Push LDL below 80 and you may never get heart disease
The 8-gram rule: LDL 200 for 40 years, 100 for 80 years, 80 for 100 years — all add up to 8 grams of cholesterol. Lowering LDL works best before a heart attack; after heart failure sets in, it no longer helps.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
The 8-gram cholesterol rule
Davidson uses a calculable rule to convince 30-year-old patients: you get heart disease once you've accumulated 8 grams of cholesterol over a lifetime. 200 mg/dL times 40 years, 100 mg/dL times 80 years, 80 mg/dL times 100 years — all 8 grams. The corollary is that if you keep LDL below 80 for life, you won't get heart disease. He also draws the boundary of the treatment window: lowering LDL before a heart attack is highly effective, lowering it after a heart attack still yields benefit, but once you reach the heart failure stage, lowering LDL gives no benefit at all.
— Michael DavidsonAn N=1 comparison between two brothers
When Davidson was 16, his father died of a heart attack at 47, and he and his 14-year-old brother were both found to have equally bad lipids. He went to medical school and took niacin, then took a statin as soon as one came on the market; his brother became a family physician and ate a strict vegetarian diet, but started statins very late and had a bypass at 44. He calls this an N of one, but considers it a fairly good control — the difference lies in the 20-to-40 window when plaque forms rapidly.
— Michael DavidsonThe first-generation CETP inhibitor died of off-target effects
Pfizer's torcetrapib raised HDL by about 75% and lowered LDL by only 10–15%, yet was halted for increased deaths. Davidson says the blood pressure rise was just the tip of the iceberg: the drug entered the adrenal glands and substantially increased aldosterone and steroidogenesis. The evidence: giving the drug to rats that lack CETP raised blood pressure substantially within an hour — showing this was an off-target effect. Subsequent companies located the structural element in the molecule that induces aldosterone and built next-generation molecules without it.
— Michael DavidsonMerck's 30,000-person study saved the target
After Lilly and Roche failed in succession, Merck ran its study out to 30,000 people with four years of follow-up, with a baseline LDL already at only 60. The result: LDL fell 17% (an absolute drop of 11 mg/dL) and relative risk fell 9%, exactly as expected. This proved that the LDL reduction from CETP inhibition translates into cardiovascular benefit. The drug ultimately never launched because the lipophilic molecule kept depositing in fat tissue and couldn't be washed out for years, also creating an environmental residue problem.
— Michael Davidson1 mmol buys a 22% drop in events
Davidson offers an extrapolatable conversion: for every 1 mmol (38.8 mg/dL) drop in LDL, major adverse cardiovascular events fall 22% in relative terms, and the relationship is essentially linear. He stresses looking at the absolute drop rather than the percentage, and used this to power the Prevail study. The other key variable is time: in the first year of treatment you can capture at most half the benefit, and the curves need time to separate — which is why he thinks Lilly's ACCELERATE study was stopped too early, and why in the same two-year window ezetimibe's curves also only separated after two years.
— Michael DavidsonGenes that purely raise HDL confer no protection
A Mendelian randomization study published in Nature around 2012 delivered the key conclusion: you cannot find any gene that ‘purely raises HDL’ and is associated with cardiovascular protection. Any HDL-raising gene linked to lower risk tends to come with lower LDL or lower triglycerides as well. Davidson concludes that a patient saying ‘my HDL is high so I'm fine’ is wrong, and that this belief was overturned more than a decade ago; what bothers him more is that doctors say it too.
— Michael DavidsonCETP loss of function erases APOE4 risk
The Bronx Aging Study found that among nearly 400 Ashkenazi Jewish centenarians, the most common gene was CETP loss of function. Subsequent data showed that in people carrying APOE4, if they also had CETP loss of function, Alzheimer's risk was substantially attenuated — Davidson says the curves essentially amount to ‘taking the E4 risk away’, down to the level of non-E4. Phenotypically these people have high HDL, very few small LDL particles and very high large LDL particles, exactly the effect of CETP inhibition.
— Michael DavidsonIf statins had been used for heart failure back then, statins would never have been invented
Davidson offers a counterfactual: if cardiologists had been forced to use lipid-lowering drugs to treat heart failure, lipid-lowering drugs might never have been developed — the heart failure trials all failed. Alzheimer's is now holding exactly that hand: there is no secondary-prevention setting like a survivable MI, so pharma can only run trials in patients with dementia, and they fail across the board, making it a pharma graveyard. Attia adds: tools developed for secondary prevention could have been taken into primary prevention and saved the most people.
— Peter Attia / Michael DavidsonThe effect is strongest in APOE4 homozygotes
The Broadway study enrolled 2,500 heart patients with a mean age of 65, with prespecified analyses comparing baseline and 12-month biomarkers. In the overall population P-tau 217 fell; the older the patient the better the effect, and the effect was better in E4 carriers. Among 4-4 homozygotes, not only was P-tau 217 more than 20% different relative to placebo, but 181, amyloid 42/40, GFAP and neurofilament light all improved. Davidson says this is the part that excites him most, and also the part where he feels the data are ‘too good to be true’ and most needs the next trial to confirm.
— Michael DavidsonThe next trial has to get in before MCI
Davidson says MCI is already too late — the equivalent of early heart failure: the brain has already shrunk, inflammation has started, tau pathology is already running. So the new study will enroll 300 ‘pre-Alzheimer's’ patients — not yet at MCI, but with tau levels already elevated based on APOE. He mentions two datasets pointing in opposite directions: the GLP-1 trials Evoke and Evoke Plus showed no effect on cognition or P-tau 217; whereas in the anti-amyloid Trailblazer trial, cognitive improvement and P-tau 217 reduction were highly correlated.
— Michael DavidsonDHA can't get into the brain because a transporter is missing
The brain is mostly DHA, yet trials of oral DHA supplements for cognition have consistently come up empty. Davidson says the new finding is the MFSD2A transporter — the channel on the blood-brain barrier that carries DHA, and its preferred substrate is DHA in the lysophosphatidylcholine form (lysoPC DHA). The body's capacity to convert ordinary DHA into that form is very limited, so taking fish oil to raise red blood cell membrane EPA/DHA from 4% to 10% will probably lower cardiovascular risk, but it can't get into the brain and can't affect Alzheimer's risk. lysoPC DHA is already on the way, and will be made as a dietary supplement rather than a drug.
— Michael DavidsonFor E4 carriers, the worst move is not to test
One of Davidson's double first cousins (his mother's brother married his father's sister, so they share identical grandparents) just died of Alzheimer's at 72 and was an E4 carrier. To the argument that ‘if you're E4, don't test, because knowing changes nothing’, both men object: knowing carrier status can drive better lifestyle, exercise, DHA supplementation, and control of diabetes, blood pressure, smoking and lipids. Attia adds: he respects the personal choice not to test, but thinks ‘knowing changes nothing’ is betting on the wrong side.
— Peter Attia / Michael DavidsonIn their own words · checked verbatim
What I like is the 8-gram rule, which is that 8 grams of cholesterol in your lifetime lead to heart disease.
Michael Davidson6:19
The blood pressure increase actually was only the tip of the iceberg.
Michael Davidson21:50
For a one millimole, which is 38.8 milligrams per deciliter of LDL, you get a 22% melt-to-brist reduction.
Michael Davidson35:18
There was no pure HDL raising gene that was associated with protection.
Michael Davidson41:29
Well, look at the curves. It basically took the E4 risk away.
Michael Davidson1:05:15
There's no ApoB in the brain. There's no ApoB. It's all ApoE and HDL in the brain.
Michael Davidson1:10:27
Alzheimer's is not a disease of old age. It's a disease of middle age that presents in old age.
Peter Attia1:19:48
It's almost like had cardiologists had the misfortune of trying to use lipid lowering therapy to treat heart failure, we may never have developed lipid lowering drugs.
Michael Davidson1:21:50
I mean, HDL fooled everybody for decades. For decades, people were chasing the wrong thing, and they had all the epidemiology to back it up. But at the end of the day, it didn't matter.
Michael Davidson1:31:11
I just lost a double first cousin. My mother's brother married my father's sister. So we have all the same grandparents. She just died at age 72 of Alzheimer's. He was an E4.
Michael Davidson1:56:00
I think when you're looking so hard for things, you start to make signal out of noise.
Peter Attia1:57:01
Figures
| LDL threshold in the 8-gram rule | 80 mg/dL | 6:19 |
| HDL increase from torcetrapib | about 75% | 26:55 |
| LDL reduction from torcetrapib | 10–15% | 26:55 |
| Merck CETP study size and follow-up | 30,000 people, 4 years | 32:03 |
| Incidence of statin intolerance in placebo-controlled trials | about 4.9% | 39:24 |
| LDL reduction in the TULIP study | 45% | 40:27 |
| HDL increase from obicetrapib | 150% | 40:27 |
| One-year MACE reduction in the Broadway study | 21% | 47:45 |
| Enrollment in the Broadway study | 2,500 patients | 1:27:02 |
| Mean age of Broadway patients | 65 years | 1:27:02 |
| P-tau 217 difference vs placebo in APOE4 homozygotes | more than 20% | 1:27:02 |
| Clinical trial adherence benchmark | 80% | 1:43:36 |
| Expected timing of Clotho human data | about a year away | 1:54:55 |
| Age at which the relative died of Alzheimer's disease | 72 years | 1:56:00 |
Glossary
- CETP / cholesteryl ester transfer protein
- The protein that transfers cholesterol from HDL to LDL; inhibiting it raises HDL and lowers LDL.
- APOE4 / apolipoprotein E4 allele
- The strongest genetic risk factor for Alzheimer's disease, conferring roughly 3-fold risk in heterozygotes and roughly 10-fold in homozygotes.
- P-tau 217 / phosphorylated tau 217
- A tau fragment measurable in plasma that reflects tau pathology in the brain; a blood biomarker for Alzheimer's.
- MFSD2A / major facilitator superfamily domain-containing protein 2A
- The channel on the blood-brain barrier that transports DHA, preferring DHA in the lysophosphatidylcholine form.
- lysoPC DHA / lysophosphatidylcholine DHA
- The form in which DHA is bound to lysophosphatidylcholine; the preferred substrate of MFSD2A.
- Lp(a) / lipoprotein(a)
- A lipoprotein associated with cardiovascular risk; the Horizon study tried to prove its causality.
How to listen
Clinicians focused on cardiovascular and Alzheimer's prevention, biopharma investors, and anyone with a family history of early heart disease or APOE4 who wants to know what to test and what to push down.
The section around 1:43:36 on clinical trial adherence and operational details — low value for non-practitioners.