More people living longer means more cancer cases. The fairer test is the death rate, adjusted for age.
Worldwide, cancer deaths more than doubled from 1980 to 2021, but the age-adjusted death rate fell by more than 20% (see chart).
The global data stops at 2021. US data runs to 2023, and ting, averting 4.8 million deaths since 1991.
Worldwide, cancer deaths more than doubled from 1980 to 2021, but the age-adjusted death rate fell by more than 20% (see chart).
The global data stops at 2021. US data runs to 2023, and ting, averting 4.8 million deaths since 1991.
2
0
1
Do cancer vaccines work yet? In the past two weeks, three trials have given three very different answers.
Yes, in melanoma. Moderna and Merck built an mRNA vaccine from each patient's tumor mutations after surgery, then gave it with Keytruda. In 1,137 patients, 82% were alive ce at two years, versus 70% on Keytruda alone. The full data were presented today at ESMO in Madrid.
Not in pancreatic cancer, at least not yet. Elicio's off-the-shelf vaccine targets seven common mutations in the KRAS cancer gene and was given after surgery and standard treatment. In 144 patients, it did not clearly reduce recurrence overall, though some subgroups did better, and the company is talking to the FDA about a phase 3.
Not yet, either, in advanced head and neck cancer. BioNTech's vaccine against two proteins from the HPV16 virus, added to Keytruda, did not clearly slow the cancer, and the trial now waits to see whether people live longer.
The pattern matters. The vaccine that worked was personalized to each patient’s tumor and used after surgery had removed the cancer doctors could see. Even then, it took about six months before the two groups pulled apart.
The clearest signal so far is this: a patient-specific vaccine can help keep melanoma from coming back after surgery. The next question is which cancers can follow.
Yes, in melanoma. Moderna and Merck built an mRNA vaccine from each patient's tumor mutations after surgery, then gave it with Keytruda. In 1,137 patients, 82% were alive ce at two years, versus 70% on Keytruda alone. The full data were presented today at ESMO in Madrid.
Not in pancreatic cancer, at least not yet. Elicio's off-the-shelf vaccine targets seven common mutations in the KRAS cancer gene and was given after surgery and standard treatment. In 144 patients, it did not clearly reduce recurrence overall, though some subgroups did better, and the company is talking to the FDA about a phase 3.
Not yet, either, in advanced head and neck cancer. BioNTech's vaccine against two proteins from the HPV16 virus, added to Keytruda, did not clearly slow the cancer, and the trial now waits to see whether people live longer.
The pattern matters. The vaccine that worked was personalized to each patient’s tumor and used after surgery had removed the cancer doctors could see. Even then, it took about six months before the two groups pulled apart.
The clearest signal so far is this: a patient-specific vaccine can help keep melanoma from coming back after surgery. The next question is which cancers can follow.
8
4
61
Millions of people take Ozempic or Mounjaro, and many wonder if these drugs might also protect memory. So far, the biggest test says no. In 3,808 people with early Alzheimer's, the drug in Ozempic did not slow the disease.
The idea came from health records. A 2024 Case Western with type 2 diabetes on semaglutide, the drug in Ozempic and Wegovy, were diagnosed with Alzheimer's about a third as often as similar people on insulin.
Novo Nordisk, the maker of Ozempic, then ran a proper trial. In two studies, people took semaglutide or a placebo for two years. Their memory and daily life declined at the same pace.
These results answer different questions. Health records can be skewed because people on insulin are often sicker to begin with, while the trials tested people who already had Alzheimer's. Mounjaro has no Alzheimer's trial results yet.
The bottom line: these drugs still help with weight, blood sugar, and in the case of Ozempic, heart attacks and strokes. But for memory, the largest trials so far came up empty.
The idea came from health records. A 2024 Case Western with type 2 diabetes on semaglutide, the drug in Ozempic and Wegovy, were diagnosed with Alzheimer's about a third as often as similar people on insulin.
Novo Nordisk, the maker of Ozempic, then ran a proper trial. In two studies, people took semaglutide or a placebo for two years. Their memory and daily life declined at the same pace.
These results answer different questions. Health records can be skewed because people on insulin are often sicker to begin with, while the trials tested people who already had Alzheimer's. Mounjaro has no Alzheimer's trial results yet.
The bottom line: these drugs still help with weight, blood sugar, and in the case of Ozempic, heart attacks and strokes. But for memory, the largest trials so far came up empty.
6
1
15
At 35, you may feel young, but the arteries feeding your heart may already be changing almost twice as fast as they will at 55.
That comes from a study of more than 25,000 microscope slides from 970 adults ages 21 to 70. Researchers at Sanford Burnham Prebys and the NIH used an 40 tissues change with age.
The coronary arteries changed fastest in the 30s. Pathologists' notes also showed early plaque rising most sharply between the 20s and 30s.
The bigger lesson: organs do not age on the same schedule. The uterus and vagina changed most in the 50s, around menopause.
This was an average across people, since each donor was seen only once, after death. Brain arteries were not included.
Your body does not age all at once. Your arteries start earlier than you think
That comes from a study of more than 25,000 microscope slides from 970 adults ages 21 to 70. Researchers at Sanford Burnham Prebys and the NIH used an 40 tissues change with age.
The coronary arteries changed fastest in the 30s. Pathologists' notes also showed early plaque rising most sharply between the 20s and 30s.
The bigger lesson: organs do not age on the same schedule. The uterus and vagina changed most in the 50s, around menopause.
This was an average across people, since each donor was seen only once, after death. Brain arteries were not included.
Your body does not age all at once. Your arteries start earlier than you think
9
4
37
Americans are flying to Shanghai for a lung cancer pill that the FDA approved 15 months ago.
Max Doppelt did it for his mother. Her lung cancer had spread to her brain. The drug most likely to help her, sunvozertinib, had been sold in China since 2023 but was unavailable in the came home with 26 boxes. Nineteen months later, she is alive and still working. He says at least 30 Americans have made the same trip.
In the trial behind the US approval, the pill shrank tumors in 46 out of every 100 patients whose cancer had worsened after chemotherapy.
The holdup is business, not medicine. Dizal, the Shanghai company that makes it, had no US sales network, according to US oncologists in May. AstraZeneca licensed the drug in July for $600 million and says US sales will start this quarter.
An @FDA approval helps American patients only when a company actually sells the drug here.
Max Doppelt did it for his mother. Her lung cancer had spread to her brain. The drug most likely to help her, sunvozertinib, had been sold in China since 2023 but was unavailable in the came home with 26 boxes. Nineteen months later, she is alive and still working. He says at least 30 Americans have made the same trip.
In the trial behind the US approval, the pill shrank tumors in 46 out of every 100 patients whose cancer had worsened after chemotherapy.
The holdup is business, not medicine. Dizal, the Shanghai company that makes it, had no US sales network, according to US oncologists in May. AstraZeneca licensed the drug in July for $600 million and says US sales will start this quarter.
An @FDA approval helps American patients only when a company actually sells the drug here.
In January 2025 my mom’s lung cancer spread to her brain. The drug with the best chance of helping her was approved in China but not here.
A doctor told me, almost as an aside, that if she were Chinese she could buy it at a pharmacy.
So I flew to Shanghai and came home with 26 boxes in my luggage. That was 19 months ago. My mom is still here, still working, and in her words has had an amazing three years.
The drug was approved in China in 2023. The FDA approved it in July 2025. It still is not for sale here. I know of at least 30 Americans who have made the same trip.
I have been scared to tell this story because I worried it could cost my mom access to the medicine keeping her alive. But the problem matters more than my silence. These drugs need to be developed and available here, and that means FDA reform.
Art Allen at KFF Health News told our story this week. It also ran in The Washington Post.
kffhealthnews.org/health-industr…
19
109
506
One night of sleep can hint at how long you might live.
When doctors suspect sleep apnea, they send you for a sleep test. You spend the night wired up in a clinic, and the main score is simple: how many times an hour you stop breathing.
At Cleveland Clinic, that number did not kely to die. An AI that read the rest of the night did. Using brain waves, heartbeat and breathing, it sorted about 9,300 patients into five risk groups.
Six years later, about 18 in 100 in the highest-risk group had died, versus 6 in 100 in the lowest. Even among people the same age and weight, with the same number of breathing stops, the highest-risk group still died at more than twice the rate.
Age, sex and weight do much of the work. At Stanford, a similar AI looked at pairs of patients where one died first and picked the right one 84 times out of 100. A simple model using only age, sex, weight and race got 79.
Both studies looked at people referred to sleep clinics, and neither tested a smartwatch or ring.
When doctors suspect sleep apnea, they send you for a sleep test. You spend the night wired up in a clinic, and the main score is simple: how many times an hour you stop breathing.
At Cleveland Clinic, that number did not kely to die. An AI that read the rest of the night did. Using brain waves, heartbeat and breathing, it sorted about 9,300 patients into five risk groups.
Six years later, about 18 in 100 in the highest-risk group had died, versus 6 in 100 in the lowest. Even among people the same age and weight, with the same number of breathing stops, the highest-risk group still died at more than twice the rate.
Age, sex and weight do much of the work. At Stanford, a similar AI looked at pairs of patients where one died first and picked the right one 84 times out of 100. A simple model using only age, sex, weight and race got 79.
Both studies looked at people referred to sleep clinics, and neither tested a smartwatch or ring.
5
6
49
Pope Leo XIV has lung cancer, despite never smoking. That is not unusual. About 1 in 4 people with lung cancer worldwide never smoked, but US and UK screening usually focuses only on smokers.
Taiwan goes further. It also scans people with a parent, sibling, or child who had lung ar, it found cancer in 1.4% of them, more than double the 0.6% seen in heavy smokers. Nearly 9 in 10 of those relatives were nonsmokers.
But broader screening has a cost. CT scans also catch cancers that would never have caused harm. In Taiwanese women, almost none of whom smoke, early lung cancers found rose more than sixfold from 2004 to 2018, while deaths stayed flat.
If your parents or siblings had lung cancer and you are over 50, ask your doctor whether a scan makes sense, even if you have never smoked.
Taiwan goes further. It also scans people with a parent, sibling, or child who had lung ar, it found cancer in 1.4% of them, more than double the 0.6% seen in heavy smokers. Nearly 9 in 10 of those relatives were nonsmokers.
But broader screening has a cost. CT scans also catch cancers that would never have caused harm. In Taiwanese women, almost none of whom smoke, early lung cancers found rose more than sixfold from 2004 to 2018, while deaths stayed flat.
If your parents or siblings had lung cancer and you are over 50, ask your doctor whether a scan makes sense, even if you have never smoked.
1
0
16
1 in 8 cancers worldwide starts with a germ you can catch.
That matters because it creates a target years before a tumor exists. Six examples of how prevention and early action can blunt cancer risk:
1. HPV and cervical cancer. In England, girls offered the vaccine at 12 or 13 ancer in their twenties.
2. HPV and throat cancer. HPV causes more than 8 in 10 back-of-throat cancers in the US. In one survey, about 1 in 60 unvaccinated young adults carried vaccine strains in the mouth, versus about 1 in 900 vaccinated ones.
3. Hepatitis B and liver cancer. In a Chinese trial, newborns who got the shot had 72% less liver cancer over 37 years.
4. H. pylori and stomach cancer. A week of antibiotics cut stomach cancers from 23 to 10 in a Korean trial of 1,838 people with an affected close relative. In China, a 10-day course for infected villagers cut risk by 14% compared with a symptom pill.
5. Hepatitis C and liver cancer. In records of 22,500 US veterans, people cured with newer pills got liver cancer at about a quarter the rate of those still infected.
6. HIV. Starting HIV pills right away, rather than waiting, cut infection-linked cancers from 23 to 6 in a trial of 4,684 adults.
Epstein-Barr virus, which causes mono, has no vaccine yet. In Hong Kong, a blood test for its DNA found 71% of the nose cancers it detected at an early stage, versus 20% before screening.
The punchline: cancers that start with a germ can often be stopped by targeting the germ.
That matters because it creates a target years before a tumor exists. Six examples of how prevention and early action can blunt cancer risk:
1. HPV and cervical cancer. In England, girls offered the vaccine at 12 or 13 ancer in their twenties.
2. HPV and throat cancer. HPV causes more than 8 in 10 back-of-throat cancers in the US. In one survey, about 1 in 60 unvaccinated young adults carried vaccine strains in the mouth, versus about 1 in 900 vaccinated ones.
3. Hepatitis B and liver cancer. In a Chinese trial, newborns who got the shot had 72% less liver cancer over 37 years.
4. H. pylori and stomach cancer. A week of antibiotics cut stomach cancers from 23 to 10 in a Korean trial of 1,838 people with an affected close relative. In China, a 10-day course for infected villagers cut risk by 14% compared with a symptom pill.
5. Hepatitis C and liver cancer. In records of 22,500 US veterans, people cured with newer pills got liver cancer at about a quarter the rate of those still infected.
6. HIV. Starting HIV pills right away, rather than waiting, cut infection-linked cancers from 23 to 6 in a trial of 4,684 adults.
Epstein-Barr virus, which causes mono, has no vaccine yet. In Hong Kong, a blood test for its DNA found 71% of the nose cancers it detected at an early stage, versus 20% before screening.
The punchline: cancers that start with a germ can often be stopped by targeting the germ.
1
0
22
Six months on, a weekly shot that shrinks your jeans size still makes more money than @OpenAI and @AnthropicAI combined.
The shots are Ozempic and Mounjaro. Last year they brought in $71B, four times the $18B the two AI labs booked.
This year AI almost caught up, mostly es exploded. It sold $11.5B from April to June alone, more than Ozempic, after $4.6B in all of 2025. The drugs still lead by about a third.
Only about 1 in 50 people with obesity worldwide takes one. Two obesity pills arrived this year: @novonordisk's Wegovy in January and @EliLillyandCo's Foundayo in April, both priced at $149 a month, and more than 30 new obesity drugs are in late-stage trials.
AI already has 1.2 billion weekly users. The drugs have barely started.
The shots are Ozempic and Mounjaro. Last year they brought in $71B, four times the $18B the two AI labs booked.
This year AI almost caught up, mostly es exploded. It sold $11.5B from April to June alone, more than Ozempic, after $4.6B in all of 2025. The drugs still lead by about a third.
Only about 1 in 50 people with obesity worldwide takes one. Two obesity pills arrived this year: @novonordisk's Wegovy in January and @EliLillyandCo's Foundayo in April, both priced at $149 a month, and more than 30 new obesity drugs are in late-stage trials.
AI already has 1.2 billion weekly users. The drugs have barely started.
A weekly jab in the belly is generating more revenue than the entire AI industry.
Ozempic + Mounjaro: $71B in 2025.
OpenAI + Anthropic: $29B.
And they've barely started. ~2% of the 800 million eligible patients can currently access them.
h/t @DrSamuelBHume
8
18
101
The oldest known land animal is Jonathan, a tortoise estimated at 194. And parts of his DNA still look young.
He was already fully grown when he was shipped to St Helena in 1882. This month, scientists read his DNA from a cheek scrape, since a blood draw was not allowed for his with four other giant tortoises, from a 5-year-old to two old males.
Most of Jonathan's DNA looks its age. The chemical tags on genes get messier as animals get older, and his are messy too.
But at 272 gene switches, his tags were as tidy as the young tortoises'. Thirty-five of those genes work in mitochondria. He also carries changes in 287 genes that the two other tortoise genomes he was checked against do not have, including genes that repair DNA and guard chromosome tips. It is one animal, so none of this proves cause.
At 194, Jonathan shows that even a very old body can keep some parts young.
He was already fully grown when he was shipped to St Helena in 1882. This month, scientists read his DNA from a cheek scrape, since a blood draw was not allowed for his with four other giant tortoises, from a 5-year-old to two old males.
Most of Jonathan's DNA looks its age. The chemical tags on genes get messier as animals get older, and his are messy too.
But at 272 gene switches, his tags were as tidy as the young tortoises'. Thirty-five of those genes work in mitochondria. He also carries changes in 287 genes that the two other tortoise genomes he was checked against do not have, including genes that repair DNA and guard chromosome tips. It is one animal, so none of this proves cause.
At 194, Jonathan shows that even a very old body can keep some parts young.
3
5
18
Aging is the biggest risk factor for disease, yet the FDA has never approved a drug to treat it.
Your risk of dying roughly doubles every 8 years from midlife on. Aging is also the top risk factor for most cancers, many forms of heart disease, and dementia. But because aging is , drugs are still tested one disease at a time.
That is where the money goes. The US National Institute on Aging put only about 7.6% of its $4.5 billion budget into the biology of aging last year. TAME, the trial built to test metformin against aging itself, has been ready for years and is still short of funding. Its team first met with the FDA in 2015, and it still has not launched.
As long as we keep calling aging natural, the drugs that might slow it will keep getting stuck in one-disease-at-a-time trials, if they are tested at all.
Your risk of dying roughly doubles every 8 years from midlife on. Aging is also the top risk factor for most cancers, many forms of heart disease, and dementia. But because aging is , drugs are still tested one disease at a time.
That is where the money goes. The US National Institute on Aging put only about 7.6% of its $4.5 billion budget into the biology of aging last year. TAME, the trial built to test metformin against aging itself, has been ready for years and is still short of funding. Its team first met with the FDA in 2015, and it still has not launched.
As long as we keep calling aging natural, the drugs that might slow it will keep getting stuck in one-disease-at-a-time trials, if they are tested at all.
The most dangerous myth in medicine is that aging is natural and therefore acceptable
8
7
50
Two obesity pills. Very different drugs.
Swallow the Wegovy pill, and only about 1% to 2% reaches your blood. Swallow Lilly's Foundayo, and most of a starting dose does.
Why? Wegovy is semaglutide, the drug in the Ozempic and Wegovy shots, turned into a daily tablet. It is a the gut mostly breaks it down. That means a helper chemical, an empty stomach and a 30-minute wait.
Foundayo is newer. It is Lilly’s small-molecule drug from the maker of Zepbound. Same target, but no helper and no food rules.
Both work. In separate trials, the Wegovy pill cut body weight by about 14% on average, and Foundayo by about 11%, versus about 2% on a dummy pill.
Novo says more than 8 in 10 new Wegovy pill prescriptions go to people who had no Ozempic-type prescription in the past year.
If that holds, the Wegovy pill is mostly reaching people who have not been on these drugs before. No head-to-head obesity trial has yet tested which pill is more effective over time.
Swallow the Wegovy pill, and only about 1% to 2% reaches your blood. Swallow Lilly's Foundayo, and most of a starting dose does.
Why? Wegovy is semaglutide, the drug in the Ozempic and Wegovy shots, turned into a daily tablet. It is a the gut mostly breaks it down. That means a helper chemical, an empty stomach and a 30-minute wait.
Foundayo is newer. It is Lilly’s small-molecule drug from the maker of Zepbound. Same target, but no helper and no food rules.
Both work. In separate trials, the Wegovy pill cut body weight by about 14% on average, and Foundayo by about 11%, versus about 2% on a dummy pill.
Novo says more than 8 in 10 new Wegovy pill prescriptions go to people who had no Ozempic-type prescription in the past year.
If that holds, the Wegovy pill is mostly reaching people who have not been on these drugs before. No head-to-head obesity trial has yet tested which pill is more effective over time.
6
0
26
You may be carrying a built-in calorie burner. It keeps you warm and may help protect you from disease.
It is called brown fat. Adults who keep more of it tend to have far less diabetes, heart disease, and depression.
Babies carry it between their shoulder blades because they g seals and reindeer in the far north run on it. Its cells are packed with mitochondria that burn sugar and fat straight into heat. Most of us lose a lot of it with age.
This week, a @Nature study from @RockefellerUniv and @MSKCancerCenter analyzed scans from more than 100,000 cancer patients. Nearly 9,000 still had brown fat. Compared with similar patients who did not, they had less than half the type 2 diabetes, over 40% less heart disease, and a third less depression. That is a link, not proof.
The good news: it can still grow in adults. In a small study, a month of regular cold raised active brown fat by almost half.
Losartan, a common blood pressure pill, also boosted its heat-making genes in mice, though it has not been tested in people.
So the fat that kept you warm as a baby may also help keep you healthier as an adult.
It is called brown fat. Adults who keep more of it tend to have far less diabetes, heart disease, and depression.
Babies carry it between their shoulder blades because they g seals and reindeer in the far north run on it. Its cells are packed with mitochondria that burn sugar and fat straight into heat. Most of us lose a lot of it with age.
This week, a @Nature study from @RockefellerUniv and @MSKCancerCenter analyzed scans from more than 100,000 cancer patients. Nearly 9,000 still had brown fat. Compared with similar patients who did not, they had less than half the type 2 diabetes, over 40% less heart disease, and a third less depression. That is a link, not proof.
The good news: it can still grow in adults. In a small study, a month of regular cold raised active brown fat by almost half.
Losartan, a common blood pressure pill, also boosted its heat-making genes in mice, though it has not been tested in people.
So the fat that kept you warm as a baby may also help keep you healthier as an adult.
5
0
19
VO2 max gets all the attention.
But there is another health metric that may matter even more, and almost nobody measures it outside research scans: blood flow to your brain.
Your brain is 2% of your body weight but uses about 20% of your oxygen at rest.
That flow moves all raises it. Two hours after a strong espresso's worth of caffeine, it was about 20% lower. Fast, heavy breathwork cut it by about 40% in lab volunteers. Deep sleep lowers it too.
Under those swings is a slow decline. In a 2025 @ScientificData atlas of more than 1,000 healthy people, children's brains got nearly twice the blood of people over 65, and people over 65 got about a quarter less than adults aged 25 to 44. Women ran 10 to 15% above men from their teens on.
Lower flow also tracked with dementia. In 4,759 Dutch adults, the quarter with the least flow were about twice as likely to be diagnosed within 7 years as the quarter with the most. In Alzheimer's, flow falls several times faster than in normal aging.
Meditation is still an open question. In an 18-month French trial, adults over 65 who meditated saw flow rise slightly in two brain regions, but not enough to rule out chance.
Fitness has stronger evidence for lifespan, and I could not find a head-to-head study on brain flow. But one trial found that aiming for a top blood pressure number below 120, rather than below 140, increased blood flow throughout the brain.
If I wanted more blood reaching my brain, that's where I'd start.
But there is another health metric that may matter even more, and almost nobody measures it outside research scans: blood flow to your brain.
Your brain is 2% of your body weight but uses about 20% of your oxygen at rest.
That flow moves all raises it. Two hours after a strong espresso's worth of caffeine, it was about 20% lower. Fast, heavy breathwork cut it by about 40% in lab volunteers. Deep sleep lowers it too.
Under those swings is a slow decline. In a 2025 @ScientificData atlas of more than 1,000 healthy people, children's brains got nearly twice the blood of people over 65, and people over 65 got about a quarter less than adults aged 25 to 44. Women ran 10 to 15% above men from their teens on.
Lower flow also tracked with dementia. In 4,759 Dutch adults, the quarter with the least flow were about twice as likely to be diagnosed within 7 years as the quarter with the most. In Alzheimer's, flow falls several times faster than in normal aging.
Meditation is still an open question. In an 18-month French trial, adults over 65 who meditated saw flow rise slightly in two brain regions, but not enough to rule out chance.
Fitness has stronger evidence for lifespan, and I could not find a head-to-head study on brain flow. But one trial found that aiming for a top blood pressure number below 120, rather than below 140, increased blood flow throughout the brain.
If I wanted more blood reaching my brain, that's where I'd start.
13
12
104
Menopause is neither natural nor universal. It occurs mainly in humans and in a few toothed whales, such as orcas.
Until now, menopause has mostly been studied in small groups. The landmark US study, SWAN, followed about 3,300 women.
@UriAlonWeizmann's lab at @WeizmannScience a big way. @glen_pridham and @Yoav_Hayut borrowed a trick astronomers use to sharpen blurry images and applied it to 300 million lab tests from more than a million women, in @NatureMedicine.
The result is a much clearer picture of what changes before menopause.
Sorted by age, estrogen fades gradually. But sorted by years from the last period, it drops in a single step. 52 of 90 measurements move with it, including bone density, liver tests, cholesterol, and blood counts. Men analyzed the same way show no step.
The shift starts 12 to 15 years before the last period, about 7 years earlier than past estrogen studies suggested. Women on hormone therapy showed smaller steps, though this was only observational.
These are averages, so we don’t yet know if each woman’s own change is this sudden.
For a typical woman, whose last period comes around 50, the hormones start shifting in her mid to late 30s.
Until now, menopause has mostly been studied in small groups. The landmark US study, SWAN, followed about 3,300 women.
@UriAlonWeizmann's lab at @WeizmannScience a big way. @glen_pridham and @Yoav_Hayut borrowed a trick astronomers use to sharpen blurry images and applied it to 300 million lab tests from more than a million women, in @NatureMedicine.
The result is a much clearer picture of what changes before menopause.
Sorted by age, estrogen fades gradually. But sorted by years from the last period, it drops in a single step. 52 of 90 measurements move with it, including bone density, liver tests, cholesterol, and blood counts. Men analyzed the same way show no step.
The shift starts 12 to 15 years before the last period, about 7 years earlier than past estrogen studies suggested. Women on hormone therapy showed smaller steps, though this was only observational.
These are averages, so we don’t yet know if each woman’s own change is this sudden.
For a typical woman, whose last period comes around 50, the hormones start shifting in her mid to late 30s.
14
30
142
Hey Lancet Commission, what exactly is "unhealthy aging"?
All aging is unhealthy aging. Aging is the number one risk factor for age-related diseases. That is why they are called age-related diseases.
You put malicious AI and nuclear war at the top of your pyramid, the threats ial." On your own chart, their error bars start at about zero. That is a shrug with a confidence interval.
Aging sits a tier lower, and its toll is not a guess. It is counted in the death records. Two in three people who died worldwide in 2023 were 60 or older, and about six in ten of those died of heart disease, stroke, cancer, or dementia.
So the two threats whose ranges start at zero got the top tier, and the one burying our parents got an adjective.
AI can help with this right now. It can speed up drug discovery, find patients for clinical trials faster, read screening scans to catch more cancers, and help bring basic health care to the 4.6 billion people who still lack it.
Point AI at aging. Then redraw the pyramid.
All aging is unhealthy aging. Aging is the number one risk factor for age-related diseases. That is why they are called age-related diseases.
You put malicious AI and nuclear war at the top of your pyramid, the threats ial." On your own chart, their error bars start at about zero. That is a shrug with a confidence interval.
Aging sits a tier lower, and its toll is not a guess. It is counted in the death records. Two in three people who died worldwide in 2023 were 60 or older, and about six in ten of those died of heart disease, stroke, cancer, or dementia.
So the two threats whose ranges start at zero got the top tier, and the one burying our parents got an adjective.
AI can help with this right now. It can speed up drug discovery, find patients for clinical trials faster, read screening scans to catch more cancers, and help bring basic health care to the 4.6 billion people who still lack it.
Point AI at aging. Then redraw the pyramid.
What will shape the health of the next generation?
A new Lancet Commission outlines 17 interlinked catastrophic threats to health and human survival between 2025 and 2100: spkl.io/60157nJvz #21stCenturyThreatsToHealth @IHME_UW @UNFPA
2
1
14
Do adults with ADHD face a higher risk of dementia? A study of 187,341 Americans age 50 and up says: maybe, but the picture is messy.
Over about 10 years of health records, about 7 in 100 people with ADHD were diagnosed with mild cognitive impairment or dementia, versus about 3 uestion is doctor visits. People with ADHD had about 9 in the year before follow-up started, compared with 3 for everyone else, so memory problems may have been spotted earlier. After matching people on age, sex, illnesses, Alzheimer's risk gene, and doctor use, the rate was still 2.3 times higher.
But early dementia can also look like ADHD, and the authors could not rule that out. When they required at least 8 years between diagnoses, the gap shrank to 1.4 times. At 9 years, it was 1.3 and could be chance.
People in the top fifth for inherited ADHD tendency, based on DNA, had a 35% higher rate than the bottom fifth.
The study could not test whether treatment changes the risk.
Bottom line: adults with ADHD had a little more than double the rate of these diagnoses, even after accounting for doctor visits, but the evidence is not settled.
Over about 10 years of health records, about 7 in 100 people with ADHD were diagnosed with mild cognitive impairment or dementia, versus about 3 uestion is doctor visits. People with ADHD had about 9 in the year before follow-up started, compared with 3 for everyone else, so memory problems may have been spotted earlier. After matching people on age, sex, illnesses, Alzheimer's risk gene, and doctor use, the rate was still 2.3 times higher.
But early dementia can also look like ADHD, and the authors could not rule that out. When they required at least 8 years between diagnoses, the gap shrank to 1.4 times. At 9 years, it was 1.3 and could be chance.
People in the top fifth for inherited ADHD tendency, based on DNA, had a 35% higher rate than the bottom fifth.
The study could not test whether treatment changes the risk.
Bottom line: adults with ADHD had a little more than double the rate of these diagnoses, even after accounting for doctor visits, but the evidence is not settled.
![A, Cumulative incidence of MCI or dementia by ADHD diagnosis (ADHD vs no ADHD). B, Cumulative incidence of MCI or dementia by ADHD PRS (low [quintile 1; Q1] vs high [quintile 5; Q5]). Hazard ratios (HRs) and P values from unadjusted Cox proportional hazards models.](/pic/orig/media%2FHUdn051WIAA3U_7.png)
3
1
9
Take blood from someone who is 101, and in a dish, it can become brain stem cells that are about 13% as old as the donor.
That is the average on an epigenetic clock, which estimates age from chemical tags on DNA, across donors up to 101.
Even better: up to six months later, the ng, showed no large DNA changes in the team's test, and could mature into neurons plus two kinds of brain support cell.
Researchers in Bonn and Aachen did it with just two genes, converting blood cells directly into neural stem cells without an embryo-like stage in between.
Cells grown longer averaged about 5% as old. Even cells from 87-year-old and 101-year-old donors showed none of the classic signs of old cells in the authors' tests.
The big takeaway is that a blood draw from someone very old could produce brain stem cells carrying that person's own DNA, but reading far younger than the donor.
That is the average on an epigenetic clock, which estimates age from chemical tags on DNA, across donors up to 101.
Even better: up to six months later, the ng, showed no large DNA changes in the team's test, and could mature into neurons plus two kinds of brain support cell.
Researchers in Bonn and Aachen did it with just two genes, converting blood cells directly into neural stem cells without an embryo-like stage in between.
Cells grown longer averaged about 5% as old. Even cells from 87-year-old and 101-year-old donors showed none of the classic signs of old cells in the authors' tests.
The big takeaway is that a blood draw from someone very old could produce brain stem cells carrying that person's own DNA, but reading far younger than the donor.
3
1
13
Half the children born with cystic fibrosis used to die before 14.
Today, half the babies diagnosed with it are expected to live to 66.
Cancer usually gets the headlines for medical progress like that. But five other diagnoses that once meant dying young have changed almost as progress happened within living memory.
1. Cystic fibrosis. In 1969, half the children born with it died before 14. By 1990, that figure had reached 28. Drugs that target the faulty protein behind the disease arrived in 2012 and 2019.
2. HIV. Without treatment, half of people died within about 10 years. A 20-year-old who started HIV pills in 2008 to 2010 and responded well is now projected to live to about 78.
3. Hepatitis C, a virus that slowly scars the liver. Treatment once meant 48 weeks of injections and pills, and it cleared the virus in 46% of people with the most common type. Now, eight weeks of pills clear it in 99% of those with that type and no advanced scarring.
4. Spinal muscular atrophy type 1, a gene disease that weakens a baby's muscles until breathing fails. Without treatment, 6 of 23 babies were alive without a permanent breathing machine at about 14 months. After a single gene-therapy infusion, 31 of 32 were.
5. Type 1 diabetes. Before insulin reached patients in 1922, most children diagnosed with it died. In Scotland in 2008 to 2010, a 20-year-old with it could expect 46 to 48 more years, still 11 to 13 short of people the same age without it.
6. Heart attack. In the 1960s, about 29 in 100 people admitted to hospital with one died there. In US hospitals in 2019 to 2020, about 5 in 100 did.
In 2024, 65% of American adults with cystic fibrosis had a job. If your child or parent gets one of these diagnoses, the odds you once heard may be obsolete. The future of biotechnology solving all of these deadly diseases looks bright.
Today, half the babies diagnosed with it are expected to live to 66.
Cancer usually gets the headlines for medical progress like that. But five other diagnoses that once meant dying young have changed almost as progress happened within living memory.
1. Cystic fibrosis. In 1969, half the children born with it died before 14. By 1990, that figure had reached 28. Drugs that target the faulty protein behind the disease arrived in 2012 and 2019.
2. HIV. Without treatment, half of people died within about 10 years. A 20-year-old who started HIV pills in 2008 to 2010 and responded well is now projected to live to about 78.
3. Hepatitis C, a virus that slowly scars the liver. Treatment once meant 48 weeks of injections and pills, and it cleared the virus in 46% of people with the most common type. Now, eight weeks of pills clear it in 99% of those with that type and no advanced scarring.
4. Spinal muscular atrophy type 1, a gene disease that weakens a baby's muscles until breathing fails. Without treatment, 6 of 23 babies were alive without a permanent breathing machine at about 14 months. After a single gene-therapy infusion, 31 of 32 were.
5. Type 1 diabetes. Before insulin reached patients in 1922, most children diagnosed with it died. In Scotland in 2008 to 2010, a 20-year-old with it could expect 46 to 48 more years, still 11 to 13 short of people the same age without it.
6. Heart attack. In the 1960s, about 29 in 100 people admitted to hospital with one died there. In US hospitals in 2019 to 2020, about 5 in 100 did.
In 2024, 65% of American adults with cystic fibrosis had a job. If your child or parent gets one of these diagnoses, the odds you once heard may be obsolete. The future of biotechnology solving all of these deadly diseases looks bright.
4
1
14
Should you chase more REM sleep to protect your brain? In 95,559 UK adults, 48 more minutes of REM per night was linked to a 46% lower dementia risk over nine years. But the link may run in reverse.
REM is the dreaming stage, when your eyes dart behind closed lids and your arms mingham study of home sleep tests pointed the same way in 2017.
Parkinson's can alter REM sleep a decade before diagnosis, and parkinsonism showed the strongest association, with an 80% lower risk. This study also did not measure REM in a sleep lab. Instead, a computer model estimated it from a week of wrist movement in @uk_biobank volunteers, and the Beijing authors call its match with lab studies modest.
The same data also showed that of the 41 strongest links between sleep duration and higher disease risk, 37 involved sleeping under five hours.
So a high REM number may reflect a healthy brain more than protect one.
REM is the dreaming stage, when your eyes dart behind closed lids and your arms mingham study of home sleep tests pointed the same way in 2017.
Parkinson's can alter REM sleep a decade before diagnosis, and parkinsonism showed the strongest association, with an 80% lower risk. This study also did not measure REM in a sleep lab. Instead, a computer model estimated it from a week of wrist movement in @uk_biobank volunteers, and the Beijing authors call its match with lab studies modest.
The same data also showed that of the 41 strongest links between sleep duration and higher disease risk, 37 involved sleeping under five hours.
So a high REM number may reflect a healthy brain more than protect one.
5
1
15
Your teeth can repair themselves, but that ability seems to fade after 60.
Each tooth has stem cells in its soft core, the pulp, that lay down new dentin when a cavity gets close. They thin out with age. In mice, they were easy to find at 3 months and barely detectable by 18, mouse. In people over 60, a deep cavity was more likely than in adults aged 18 to 40 to progress to irreversible pulpitis, an inflammation that usually means a root canal.
A team in China traced the decline to one protein, NFATC1. Without it, the stem cells stop dividing and turn senescent, meaning they stay alive but stop doing their job. In mice engineered to lack NFATC1, the drug pair dasatinib and quercetin, which clears senescent cells, brought dentin repair back to control levels within two weeks of injury.
If you are over 60, that is one more reason to get a cavity filled early. The cells that would wall it off may no longer be up to the job.
Each tooth has stem cells in its soft core, the pulp, that lay down new dentin when a cavity gets close. They thin out with age. In mice, they were easy to find at 3 months and barely detectable by 18, mouse. In people over 60, a deep cavity was more likely than in adults aged 18 to 40 to progress to irreversible pulpitis, an inflammation that usually means a root canal.
A team in China traced the decline to one protein, NFATC1. Without it, the stem cells stop dividing and turn senescent, meaning they stay alive but stop doing their job. In mice engineered to lack NFATC1, the drug pair dasatinib and quercetin, which clears senescent cells, brought dentin repair back to control levels within two weeks of injury.
If you are over 60, that is one more reason to get a cavity filled early. The cells that would wall it off may no longer be up to the job.
6
4
28
Why can a watch count your heartbeats but can't read your blood sugar? 25 years of trying says: it is brutally hard.
Light struggles to pick sugar out of water, fat, and skin.
Outside the skin
-> 2001. The FDA-approved GlucoWatch pulled sugar through the skin with a current. It irre by 2007.
-> 2014. Google's tear-reading contact lens was shelved by 2018.
-> 2023. Rockley, which aimed infrared at the wrist, went bankrupt.
-> 2024. The FDA said no watch or ring can read blood sugar on its own.
Under the skin
-> 1999. The first monitor put a thin wire under the skin. An enzyme on it, glucose oxidase, reacts with sugar in the fluid between cells, making a tiny current. More sugar, more current.
-> 2017. People could dose insulin from them, no finger prick.
-> 2024. Over-the-counter monitors arrived for people not on insulin.
-> 2026. Biolinq Shine's seven microneedles reach just 0.25 mm, and a light shows your range.
Before blood sugar rises, Insulin
No wearable reads insulin yet. Google's Pixel Watch estimates insulin resistance from watch data, weight, and age, catching 6 in 10 cases in testing.
People who later get diabetes have a third less insulin sensitivity 13 years before diagnosis, on average, and their blood sugar climbs sharply only in the last 3 years.
Glucose sensors were originally developed to dose insulin. The next goal is catching the slide toward diabetes years earlier.
Light struggles to pick sugar out of water, fat, and skin.
Outside the skin
-> 2001. The FDA-approved GlucoWatch pulled sugar through the skin with a current. It irre by 2007.
-> 2014. Google's tear-reading contact lens was shelved by 2018.
-> 2023. Rockley, which aimed infrared at the wrist, went bankrupt.
-> 2024. The FDA said no watch or ring can read blood sugar on its own.
Under the skin
-> 1999. The first monitor put a thin wire under the skin. An enzyme on it, glucose oxidase, reacts with sugar in the fluid between cells, making a tiny current. More sugar, more current.
-> 2017. People could dose insulin from them, no finger prick.
-> 2024. Over-the-counter monitors arrived for people not on insulin.
-> 2026. Biolinq Shine's seven microneedles reach just 0.25 mm, and a light shows your range.
Before blood sugar rises, Insulin
No wearable reads insulin yet. Google's Pixel Watch estimates insulin resistance from watch data, weight, and age, catching 6 in 10 cases in testing.
People who later get diabetes have a third less insulin sensitivity 13 years before diagnosis, on average, and their blood sugar climbs sharply only in the last 3 years.
Glucose sensors were originally developed to dose insulin. The next goal is catching the slide toward diabetes years earlier.
3
1
15
A new drug launched in the U.S. now has a median list price of $254,213 a year, up sharply from $38,857 for drugs launched between 2008 and 2013.
So what is driving the jump? Partly, the mix of drugs has changed. Cell and gene therapies launched at nearly 10 times the price of hich make up most pills. Even after adjusting for that shift, launch prices were rising about 8% a year before the Inflation Reduction Act.
The law does not appear to have changed that. The study found no clear sign it pushed launch prices up or slowed the climb. That makes sense: Medicare can negotiate prices only after drugs have been on the market for years, and it does not limit launch prices.
A second study points to another force: longer protection from generics. Small-molecule drugs approved in 2019 had an average of 6.9 listed patents, compared with 2.1 for those approved in 1990. Patents listed after approval added a median of 4.7 years of extra protection, though a quarter added less than three months.
These are list prices before insurer rebates. What patients actually pay depends on their plan.
For Medicare patients, out-of-pocket spending on covered Part D drugs is capped at $2,100 in 2026. If you have private insurance, ask what you would owe before starting a new drug, because the launch price is where bargaining starts.
So what is driving the jump? Partly, the mix of drugs has changed. Cell and gene therapies launched at nearly 10 times the price of hich make up most pills. Even after adjusting for that shift, launch prices were rising about 8% a year before the Inflation Reduction Act.
The law does not appear to have changed that. The study found no clear sign it pushed launch prices up or slowed the climb. That makes sense: Medicare can negotiate prices only after drugs have been on the market for years, and it does not limit launch prices.
A second study points to another force: longer protection from generics. Small-molecule drugs approved in 2019 had an average of 6.9 listed patents, compared with 2.1 for those approved in 1990. Patents listed after approval added a median of 4.7 years of extra protection, though a quarter added less than three months.
These are list prices before insurer rebates. What patients actually pay depends on their plan.
For Medicare patients, out-of-pocket spending on covered Part D drugs is capped at $2,100 in 2026. If you have private insurance, ask what you would owe before starting a new drug, because the launch price is where bargaining starts.
1
1
10
Your organs do not all age at the same speed, and now a blood test can tell them apart.
@wysscoray's Stanford lab built protein clocks for each organ and tested them in 44,498 UK Biobank volunteers followed for up to 17 years. The signal was striking: people whose brains read imer’s rate, versus 0.35% for those whose brains read youngest. People with 8 or more old-reading organs had about 8 times the risk of dying.
Eric Topol calls organ clocks "a breakthrough in their own right." The big question now is whether lowering an organ's age can change longevity outcomes.
@wysscoray's Stanford lab built protein clocks for each organ and tested them in 44,498 UK Biobank volunteers followed for up to 17 years. The signal was striking: people whose brains read imer’s rate, versus 0.35% for those whose brains read youngest. People with 8 or more old-reading organs had about 8 times the risk of dying.
Eric Topol calls organ clocks "a breakthrough in their own right." The big question now is whether lowering an organ's age can change longevity outcomes.
Here's our review of molecular aging clocks. Organ clocks are a major advance
@NatureMedicine with @wysscoray
nature.com/articles/s4159…
6
4
21
Do statins protect your brain? Depends who you ask.
Yesterday, a UK Biobank study found 28% less Alzheimer's in statin users. Plenty of studies echo that result. But another team using the same UK Biobank found 19% more.
Then there's the trial. In August, 9,971 Australians over or a placebo for six years. Major heart events fell 30%. Dementia didn't change.
My take is that the records mostly capture people who already look after themselves. Statins have a clear place for your heart, and that part is proven.
Yesterday, a UK Biobank study found 28% less Alzheimer's in statin users. Plenty of studies echo that result. But another team using the same UK Biobank found 19% more.
Then there's the trial. In August, 9,971 Australians over or a placebo for six years. Major heart events fell 30%. Dementia didn't change.
My take is that the records mostly capture people who already look after themselves. Statins have a clear place for your heart, and that part is proven.
12
1
25
Most people think a hoarse voice means a throat problem. A study of more than 700,000 Americans suggests it can also be a brain-health clue.
Drexel researchers found that people 50 and older with a voice disorder and no hearing loss were 58% more likely to later receive a nosis than people seen for a routine checkup.
You might assume a weak voice cuts people off and that isolation wears on the brain. But the brain may change first, and the voice follows. In 20 people followed for years before they developed Parkinson's or a related disease, doctors rated their speech as quieter and flatter, and their faces as less expressive, from an estimated 10 years before diagnosis.
No trial has shown that voice therapy protects memory.
If your voice has not cleared or improved in four weeks, have your vocal cords checked and mention any memory slips while you are there.
Drexel researchers found that people 50 and older with a voice disorder and no hearing loss were 58% more likely to later receive a nosis than people seen for a routine checkup.
You might assume a weak voice cuts people off and that isolation wears on the brain. But the brain may change first, and the voice follows. In 20 people followed for years before they developed Parkinson's or a related disease, doctors rated their speech as quieter and flatter, and their faces as less expressive, from an estimated 10 years before diagnosis.
No trial has shown that voice therapy protects memory.
If your voice has not cleared or improved in four weeks, have your vocal cords checked and mention any memory slips while you are there.
1
0
6
If one of your parents made it to 100, people probably tell you you've got great genes. For your heart, that may be true. For cancer, not so much.
In UK Biobank, children of centenarians were less likely to have high blood pressure (36% vs. 47%) or heart disease (10% vs. 16%). groups.
The gap held after adjusting for smoking, exercise, education, and neighborhood deprivation. Two US cohorts also found lower blood pressure. The study did not test genes.
A 2019 study of a million parents' lifespans points the same way. The gene variants most tied to lifespan are linked to heart disease, dementia, and lung cancer from smoking, not most other cancers.
Bottom line: having long-lived parents may be a clue for better heart health. They are not a free pass to skip your colonoscopy or mammogram.
In UK Biobank, children of centenarians were less likely to have high blood pressure (36% vs. 47%) or heart disease (10% vs. 16%). groups.
The gap held after adjusting for smoking, exercise, education, and neighborhood deprivation. Two US cohorts also found lower blood pressure. The study did not test genes.
A 2019 study of a million parents' lifespans points the same way. The gene variants most tied to lifespan are linked to heart disease, dementia, and lung cancer from smoking, not most other cancers.
Bottom line: having long-lived parents may be a clue for better heart health. They are not a free pass to skip your colonoscopy or mammogram.
3
0
6
One correction, because it matters for the women reading this: the site does show women wearing Temple, like this one.
Nothing is hidden either. The headset study has "in healthy young men” right in its title on our science page: temple.com/science-and-re…
And a good share of the people builseveral leading their own areas, some of whom Deepinder has featured in his own posts.
Fair point on making who each study tested impossible to miss, and that goes to the team. You're right that a woman should be measured against her own pattern, and we want to go much further than the cycle. Women are more than their reproductive years. Brain aging, heart and blood vessel aging, metabolism: all of it plays out differently in women, through the cycle, menopause and the decades after. That whole lifetime is what we're building Temple to follow.
Keep pushing.
Nothing is hidden either. The headset study has "in healthy young men” right in its title on our science page: temple.com/science-and-re…
And a good share of the people builseveral leading their own areas, some of whom Deepinder has featured in his own posts.
Fair point on making who each study tested impossible to miss, and that goes to the team. You're right that a woman should be measured against her own pattern, and we want to go much further than the cycle. Women are more than their reproductive years. Brain aging, heart and blood vessel aging, metabolism: all of it plays out differently in women, through the cycle, menopause and the decades after. That whole lifetime is what we're building Temple to follow.
Keep pushing.
0
0
0
What happened at EASD in Milan last week? A lot. The conference centered on how newer diabetes treatments are moving to the front of care, while metformin still held its own in a head-to-head test.
New US and European guidance says people with type 2 diabetes can start on an pill like Farxiga at diagnosis, and for people with heart disease, kidney disease or heart failure, the two may be combined early. For most of the past 20 years, metformin has been the default first pill.
Then came a Swedish trial of Farxiga as the first drug in 2,072 people with newly diagnosed type 2 diabetes who did not have major heart or kidney disease. After 3.8 years, it did not beat metformin on deaths, heart events, or eye, kidney and foot damage.
Other takeaways from EASD 2026 in Milan:
1. Orforglipron, Lilly's daily GLP-1 pill, matched insulin on heart safety in 2,749 people with type 2 diabetes and heart or kidney disease, with fewer dangerous lows.
2. Retatrutide and survodutide, two weekly obesity shots, produced major weight loss in trials.
3. A Nordic registry study found semaglutide was not linked to more pancreatic cancer, but that was not proof.
4. In Mauritius, diabetes fell from 23.3% of adults in 2009 to 19.9% in 2021, according to national surveys.
5. Type 1 screening found early disease in about 2 in 100 children with an affected parent or sibling, and 3 in 1,000 without one.
6. Diamyd did not delay type 1 overall, but may have helped a subgroup with one immune gene type.
Key takeaway: if you have type 2 diabetes plus heart, kidney, or heart failure risk, ask now whether a GLP-1 drug or Farxiga fits your plan. If not, metformin is still very much in the game. And for families worried about type 1, antibody screening is becoming more relevant.
New US and European guidance says people with type 2 diabetes can start on an pill like Farxiga at diagnosis, and for people with heart disease, kidney disease or heart failure, the two may be combined early. For most of the past 20 years, metformin has been the default first pill.
Then came a Swedish trial of Farxiga as the first drug in 2,072 people with newly diagnosed type 2 diabetes who did not have major heart or kidney disease. After 3.8 years, it did not beat metformin on deaths, heart events, or eye, kidney and foot damage.
Other takeaways from EASD 2026 in Milan:
1. Orforglipron, Lilly's daily GLP-1 pill, matched insulin on heart safety in 2,749 people with type 2 diabetes and heart or kidney disease, with fewer dangerous lows.
2. Retatrutide and survodutide, two weekly obesity shots, produced major weight loss in trials.
3. A Nordic registry study found semaglutide was not linked to more pancreatic cancer, but that was not proof.
4. In Mauritius, diabetes fell from 23.3% of adults in 2009 to 19.9% in 2021, according to national surveys.
5. Type 1 screening found early disease in about 2 in 100 children with an affected parent or sibling, and 3 in 1,000 without one.
6. Diamyd did not delay type 1 overall, but may have helped a subgroup with one immune gene type.
Key takeaway: if you have type 2 diabetes plus heart, kidney, or heart failure risk, ask now whether a GLP-1 drug or Farxiga fits your plan. If not, metformin is still very much in the game. And for families worried about type 1, antibody screening is becoming more relevant.
1
1
8
Zombie cells are longevity science's favorite target. They're worn-out cells that stop dividing, refuse to die, and leak inflammation into nearby tissue.
A @UConn team combined 38 blood proteins linked to these cells into a single score. In @uk_biobank adults followed for about dard-deviation increase in the score was associated with a 41% higher risk of death.
In MEDEX trial, an 18-month exercise trial in older adults, the score increased less in people assigned to exercise. The difference between groups was not large enough to rule out chance.
Where these proteins originate in the body is still unknown.
A @UConn team combined 38 blood proteins linked to these cells into a single score. In @uk_biobank adults followed for about dard-deviation increase in the score was associated with a 41% higher risk of death.
In MEDEX trial, an 18-month exercise trial in older adults, the score increased less in people assigned to exercise. The difference between groups was not large enough to rule out chance.
Where these proteins originate in the body is still unknown.
2
0
12
A daily pill from @MonteRosaTx cut CRP, the standard blood test for inflammation, by 85% in four weeks in the company's trial of 108 adults at increased risk of heart disease. Other markers dropped too:
-> CRP: 85%
-> IL-6, an inflammation signal: 54%
-> Fibrinogen, a clotting protein: ted fat particle: 24%
Why it matters: statins lower heart risk, but inflammation is a separate risk that can stay high even on them. In large trials of people already on cholesterol treatment, those with the highest CRP had 2.7x the heart-death rate of those with the lowest.
MRT-8102 works by breaking down NEK7, a protein that helps switch on one of the body's main inflammation alarms.
Two injected anti-inflammatory drugs tried this first. @Novartis's canakinumab cut heart events 15% in 2017 but raised fatal infections, and @novonordisk's ziltivekimab lowered these markers but failed to cut heart attacks this July.
MRT-8102 has no heart attack results yet.
-> CRP: 85%
-> IL-6, an inflammation signal: 54%
-> Fibrinogen, a clotting protein: ted fat particle: 24%
Why it matters: statins lower heart risk, but inflammation is a separate risk that can stay high even on them. In large trials of people already on cholesterol treatment, those with the highest CRP had 2.7x the heart-death rate of those with the lowest.
MRT-8102 works by breaking down NEK7, a protein that helps switch on one of the body's main inflammation alarms.
Two injected anti-inflammatory drugs tried this first. @Novartis's canakinumab cut heart events 15% in 2017 but raised fatal infections, and @novonordisk's ziltivekimab lowered these markers but failed to cut heart attacks this July.
MRT-8102 has no heart attack results yet.
5
8
34
Up to 1 in 4 people carry APOE4, the biggest common genetic risk for Alzheimer's. One copy roughly triples the odds. After a month on an old pill, carriers had more blood flow in parts of the brain than before.
The pill is rapamycin, a transplant drug first found in Easter n's team at @Mizzou gave healthy adults aged 45 to 65 a daily 1 mg dose, with no placebo group.
In the 9 carriers, blood flow rose more than 15% in several brain regions, while in the 14 people without APOE4 it didn't clearly change. An earlier look at the data found the biggest rise in the 4 women carriers, about 35% across the brain, too few to be sure.
Almost two-thirds of Americans with Alzheimer's are women. Lin's next trial, due in December, will give rapamycin or a placebo to 225 women. Whether more blood flow means less Alzheimer's is still unknown.
The pill is rapamycin, a transplant drug first found in Easter n's team at @Mizzou gave healthy adults aged 45 to 65 a daily 1 mg dose, with no placebo group.
In the 9 carriers, blood flow rose more than 15% in several brain regions, while in the 14 people without APOE4 it didn't clearly change. An earlier look at the data found the biggest rise in the 4 women carriers, about 35% across the brain, too few to be sure.
Almost two-thirds of Americans with Alzheimer's are women. Lin's next trial, due in December, will give rapamycin or a placebo to 225 women. Whether more blood flow means less Alzheimer's is still unknown.
4
1
27
You lost more than half your deep sleep before you turned 25.
A 5-year-old gets about three hours a night. A 25-year-old gets about 74 minutes. An 85-year-old gets about 22 minutes, with nearly two hours lying awake in bed. @WilliamWallace made this video from 65 sleep studies.
eople sell for deep sleep?
Sleep labs have tested most of them, usually in people around 60:
-> Magnesium: about 6 extra minutes, in a 2002 study of 12 people.
-> Glycine and melatonin: no added minutes.
-> Timed pink noise: bigger deep-sleep waves, but no extra time.
-> Blue-blocking glasses: one lab test in 13 teenage boys found no change.
-> Red light: I found no lab tests.
-> Best evidence: CBT for insomnia, a talking therapy, raised deep sleep in 46 adults averaging 61 and beat a sleeping pill.
But age may not explain the whole curve.
Studies of people over 60 alone did not find a clear drop in deep sleep. And the relationship goes both ways: one night of disrupted deep sleep raised amyloid, the Alzheimer's protein, in healthy adults, while in older adults a shrinking brain goes with weaker deep sleep.
Bottom line: if someone says they can give you those minutes back, ask for one number.
How many minutes of deep sleep did it add in a sleep lab in people your age?
A 5-year-old gets about three hours a night. A 25-year-old gets about 74 minutes. An 85-year-old gets about 22 minutes, with nearly two hours lying awake in bed. @WilliamWallace made this video from 65 sleep studies.
eople sell for deep sleep?
Sleep labs have tested most of them, usually in people around 60:
-> Magnesium: about 6 extra minutes, in a 2002 study of 12 people.
-> Glycine and melatonin: no added minutes.
-> Timed pink noise: bigger deep-sleep waves, but no extra time.
-> Blue-blocking glasses: one lab test in 13 teenage boys found no change.
-> Red light: I found no lab tests.
-> Best evidence: CBT for insomnia, a talking therapy, raised deep sleep in 46 adults averaging 61 and beat a sleeping pill.
But age may not explain the whole curve.
Studies of people over 60 alone did not find a clear drop in deep sleep. And the relationship goes both ways: one night of disrupted deep sleep raised amyloid, the Alzheimer's protein, in healthy adults, while in older adults a shrinking brain goes with weaker deep sleep.
Bottom line: if someone says they can give you those minutes back, ask for one number.
How many minutes of deep sleep did it add in a sleep lab in people your age?
14
23
148
A female Greenland shark is not old enough to breed until she is about 150. One lived at least 272 years.
A 2026 review lists 101 species that can live 250 years or more. Only 11 are animals. Ninety are plants, and not one is a mammal. The human record is 122.
So how do you ng-lived species tend to carry extra DNA-repair genes, and some keep making new cells for life.
A hydra, a tiny freshwater animal, replaces its body cells within weeks and, in the lab, gets no likelier to die as it ages. Across 61 plant species, trees carry more copies of PARP, a family of DNA-repair genes, than herbs do.
Repair and renewal go together. Every cell division can miscopy DNA, and a mistake in the wrong gene can start a cancer. Our hearts renew slowly, replacing about 1% of their muscle cells a year at 25 and 0.45% at 75.
One line of human cells already holds its DNA to a higher standard. The cells that make sperm pick up roughly 70 to 80 times fewer mutations per division than skin cells grown in a dish.
The biologist Tom Kirkwood proposed why in 1977. Natural selection keeps DNA copying accurate in the cells that get passed on, while the rest of the body only has to keep working for about as long as the animal would survive in the wild.
The Greenland shark’s bargain is stark: repair the DNA, preserve the body, and wait 150 years to reproduce.
A 2026 review lists 101 species that can live 250 years or more. Only 11 are animals. Ninety are plants, and not one is a mammal. The human record is 122.
So how do you ng-lived species tend to carry extra DNA-repair genes, and some keep making new cells for life.
A hydra, a tiny freshwater animal, replaces its body cells within weeks and, in the lab, gets no likelier to die as it ages. Across 61 plant species, trees carry more copies of PARP, a family of DNA-repair genes, than herbs do.
Repair and renewal go together. Every cell division can miscopy DNA, and a mistake in the wrong gene can start a cancer. Our hearts renew slowly, replacing about 1% of their muscle cells a year at 25 and 0.45% at 75.
One line of human cells already holds its DNA to a higher standard. The cells that make sperm pick up roughly 70 to 80 times fewer mutations per division than skin cells grown in a dish.
The biologist Tom Kirkwood proposed why in 1977. Natural selection keeps DNA copying accurate in the cells that get passed on, while the rest of the body only has to keep working for about as long as the animal would survive in the wild.
The Greenland shark’s bargain is stark: repair the DNA, preserve the body, and wait 150 years to reproduce.
4
7
35
A Stanford lab built a drug company run by AI agents.
An AI chief scientist gets a question, breaks it into parts for genetics, safety, chemistry, and trial-design agents, then stitches the answers together. Humans set the questions and check the work.
Why start there? Because r human trials fail. So the lab trained its company on the history of failure. More than 37,000 agents read 55,984 completed drug trials and logged what happened. It took less than a week.
One finding stood out: drugs aimed at genes used by only a few cell types were 48% more likely to reach market and caused fewer side effects. The likely reason is simple. If a protein is active in fewer places, the drug has fewer places to do harm. The 48% comes from completed trials, and no drug has yet been selected under this rule and tested in people.
Then the lab asked its company for a lung cancer drug using only data available before 2025.
The agents pointed to an antibody-drug combo that targets B7-H3. But industry had already moved. Daiichi Sankyo built that exact kind of drug and first gave it to patients in 2019, more than five years before the cutoff. With Merck, it is now awaiting an FDA decision by 10 October for people whose small cell lung cancer has returned after chemotherapy.
The takeaway is not that AI found a new drug first. It is that AI can already sift the past fast enough to surface the same bets pharma eventually makes, and maybe help find the next ones sooner.
An AI chief scientist gets a question, breaks it into parts for genetics, safety, chemistry, and trial-design agents, then stitches the answers together. Humans set the questions and check the work.
Why start there? Because r human trials fail. So the lab trained its company on the history of failure. More than 37,000 agents read 55,984 completed drug trials and logged what happened. It took less than a week.
One finding stood out: drugs aimed at genes used by only a few cell types were 48% more likely to reach market and caused fewer side effects. The likely reason is simple. If a protein is active in fewer places, the drug has fewer places to do harm. The 48% comes from completed trials, and no drug has yet been selected under this rule and tested in people.
Then the lab asked its company for a lung cancer drug using only data available before 2025.
The agents pointed to an antibody-drug combo that targets B7-H3. But industry had already moved. Daiichi Sankyo built that exact kind of drug and first gave it to patients in 2019, more than five years before the cutoff. With Merck, it is now awaiting an FDA decision by 10 October for people whose small cell lung cancer has returned after chemotherapy.
The takeaway is not that AI found a new drug first. It is that AI can already sift the past fast enough to surface the same bets pharma eventually makes, and maybe help find the next ones sooner.
4
0
11
There's still no proven test for how fast your body is aging.
Scientists call that a biomarker of aging. The drawing below, from a 2017 Karolinska Institutet review (@karolinskainst), shows three people the same age with different biological ages. A good test would tell them t any test is only useful up to about midlife. By old age, people differ so much that the same score tells you less.
Any test has to beat plain age. In 1825, the actuary Benjamin Gompertz showed that an adult's risk of dying rises steadily with age, and in the UK today it roughly doubles every seven to eight years. The first DNA clocks were trained to guess your age from chemical tags on DNA in blood. In a 2019 analysis of older Scots, clocks that got better at guessing age got worse at predicting who died.
The Dunedin Study in New Zealand measured speed instead. It tracked 19 signs of wear in the heart, lungs, kidneys, gums, and immune system in about 1,000 people at ages 26, 32, 38, and 45. Scaled so the average person aged one biological year per calendar year, the slowest aged 0.4 and the fastest 2.4.
In 2022, Dan Belsky (@danbelsky) and the Dunedin team found the DNA tags in blood that best matched each person's pace, so one blood sample now estimates it. That test is DunedinPACE, and 1.0 is the group's average pace. Higher readings predicted who died sooner in two US studies, and a two-year calorie-restriction trial slowed the pace it reports by 2% to 3%. But it was built from one group of New Zealanders, and no trial has yet shown that slowing it cuts illness or death.
Watches, rings, and apps now claim to show your age too. I couldn't find one whose maker has published a test showing that number predicts illness or death better than age.
Proving any age number works means measuring people now and waiting years to see who gets ill. That is why life insurers, who bet money on when you'll die, don't price on these age scores. They ask your age, whether you smoke or drink, your height and weight, and what illnesses you and your family have had.
Scientists call that a biomarker of aging. The drawing below, from a 2017 Karolinska Institutet review (@karolinskainst), shows three people the same age with different biological ages. A good test would tell them t any test is only useful up to about midlife. By old age, people differ so much that the same score tells you less.
Any test has to beat plain age. In 1825, the actuary Benjamin Gompertz showed that an adult's risk of dying rises steadily with age, and in the UK today it roughly doubles every seven to eight years. The first DNA clocks were trained to guess your age from chemical tags on DNA in blood. In a 2019 analysis of older Scots, clocks that got better at guessing age got worse at predicting who died.
The Dunedin Study in New Zealand measured speed instead. It tracked 19 signs of wear in the heart, lungs, kidneys, gums, and immune system in about 1,000 people at ages 26, 32, 38, and 45. Scaled so the average person aged one biological year per calendar year, the slowest aged 0.4 and the fastest 2.4.
In 2022, Dan Belsky (@danbelsky) and the Dunedin team found the DNA tags in blood that best matched each person's pace, so one blood sample now estimates it. That test is DunedinPACE, and 1.0 is the group's average pace. Higher readings predicted who died sooner in two US studies, and a two-year calorie-restriction trial slowed the pace it reports by 2% to 3%. But it was built from one group of New Zealanders, and no trial has yet shown that slowing it cuts illness or death.
Watches, rings, and apps now claim to show your age too. I couldn't find one whose maker has published a test showing that number predicts illness or death better than age.
Proving any age number works means measuring people now and waiting years to see who gets ill. That is why life insurers, who bet money on when you'll die, don't price on these age scores. They ask your age, whether you smoke or drink, your height and weight, and what illnesses you and your family have had.
3
1
13
Knees are built to last, but cartilage barely gets replaced. That is why osteoarthritis has become a global problem: 595 million people live with it, and for decades, the options were mostly painkillers, steroid shots, or a new knee. Now, 5 drugs are in the running. Here is who Semaglutide (Wegovy, Novo Nordisk). In 407 adults with obesity and knee arthritis, average age 56, knee pain fell 42 points on a 0 to 100 pain score against 28 on placebo over 68 weeks, while they lost 14% of their weight. Not licensed for arthritis yet.
2. Lorecivivint (Biosplice). One shot into the knee that dampens Wnt, a signal that drives cartilage breakdown. Tested in moderate arthritis. One Phase 3 missed its pain goal, and a two-year follow-up found pain 5 points lower than placebo on a 100-point score. Biosplice filed for FDA approval in January.
3. LEVI-04 (Levicept). A monthly IV drip that mops up neurotrophin-3, a protein that turns up pain nerves. In 518 people with painful knee arthritis, with an average age of 64, pain eased by 0.8 points more than placebo on a 10-point scale after 4 months. It caused no extra joint damage, unlike earlier drugs aimed at pain nerves.
4. PCRX-201 (Pacira). A one-time gene therapy that makes the knee produce its own anti-inflammatory protein. In 72 people with moderate to severe arthritis, relief held for 3 years, with no placebo group yet. Phase 2 is underway.
5. Sprifermin. In 549 people aged 40 to 85 with moderate arthritis, it thickened cartilage by 0.05 mm more than placebo at 2 years, about half the thickness of a sheet of paper, and the gain was still there at 5 years. No trial is active now.
Only sprifermin clearly thickened cartilage, and its pain relief did not exceed placebo. Among the placebo-controlled trials, weight loss delivered the biggest pain drop.
2. Lorecivivint (Biosplice). One shot into the knee that dampens Wnt, a signal that drives cartilage breakdown. Tested in moderate arthritis. One Phase 3 missed its pain goal, and a two-year follow-up found pain 5 points lower than placebo on a 100-point score. Biosplice filed for FDA approval in January.
3. LEVI-04 (Levicept). A monthly IV drip that mops up neurotrophin-3, a protein that turns up pain nerves. In 518 people with painful knee arthritis, with an average age of 64, pain eased by 0.8 points more than placebo on a 10-point scale after 4 months. It caused no extra joint damage, unlike earlier drugs aimed at pain nerves.
4. PCRX-201 (Pacira). A one-time gene therapy that makes the knee produce its own anti-inflammatory protein. In 72 people with moderate to severe arthritis, relief held for 3 years, with no placebo group yet. Phase 2 is underway.
5. Sprifermin. In 549 people aged 40 to 85 with moderate arthritis, it thickened cartilage by 0.05 mm more than placebo at 2 years, about half the thickness of a sheet of paper, and the gain was still there at 5 years. No trial is active now.
Only sprifermin clearly thickened cartilage, and its pain relief did not exceed placebo. Among the placebo-controlled trials, weight loss delivered the biggest pain drop.
4
4
43
A blood test now spots Alzheimer's 91% of the time in people with memory problems. Family doctors get it right only 61% of the time.
Here are 8 tests and drugs changing Alzheimer's care and who each one is for.
TESTS THAT FIND IT EARLIER
1. p-tau217 blood tests. p-tau217 is a lzheimer's builds. The 91% figure came from 1,213 Swedish patients with memory complaints. FDA has cleared four such tests since May 2025, but none for people without symptoms.
2. The same test as a clock. In people with no symptoms, it predicted when memory problems would start within about 3 to 4 years. Research use only for now.
TREATMENTS ONCE MEMORY PROBLEMS START
For early Alzheimer's, mild memory loss or mild dementia, with confirmed plaque, the sticky protein clumps:
3. Lecanemab (Leqembi, Eisai/Biogen). Tested in people aged 50 to 90, average 71. Slowed decline 27% over 18 months. Since July, it can be self-injected at home from the first dose.
4. Donanemab (Kisunla, Lilly). Tested in people aged 60 to 85, average 73. Slowed decline 22%, and 35% in an earlier-stage group.
5. Trontinemab (Roche). Built to cross the brain's protective barrier more easily. In an early trial, 91% of patients had plaque below the level a scan counts as positive by 28 weeks. Phase 3 reports in 2028.
PREVENTION BEFORE SYMPTOMS
6. The same antibodies, given earlier. Lilly and Roche are testing them in people aged 55 to 80 with plaque but little or no memory loss. A separate trial doses people aged 18 who carry a rare gene that causes Alzheimer's in midlife.
7. Obicetrapib (Ubeslo, NewAmsterdam/Menarini), a cholesterol pill, was approved in Europe on Monday. In a heart trial, 29 people with two copies of ApoE4, the strongest common risk gene, saw p-tau217 fall 8% on the pill and rise 13% on placebo over a year. A 400-person trial in ages 55 to 85 starts now.
8. Shingles vaccine. Wales offered it by birth date. People born just after the cutoff, then about 80, had 20% fewer dementia diagnoses over 7 years than those born just before. Two randomized trials are now recruiting about 196,000 people, aged 76 and up in Finland and 65 and up in Denmark.
Bottom line: if memory problems have started, ask whether a p-tau217 blood test fits you. If they have not and Alzheimer’s runs in your family, ask a memory clinic about the prevention trials above.
Here are 8 tests and drugs changing Alzheimer's care and who each one is for.
TESTS THAT FIND IT EARLIER
1. p-tau217 blood tests. p-tau217 is a lzheimer's builds. The 91% figure came from 1,213 Swedish patients with memory complaints. FDA has cleared four such tests since May 2025, but none for people without symptoms.
2. The same test as a clock. In people with no symptoms, it predicted when memory problems would start within about 3 to 4 years. Research use only for now.
TREATMENTS ONCE MEMORY PROBLEMS START
For early Alzheimer's, mild memory loss or mild dementia, with confirmed plaque, the sticky protein clumps:
3. Lecanemab (Leqembi, Eisai/Biogen). Tested in people aged 50 to 90, average 71. Slowed decline 27% over 18 months. Since July, it can be self-injected at home from the first dose.
4. Donanemab (Kisunla, Lilly). Tested in people aged 60 to 85, average 73. Slowed decline 22%, and 35% in an earlier-stage group.
5. Trontinemab (Roche). Built to cross the brain's protective barrier more easily. In an early trial, 91% of patients had plaque below the level a scan counts as positive by 28 weeks. Phase 3 reports in 2028.
PREVENTION BEFORE SYMPTOMS
6. The same antibodies, given earlier. Lilly and Roche are testing them in people aged 55 to 80 with plaque but little or no memory loss. A separate trial doses people aged 18 who carry a rare gene that causes Alzheimer's in midlife.
7. Obicetrapib (Ubeslo, NewAmsterdam/Menarini), a cholesterol pill, was approved in Europe on Monday. In a heart trial, 29 people with two copies of ApoE4, the strongest common risk gene, saw p-tau217 fall 8% on the pill and rise 13% on placebo over a year. A 400-person trial in ages 55 to 85 starts now.
8. Shingles vaccine. Wales offered it by birth date. People born just after the cutoff, then about 80, had 20% fewer dementia diagnoses over 7 years than those born just before. Two randomized trials are now recruiting about 196,000 people, aged 76 and up in Finland and 65 and up in Denmark.
Bottom line: if memory problems have started, ask whether a p-tau217 blood test fits you. If they have not and Alzheimer’s runs in your family, ask a memory clinic about the prevention trials above.
11
36
146
Yes, sadly, Opus 5 is the only usable model for bio/med from Anthropic, since we cannot even use the latest Opus 5.5 model.
Opus 5.5 has the same biology safeguards as Fable 5.1; when they trip, questions are handed to the older Opus 5. The only way past them is for a vetted organization to apply tification Program, so a doctor, student, or independent scientist asking an ordinary biology question is treated as a risk by default.
anthropic.com/claude-opus-5-…
Opus 5.5 has the same biology safeguards as Fable 5.1; when they trip, questions are handed to the older Opus 5. The only way past them is for a vetted organization to apply tification Program, so a doctor, student, or independent scientist asking an ordinary biology question is treated as a risk by default.
anthropic.com/claude-opus-5-…
1
0
1
The plaque most likely to kill you can start shrinking in weeks.
Soft plaque is the dangerous kind: a fatty core under a thin cap. If that cap tears, a clot can block the artery. A calcium scan will miss it. A CT angiogram can see it.
That matters because treatment can work ot of evolocumab (Repatha) reduced soft plaque by 13% in about five weeks. A Dutch trial saw a similar early drop with a strong statin plus a dummy shot, suggesting the LDL drop was doing the work.
The benefit keeps building.
After a year, the injection thickened the cap twice as much as a dummy shot. And over five years in high-risk people, it cut heart attacks, strokes, and heart-death risk from 8.0% to 6.2%.
Inclisiran (Leqvio) also showed promise: in Tokyo, the twice-yearly shot was linked to 17% less soft plaque over 15 months, though that study had no placebo group.
The main takeaway is that you should ask your doctor for your LDL target and hit it early. Soft plaque can start shrinking within weeks and continue to improve for years.
Soft plaque is the dangerous kind: a fatty core under a thin cap. If that cap tears, a clot can block the artery. A calcium scan will miss it. A CT angiogram can see it.
That matters because treatment can work ot of evolocumab (Repatha) reduced soft plaque by 13% in about five weeks. A Dutch trial saw a similar early drop with a strong statin plus a dummy shot, suggesting the LDL drop was doing the work.
The benefit keeps building.
After a year, the injection thickened the cap twice as much as a dummy shot. And over five years in high-risk people, it cut heart attacks, strokes, and heart-death risk from 8.0% to 6.2%.
Inclisiran (Leqvio) also showed promise: in Tokyo, the twice-yearly shot was linked to 17% less soft plaque over 15 months, though that study had no placebo group.
The main takeaway is that you should ask your doctor for your LDL target and hit it early. Soft plaque can start shrinking within weeks and continue to improve for years.
10
17
132
Metformin or Jardiance: which is likelier to help you live longer and healthier?
Most longevity folks would pick metformin. But the evidence is not as strong as the hype. In a 21-year US study of 3,234 people with prediabetes, those first given metformin had no fewer heart art deaths than those first given a placebo.
Jardiance and Farxiga are SGLT2 inhibitors: pills that help the kidneys flush extra sugar into the urine. In 13 large trials of people with diabetes, heart failure, or kidney disease, this class cut heart-failure hospital stays and slowed kidney damage. Deaths fell in some trials. The strongest example: Jardiance in 7,020 people with diabetes and heart disease, most already on metformin, reduced deaths from 8.3% to 5.7% over three years.
But there is a catch: these were already sick people. None of these drugs has been shown to slow aging in healthy people.
From 2027, an @ARPA_H-funded trial will test Farxiga, Ozempic, rapamycin, or a placebo in 726 generally healthy 60- to 65-year-olds, then measure how well their bodies and minds keep working.
Metformin is not one of the four. Its own aging trial, TAME, is still trying to raise the money to start.
Bottom line: metformin has the bigger longevity brand, but SGLT2 inhibitors have the stronger real-world track record so far. The science still does not prove any of these drugs slows ageing in healthy people.
Most longevity folks would pick metformin. But the evidence is not as strong as the hype. In a 21-year US study of 3,234 people with prediabetes, those first given metformin had no fewer heart art deaths than those first given a placebo.
Jardiance and Farxiga are SGLT2 inhibitors: pills that help the kidneys flush extra sugar into the urine. In 13 large trials of people with diabetes, heart failure, or kidney disease, this class cut heart-failure hospital stays and slowed kidney damage. Deaths fell in some trials. The strongest example: Jardiance in 7,020 people with diabetes and heart disease, most already on metformin, reduced deaths from 8.3% to 5.7% over three years.
But there is a catch: these were already sick people. None of these drugs has been shown to slow aging in healthy people.
From 2027, an @ARPA_H-funded trial will test Farxiga, Ozempic, rapamycin, or a placebo in 726 generally healthy 60- to 65-year-olds, then measure how well their bodies and minds keep working.
Metformin is not one of the four. Its own aging trial, TAME, is still trying to raise the money to start.
Bottom line: metformin has the bigger longevity brand, but SGLT2 inhibitors have the stronger real-world track record so far. The science still does not prove any of these drugs slows ageing in healthy people.
22
52
232
By 75, your kidneys may still look nearly normal on a scan, even though they have lost nearly half their filtering units.
Researchers at @MayoClinic and @ClevelandClinic studied 1,638 living kidney donors, all of them screened as healthy enough to give one away. From CT scans y counted the filtering units, called glomeruli, in each kidney.
Donors aged 18 to 29 had about 990,000. Donors aged 70 to 75 had about 520,000, which is 48% fewer, while the kidney's outer layer had shrunk by only 16%.
Adults cannot grow new filters. What you have left is what you finish with, which is why doctors cut so many drug doses as people age and why the same dose of a common painkiller lands differently at 75 than at 35.
Two drug classes now slow that decline once kidney disease has started. SGLT2 inhibitors, first sold for diabetes, cut the combined risk of major kidney decline, kidney failure, or death from kidney or heart causes by 39% in DAPA-CKD, in people with and without diabetes. Semaglutide cut major kidney events by 24% in FLOW.
These were different people at different ages, each measured once, so this is a snapshot of a population rather than a life history of a single kidney.
Researchers at @MayoClinic and @ClevelandClinic studied 1,638 living kidney donors, all of them screened as healthy enough to give one away. From CT scans y counted the filtering units, called glomeruli, in each kidney.
Donors aged 18 to 29 had about 990,000. Donors aged 70 to 75 had about 520,000, which is 48% fewer, while the kidney's outer layer had shrunk by only 16%.
Adults cannot grow new filters. What you have left is what you finish with, which is why doctors cut so many drug doses as people age and why the same dose of a common painkiller lands differently at 75 than at 35.
Two drug classes now slow that decline once kidney disease has started. SGLT2 inhibitors, first sold for diabetes, cut the combined risk of major kidney decline, kidney failure, or death from kidney or heart causes by 39% in DAPA-CKD, in people with and without diabetes. Semaglutide cut major kidney events by 24% in FLOW.
These were different people at different ages, each measured once, so this is a snapshot of a population rather than a life history of a single kidney.
5
2
24
If metformin really slowed aging, you would expect it to change the one outcome that matters most: who dies. Somebody finally tested that in older people without diabetes, and the result was a flop.
Metformin has been in pharmacies since the 1950s and is the most-cited candidate internet.
The case for it is not nothing. In the 1998 UKPDS trial, overweight people with diabetes on metformin had 36% fewer deaths over about 10 years than people managed on diet alone. @NirBarzilaiMD's TAME trial has aimed to test it properly in more than 3,000 healthy adults aged 65 to 79 for years, and is still raising the money to launch.
Met-HeFT, run from Aarhus in Denmark (@AUHdk), got partway there. It enrolled 940 people with weak hearts, average age 70, with only about 10% carrying a diabetes diagnosis and the rest running high blood sugar or obesity, then gave them metformin or placebo and followed them for 3.7 years.
Death, worsening heart failure, heart attack or stroke hit 25.1% on metformin and 23.4% on placebo.
Metformin also failed to cut new diabetes, and it caused no harm.
So the takeaway is simple: in a real-world test outside diabetes, metformin did not deliver a longevity win.
One disease group, one country, and a heart endpoint rather than an aging one. The bigger test outside diabetes is the VA's IMPACT trial in 7,410 people with prediabetes and artery disease, reporting around 2029.
Metformin has been in pharmacies since the 1950s and is the most-cited candidate internet.
The case for it is not nothing. In the 1998 UKPDS trial, overweight people with diabetes on metformin had 36% fewer deaths over about 10 years than people managed on diet alone. @NirBarzilaiMD's TAME trial has aimed to test it properly in more than 3,000 healthy adults aged 65 to 79 for years, and is still raising the money to launch.
Met-HeFT, run from Aarhus in Denmark (@AUHdk), got partway there. It enrolled 940 people with weak hearts, average age 70, with only about 10% carrying a diabetes diagnosis and the rest running high blood sugar or obesity, then gave them metformin or placebo and followed them for 3.7 years.
Death, worsening heart failure, heart attack or stroke hit 25.1% on metformin and 23.4% on placebo.
Metformin also failed to cut new diabetes, and it caused no harm.
So the takeaway is simple: in a real-world test outside diabetes, metformin did not deliver a longevity win.
One disease group, one country, and a heart endpoint rather than an aging one. The bigger test outside diabetes is the VA's IMPACT trial in 7,410 people with prediabetes and artery disease, reporting around 2029.
8
9
49
What if CAR-T could be made inside the body instead of in a factory? Since 2025, Big Pharma has bet up to $3.4 billion on that possibility.
A three-patient trial in Shanghai just offered an early glimpse of what that could look like.
Traditional CAR-T is powerful but ove a patient’s T cells, engineer them in a lab to kill B cells, then return them after chemotherapy. In a German series of 15 people with lupus, myositis, or scleroderma, every patient came off immune drugs. But treatment takes weeks, and this summer Novartis halted eight autoimmune CAR-T trials after three deaths from a severe immune reaction.
Researchers at Ruijin Hospital and Abogen skipped the cell factory. They gave patients an injection.
ABO2203 packages mRNA in tiny fat particles. Once injected, the mRNA instructs the body to make a protein that binds a T cell at one end and a B cell at the other. In three patients with lupus, Sjogren's, or antiphospholipid syndrome whose immune systems were destroying their platelets, B cells vanished, platelet counts recovered over six months, and side effects remained mild to moderate. There was no cytokine storm. When B cells returned, they were young, naive cells.
It is only three patients at one hospital. But it helps explain why AbbVie, AstraZeneca, and Gilead are betting on reprogramming immune cells inside the body. If the approach holds up, the body itself could become the cell factory.
A three-patient trial in Shanghai just offered an early glimpse of what that could look like.
Traditional CAR-T is powerful but ove a patient’s T cells, engineer them in a lab to kill B cells, then return them after chemotherapy. In a German series of 15 people with lupus, myositis, or scleroderma, every patient came off immune drugs. But treatment takes weeks, and this summer Novartis halted eight autoimmune CAR-T trials after three deaths from a severe immune reaction.
Researchers at Ruijin Hospital and Abogen skipped the cell factory. They gave patients an injection.
ABO2203 packages mRNA in tiny fat particles. Once injected, the mRNA instructs the body to make a protein that binds a T cell at one end and a B cell at the other. In three patients with lupus, Sjogren's, or antiphospholipid syndrome whose immune systems were destroying their platelets, B cells vanished, platelet counts recovered over six months, and side effects remained mild to moderate. There was no cytokine storm. When B cells returned, they were young, naive cells.
It is only three patients at one hospital. But it helps explain why AbbVie, AstraZeneca, and Gilead are betting on reprogramming immune cells inside the body. If the approach holds up, the body itself could become the cell factory.
4
3
25
Stanford found 29,959 unknown forms of life in human mouths and guts.
They are loops of RNA that replicate inside gut bacteria and encode a protein family found nowhere else in biology. About half of mouth samples contain them. The team named them obelisks because they had em.
No one knows what they do. No disease is linked to them, which is the only reason you have not heard of them.
They are loops of RNA that replicate inside gut bacteria and encode a protein family found nowhere else in biology. About half of mouth samples contain them. The team named them obelisks because they had em.
No one knows what they do. No disease is linked to them, which is the only reason you have not heard of them.
1
2
19
A negative cancer blood test does not mean you are cancer-free. The company behind the $949 @GrailBio test says a negative result “does not rule out cancer,” and its largest trial missed most cancers diagnosed within a year.
The test looks for bits of DNA tumors shed into the 000 people aged 50 to 77 were split so half were offered it yearly for three years.
A positive result was cancer about half the time. The rest had hospital checks that found no cancer at the time.
The trials were built for people 50 and over, so nobody knows how well it works at 30 or 40.
Most cancers it found were types with no routine screening, and more were caught early. But the trial missed its main goal, fewer people being diagnosed late, and nobody knows yet if it saves lives.
@FDA advisers vote today on whether to approve it for adults 50 and older, alongside mammograms and colonoscopies.
The test looks for bits of DNA tumors shed into the 000 people aged 50 to 77 were split so half were offered it yearly for three years.
A positive result was cancer about half the time. The rest had hospital checks that found no cancer at the time.
The trials were built for people 50 and over, so nobody knows how well it works at 30 or 40.
Most cancers it found were types with no routine screening, and more were caught early. But the trial missed its main goal, fewer people being diagnosed late, and nobody knows yet if it saves lives.
@FDA advisers vote today on whether to approve it for adults 50 and older, alongside mammograms and colonoscopies.
7
8
31
Tumor cells are literally handing their mitochondria to the T cells meant to kill them.
The T cells can't digest them, so the cancer's mitochondria take over, the T cells go senescent, and killing stops. Researchers spotted it because the T cells carried mitochondrial DNA those of the tumor.
Patients whose tumors had those mutations did worse on immunotherapy.
The T cells can't digest them, so the cancer's mitochondria take over, the T cells go senescent, and killing stops. Researchers spotted it because the T cells carried mitochondrial DNA those of the tumor.
Patients whose tumors had those mutations did worse on immunotherapy.
12
49
234
What if you could stay sharp on less sleep without paying for it later?
That is the promise of a new class of wakefulness drugs.
Old stimulants make you push harder. Adderall and Ritalin sit in the same legal tier as oxycodone, and modafinil still carries a rare skin-reaction work differently.
Your brain stays awake on a signal called orexin. Narcolepsy happens when the cells that make it die. Oveporexton gives that signal back, and the FDA approved it in August.
The early head-to-head data looks promising.
In sleep-deprived men, oveporexton and modafinil both held up for four hours. By hour eight, modafinil had faded to 20 minutes on a 40-minute test, while the orexin drug held at 37.
It is not free.
The approved pill gives most patients insomnia and makes them pee more. But it restores a missing signal instead of whipping what is left.
And the pipeline is growing.
Other orexin pills are now in trials for ADHD, idiopathic hypersomnia, and the fatigue of Parkinson's.
That is the promise of a new class of wakefulness drugs.
Old stimulants make you push harder. Adderall and Ritalin sit in the same legal tier as oxycodone, and modafinil still carries a rare skin-reaction work differently.
Your brain stays awake on a signal called orexin. Narcolepsy happens when the cells that make it die. Oveporexton gives that signal back, and the FDA approved it in August.
The early head-to-head data looks promising.
In sleep-deprived men, oveporexton and modafinil both held up for four hours. By hour eight, modafinil had faded to 20 minutes on a 40-minute test, while the orexin drug held at 37.
It is not free.
The approved pill gives most patients insomnia and makes them pee more. But it restores a missing signal instead of whipping what is left.
And the pipeline is growing.
Other orexin pills are now in trials for ADHD, idiopathic hypersomnia, and the fatigue of Parkinson's.
9
7
49
A woman with type 1 diabetes stopped needing insulin 75 days after a transplant, and a year later, she still does not.
Type 1 diabetes happens because the immune system destroys the islet cells in the pancreas that make insulin. Transplanting islets from a dead donor has worked are scarce and the recipient has to take immunosuppressants for life.
The obvious fix is to grow islets from the patient's own cells. The obvious problem is that the standard method uses Yamanaka factors, delivered by a virus, including one that is a cancer gene.
This team skipped all of that. They reprogrammed her own cells back into stem cells with small molecules alone, with no virus and no Yamanaka factors, and then turned them into islets.
The surgeon placed them under the sheath of muscle in her abdomen instead of in the liver, where islets usually go, because a graft you can see on a scan is a graft you can cut out if it misbehaves.
Her time in the healthy glucose range went from 43% to over 98%.
One patient, published in @CellCellPress. She was already on immunosuppressants for a previous liver transplant, so this does not yet tell us whether autologous islets can escape the autoimmune attack that caused the disease in the first place.
This is an impressive proof of concept, but the real test is whether it works in patients who are not already immunosuppressed.
Type 1 diabetes happens because the immune system destroys the islet cells in the pancreas that make insulin. Transplanting islets from a dead donor has worked are scarce and the recipient has to take immunosuppressants for life.
The obvious fix is to grow islets from the patient's own cells. The obvious problem is that the standard method uses Yamanaka factors, delivered by a virus, including one that is a cancer gene.
This team skipped all of that. They reprogrammed her own cells back into stem cells with small molecules alone, with no virus and no Yamanaka factors, and then turned them into islets.
The surgeon placed them under the sheath of muscle in her abdomen instead of in the liver, where islets usually go, because a graft you can see on a scan is a graft you can cut out if it misbehaves.
Her time in the healthy glucose range went from 43% to over 98%.
One patient, published in @CellCellPress. She was already on immunosuppressants for a previous liver transplant, so this does not yet tell us whether autologous islets can escape the autoimmune attack that caused the disease in the first place.
This is an impressive proof of concept, but the real test is whether it works in patients who are not already immunosuppressed.
2
7
40
Everyone tells you to sleep to protect your brain from dementia because deep sleep supposedly washes out Alzheimer's proteins.
That idea dates to 2013: during deep sleep, the spaces around the brain's blood vessels widen, fluid moves through them, and waste, including amyloid, of the support came from anatomy and tracer studies.
But a group in @NatureNeuro measured what actually came out. They injected a dye into mouse brains and counted how much left. Awake mice cleared 70 to 80% at peak; asleep mice cleared less. Mice under three separate anaesthetics cleared less too, and the deeper the sleep ran on the brainwave trace, the less came out.
If that holds, the washing is not happening because the animal is asleep. It may be happening while it is awake, with sleep protecting the brain some other way.
That is not a reason to sleep less. Poor sleep still tracks with dementia in people, and none of this changes that.
The bigger point is that the field does not agree on the measurement. Other labs using a different method get the opposite result, so the real fight is over whose measurement is right, and that has to be settled in something bigger than a mouse.
That idea dates to 2013: during deep sleep, the spaces around the brain's blood vessels widen, fluid moves through them, and waste, including amyloid, of the support came from anatomy and tracer studies.
But a group in @NatureNeuro measured what actually came out. They injected a dye into mouse brains and counted how much left. Awake mice cleared 70 to 80% at peak; asleep mice cleared less. Mice under three separate anaesthetics cleared less too, and the deeper the sleep ran on the brainwave trace, the less came out.
If that holds, the washing is not happening because the animal is asleep. It may be happening while it is awake, with sleep protecting the brain some other way.
That is not a reason to sleep less. Poor sleep still tracks with dementia in people, and none of this changes that.
The bigger point is that the field does not agree on the measurement. Other labs using a different method get the opposite result, so the real fight is over whose measurement is right, and that has to be settled in something bigger than a mouse.
6
1
39
Cirrhosis was permanent. Deafness from a broken OTOF gene was permanent. Hands paralyzed for thirty years were permanent. Those sentences were all in the textbook, and all three are now wrong. Eight things medicine once called irreversible, ranked by what came back:
1. Hearing. uts a working OTOF gene into the inner ear so hair cells that were always there but silent can finally pass sound on. 16 of 20 children hit the 24-week target, and 5 of the 12 followed to week 48 reached normal hearing. @FDA gave it accelerated approval on 23 April 2026, which means the benefit still has to be confirmed. A Chinese team got there first: five of six children, thresholds down 40 to 57 decibels.
2. Cirrhosis. 71 of 96 patients with cirrhosis from hepatitis B no longer had it after five years on tenofovir, on biopsies taken before and after. Published in 2012 and still the most solid finding here.
3. Hand function after spinal cord injury. ARC-EX (@onwdempowered) runs current through the skin over the neck during rehab. 72 percent of 60 people beat the meaningful-change threshold on both strength and function, and the company reports gains in people up to 34 years after their injury. Authorised 19 December 2024.
4. Sight in one specific blindness. Luxturna, approved 19 December 2017, one injection under the retina.
5. Insulin independence. Zimislecel (Vertex), islet cells grown from stem cells and infused into the liver's blood supply. Ten of twelve people off insulin entirely at one year. It costs them lifelong immunosuppression, two people in the study died, and it is still in Phase 3.
6. Heart muscle. Carbon-14 from nuclear tests acts as a birth date stamp in heart cells, and in patients whose hearts recovered on a mechanical pump, new muscle was being made at more than six times the healthy rate.
7. Crude vision in one blind man via an algal protein plus goggles.
8. Teeth, so far in mice and ferrets, with a Japanese trial being set up.
1 to 4 are approved or authorized. Regeneron's OTOF gene therapy reached the American market 27 months after the Chinese team published theirs.
1. Hearing. uts a working OTOF gene into the inner ear so hair cells that were always there but silent can finally pass sound on. 16 of 20 children hit the 24-week target, and 5 of the 12 followed to week 48 reached normal hearing. @FDA gave it accelerated approval on 23 April 2026, which means the benefit still has to be confirmed. A Chinese team got there first: five of six children, thresholds down 40 to 57 decibels.
2. Cirrhosis. 71 of 96 patients with cirrhosis from hepatitis B no longer had it after five years on tenofovir, on biopsies taken before and after. Published in 2012 and still the most solid finding here.
3. Hand function after spinal cord injury. ARC-EX (@onwdempowered) runs current through the skin over the neck during rehab. 72 percent of 60 people beat the meaningful-change threshold on both strength and function, and the company reports gains in people up to 34 years after their injury. Authorised 19 December 2024.
4. Sight in one specific blindness. Luxturna, approved 19 December 2017, one injection under the retina.
5. Insulin independence. Zimislecel (Vertex), islet cells grown from stem cells and infused into the liver's blood supply. Ten of twelve people off insulin entirely at one year. It costs them lifelong immunosuppression, two people in the study died, and it is still in Phase 3.
6. Heart muscle. Carbon-14 from nuclear tests acts as a birth date stamp in heart cells, and in patients whose hearts recovered on a mechanical pump, new muscle was being made at more than six times the healthy rate.
7. Crude vision in one blind man via an algal protein plus goggles.
8. Teeth, so far in mice and ferrets, with a Japanese trial being set up.
1 to 4 are approved or authorized. Regeneron's OTOF gene therapy reached the American market 27 months after the Chinese team published theirs.
6
37
123