Longevity Doctor | Medical Concierge | Consultant in Longevity and Personalized Medicine | Entepreneur | Speaker | Author | Lecturer | Anesthesiologist
Your friends are an anti-aging treatment. I have been saying it for years, and now there is a number.
MIDUS cohort, 2,117 Americans. Researchers built a cumulative social advantage score from parental warmth in childhood, community ties, and current emotional support, then they ran the aging clocks.
The result: higher social advantage tracked with slower DunedinPACE (the pace-of-aging speedometer), slower GrimAge (the strongest mortality clock we have), and lower IL-6, a core inflammation molecule.
Two details make this finding solid rather than merely pleasant:
1. The signal runs through inflammation, which gives it a plausible and measurable mechanism.
2. First-generation clocks, the ones trained to guess your calendar age, see nothing, while only the clocks that capture current aging pace pick it up. That is exactly what you would expect if social connection changes how fast you are aging rather than how old you already are.
For scale: omega-3s, the single best supplement result in randomized trials, buy you 2.9 to 3.8 months over 3 years (DO-HEALTH, n=777). Cumulative social advantage sits in the same league, it costs nothing, and it compounds.
The bricks are small, they are additive, and you stack them for 30 years. The most profitable one this weekend might not be in a pharmacy.
Full article below.
Nobody can erase a car's mileage. Everybody can lift their foot.
That image, from Steve Horvath himself, fixes almost every misunderstanding about biological age, so let me unpack it.
First-generation clocks (Horvath 2013, Hannum 2013) are odometers. Trained to guess your calendar age from DNA methylation, they tell you how many kilometers you have driven. The accuracy is impressive (r = 0.96 with civil age), and it is almost useless for decisions, because a backward-looking number does not tell you what to change.
Second-generation clocks (PhenoAge 2018, GrimAge 2019) are odometers corrected by the state of the engine. They were trained on clinical markers and mortality instead of birthdays, and each additional year of PhenoAge carries +4.5 % all-cause mortality risk. They do better, and they still look backwards.
DunedinPACE (2022) is the speedometer. Built on 19 biomarkers tracked for 19 years in the Dunedin cohort, it measures how fast you are aging right now. A value of 1.0 means one biological year per calendar year. At 1.2, you are aging 20 % too fast. At 0.85, you have lifted your foot.
Here is why it matters: in the CALERIE randomized trial, 2 years of caloric restriction slowed DunedinPACE by 2 to 3 % while PhenoAge and GrimAge did not move at all. The same trial, the same blood, and 2 opposite verdicts. Every headline about "reversing your biological age" should be read with that dissociation in mind.
The speedometer responds within 12 to 24 months. That is your window, and that is where your leverage is: vegetables, omega-3s, strength, sleep, and people you love, stacked for 30 years.
That is the program that will let me ski at 90. Full article below.
"Biological age: 38." He is 52. He was thrilled.
I had to tell him the number means almost nothing on its own. Here is what Steve Horvath, the man who invented the first epigenetic clock, spent 2 hours 39 minutes explaining about his own creation, and what it changes for you.
1. Biological age is not one thing. It is at least 5 things: functional age (can you get out of a chair without using your hands), molecular age, organ-by-organ age, mortality risk, and your current pace of aging. These 5 dimensions do not move together, and each clock captures exactly one of them. A single number on a printout collapses all of that.
2. There are dozens of epigenetic clocks, and they disagree. In the landmark reliability study (Higgins-Chen, Nature Aging 2022), the same blood sample, run twice on the same machine, differed by up to 9 years depending on the clock. PhenoAge was the least reliable, with a median deviation of 2.4 years between technical replicates. PC-corrected clocks bring that error down to 0.3 to 0.8 years, and most consumer tests do not use that correction.
3. It gets worse: guessing your calendar age and predicting your death are 2 different skills. Across 39 aging biomarkers evaluated on more than 20,000 people, the correlation between chronological-age accuracy and mortality prediction was R = 0.12, which is to say there is none. "96 % accuracy" is a marketing argument, not a clinical one.
4. The only clock that reliably moves in randomized human trials is DunedinPACE, a speedometer rather than an odometer. It measures how fast you are aging right now, and it responds within 12 to 24 months. That is where your leverage is.
5. What moves it: caloric restriction slowed it by 2 to 3 % over 2 years (CALERIE, n=220). One gram of omega-3s per day avoided 2.9 to 3.8 months of aging over 3 years (DO-HEALTH, n=777). Vitamin D alone did nothing, and exercise alone showed nothing visible in blood, but all 3 together became additive. That word is the whole strategy.
You stack small, verified, additive bricks for 30 years. None of them comes on a 300-franc printout.
Full article below.
In older adults, a high GDF-15 level goes together with frailty and with a higher risk of dying in the years that follow. That sentence needs reading carefully, because these are associations measured in large cohorts and they do not prove the molecule causes anything.
What it describes is simple. When a cell has to spend far more energy to keep doing its usual job, it releases this signal into the blood. What it burns to hold its position is not spent on repairing itself. Maintain an organism that way for years and the reserves run short.
Many people watch exactly this happen to a parent without having a word for it. Nothing specific is wrong, every test comes back acceptable, and yet something visibly costs more than it used to.
No treatment targets this marker today, and that is not what should be expected from it. What it indicates is where to look, which is the strain that lasted long enough to become that person's ordinary life.
None of this replaces medical advice. If the question comes up for a relative, it belongs in a conversation with their doctor.
Full article, with the studies, in the reply below.
A marker that rises in almost every disease looks useless. It is the opposite, and the misreading is worth understanding.
GDF-15 climbs in cardiovascular, renal and metabolic disease, in cancer and in severe infection. A signal that points everywhere appears to point nowhere, so it sat in a blind spot for years while more specific markers got the attention.
What it measures is not a disease. It is what a cell spends to keep doing its usual job once the demand exceeds what it can deliver. Human fibroblasts held under chronic stress in culture raised their energy expenditure by 61.9 percent. Cells in a dish are not a human being, and the figure is still worth sitting with, because what a cell burns to hold its position is not spent on repairing itself.
The links between this marker and disease are associations measured in cohorts. They do not establish that the molecule triggers anything, and treating it as a cause is how the next decade of supplements will be sold.
Pushing the number down would therefore be the wrong goal. You would be working on the smoke while the fire keeps burning.
The useful question is not how to lower this marker. It is what has been keeping it high for months.
Full article, with the studies, in the reply below.
There is a molecule in your blood that rises whenever your cells can no longer keep up. It is called GDF-15, and I spent a good part of this week reading about it, because it behaves unlike most of what we measure.
Levels climb in cardiovascular and kidney disease, and the same is true in cancer, in diabetes and in severe infection. In older adults, high values go together with frailty and with a higher risk of dying in the years that follow. Those are associations rather than proof of cause, and a marker that moves in nearly every illness looks useless at first, since a signal pointing everywhere seems to point nowhere.
That is exactly where the reading goes wrong.
GDF-15 is not the disease. It is what a cell releases when what is being asked of it exceeds what it can deliver. Pregnancy shows this more clearly than anything else, because the fetus and the placenta pour GDF-15 into the mother's blood, and that surge is now understood to drive the nausea and vomiting of early pregnancy. Its severe form, hyperemesis gravidarum, has cost women their lives, so this is not an abstract laboratory curiosity.
One laboratory number stayed with me. Human fibroblasts kept under chronic stress increased their energy expenditure by 61.9 percent. The job stays the same and the cost of holding it goes up. GDF-15 is the trace that spending leaves in the blood.
Which is why trying to push the number down would be the wrong goal. You would be working on the smoke while the fire keeps burning, since the molecule is a witness and not the culprit.
The lever, unfortunately, does not come in a bottle. What usually keeps this signal high is a situation that has been wearing someone down for months, the one they stopped calling a problem because it became normal. Working on that is the part no capsule can do for you.
Full article, with the studies, in the reply below.
Attention pour ce 12 août 2026, l'éclipse solaire qui sera visible en Europe.
Si :
> Vous Fixer le Soleil quelques secondes
> Vous pouvez brûler DEFINITEVEMENT votre rétine sans douleur.
Prenez des lunettes d’éclipse certifiées ISO 12312-2.
Les lunettes de soleil ne suffisent pas. Vous pouvez vraiment y laisser une partie de votre vue.
Le levier le mieux mesuré ne coûte rien.
Il coûte du temps.
C'est l'endurance. Onze essais randomisés ont réuni 1 147 hommes présentant un trouble de l'érection : 30 à 60 minutes à intensité modérée, 3 à 5 fois par semaine. Le score de fonction érectile progresse de 2,8 points en moyenne.
Le gain monte avec la sévérité de départ : 2,3 points dans les formes légères, 4,9 dans les sévères.
C'est un effet mesuré, pas une garantie. Mais ce qui entretient vos artères entretient le reste.
Une semaine d'abstinence ferait grimper la testostérone de 45%.
L'étude qui l'affirmait a été rétractée par la revue.
Elle date de 2003, retirée pour publication redondante, jamais reproduite. Aucun autre travail ne retrouve d'effet sur la testostérone ni le cortisol.
Dans un essai, 70 couples ont doublé leur fréquence sur 90 jours : ils se sont dits légèrement moins heureux, avec moins de désir.
Ces études sont petites. La détente ressentie est réelle, sans explication hormonale démontrée.