Peter Attia MD240 ‒ The confusion around HDL and its link to cardiovascular disease | Dan Rader, M.D.
CHAPTERS
- 0:11 – 5:17
Introduction and Why HDL Is the Harder Half of the Lipid Story
Peter Attia introduces Dan Rader and explains why this is the first dedicated HDL episode: HDL biology is far more complex and less understood than apoB biology. He previews the topics ahead, from HDL's origin and measurement to CETP inhibitors and neurodegeneration.
- •Rader studies the metabolism and function of HDL
- •The apoB side has been easier to manipulate than the apoA side
- •Episode roadmap: biogenesis, measurement, CETP inhibitors, brain
- •Attia objects to calling HDL 'the good cholesterol'
- 5:17 – 10:24
What Lipoproteins Are and the ApoA vs. Lp(a) Naming
Rader explains that lipoproteins are lipid transport vehicles, contrasting apoB particles with HDL, which is defined by apoA-I. The two clarify that uppercase ApoA has nothing to do with the lowercase 'a' in Lp(a).
- •Lipids need lipoproteins to travel in the blood
- •ApoB particles carry triglycerides for energy and drive atherosclerosis
- •HDL is characterized by apoA-I rather than apoB
- •Uppercase ApoA-I differs from lowercase apo(a) in Lp(a)
- 10:24 – 14:12
HDL Biogenesis: ApoA-I, ABCA1, Tangier Disease and LCAT
Rader walks through how HDL forms: apoA-I is secreted nearly free by gut and liver, acquires lipid via ABCA1, and matures when LCAT esterifies cholesterol into the core. Rare human deficiencies of ABCA1 (Tangier disease) and LCAT show the importance of each step.
- •ApoA-I, unlike apoB, can exchange between particles and several copies sit on each HDL
- •ABCA1 loads nascent apoA-I with phospholipid and cholesterol
- •Tangier disease causes near-absent HDL and cholesterol buildup in macrophages
- •LCAT creates cholesteryl esters that form the mature HDL core
- •LCAT deficiency leads to low HDL and kidney disease
- 14:12 – 25:50
Why 'High' vs 'Low' Density, and HDL Across Species
Rader explains the historical, centrifugation-based origin of density names and how much smaller HDL is than LDL. The conversation then turns to HDL being the main lipoprotein in mice and other mammals, and why humans clear apoB less efficiently.
- •Density names come from how lipoproteins spin in a centrifuge
- •HDL is roughly one-fifth to one-tenth the size of LDL
- •Around 90% of mouse blood cholesterol is in HDL
- •Other mammals make apoB too but clear it far more efficiently
- •Humans need apoB for energy but it is now mostly a problem
- 25:50 – 31:45
ApoA Family, HDL as a Platform, and Confusing Subfraction Names
Rader describes other apoAs like apoA-5 and frames HDL as a platform carrying many proteins and lipids. He clarifies that HDL2/HDL3 numbering is unrelated to apoA numbering and says HDL fractionation adds little clinical value.
- •ApoA-4 and apoA-5 exist; apoA-5 promotes lipoprotein lipase activity
- •HDL is a dynamic platform, unlike the monogamous apoB
- •HDL2/HDL3 naming has no relationship to apoA-1/2/4/5
- •Fractionating HDL isn't clinically useful for predicting risk
- •Attia stopped using HDL fractionation metrics around 2016
- 31:45 – 36:15
HDL Remodeling: Lipases, CETP, and What a Lipid Panel Measures
Rader describes how hepatic and endothelial lipase and CETP remodel HDL, with CETP-deficient people having very high HDL cholesterol. Attia pauses to clarify what a standard lipid panel reports and that HDL-C, HDL-P and apoA-I are the actual lab metrics.
- •Hepatic and endothelial lipase modify HDL phospholipids and size
- •CETP moves cholesteryl esters from HDL to apoB lipoproteins
- •Mice lack CETP and so have high HDL cholesterol
- •HDL is a lipoprotein; HDL-C, HDL-P and apoA-I are the measurements
- 36:15 – 49:11
CETP Inhibitor Trials and the Lesson of Hard Outcomes
Starting from the Framingham observation that high HDL-C predicted lower risk, the two trace torcetrapib's failure and the later CETP inhibitors that raised HDL without meaningful benefit. Attia uses triparanol and clomiphene/desmosterol as cautionary examples, and Rader notes obicetrapib is still in development mainly for LDL and apoB lowering.
- •Framingham data fueled the 'good cholesterol' idea
- •Torcetrapib raised HDL but increased deaths, likely from off-target effects
- •Later CETP inhibitors were flat, stopped early, or gave only a 9% reduction
- •HDL cholesterol itself is not directly causally protective
- •Obicetrapib is pursued for LDL and apoB lowering
- 49:11 – 58:04
SR-B1 Deficiency, High HDL Paradox, and Clinical Interpretation
Rader introduces SR-B1, the HDL receptor, whose deficiency produces very high HDL yet greater heart disease risk, much like a locked room. He and Attia discuss genetics and Mendelian randomization, and Rader advises that a high HDL should never be a reason to withhold otherwise indicated treatment.
- •SR-B1 pulls cholesteryl ester from HDL into the liver
- •SR-B1 deficiency raises HDL but increases atherosclerosis risk
- •HDL levels are polygenic and mostly lack a single gene cause
- •Mendelian randomization doesn't show HDL-C is causally protective
- •High HDL is never a reason not to treat; low HDL may tip borderline decisions
- 58:04 – 1:08:19
HDL Cholesterol as an Integrator: Insulin Resistance and Triglycerides
Rader likens HDL cholesterol to an HbA1c for cardiovascular risk factors, integrating postprandial triglyceride handling over time. They discuss the high-fat milkshake experiments, the role of insulin resistance and adiponectin, and the limited clinical utility of adiponectin.
- •HDL-C acts like an integrator of triglyceride metabolism, insulin resistance and inflammation
- •Postprandial triglyceride excursions lower HDL modestly acutely and substantially chronically
- •Fasting triglycerides also affect HDL metabolism
- •Adiponectin may offer a separate route linking insulin resistance to HDL
- •Rader is skeptical of adiponectin as a clinical marker
- 1:08:19 – 1:12:10
Niacin: Raising HDL Without Benefit
Rader recounts prescribing niacin because it raised HDL and lowered triglycerides, apoB and Lp(a), until the AIM-HIGH and Merck trials showed minimal benefit. He describes telling patients to stop, and says the mechanism of its HDL raising beyond triglyceride lowering is not understood.
- •Niacin modestly lowers triglycerides, apoB, LDL and Lp(a) by about 15-20%
- •Two large trials showed disappointing effects on events
- •Rader described this as a humbling reversal with patients
- •How niacin raises HDL beyond triglyceride lowering remains unclear
- 1:12:10 – 1:23:22
Cholesterol Efflux and Reverse Cholesterol Transport
Starting from the foam cell, Rader explains efflux of cholesterol from macrophages to apoA-I/HDL and the broader process of reverse cholesterol transport back to the liver. The mature HDL measured as HDL-C is probably not the particle driving efflux, which is why raising HDL-C hasn't helped, and function may differ between people with the same HDL-C.
- •Foam cells are macrophages overloaded with cholesterol and an early feature of atherosclerosis
- •Only the liver can metabolize cholesterol, so other cells must efflux it
- •ABCA1 and apoA-I drive efflux; HDL carries cholesterol back to the liver
- •Mouse studies linked more efficient reverse cholesterol transport to less atherosclerosis
- •Static HDL-C misses function and flux
- 1:23:22 – 1:33:57
Measuring HDL Function: Efflux Capacity, HDL-P, and CSL112
Rader describes the ex vivo cholesterol efflux assay and says it predicted incident cardiovascular events better than HDL-C, though causality remains unproven. They discuss prospects for a clinical assay, HDL-P as another static measure, and the CSL112 trial of infused apoA-I as a test of the efflux hypothesis.
- •Efflux assay uses isolated HDL on labeled macrophages
- •Efflux capacity predicted events in a large prospective cohort beyond HDL-C
- •A clinical assay may be available in two to three years
- •HDL particle number is somewhat better than HDL-C but still static
- •CSL112 infusion trial in acute coronary syndrome tests the efflux hypothesis
- 1:33:57 – 1:43:06
ApoE, ApoA-I and HDL in Neurodegenerative Disease
Attia asks about HDL and neurodegeneration, and Rader introduces apoE and its isoforms, with apoE4 the major genetic risk for Alzheimer's and apoE2 protective. ApoA-I appears in the CSF without being made in the brain and observational data suggest it is protective, plausibly through ABCA1 and ABCA7. Blood and CSF levels poorly correlate.
- •ApoE2 impairs remnant clearance but is protective against Alzheimer's
- •ApoE4 is the major dose-dependent genetic risk for Alzheimer's
- •ApoA-I is found in the brain and CSF but isn't made there
- •ABCA1 and ABCA7 are Alzheimer's risk genes, making a role for apoA-I plausible
- •Blood and CSF levels of apoA-I and apoE correlate poorly
- 1:43:06 – 1:54:09
Other HDL Functions and the Future of HDL Research
Rader notes HDL effects on nitric oxide and skeletal muscle insulin sensitivity that are real in the lab but of uncertain human relevance. Both discuss the prospects for functional risk assays and for interventions hinging on the CSL112 trial. Rader suggests brain lipid metabolism rather than HDL per se is the next frontier.
- •HDL can promote nitric oxide production via SR-B1 in endothelial cells
- •HDL may promote insulin sensitivity in muscle, with unclear human relevance
- •Over 50% chance of a clinically useful function assay emerging, per Rader
- •A positive CSL112 trial would revive efflux as a target; a negative one would likely end it
- •Lipid metabolism in the brain is a promising frontier
