Canagliflozin (SGLT2 inhibitor)

How the drug works

In the kidney, sugar that gets into the urine is normally almost completely reabsorbed; a transporter called SGLT2 does this. Canagliflozin blocks it, the sugar stays in the urine and is excreted.

Depending on blood sugar levels, that amounts to around 50 to 80 grams a day, that is, 200 to 300 kilocalories. This is exactly what makes the substance attractive to the scene — and exactly what does not explain the longevity finding.

The mouse finding, twice

In the Interventions Testing Program, treatment initially began at 7 months, roughly middle adulthood in mice. In males, median lifespan rose by 14 percent, and the upper end shifted as well: the age that only 10 percent of the animals reach was 9 percent higher.

Then a second start was made, this time only at 16 months — roughly the fifties in humans. Again 14 percent. This is the strongest single finding for the substance, because with most longevity candidates you have to start early.

Why the weight explanation does not hold

One would assume the drug works via weight loss. The numbers say the opposite. The females became considerably lighter, 17 to 19 percent across several measurement points, and their fat mass fell by 41 percent. They did not live one bit longer: 1 percent, not significant.

The males became only 5 to 8 percent lighter, their fat mass did not change significantly — and they lived 14 percent longer. Exactly the other way round, then. Anyone who says canagliflozin works via calorie loss has to explain why the animals with the greatest loss got nothing out of it. Incidentally, the long-term value HbA1c did not change significantly in either sex.

Sex is not a side issue

The difference between males and females is not an isolated case, but one of the underestimated findings of the program. In the same round of publications there was an estrogen derivative, 16-hydroxyestradiol: 15 percent more lifespan in males and 7 percent less in females. The same substance, opposite directions.

If you cannot even extrapolate from mouse data to the other sex of the same species, the leap to humans is correspondingly larger.

What the studies show

CANVAS — 10,142 people with type 2 diabetes

The large endpoint study on canagliflozin: 10,142 people with type 2 diabetes and high cardiovascular risk. The combined endpoint of cardiac death, heart attack and stroke occurred at 26.9 per 1,000 patient-years compared with 31.5 under placebo, p equal to 0.02. Lower-limb amputations: 6.3 per 1,000 patient-years compared with 3.4 under placebo, that is, almost a doubling.

CREDENCE — 4,401 people with type 2 diabetes and kidney disease

CREDENCE tested canagliflozin itself in 4,401 people with type 2 diabetes and chronic kidney disease with increased protein excretion. The combined endpoint of kidney failure, doubling of creatinine and death from renal or cardiovascular causes occurred at 43.2 per 1,000 patient-years compared with 61.2 under placebo, hazard ratio 0.70, confidence interval 0.59 to 0.82. The study was stopped early after a mean of 2.6 years because of the clear benefit; here there was no difference in amputations and fractures.

EMPA-KIDNEY — class evidence: 6,609 people with chronic kidney disease

EMPA-KIDNEY tested the related drug empagliflozin in 6,609 people with chronic kidney disease, 54 percent of whom did not have diabetes. The combined endpoint of progression of kidney disease or cardiovascular death occurred in 13.1 percent compared with 16.9 percent under placebo, hazard ratio 0.72, confidence interval 0.64 to 0.82, p-value below 0.001, after a mean of 2 years.

Interventions Testing Program — early and late start

In the ITP, canagliflozin was first given from an age of 7 months: median lifespan of males rose by 14 percent, and the age that only 10 percent of the animals reach was 9 percent higher. In a second round, treatment began only at 16 months, with the same result. In females the effect was 1 percent, even though they lost 17 to 19 percent of their weight and 41 percent of their fat mass.

Where the data stop

For the question of whether a healthy person lives longer with an SGLT2 inhibitor, there are no data. Not few — none. No one has ever given a healthy fifty-year-old an SGLT2 inhibitor for 20 years and looked at how long they live. The human studies run for around 2 years and measure disease endpoints in sick people. This is the same gap as with acarbose: excellent data on a different question.

It also remains unclear why the mouse finding comes about. Attention focuses on the blood sugar peak after feeding and on insulin signals that run differently in male mice; no one knows for sure. And if you do not know why something works, you also do not know in whom it does not work. What is usable, by contrast, is the common denominator of both ITP hits: that the blood sugar peak after a meal is a more plausible target than the fasting value can be taken seriously even without a pill.

Status, approval and legal

Canagliflozin has been approved for type 2 diabetes for a good decade and was known in Germany as Invokana; the class is additionally regarded as a heart and kidney drug. There is no approval for aging, and the drug is prescription-only. Anyone who obtains it bypassing a doctor gets a substance whose main finding depends on one sex of another species and whose best-known side effect sits poorly with the dietary habits of the target group.

Safety

Two things are in the foreground. First, the known side effects: genital fungal infections are very common, because you excrete sugar, and sugar is food. Added to this are frequent urination and fluid deficiency. In CANVAS there were more lower-limb amputations, 6.3 per 1,000 patient-years compared with 3.4 under placebo; the absolute numbers are small, they were diabetics with an already high amputation rate, and in later analyses the signal could not be consistently confirmed.

Second, euglycemic ketoacidosis, an acidification of the blood that can become life-threatening and in which blood sugar looks normal — the classic warning light does not come on. In EMPA-KIDNEY there were 6 cases under empagliflozin versus 1 under placebo, so it is rare among more than 6,000 people. The triggers are, of all things, things that are popular in this scene: prolonged fasting, very low-carbohydrate diets, hard exertion, surgery. Anyone who takes an SGLT2 inhibitor and does keto at the same time or fasts for three days combines two things that are harmless individually and not together.

Sources

• Neal et al., N Engl J Med 2017 (CANVAS program, canagliflozin in type 2 diabetes)
• Perkovic et al., N Engl J Med 2019 (CREDENCE, canagliflozin in type 2 diabetes and kidney disease)
• EMPA-KIDNEY Collaborative Group, N Engl J Med 2023 (EMPA-KIDNEY, empagliflozin in chronic kidney disease; class evidence, not canagliflozin)
• Miller et al., JCI Insight 2020 (canagliflozin and lifespan in the Interventions Testing Program)
• Miller et al., GeroScience 2024 (late start of treatment with canagliflozin in mice, Interventions Testing Program)
• Rosenstock and Ferrannini, Diabetes Care 2015 (euglycemic ketoacidosis under SGLT2 inhibitors)