Some drugs lengthen mouse lives, but proven prevention still ranks higher

Key takeaways

• In a study of 216,917 U.S. adults, 70% of women who never smoked lived from 25 to 80, compared with 38% of women who smoked; for men, the figures were 61% and 26%¹.
• The National Institute on Aging’s Interventions Testing Program tests compounds in mice at three labs at once, and by May 2024 its researchers counted seven drugs that lengthened life by at least 10% in one or both sexes²⁷.
• Resveratrol and metformin taken alone didn’t lengthen mouse lives in that program, and a diabetes drug that helped male mice shortened the lives of females who started it late in life⁴⁵⁷.
• The FDA accepts a lab marker as proof of benefit only after trials tie it to real outcomes, the way lowering blood pressure was tied to fewer strokes¹⁰¹¹.
• Before you buy or join anything sold for longevity, place it on the evidence ladder below and ask the seven questions, starting with whether anyone has tested it in people.

At the University of Michigan, male mice fed meclizine, a drug used since 1955 for nausea and dizziness, lived 11% longer at the median than mice on plain chow. At the Jackson Laboratory in Bar Harbor, Maine, the gain was 15%³.

At the third lab, in San Antonio, the gain was zero³.

Pooled across all three sites, meclizine added 8% to the median life of male mice and nothing for females, researchers reported in late 2023³. It’s a real result, though a modest one, and the team said tests of higher doses were under way³.

That split is the reason the program exists. The National Institute on Aging (NIA) built its Interventions Testing Program so that no single lab can crown a longevity drug on its own²³⁴.

It’s also a useful lesson for anyone reading longevity ads. Some anti-aging science is careful and real. None of the mouse winners has yet been shown to add years to human lives. The steps that have been shown to do that are older, cheaper and less exciting.

The proven steps are counted in human years and lives

The strongest longevity evidence doesn’t come from labs. It comes from following very large numbers of people for many years.

Not smoking. Researchers led by Prabhat Jha linked national health survey data on 216,917 U.S. adults, collected from 1997 to 2004, to death records¹. People who smoked died at about three times the rate of people who never had, and they lost more than a decade of life¹.

Put in absolute terms, 70 of every 100 women who never smoked lived from 25 to 80, compared with 38 of every 100 women who smoked. For men, it was 61 versus 26¹. Our guide to quitting tobacco covers how many years quitting wins back at each age.

Moving. In 2012, a team led by Steven Moore of the National Cancer Institute pooled six studies of 654,827 adults followed for a median of 10 years⁸. Compared with no leisure activity, up to about 75 minutes of brisk walking a week went with 1.8 more years of life after 40⁸.

About 150 to 299 minutes a week went with 3.4 more years, and 450 minutes or more with 4.5⁸. People reported their own activity, and the study watched rather than assigned it, so it can’t prove cause⁸. Our article on small doses of activity covers newer data.

Vaccines. A 2024 study in The Lancet, led by the World Health Organization, estimated that vaccines against 14 diseases saved 154 million lives worldwide over 50 years, 101 million of them infants⁹. Each life saved meant an average of 66 years of full health⁹.

That’s a global model, not a U.S. adult figure, and it leans heavily on measles shots in children⁹. Our guide to vaccination schedules explains the adult schedules and the current federal debate.

Blood pressure and screening. Controlling blood pressure and getting a few well-chosen cancer screenings also earn their place. Our articles on changeable risk factors and on screening benefits and harms give the numbers, and companion articles in this series cover blood pressure in depth.

A federal mouse lab built to avoid fooling itself

A mouse study run in one lab, on one inbred strain, can reflect that lab’s housing conditions or that strain’s quirks. The NIA program, running since 2004, was designed around both problems²⁴.

Each compound is tested at the same time at three sites: the Jackson Laboratory, the University of Michigan and the University of Texas Health Science Center at San Antonio²⁴.

The mice are a mixed breed called UM-HET3, so each animal is genetically different²⁴. A group of them looks more like a crowd than a set of twins.

Anyone can nominate a compound, including academic and company scientists, and a committee picks a few each year⁴. Each study is large enough to have an 80% chance of spotting a 10% change in lifespan, even if one site’s data are lost⁴. Every treated group is compared with its own untreated controls, and the results are pooled³.

The program tests prescription drugs, but also foods, supplements, plant extracts and hormones². That makes it something like a fair public tryout for longevity candidates, natural or not.

What made mice live longer, and what didn’t

Here are some of the program’s results, as reported by its own researchers. “Median” means the age by which half the mice had died.

Compound	What it is	Result in the program’s mice
Rapamycin	A drug also called sirolimus	Longer average and maximum life in both sexes at all three sites, whether started at 9 or 20 months of age4
Rapamycin plus metformin	Two drugs combined	23% longer median life in both sexes; unclear how much metformin added to rapamycin alone5
Acarbose	A diabetes drug that slows starch digestion	22% longer median life in males and 5% in females when started young6
17-alpha-estradiol	A weakly estrogenic hormone	19% longer median life, males only5
Canagliflozin	An FDA-approved diabetes drug	14% longer median life in males; no gain in females6
Astaxanthin	A natural pigment with antioxidant activity	12% longer median life, males only3
Meclizine	A nausea and dizziness drug	8% longer median life, males only3
Glycine	An amino acid	A gain of under 5%6

The misses matter as much. Resveratrol failed at two doses⁴. So did metformin taken alone, fish oil, nicotinamide riboside, fisetin and alpha-ketoglutarate²³⁵⁷.

By the researchers’ own count in May 2024, seven drugs tested in the program had lengthened life by at least 10% in one or both sexes⁷. That’s a real scientific achievement. It is also a short list after two decades of testing.

The list keeps moving. The program’s database now includes a 2025 paper reporting longer lives in male, but not female, mice given three more compounds: epicatechin, halofuginone and mitoglitazone².

A win in mice isn’t a prescription for you

Look again at the table, and a pattern jumps out: many of the gains showed up mainly or only in males.

Canagliflozin is the sharpest example. When the program started it in older mice, at 16 months, males lived 14% longer, and females lived 6% shorter⁷. Older females had about 20 times the drug level in their blood that young males did⁷.

Dose is a second trap. In one study, the mouse food meant to carry dimethyl fumarate held only about 35% of the planned amount, which may explain its null result³. The authors of the meclizine and astaxanthin study also wrote that they don’t know whether their doses were the best ones³.

The canagliflozin team wrote that drugs like these may offer a starting point for human anti-aging research⁶. For metformin, the program’s authors described the human evidence, mostly observational reports in people with diabetes, as suggestive enough to prompt calls for large clinical trials⁵.

Some of these compounds are already human medicines, such as acarbose and canagliflozin for diabetes⁶. They were tested in people to control blood sugar, not to slow aging. Our article on animal studies explains how rarely animal results carry over to approved treatments, and our longevity-clinics guide covers what human trials of rapamycin have found so far.

The longevity evidence ladder starts with what a study counted

Many longevity pitches don’t promise more years. They promise a better number: a lower “biological age,” a higher NAD+ level, a calmer inflammation marker.

The FDA calls a number like that a surrogate endpoint: a stand-in for how patients feel, function or survive¹⁰. Some stand-ins have earned trust. Many trials showed that lowering systolic blood pressure, the top number, reduces strokes, so the agency accepts blood pressure as a substitute for stroke in drug trials¹⁰.

The FDA calls those “validated” surrogates, backed by trial data showing that a change in the marker predicts real benefit¹⁰¹¹. Others are only “reasonably likely” to predict benefit and still need confirming trials¹⁰.

Even validated markers can fail. The agency notes that some treatments improved a trusted marker in one setting and proved useless or harmful in another¹¹. A marker with no outcome trials behind it is a hypothesis, not a result. Our article on biological-age tests shows how much one popular marker can wobble.

The ladder

This tool is ours. Find the rung a claim stands on, then treat it the way the last column suggests.

Rung	What it means	Example	How to treat it
1. Human trials that count outcomes	Chance decides who gets the treatment, and the study counts illness, disability or death	Blood pressure drug trials that showed fewer strokes10	Act on it, with your clinician
2. Human trials of a marker	The treatment moves a number, not yet an outcome	A drug judged on a “reasonably likely” marker10	Promising; wait for outcome trials unless the marker is validated
3. Human observational studies	Researchers watch people who chose a habit	Activity and longer life in 654,827 adults8	Strong when results are large, consistent and dose-related, as with smoking1
4. Animal studies	Tests in mice, worms or flies	Rapamycin and acarbose in the NIA program46	A reason for human trials, not for buying
5. Cell and lab-dish studies	A compound changes cells in a dish	“Activates longevity genes” in a lab	Interesting biology; no health claim yet
6. Stories and testimonials	One person’s experience	A clinic’s before-and-after video	A reason to look for rungs 1 to 3

Two cautions. Observational studies can mislead when the people who choose a habit differ in other ways, as our article on observational studies explains. And a claim can quote a higher rung than it stands on, such as citing a mouse study in a headline about people.

Check a claim yourself in 10 minutes

Two free tools can show where a claim stands.

The trial registry. ClinicalTrials.gov lists studies by an ID that starts with “NCT”¹². Each record shows the study design, how many people enrolled, the main outcome the researchers chose in advance and the eligibility rules¹².

Check the results tab. “No Study Results Posted” means nothing has been submitted, and “Results Submitted” means results are waiting to be posted¹². Sponsors supply the information, and a listing doesn’t mean the government has reviewed the study¹³. Our article on how clinical trials work has a fuller walkthrough.

The preprint label. A preprint is a paper posted online before other scientists have reviewed it. The health-science server medRxiv, launched in June 2019 and run by Cold Spring Harbor Laboratory with BMJ and Yale, screens for basics but doesn’t peer review¹⁴.

Its own caution says preprints shouldn’t guide clinical practice or health-related behavior, and the news shouldn’t report them as established¹⁴. If an ad cites a link with “rxiv” in it, treat the claim as a draft.

Seven questions for any longevity claim

These questions, ours rather than a validated test, build on the ladder.

1. What rung is it on? Ask for the human study, not the mechanism.
2. Did it count something you’d feel? Illness, disability or death beat a lab number, unless the number is validated¹⁰.
3. Who was studied? Sex, age and health matter. One drug helped male mice and harmed older females⁷.
4. Did anyone repeat it? Meclizine gained 15% at one site and nothing at another³.
5. Is the dose what you’d take? Lab doses can miss the target, as one mouse diet did³.
6. Is it registered, and are results posted? If a study ended years ago and still shows no results, ask why¹².
7. What will it cost you? Count the money and the attention, then name the proven step that money or time could buy instead.

The hidden side effect is opportunity cost

Every hour and dollar spent on an experiment has another possible use. For most people, the payoff from the proven steps is far larger and far better documented.

Consider the walking data. Even up to 75 minutes of brisk walking a week went with 1.8 more years of life in the pooled study, and walking costs nothing⁸. No supplement or drug on the mouse list has shown anything comparable in people.

None of this makes curiosity foolish. Research like the NIA’s is how the next proven treatment will be found, and some of its candidates come from plants and foods². Experimental doesn’t mean worthless. It means unfinished.

Natural health has long put prevention first and asked hard questions of anyone with something to sell. Those instincts fit the evidence well here. The habits with the best proof are the ones the tradition has always valued: not smoking, moving every day, eating real food, and protecting the young with proven care. A later article on preventive care on a budget covers how to get the proven services cheaply.

What the evidence says

Claim	Evidence level	Why
Not smoking adds years of life	Strong evidence	Survival to 80 roughly doubled in never-smokers in national data1
Regular activity goes with a longer life	Strong evidence	654,827 adults; benefit rose with dose, though the data are observational8
Vaccination has saved lives on a vast scale	Strong evidence	154 million lives in 50 years by WHO-led estimate9
Rapamycin, acarbose and similar drugs are strong candidates for human aging trials	Promising but early	Repeated in mice at three sites; no human lifespan data46
Resveratrol extends mouse lifespan	Evidence says no, or unsafe	Failed at two doses in the NIA program’s mice4
Metformin alone extends lifespan	Not enough evidence in people	It failed alone in the program’s mice, and the human data are mostly observational5
Any drug or supplement has been shown to slow human aging	Not enough evidence	Mouse wins haven’t been confirmed in human outcome trials6

Questions readers ask

The bottom line

Real science is extending lifespan in mice, and the NIA’s three-lab program is one of the most careful efforts anywhere. But most of its wins are modest, many are limited to one sex, and none has been shown to add years to a human life.

The steps with that proof are the old ones. Put not smoking, daily movement, vaccines and well-chosen screening first. Use the ladder and the seven questions on everything else, and treat experiments as unfinished work.

Related reading

• CDC tied HIV infections to spa facials, so score risky treatments before you pay
• Only 1 in 20 treatments tested in animals wins approval for people
• How to sort longevity clinics that mix proven care with pricey experiments
• How to decode “reverse your age” ads that outrun the science
• Why observational studies got smoking right and hormone therapy wrong

Sources

1. Jha P, Ramasundarahettige C, Landsman V, et al. “21st-Century Hazards of Smoking and Benefits of Cessation in the United States.” New England Journal of Medicine, 2013;368:341-350 (copy hosted by Disease Control Priorities). dcp-3.org
2. The Jackson Laboratory, Mouse Phenome Database. “NIA Interventions Testing Program (ITP).” Project page, accessed October 2026. phenome.jax.org
3. Harrison DE, Strong R, Miller RA, et al. “Astaxanthin and meclizine extend lifespan in UM-HET3 male mice; fisetin, SG1002 (hydrogen sulfide donor), dimethyl fumarate, mycophenolic acid, and 4-phenylbutyrate do not significantly affect lifespan in either sex at the doses and schedules used.” GeroScience, 2024;46:795-816. link.springer.com
4. Miller RA, University of Michigan. “Interventions Testing Program (ITP).” Program description and early results. websites.umich.edu
5. Strong R, Miller RA, Antebi A, et al. “Longer lifespan in male mice treated with a weakly estrogenic agonist, an antioxidant, an α-glucosidase inhibitor or a Nrf2-inducer.” Aging Cell, 2016;15(5):872-884 (author manuscript, University of East Anglia). ueaeprints.uea.ac.uk
6. Miller RA, Harrison DE, Allison DB, et al. “Canagliflozin extends life span in genetically heterogeneous male but not female mice.” bioRxiv preprint, 2020. biorxiv.org
7. Miller RA, Harrison DE, Cortopassi GA, et al. “Lifespan effects in male UM-HET3 mice treated with sodium thiosulfate, 16-hydroxyestriol, and late-start canagliflozin.” GeroScience, 2024. link.springer.com
8. Moore SC, et al. “Leisure Time Physical Activity of Moderate to Vigorous Intensity and Mortality: A Large Pooled Cohort Analysis.” PLOS Medicine, 2012. journals.plos.org
9. World Health Organization (via the Partnership for Maternal, Newborn and Child Health). “Global immunization efforts have saved at least 154 million lives over the past 50 years.” News release, April 24, 2024. pmnch.who.int
10. U.S. Food and Drug Administration. “Surrogate Endpoint Resources for Drug and Biologic Development.” FDA, content current as of July 24, 2018. fda.gov
11. U.S. Food and Drug Administration. “FDA Facts: Biomarkers and Surrogate Endpoints.” FDA, content current as of December 21, 2017. fda.gov
12. National Library of Medicine. “How to Read a Study Record.” ClinicalTrials.gov help. classic.clinicaltrials.gov
13. National Library of Medicine. “ClinicalTrials.gov” modernization fact sheet. NIH, October 2023. nlm.nih.gov
14. Ross JS, Krumholz HM. “medRxiv: The Preprint Server for Health Sciences.” NIH Collaboratory Grand Rounds slides, July 12, 2019 (hosted by Duke Clinical Research Institute). dcricollab.dcri.duke.edu
15. American Heart Association. “Heart Attack, Stroke and Cardiac Arrest Symptoms.” heart.org, accessed October 2026. heart.org