Peter Attia MD210 - Lp(a) and its impact on heart disease | Benoît Arsenault, Ph.D.& Peter Attia, M.D.
CHAPTERS
- 0:02 – 5:08
How Benoît Came to Study Lp(a)
Peter introduces the topic and asks where Benoît's interest in Lp(a) came from. Benoît describes his postdoc work, the earlier negative studies, and how the TNT trial and a statin-response genetics consortium pointed to Lp(a) as a strong residual-risk marker.
- •Benoît began Lp(a) research during his postdoc years at Laval and in Amsterdam
- •Genetic association studies from 2009 to 2011 revived a field hurt by negative studies
- •In the TNT trial, Lp(a) was the strongest of 18 biomarkers predicting residual cardiovascular risk
- •A statin pharmacogenetics consortium found Lp(a) was the top genetic factor explaining statin response
- 5:08 – 15:04
Epidemiology, Early Negative Studies and Lp(a) Basics
Benoît describes how common elevated Lp(a) is and why early epidemiology studies came out negative. Peter then explains the gene, apolipoprotein(a), and how it attaches to LDL.
- •About 20% of the world population has Lp(a) high enough to raise risk, with levels varying by ancestry
- •Lp(a) was discovered in 1963 and linked to cardiovascular events in European cohorts in the late 1970s and early 1980s
- •Early assays were poor, partly because of the isoform copy-number variation, so many studies were negative
- •Peter explains that apo(a) binds LDL to make a more harmful particle
- 15:04 – 21:54
The Measurement Problem and Mendelian Randomization
The two discuss why Lp(a) is hard to measure compared with apoB, since apo(a) has no single molar weight. They then cover the 2009 genetic studies and why Mendelian randomization supports a causal role for Lp(a).
- •ApoB counts particles cleanly because each LDL has one apoB100, but apo(a) size varies
- •Benoît favors nanomoles per liter over milligrams per deciliter, using immunoturbidimetric assays rather than NMR
- •Three 2009 genetic studies, including PROCARDIS, showed a dose-response between Lp(a)-raising variants and heart disease
- •Cis-acting variants and the absence of other effects make the genetic instruments reliable for causal inference
- 21:54 – 26:12
Clinical Lp(a) Tests and Measuring Once
Peter asks about the commercial assays used today. Benoît explains why particle-based measurements beat cholesterol content, how mg/dL and nmol/L relate, and why Lp(a) needs to be measured only once.
- •Lp(a) cholesterol content is discordant with particle number, even more than LDL-C is with apoB
- •Most labs report the mass of the whole particle in mg/dL
- •A high mg/dL result will generally also be high in nmol/L; 50 mg/dL is roughly 125 nmol/L
- •Lp(a) is stable over time, so most guidelines suggest measuring it once in a lifetime
- 26:12 – 35:47
Production of Lp(a) and Why Statins Don't Lower It
Benoît explains that Lp(a) is made in the liver and that its levels are set mainly by production rate. They discuss why statins fail to lower it, and why statins are still beneficial for people with high Lp(a).
- •Apo(a) is expressed in the liver and probably binds apoB within liver cells
- •Catabolism is mostly hepatic, and the receptor involved remains unclear
- •Statins raise LDL receptor density but do not lower Lp(a); studies show about a 10% increase on average
- •Trials such as the Heart Protection Study suggest statins help patients with high Lp(a)
- 35:47 – 46:00
Why Lp(a) Is More Atherogenic: Oxidized Phospholipids and Structure
Peter asks what makes Lp(a) more virulent than a regular apoB particle. Benoît points to oxidized phospholipids, plasminogen homology, and the kringle structure of apo(a), and the two speculate about free apo(a).
- •Lp(a) carries many more oxidized phospholipids than LDL, driving inflammatory and calcifying signals
- •Its sequence homology with plasminogen suggests effects on thrombosis and fibrinolysis
- •Apo(a) arose from a duplication of the plasminogen gene, and kringle IV type 2 repeats cause isoform size variation
- •Cysteine residues on kringle IV type 9 form the disulfide bond with apoB
- •Lp(a) also carries autotaxin and a distinct proteome
- 46:00 – 53:07
Lp(a) and Aortic Valve Stenosis
Peter asks about the hazard ratio for aortic stenosis. Benoît describes the genetic evidence, how risk scales with Lp(a) level, and imaging findings of early valve changes, and Peter stresses early detection.
- •Lp(a) is probably an initiator of aortic stenosis, per a 2013 GWAS of valve calcification
- •Risk rises about 50% to 100% at around 50 mg/dL and much more at very high levels
- •Sodium fluoride PET/CT shows valve changes before calcification appears
- •Progression may be faster in younger patients with high Lp(a)
- •Peter notes outcomes are better with earlier intervention on aortic stenosis
- 53:07 – 1:03:01
Particle Number Over Isoform Size, and Inheritance
Benoît explains that genetic studies resolved the debate about isoform size versus concentration in favor of particle number. They then cover autosomal dominant inheritance, how unpredictable children's levels are, and when to measure.
- •Isoform size matters only because it relates to Lp(a) level, as shown by genetic discordance analyses and the deCODE cohort
- •This mirrors apoB, where particle number is what counts
- •Inheritance is autosomal dominant, but about 2,000 variants make offspring levels hard to predict
- •The gene is fully expressed by about age two, and levels rise only slowly through adulthood
- 1:03:01 – 1:08:46
Guidelines, Screening and Extreme Cases
The two compare guideline recommendations on Lp(a) testing across the US, Canada and Europe. They also discuss the link with familial hypercholesterolemia and extreme cases.
- •American guidelines recommend measuring Lp(a) mainly after disease or family history appears, while Canadian guidelines advise testing everyone once
- •European guidelines advise universal testing partly to find very high Lp(a) that can mimic familial hypercholesterolemia
- •Peter argues testing is cheap and should be done on everyone, before age 18
- •Benoît mentions case reports of children with FH and high Lp(a) who needed liver transplants
- 1:08:46 – 1:20:40
Niacin, CETP Inhibitors and the Size of Reduction Needed
Peter asks about niacin as a therapy for Lp(a). Benoît explains why outcome trials don't support it, how Mendelian randomization suggests a large reduction is needed, and what CETP inhibitor trials showed about apoB.
- •Niacin lowers Lp(a) by about 20 to 30% but AIM-HIGH and HPS2-THRIVE showed no cardiovascular benefit
- •Mendelian randomization modeling suggests a large Lp(a) reduction, perhaps 50 to 100 mg/dL, is needed for statin-like benefit
- •Benoît cautions that comparing lifelong genetic effects with short trials has caveats
- •In the REVEAL trial of the CETP inhibitor anacetrapib, benefit tracked apoB lowering, not HDL raising
- 1:20:40 – 1:27:23
Prevalence, Penetrance and Physician Awareness
Peter and Benoît discuss how common high Lp(a) is across populations, and that it is the most prevalent form of dyslipidemia, although not everyone with high levels has events. They consider why physician awareness is low and why testing is worthwhile even without a specific drug.
- •Lp(a) is distributed differently by ancestry, but risk is proportional to level
- •Benoît says it is by far the most prevalent form of dyslipidemia, though penetrance is not 100%
- •Residual inflammation may influence risk in people with high Lp(a)
- •Benoît sees no evidence that awareness is higher in Canada or Europe, and says guidelines take time to reach practice
- •In EPIC-Norfolk, managing the Life's Simple 7 risk factors was associated with about two-thirds lower risk in high Lp(a)
- 1:27:23 – 1:37:45
Why Disease Varies Among People with High Lp(a)
Peter describes patients with modest Lp(a) and devastating disease, and others with very high levels but little disease. Benoît discusses his lab's study of isolated Lp(a) particles and candidate genes such as ApoE, CETP and ApoH.
- •Peter describes patients whose disease severity doesn't match their Lp(a) level
- •Benoît's lab isolated Lp(a) from matched aortic stenosis patients and controls, hypothesizing differences in the particle's proteome and adhesion molecules
- •Isolating Lp(a) requires chromatography because it has LDL's size and HDL's density
- •The best GWAS on Lp(a) levels implicated ApoE, CETP and ApoH, but not PCSK9 or the LDL receptor
- 1:37:45 – 1:48:05
Lp(a) in Plaque, Disease Associations and Oxidized Phospholipid Assays
The conversation covers how Lp(a) enters tissue and whether it drives initiation or progression of atherosclerosis. Benoît then ranks the diseases most associated with high Lp(a) and discusses oxidized phospholipid assays.
- •Lp(a) is found in plaques and valves, and Benoît says it acts both in initiation and progression
- •FDG-PET shows more inflammation in the arteries of high Lp(a) people without disease
- •The strongest associations are aortic stenosis, PAD and MI, then ischemic stroke and kidney disease, with no signal for deep vein thrombosis
- •Oxidized phospholipid assays do not clearly predict beyond Lp(a) level, so apoB and Lp(a) remain the markers to measure
- 1:48:05 – 2:05:56
The PCSK9 Story and PCSK9 Inhibitor Trials
Benoît tells the story of PCSK9's discovery in a family with FH and explains how it degrades LDL receptors. They discuss loss-of-function carriers with very low LDL and the FOURIER and ODYSSEY trials.
- •PCSK9 was discovered in 2003 in a French FH family with no LDL receptor mutation
- •It lowers LDL receptor density by degradation inside the cell, by binding at the surface, and by blocking recycling
- •Loss-of-function variants give low LDL and protection from cardiovascular disease without other health problems
- •FOURIER and ODYSSEY showed benefit on top of statins, with relative risk reduction of about 15% in a short trial
- 2:05:56 – 2:11:56
PCSK9 Inhibitors and Lp(a), and Personal Management Approaches
Benoît explains how PCSK9 influences apo(a) production, and why PCSK9 inhibitors lower Lp(a) only modestly and variably. Peter and Benoît then describe how they manage their own or their patients' risk.
- •PCSK9 inhibitors lower Lp(a) by roughly 25 to 30% on average, but only about 15% in people with high Lp(a)
- •The ENIGMA trial showed no change in arterial wall inflammation despite LDL lowering
- •Peter targets apoB of 30 to 40 mg/dL and uses PCSK9 inhibitors liberally until better drugs arrive
- •Benoît measured his Lp(a) at 200 nmol/L and started a statin before turning 40
- 2:11:56 – 2:24:36
Antisense Oligonucleotides and the HORIZON Trial
Benoît describes upcoming therapies that target the Lp(a) gene, including the Ionis antisense compound and Amgen's olpasiran. The conversation covers phase 2 results, the HORIZON outcomes trial and its timeline, and Peter closes with a public service message.
- •Ionis's antisense oligonucleotide lowered Lp(a) about 80% on average in phase 2, with 90% of patients below 50 mg/dL
- •The HORIZON secondary prevention trial is recruiting just under 8,000 patients and results are expected around 2025
- •Benoît says delays were partly due to COVID
- •Amgen and Silence Therapeutics also have Lp(a)-lowering siRNA compounds
- •Peter urges listeners to get Lp(a) checked, keep apoB low and consider an echocardiogram
