongest, citing mass balance, genetics, and clinical evidence; Feldman argues an energy-traffic model may explain the phenotype and that some high-LDL low-TG/high-HDL people may not be at elevated risk. They agree more data are needed, especially on kinetics, remnants, and long-term outcomes.

Main Topics: The lean mass hyper-responder phenotype (Priority: 5/5): Feldman describes a subset of low-carb individuals—often lean, fit, and metabolically healthy—who develop very high LDL-C and LDL-P alongside high HDL-C and low triglycerides. LDL causality vs. alternative lipid models (Priority: 5/5): Attia argues LDL particles are necessary for atherosclerosis and that the burden of evidence supports causality; Feldman questions whether the phenotype can be explained by energy trafficking rather than excess risk. Mass balance and cholesterol synthesis (Priority: 5/5): A central dispute is how the extra LDL cholesterol appears. Attia insists any model must respect mass balance and that Feldman’s model does not explain the increased cholesterol pool; Feldman points to synthesis/absorption changes and diet responsiveness. Lipoprotein kinetics, remnants, and ApoC3 (Priority: 4/5): They discuss VLDL-to-LDL conversion, remnant cholesterol, ApoC3, LDL residence time, and why standard clinical tests may not distinguish physiologic from pathologic remnants. Genetics and natural experiments (Priority: 4/5): Attia emphasizes Mendelian randomization and genetic disorders like FH and PCSK9 variants as strong evidence for LDL’s role; Feldman wants non-genetic, phenotype-based data in low-carb populations. Clinical decision-making under uncertainty (Priority: 4/5): Both acknowledge that patients must make treatment decisions without perfect certainty. Attia stresses probability and risk reduction; Feldman stresses the need to identify who among hyper-responders is truly at risk. Need for better biomarkers and studies (Priority: 3/5): They call for better assays and data on ApoC3, LDL triglyceride content, oxidized LDL, and large retrospective datasets stratified by HDL, triglycerides, and LDL-P.

Key Arguments: Attia argues LDL is necessary but not sufficient for atherosclerosis, and lowering LDL lowers cardiovascular risk even if it does not eliminate all risk. Attia says Feldman’s model fails mass balance: the phenotype has more LDL-C and LDL-P, so the question is where the extra cholesterol comes from. Feldman argues low-carb, lean, highly active people may be trafficking more fat-based energy through VLDL/LDL, producing the observed lipid pattern. Attia counters that insulin-sensitive people generally have lower VLDL triglyceride export and lower ApoC3, which should shorten residence time rather than explain higher LDL. Attia cites genetic natural experiments (FH, PCSK9, ApoC3) and pharmacologic trials as evidence that lifelong lower LDL is protective. Feldman argues that many genetic studies confound lipid metabolism with tissue uptake/clearance effects and do not isolate the exact phenotype he wants to study. Both agree that standard lipid panels are insufficient; ApoB/LDL-P, sterols, and other markers provide more insight than total cholesterol alone. Attia emphasizes that clinical decisions must be made probabilistically, not with certainty, and that high LDL in a hyper-responder should not be dismissed automatically. Feldman argues that some hyper-responders may have benign or even beneficial lipid patterns, but he wants data to identify who is truly at risk. Attia warns that the low-carb community may be overgeneralizing from nutrition skepticism to dismissing strong lipidology evidence.

Data Points: Pre-low-carb lipid panel: Total cholesterol 186 mg/dL; LDL-C 131 mg/dL; HDL-C 40 mg/dL; triglycerides 80 mg/dL - Feldman’s baseline lipid values before starting a low-carb diet Feldman A1c: 6.1% - Triggered his decision to pursue low-carb eating to avoid type 2 diabetes Feldman post-low-carb total cholesterol: 329 mg/dL - His cholesterol rose substantially after starting low-carb/keto Feldman post-low-carb LDL-C: ~200-250 mg/dL - Reported as a major rise after low-carb diet initiation Lean mass hyper-responder pattern: LDL-C ≥200 mg/dL, HDL-C ≥80 mg/dL, triglycerides ≤70 mg/dL - Feldman’s proposed phenotype definition Example patient total cholesterol: 504 mg/dL - Attia’s case example of a low-carb patient with extreme lipids Example patient LDL-C: 362 mg/dL - Direct LDL-C measurement in Attia’s case example Example patient HDL-C: 94 mg/dL - Attia’s case example Example patient triglycerides: 125 mg/dL - Attia’s case example; noted as higher than typical lean mass hyper-responder values Example patient ApoB: 283 mg/dL - Attia’s case example Example patient LDL-P: >3,500 nmol/L - Attia’s case example; assay ceiling exceeded Example patient small LDL-P: 1,483 nmol/L - Attia’s case example Example patient small dense LDL-C: 47 mg/dL - Attia’s case example Garvey study LDL-P in insulin-sensitive people: ~1,200 nmol/L - Attia cites this as a comparator Garvey study VLDL-P in insulin-sensitive people: ~80 nmol/L - Attia cites this as a comparator Garvey study LDL-P in insulin-resistant non-diabetic people: ~1,435 nmol/L - Attia cites this as a comparator Garvey study VLDL-P in insulin-resistant non-diabetic people: ~84 nmol/L - Attia cites this as a comparator Garvey study LDL-P in type 2 diabetes: ~1,600 nmol/L - Attia cites this as a comparator Garvey study VLDL-P in type 2 diabetes: ~100 nmol/L - Attia cites this as a comparator Framingham offspring / Sachs finding: 38% de novo secreted; 62% from IDL/VLDL pathways - Attia cites LDL origin kinetics in low-triglyceride patients LDL half-life: ~1 day - Attia notes kinetic studies distinguish half-life from residence time Cholesterol molecules per LDL particle: ~1,500 - Attia contrasts LDL’s capacity with HDL Cholesterol molecules per HDL particle: ~50 - Attia uses this to illustrate LDL’s cargo capacity Dietary cholesterol contribution: ~15% or less - Attia says most circulating cholesterol is endogenous/recirculated ApoC3 effect: Increases residence time of ApoB particles - Discussed as a key pathologic factor, especially in insulin resistance PCSK9 gain-of-function prevalence in FH: ~3% to 5% - Attia cites this as a natural experiment ApoC3 hypofunction and longevity: Associated with net longevity benefit - Attia references genetic longevity studies Feldman self-experiment blood draws: 100th blood draw - He emphasizes extensive self-testing Feldman exercise kinetics experiment: 3 blood tests/day for 3 days - Used to track lipid changes around workouts Feldman fasting experiment: 1 week fast - He reports LDL-C dropping from 64 to 37 mg/dL after fasting Feldman LDL-P change with keto: 920 to 1,380 nmol/L - He reports LDL-P rising after a week of keto in one experiment Saturated fat reduction target: ~25 g/day - Attia describes a patient whose LDL improved when saturated fat was reduced Patient response to MUFA substitution: LDL-P fell to ~1,300 - Attia reports improvement after replacing saturated fat with monounsaturated fat CIMT follow-up: 4 tests in a row showing regression - Feldman cites serial carotid intima-media thickness improvement while LDL remained high

Pivotal Quotes: "The lower the LDL, the lower the risk of cardiovascular disease, all other things equal." — Peter Attia: Attia’s core position on LDL causality and risk reduction "I think the Occam's razor answer is they're making a boatload more cholesterol." — Peter Attia: Attia’s rebuttal to Feldman’s energy-traffic model "I have an energy model that a lot of people are utilizing probably overly simplistically." — Dave Feldman: Feldman describing his hypothesis for the hyper-responder phenotype

Implications: Listeners with very high LDL on low-carb diets should not assume safety from high HDL/low TG alone. The episode underscores the need for ApoB/LDL-P, better kinetic studies, and individualized risk assessment before changing diet or starting lipid-lowering therapy.