Canagliflozin Added 14 Percent to Male Mouse Lifespan and Nothing to Female. Every Large Human Trial of the Drug Class Was Measuring Something Else.
Most longevity drugs arrive with a mechanism and no lifespan data. Canagliflozin arrived the other way around. The National Institute on Aging runs a programme called the Interventions Testing Program, which tests candidate compounds in genetically heterogeneous mice at three independent laboratories simultaneously, specifically so that a result at one site cannot become a field-wide belief. Most compounds fail there. Canagliflozin did not. It added 14 percent to male median lifespan, the effect showed up at all three sites, and it worked again in a second cohort that did not start the drug until the mice were 16 months old. It also did nothing for female mice in the first cohort and appeared to shorten their lives by roughly 6 percent in the second. Nobody has a settled explanation for that. Meanwhile the human evidence for this drug class is among the largest in modern cardiology, running to tens of thousands of randomized participants, and almost none of it is about aging. It was collected in people who already had cardiovascular disease, heart failure or diabetic kidney disease, and it measured how many of them died or went to hospital.
Key Takeaways: What the SGLT2 Inhibitor Aging Evidence Actually Shows
Canagliflozin is a sodium-glucose cotransporter 2 inhibitor, a drug that blocks glucose reabsorption in the kidney so that roughly 75 grams of glucose leave the body in the urine each day. The Interventions Testing Program fed canagliflozin at 180 parts per million of the diet to genetically heterogeneous UM-HET3 mice starting at 7 months of age and reported in JCI Insight in 2020 that male median lifespan rose 14 percent at p below 0.001, that the 90th percentile survival age rose 9 percent, and that the effect appeared in parallel at all three test sites, with no lifespan benefit in females. A second Interventions Testing Program cohort that did not start canagliflozin until 16 months of age, reported in 2024, again added roughly 14 percent to male median lifespan and was associated with roughly a 6 percent decrease in female lifespan, making canagliflozin one of the few compounds in the programme to extend lifespan when started in mid-life. Both sexes showed lower fasting glucose and improved glucose tolerance on the drug, so glycemic control alone does not explain why only males lived longer. A follow-up pathology analysis published in 2022 reported that canagliflozin reduced age-related lesions in the heart, kidney, liver and adrenal gland of male mice, and a separate analysis reported improved insulin sensitivity in the hypothalamus and hippocampus of 30-month-old males. EMPA-REG OUTCOME randomized 7,020 people with type 2 diabetes and established cardiovascular disease to empagliflozin or placebo and reported cardiovascular death down 38 percent, all-cause death down 32 percent, and heart failure hospitalization down 35 percent. DAPA-HF randomized 4,744 heart failure patients to dapagliflozin and reported a hazard ratio of 0.74 for the primary composite, with a hazard ratio of 0.73 in the subgroup who did not have diabetes. The CANVAS programme randomized 10,142 people to canagliflozin, cut major cardiovascular events by 14 percent, and reported lower-limb amputations at 6.3 against 3.4 per 1,000 patient-years for a hazard ratio of 1.97, a signal that was not reproduced in the CREDENCE trial of the same drug. An energy balance analysis published in Diabetes Care in 2015 calculated that the roughly 300 kilocalories excreted daily should produce about 11 kilograms of weight loss by week 90, against the roughly 3 kilograms actually observed, because food intake rises to compensate. No randomized trial has tested any SGLT2 inhibitor as an aging intervention in healthy non-diabetic middle-aged men.
What Does an SGLT2 Inhibitor Actually Do in the Body?
An SGLT2 inhibitor makes the kidney stop reclaiming sugar, so the sugar leaves in the urine. Under normal conditions the kidney filters something like 180 grams of glucose a day and reabsorbs essentially all of it, mostly through a transporter called sodium-glucose cotransporter 2 sitting in the proximal tubule. Blocking that transporter breaks the recycling step. Roughly 75 grams of glucose per day exits the body instead, along with about 400 millilitres of extra urine and a quantity of sodium. The direct consequences follow from that single mechanical fact. Blood glucose falls without any involvement of insulin, which is why these drugs do not cause hypoglycemia on their own. Insulin levels drop and glucagon rises, which shifts the body slightly toward fat oxidation and raises circulating ketones. Plasma volume falls a little, blood pressure comes down a few points, and erythropoietin rises, which lifts hematocrit over the following months. Canagliflozin, dapagliflozin, empagliflozin and ertugliflozin all work this way and differ mainly in selectivity and half-life. The drugs were approved for type 2 diabetes and turned out to do things for the heart and kidney that had very little to do with glucose, which is the part that made them interesting to everyone else.
Why Did Canagliflozin Extend Lifespan in Male Mice Only?
Nobody knows, and the honest state of the field is that the sex split is unexplained rather than explained away. The Interventions Testing Program result itself is unusually solid. Genetically heterogeneous UM-HET3 mice received canagliflozin at 180 parts per million of the diet from 7 months of age, and male median survival rose 14 percent at p below 0.001, with the 90th percentile survival age rising 9 percent and the same direction of effect appearing at each of the three geographically separate test sites. That design exists precisely to kill single-laboratory artefacts, and canagliflozin survived it. Female mice on the identical diet gained nothing. The obvious hypothesis, that the drug simply worked better metabolically in males, does not hold up: both sexes showed lower fasting glucose and better glucose tolerance on canagliflozin. Something downstream of glycemic control is doing the work, and it is sex-dependent. Candidate explanations in the literature include differences in drug exposure, differences in how male and female mice handle the insulin and IGF-1 axis, and the fact that male UM-HET3 mice die of somewhat different pathologies than females do, which leaves more room for an intervention to matter. None of these has been demonstrated. What can be said is narrower and more useful: the compound has a reproducible lifespan effect in one sex of one species, and the mechanism of that effect is not established.
The Only Lifespan Data Canagliflozin Has, and It Splits by Sex
Percent change in survival for genetically heterogeneous UM-HET3 mice fed canagliflozin at 180 parts per million of the diet in the National Institute on Aging Interventions Testing Program. The first cohort, reported in JCI Insight in 2020, started the drug at 7 months of age and gained 14 percent of median lifespan in males at p below 0.001, with the 90th percentile survival age rising 9 percent, and the effect appeared at all three independent test sites. Females gained nothing. A second cohort that started the drug at 16 months of age, reported in 2024, again added 14 percent to male median lifespan and this time subtracted roughly 6 percent from female lifespan.
Does Canagliflozin Still Work If You Start It Late in Life?
Yes, at least in male mice, and this is the finding that separates canagliflozin from most of the field. The Interventions Testing Program ran a second cohort in which canagliflozin was not introduced until the mice were 16 months old, which in a UM-HET3 mouse is somewhere past middle age. Male median lifespan again rose by roughly 14 percent. That makes canagliflozin one of only a handful of compounds the programme has found to extend lifespan from a late start, which matters enormously for anyone reasoning about human application, since no adult reading about a geroprotector has the option of starting at the mouse equivalent of adolescence. The same late-start cohort reported roughly a 6 percent decrease in female median lifespan. That is a stronger and less comfortable version of the sex split seen in the first cohort, where females simply showed no effect. Taken together, the two cohorts say the drug does something real to the aging trajectory of a male mouse and something possibly harmful to a female one. Extrapolating either half of that to a human is unwarranted, and the second half is the reason the extrapolation deserves more caution than the enthusiasm around the first half usually carries.
Did the Treated Mice Actually Age More Slowly, or Just Die Later?
The pathology suggests the animals genuinely aged differently, not that they simply avoided one cause of death. A follow-up analysis published in 2022 examined tissue from the Interventions Testing Program mice and reported that canagliflozin retarded age-related lesions in the heart, kidney, liver and adrenal gland of male mice. Lesion burden spread across four unrelated organ systems is the kind of pattern researchers look for when distinguishing a genuine slowing of aging from a drug that happens to prevent one specific fatal disease. A separate study of aged UM-HET3 mice reported that canagliflozin improved insulin sensitivity in the hypothalamus and the hippocampus of 30-month-old males, which puts a central nervous system component into the picture alongside the peripheral metabolic effects. There is a limit to how far this reasoning travels. Slowing organ pathology in a mouse that was fed a drug from 7 months of age tells you about mouse biology under laboratory conditions with a defined diet. It does not establish that the same tissue-level protection occurs in a human, and no human study has attempted to measure it.
What Did the Large Human Trials Actually Test, and What Did They Find?
The large human trials tested whether people who were already seriously ill had fewer cardiovascular and kidney events, and the answer was yes, repeatedly. EMPA-REG OUTCOME randomized 7,020 people with type 2 diabetes and established cardiovascular disease to empagliflozin or placebo over a median of roughly three years and reported cardiovascular death down 38 percent, all-cause death down 32 percent, and hospitalization for heart failure down 35 percent. Those numbers were large enough that cardiologists initially struggled to explain them, since the glucose-lowering effect of the drug is modest and the survival curves separated within months, far faster than atherosclerosis changes. The CANVAS programme randomized 10,142 people with type 2 diabetes to canagliflozin and reported major adverse cardiovascular events down 14 percent at a hazard ratio of 0.86. CREDENCE randomized 4,401 people with diabetic kidney disease to canagliflozin and reported the primary kidney composite down 30 percent at a hazard ratio of 0.70. DAPA-HF randomized 4,744 patients with heart failure and reduced ejection fraction to dapagliflozin and reported a hazard ratio of 0.74 for the composite of cardiovascular death, heart failure hospitalization and urgent heart failure visits. Every one of those trials enrolled people carrying a diagnosis. The entry criteria were the disease. Not one of them measured a marker of aging, a functional outcome of aging, or lifespan in a person without established illness.
What the Large Human Trials Measured
Hazard ratios from the outcome trials that built the clinical case for SGLT2 inhibitors. A value below the dashed line at 1.0 means the drug arm had fewer events. EMPA-REG OUTCOME randomized 7,020 people with type 2 diabetes and established cardiovascular disease to empagliflozin or placebo and reported cardiovascular death down 38 percent, all-cause death down 32 percent and heart failure hospitalization down 35 percent. CANVAS randomized 10,142 people to canagliflozin and found major cardiovascular events down 14 percent alongside roughly double the rate of lower-limb amputation, at 6.3 against 3.4 per 1,000 patient-years. DAPA-HF randomized 4,744 heart failure patients and the benefit held in the participants who did not have diabetes. Every one of these trials enrolled people who already had a diagnosed disease. None of them measured aging.
Do SGLT2 Inhibitors Work in People Who Do Not Have Diabetes?
For heart failure, yes, and this is the single most important extension of the human evidence beyond diabetes. DAPA-HF prespecified the analysis and found that in the participants who did not have diabetes at baseline, the hazard ratio for worsening heart failure or cardiovascular death was 0.73, essentially identical to the 0.74 seen in the trial overall. That result decoupled the benefit from glucose lowering in a way no subgroup analysis of a diabetes trial could. Later trials extended the pattern into heart failure with preserved ejection fraction and into chronic kidney disease without diabetes, and the drug class is now used in cardiology and nephrology on the strength of it. What that does not establish is anything about a healthy person. The non-diabetic participants in DAPA-HF still had heart failure, which is a disease with a five-year mortality comparable to many cancers. Demonstrating that a drug helps a sick person who does not happen to have diabetes is a different claim from demonstrating that it helps a metabolically healthy 45-year-old man who lifts three times a week and wants to be functional at 80. The second claim has never been tested.
Why Is the Weight Loss So Much Smaller Than the Calories Excreted?
Because appetite rises to replace almost all of it, and this is one of the better-documented compensations in metabolic medicine. Ferrannini and colleagues published an energy balance analysis in Diabetes Care in 2015 that quantified the problem directly. SGLT2 inhibition produces a net loss of roughly 75 grams of glucose per day, which is an energy loss of about 300 kilocalories. Fed into a validated dynamic model of human energy balance, a sustained deficit of that size predicts something on the order of 11 kilograms of weight loss by week 90. Trial participants lost roughly 3. The model attributes the gap to an adaptive increase in energy intake of a few hundred calories per day that appears early and does not fade. The practical shape of this is familiar to anyone who has watched the weight curves in these trials: rapid loss over the first four to eight weeks, much of it fluid from the osmotic diuresis, then a slow drift, then a plateau that arrives well before anyone expected it. Anyone approaching an SGLT2 inhibitor as a weight loss tool is approaching it as the weakest available option, and the trials say so plainly.
The Calories Leave. The Weight Mostly Does Not.
Modelled against observed weight loss on an SGLT2 inhibitor, based on the energy balance analysis published in Diabetes Care in 2015 by Ferrannini and colleagues. SGLT2 inhibition dumps roughly 75 grams of glucose into the urine each day, which is an energy loss of about 300 kilocalories daily. Fed into a validated model of human energy balance, that deficit should produce roughly 11 kilograms of weight loss by week 90. Trial participants lost about 3. The gap is explained by an adaptive rise in food intake of a few hundred calories a day that begins early and persists. The two endpoints come from the published analysis and the intermediate points are interpolated to show the shape of the divergence.
- Predicted from glucose excreted
- Actually observed in trials
What Happens to Muscle Mass on an SGLT2 Inhibitor?
Some lean mass comes off, the amount is small, and it varies by which drug is used. A network meta-analysis of body composition outcomes in type 2 diabetes trials found that dapagliflozin reduced fat mass by roughly 1.75 kilograms along with reductions in percentage body fat, visceral fat, subcutaneous fat and waist circumference. On lean mass the same analysis separated the agents: dapagliflozin and tofogliflozin looked neutral, while empagliflozin, ipragliflozin and canagliflozin were associated with modest reductions. The typical pattern across trials is a decline in body weight over the first two to four weeks that includes body water, followed by fat mass reduction with small lean mass reductions that persist. Several studies report that the lean-to-total mass ratio improves even when absolute lean mass falls, because fat is lost faster. For an older man, the number that matters is not the ratio. Absolute skeletal muscle is what carries functional capacity into the eighth decade, and a drug that removes even a small amount of it while a man is not actively resistance training is working against the thing that predicts how he ages. The trials were not designed to answer whether resistance training offsets the effect, and no study has tested an SGLT2 inhibitor added to a supervised training programme the way metformin has been tested.
What Are the Documented Risks for a Man Considering This Class?
The main risks are genital fungal infection, volume depletion, a rare form of ketoacidosis that occurs at normal blood glucose, and an amputation signal specific to one trial. Genital mycotic infection is the most common reason people discontinue, and it is driven by the mechanism itself, since glucose in the urine feeds yeast. A cohort analysis of people starting these drugs put the one-year absolute risk at 2.7 percent for men with no prior genital infection, 10.8 percent for women with no prior infection, 12.1 percent for men with a prior infection and 23.7 percent for women with a prior infection. The history matters more than the sex. Euglycemic diabetic ketoacidosis is uncommon but genuinely dangerous, because blood glucose stays near normal while ketones climb, which delays diagnosis. Its documented triggers are the conditions a fitness-oriented man is most likely to create deliberately: prolonged low carbohydrate intake, fasting, dehydration, hard prolonged exercise, illness and surgery. The osmotic diuresis also produces volume depletion, orthostatic symptoms and occasionally acute kidney injury in people who are already dry. The CANVAS programme reported lower-limb amputations at 6.3 against 3.4 per 1,000 patient-years on canagliflozin, mostly at the toe or metatarsal, for a hazard ratio of 1.97. CREDENCE, testing the same drug in a sicker kidney population, did not reproduce that finding, and the signal remains unresolved rather than confirmed or dismissed.
One-Year Absolute Risk of Genital Infection
Sugar in the urine feeds yeast, and genital mycotic infection is the most common reason people stop taking an SGLT2 inhibitor. Absolute one-year risk from a cohort analysis of people starting these drugs, split by sex and by whether the person had a documented genital infection beforehand. A man with no infection history sits at 2.7 percent, which is the lowest cell in the table and the reason this side effect reads differently for men than the package insert suggests. A prior infection raises a man's risk to 12.1 percent, above the risk of a woman without that history.
How Should a Man Read the SGLT2 Evidence Right Now?
As an unusually good animal result attached to a human record that answers a different question. The canagliflozin lifespan data is stronger than what most longevity compounds have: three independent laboratories, a large effect in males, and a repeat performance from a late start, which is the version of the result that would matter to an adult human. The human data is also stronger than what most longevity compounds have, in absolute terms, since tens of thousands of people have been randomized and followed for years. The problem is that these two strong bodies of evidence do not overlap. The mouse work measured lifespan in animals with no disease. The human work measured events in people who all had disease. The bridge between them, a randomized trial testing whether an SGLT2 inhibitor changes any aging-relevant outcome in a metabolically healthy person, does not exist yet, though smaller studies in older adults with obesity and prediabetes have started to probe the edge of that question. A man reading this should also weigh the sex asymmetry seriously rather than filing it as a curiosity, because a compound that adds 14 percent to one sex and subtracts 6 percent from the other is telling you the mechanism is conditional on biology that has not been identified. These are prescription medications with real adverse effects, and any decision about them belongs with a physician who knows the individual case.
The SGLT2 Aging Record on One Axis
Every strand of the SGLT2 inhibitor aging case arranged as a single schematic map. Cyan marks what the evidence supports, amber marks findings that are real but narrower or stranger than they sound, slate marks documented downsides, and dark grey marks claims the evidence does not carry. This is the rare geroprotector with a clean mouse lifespan result that replicated across sites and repeated on a late start. It is also a drug whose entire human record was collected in people who were already sick, and whose one lifespan finding does not apply to half the population.
Frequently Asked Questions
Have more questions? Contact us or explore our full peptide intelligence platform.
Related Articles
Does Telomere Length Determine Lifespan? Short Telomeres Track With Earlier Death. The Genetics Say Longer Ones Raise Cancer Risk.
Does telomere length determine lifespan? Short telomeres track with earlier death, yet genetic studies show longer ones raise the risk of several cancers.
Low-Dose Aspirin Added 8 Percent to Male Mouse Lifespan. In 19,114 Healthy Older Adults, It Added No Healthy Years and Deaths Ran Higher.
Does low-dose aspirin help longevity? It extended male mouse lifespan 8%, but the ASPREE trial in 19,114 healthy older adults found no gain and more bleeding.
Is NAD+ a Peptide? No. It Is a Coenzyme Built From Two Nucleotides, and the Peptide Label Comes From the Clinic, Not the Chemistry
Is NAD+ a peptide? No. NAD+ is a coenzyme built from two nucleotides. What NAD+ is used for, what IV NAD+ studies found, and how it differs from SS-31.