多享7年健康寿命：我们能从超级老人身上学到什么

American healthcare is in crisis. We have a path to prevention. It isn't reversing aging; it's just preventing the age-related morbidities of the big three. It's been a fantasy for millennia. There should be a reboot—a new standard of care based on intelligent partitioning of risk. Seven years more of health span, free of the major three diseases. Seven years. Who wouldn't take 7 years? I hope we'll seize this opportunity, because we may never get another one like this for a long time.

So, you've written this really exciting book, Super Agers: An Evidence-Based Path to Longevity, and I think it's a very timely topic. I was curious if you could set the stage for why you wanted to write it and how you see it in the context of the other books that have been coming out recently as well.

There were a few things that came together. We had done a big study we called the Wellderly, where we basically found very little in the genomes of people who had gotten to the age of 87, on average, without ever having had an age-related disease. So that was one thing that was part of it.

The second was that I got inspired by a patient I saw recently who was 98 and had never been sick. Her name was Lee Rissolo, and her relatives—her parents, uncles, and aunts—had died in their 50s and 60s. She was the outlier, and I thought, “Why?”

Then there were the books that came out. I had patients coming to me; they wanted me to write a prescription for rapamycin or order a total-body MRI. I said, “Wait, we've got to get the story straight.” These 3 things together were the impetus: Why don't I really get deep into everything we know today and see if I could lay out some blueprints for where we can go?

It's coming into a world where American healthcare is in crisis, and I was curious to get your take on where we are now in healthcare in the US and where you think we could get to.

There is this bifurcation, as I see it. You could call it the grand slam, where you get reversal of aging so you keep people healthier body-wide. That's where we see all these remarkable investments in companies like Altos, reprogramming, senolytics, and a long list of other approaches. But they're really focused on a monumental task that hasn't been shown in people, but rather in rodents.

The other side of this is that we made these big strides in the science of aging, with all these layers of data that use the metrics of aging. Why don't we use that to prevent age-related diseases—cancer, cardiovascular disease, and neurodegenerative disease? We've never done that in medicine to any appreciable extent, and this is the opportunity because we have a path to preventing disease. It isn't reversing aging; it's just preventing the age-related morbidities of the big 3.

I think this is something that a lot of people may not realize: The big 3 that you mentioned—cancer, heart disease, and Alzheimer's and dementia—are greatly exacerbated by age. It's interesting because, if you ever wanted to have something that could be a cure for multiple diseases, which would be one of the holy grails of medicine, it would be understanding the biology of aging. Where are we now in terms of things that we can use today?

It takes 20 years to get these diseases, with rare exceptions. Heart disease, almost all cancers, and neurodegenerative diseases are incubating for a very long time. They all have a common thread: a defective immune system and inflammation underpinning them.

They are variably preventable. Cardiovascular disease is 80–90% preventable through lifestyle and related factors—modifiable factors like your LDL cholesterol, that kind of thing. For cancer and neurodegenerative disease, based on what we know today about lifestyle factors, about half can be prevented.

So we have some knowledge about averting these diseases, but we have a lot more with all these clocks and new layers of data that are really changing the face of our understanding of the biology of aging.

In your book, you outline the 5 dimensions of health. I was wondering if you could walk us through them.

Sure. The first and most important one is AI, because you need that to pull all this other data we're going to talk about together. This moment is so exciting because we have multimodal AI—not only large language models, but large reasoning models.

Especially when you're talking about AI, it's all the things people have seen with generative AI and so on, but also just the ability to understand all this data that you're measuring from people.

Yes, because the other 4 are such big domains or dimensions. The omics include not just gene sequencing or arrays, but all the proteins—all the proteomic panels that we can get, which we never could get before inexpensively. It includes the gut microbiome, metabolome, and certainly the epigenome, or epigenetics. The omics are rich. We are now moving toward things like the virtual cell.

Then there are cells that have become a living drug, where we can reset the immune system and cure autoimmune diseases like we've never done before.

Could you give examples of that?

In the last couple of years, we've seen unprecedented cures. We've never had anything like this for lupus, progressive systemic sclerosis, and even cases of multiple sclerosis and dermatomyositis. Basically, it's depletion of all the B cells, and when they come back, they have forgotten what they were attacking. It's amazing—really amazing.

That leads to the autoimmune reaction. But the bigger lesson is that we've learned how to control our immune system like a rheostat, and we're going to keep getting better and better as we measure our immunome.

When you can quash an autoimmune disease, or when you're trying to cure a cancer by doing whatever it takes to keep bringing up the immune system specific to the tumor, the immune system is fundamental. That also now involves cells and vaccines. So vaccines now are capable of cures of pancreatic cancer, kidney cancer with these personalized vaccines using the proteins of the person's tumor.

Yes, and these are in clinical trials right now.

I mean, this is stuff we've never seen. That's just a frontrunner of what vaccines to treat cancer can do. We're going to be using vaccines to prevent cancer. Again, as we get older, some of us—especially those whose immune systems are becoming senescent and weak—could receive a vaccine before there is any cancer, before there's anything else, to prop the immune system up.

We also have drugs to modulate our immune system well beyond checkpoint inhibitors. Whether it's antibody-drug conjugates or tumor-infiltrating lymphocytes, and all these different ways, it's hard to imagine that in the future we're going to lose people to cancer because we'll be able to bring their immune system to the highest level when we need it—but more importantly, prevent the cancer. We can do that now. That's what's exciting.

If we put all this together, what does this mean for the individual? How would their life change? What should people be doing?

I call it “lifestyle plus.” It's a lot bigger than diet, sleep, and exercise, which we can drill down on. It's also involving environmental burdens: air pollution, plastics, microplastics, nanoplastics, and forever chemicals. Then there are other things, like time in nature.

If each of us pulled out all the stops for the lifestyle factors—which is a long list—that would help, but lifestyle factors alone aren't going to be the only way to prevent the big 3 age-related diseases.

You described a large range of things, from the most science-fiction-like drugs that are in trials for preventing cancer to lifestyle. I could also imagine AI coming into this, because one of the things AI is very good at is taking a set of data and masking out the last bit.

You could have someone's health records over 30 years and train on that, except for the last 5 years, and see if you can predict the last 5 from the first 25. Once it gets really good at that, you can take my records and say, “Hey, look, Vijay, if you don't do anything, this is where you're going to be, and we have 99% confidence in this.” That would be pretty chilling.

You're exactly right, because the pinpointing of the timing here is so extraordinary. For example, with p-tau217, which is modifiable by lifestyle, you check it again in 6 months or a year. Now you have 2 data points, and you can say, with all the other data that's available, whether you're going to see mild cognitive impairment 18 years from now, 12 years from now, or 4 years from now. Unless these steps are taken, that is fully dependent on AI models that can take all this data. If we didn't have the science of aging and AI, we'd be nowhere. We wouldn't be talking about this today. I wouldn't have written a book.

It's important for people who aren't familiar with the term “health span”: It's basically not just lifespan, but how long you can be healthy. I don't think we really want to get to some age and be demented or compromised.

What we're talking about is that, if you don't have heart disease, cancer, or neurodegenerative disease, you're pretty darn intact. You may have some achy joints and other matters, but those are the things that really interrupt and end our health span.

Maybe let's turn to another aspect of it, which is the chronic-disease aspect. When we're talking about chronic disease, we're typically talking about diabetes, heart disease, and cancer. How do we start to make an impact in that? I don't know if you want to pick one, or if you want to start with cancer.

I think we can make a huge impact in cancer because we have simple polygenic risk scores for all the common cancers. That's one layer of data to say you're at higher risk, and we have multicancer early-detection tests that can pick up microscopic cancer.

Why would people get a total-body MRI when you could find microscopic cancer, not a mass on an MRI, which may or may not be cancer? So we have some tools for cancer, but the one thing that I think is unanticipated is the GLP-1 drugs—the Ozempic, you know, Zepbound world. Yes, it's the most momentous drug class in medical history, and we've only seen part of the story so far.

In the book, I write about how it took 20 years to figure out that it wasn't just about diabetes, which is amazing. What if we had an AI today and said, “Should we test this for obesity?” The developers, Novo Nordisk and later Eli Lilly, only saw 3 or 4 pounds that people with type 2 diabetes would lose with these drugs. This woman in Norway, a scientist named Luden, kept pushing: “We've got to try it in obesity.” They wouldn't listen to her because they said diabetics weren't losing weight. They finally did it, and everyone knows the story: 20, 30, 50, 80 pounds of weight loss.

Now, when you lose that much weight, for people who are obese, you reduce the risk of cancer, heart disease, and neurodegenerative disease. It wouldn't be surprising to me that now, with pills that are remarkably effective substitutes for injections and can be made much less expensively, a large proportion of the population would be taking one of these drugs or even their successors—that is, drugs that are even more potent and potentially have fewer side effects.

So we have a drug class now added to lifestyle factors that we didn't have before.

Right. As you know, they are in big trials for preventing Alzheimer's in people who are not overweight.

Yes.

We're going to be doing a long COVID trial in people who are not overweight. The effects are really quite extraordinary. The ability to crack obesity—we would have been happy just to do that.

Yes. But all the other things that are coming from it—who would have thought that you could treat, or prevent, addiction?

That's remarkable. Some of the secrets of the gut-brain axis are tied into the immune system and the science of aging. This is what's given us this newfound potential to change. We don't have to rely only on drugs, but there's this interdependence, as we discussed.

Well, I think having lifestyle infrastructure with these drugs—that combination is particularly interesting, because you can make sure that you can lose weight while keeping muscle. Hopefully, patients can go off the drugs, at least for some periods of time, and not rebound.

We don't have encouraging data at the moment, because at least half of people gain the weight back when they stop. That's not good. But I do think that we'll come up with ways to hopefully not rely on such a long-term commitment.

The results on muscle mass—we'd been very worried about that. I think when people combine taking the drugs with strength training, and we do know there's muscle-mass loss just with weight loss alone, it looks encouraging, even though the companies have been acquiring muscle-making drugs that may not prove to be particularly necessary.

And so what else would you put into the chronic bucket? I think one of the things that you've written about is AI plus all the things you can track.

Yeah. I think the ability to look at the organ clocks, which was initially reported here at Stanford by Tony Wyss-Coray and his colleagues and is now validated and replicated by multiple groups, is remarkable. We can look at the brain, the heart, the immune system, and other vital organs, and we can say, “One organ of yours is 5 years out of pace with your real age.” Then we can integrate that with these other layers of data. If that's the case, what about your polygenic risk score? Is there anything pointing to that disease or organ?

We can look at your whole-body aging epigenetic Horvath clock. We can also look at specific proteins—for example, p-tau217 for the brain. What's amazing about that protein, which we can get now and which isn't that expensive, is that it gives us over a 20-year warning about mild cognitive impairment. It's modifiable by exercise and lifestyle. We've seen people in studies whose levels drop more than 50%, even up to 80%.

It's intriguing that it's not binary, either. You could track the gradient.

Exactly. It would get particularly scary if it's increasing. We're talking about assessing this in people without symptoms who are at high risk. I don't recommend any of these things that we're talking about until you know you have an increased risk. But once you do, then you say, “I can do something about it,” and change the course of what otherwise would be that person's natural history.

The molecular clocks—this collection of proteins—is something else that's striking. With Olink and SomaLogic, there are between 6,000 and 11,000 plasma proteins. What we've learned from them, including the fact that there are 3 bursts of aging during our life, shows that aging is not just a linear story. We're learning about the underpinnings of diseases, but most importantly, we have these organ clocks that are inexpensive to obtain.

When you start having genes and proteins and these other layers of data, that's when you find out what is making us unique and what we are at risk for during our extended lifetime, and therefore what we should do to change it and improve.

Yeah, there are a lot of enemies of the future, you know. Maybe a nicer way to put it is that people could be skeptical. They're used to operating a certain way, and they have a certain belief that this isn't going to work, or whatever the reason. What would you tell your fellow clinical colleagues to try to change their mindset from a sick-care mindset to a preventive mindset?

Yeah, I mean, to me, it's all about compelling data. For example, the Alzheimer's drugs, which don't really work and are very risky—the reason they were ultimately bought into by the FDA was because the amyloid came out on the scans, right? And there was a little bit of cognitive-score improvement.

But here we have metrics that are extraordinary to help us as a bridge to compelling evidence. Ultimately, you want to say we prevented these diseases in people who had a definition of their risk and then active surveillance and prevention. Pull out all the stops, right?

For example, speaking about waste, we do mass screening for cancer. We treat everyone the same, based on their age, and that's the only criterion for screening: age. We only pick up 14% of cancers from that mass screening, which costs hundreds of billions of dollars a year.

What about the fact that 88% of women will never have breast cancer? Why do 100% of women have to go through this? Especially with Bayes' rule, you could actually use these as priors that you could measure, and we don't do it.

Why don't we take the risk profile and say, you know what? For a woman, or for a person having a colonoscopy, you don't really ever have to have it, or you can have it once in your lifetime or twice, whatever. We don't treat people as human beings with particular aspects that we can define today. Why? Why is that? We're ingrained in stupidity.

Maybe when these mass-screening programs started, that was the best we could do.

Yeah, but we've known about polygenic risk scores, and we now know about all these other ways to assess risk. Then, with the AI part of it added on, we have to do better. Just having the screening part cleaned up would save a tremendous amount of money.

So let's shift gears to talking about the future. What do you think—let's assume things work out well—that people will see? What is the best-case scenario that you think is plausible? What's the science that's coming on the horizon? Let's say we all decide to make this shift toward prevention. What do you think we will get from it in our next 5 to 10 years?

Well, I think we'll start to see that people are eventually getting to much older ages than we are now without these 3 major diseases. I think that's a gradual thing. It's not like we're going to see a light switch here, but that would be the trend. We will see countries implement it because they don't have the obstacles that we have. We'll see much less of that, and we'll gradually see this curve bend toward people who are older and healthier.

We're not talking about curing. We're talking about prevention, which is a lot better than curing. But it takes time to see the benefit.