Longevity supplement stacks: what happens when you combine unproven things

Why stacks exist

The reasoning is straightforward and superficially sound. Aging has multiple recognized mechanisms, no single compound addresses all of them, so covering several mechanisms should be better than covering one.^[1]

The flaw is in the premise rather than the logic. Covering several mechanisms is only better if covering one does anything, and for most of these compounds that has not been established. Multiplying an unproven effect by five gives you five unproven effects and a larger bill.

What each common component has actually shown

Early / mixed evidence Take a typical stack apart and the picture is consistent. NAD precursors reliably raise blood NAD, and a trial that raised NAD in human skeletal muscle found no improvement in muscle function.^[2]

Spermidine has excellent mechanistic credentials and a delivery problem: a placebo-controlled study found supplementation did not raise plasma spermidine.^[3] Fisetin's reputation rests on a mouse study whose lifespan result the rigorous multi-site testing program did not reproduce.^[4]

Things often look good in mice and then turn out to be bad in people.

Urolithin A is the best of them, with randomized human trials reporting improved muscle endurance and biomarker changes, effect sizes that are modest, and a systematic review concluding that healthspan impact remains unestablished.^[5]

And creatine, which has the strongest evidence of the lot for preserving muscle and strength in older adults, is usually absent because it costs pennies and sounds like something for teenagers.^[6]

The attribution problem

Suppose your stack works. Which component did it? You cannot know, and neither can anyone else, because you changed several variables at once with no control and no measurement.

This is not an abstract complaint. It means a stack cannot accumulate evidence over time, no matter how long you take it or how carefully you feel. Personal experience with a multi-component regimen is uninterpretable by construction, which is precisely why trials test one thing at a time.

The measurement problem

Even if you were willing to attribute loosely, there is nothing to attribute. Autophagy cannot be measured in a living person. Senescent cell burden requires tissue analysis. Tissue NAD requires a biopsy. Biological age tests have test-retest error comparable in size to the changes people celebrate.

So the stack is unfalsifiable in practice. Nothing you can buy will tell you it is not working, which is an uncomfortable property for something you intend to take for decades. Our page on what biological age tests can and cannot measure covers that in detail.

What large supplement trials have generally found

Less than expected, repeatedly. The largest randomized trials of individual supplements in general populations have often failed to find benefits that observational data predicted, including the vitamin D and omega-3 arms of a very large trial that missed their primary endpoints.^[7]

That history is directly relevant. The compounds in longevity stacks have far less human data than vitamin D had when those trials began, and the base rate for supplements surviving proper testing is not encouraging.

Sorry to be cynical: people are very easy to fool.

Interactions nobody has modeled

Resveratrol affects blood clotting and interacts with anticoagulants. Several compounds affect glucose handling in different directions. Berberine lowers blood sugar meaningfully. None of these combinations has been studied for interaction, and each added component increases the number of untested pairings geometrically.

The risk is still low in absolute terms, because most of these compounds are mild. It is not zero, and it is invisible.

A defensible way to think about longevity supplement stacks

Decide what you are buying. If it is a small, affordable bet on two or three plausible mechanisms, taken with clear eyes about the evidence, that is a legitimate personal choice and this page is not an argument against it.

If it is a twelve-item protocol costing several hundred dollars a month, purchased in the belief that it is doing something demonstrated, the evidence does not support that belief for any component of it. Our reviews of the supplement evidence compound by compound and the NAD category specifically apply the same standard, and the Blueprint page covers what happens when this approach is taken to its limit.

Frequently asked questions

What is a longevity supplement stack?

A combination of compounds taken together on the reasoning that each targets a different mechanism of aging. Typical stacks include an NAD precursor, resveratrol, spermidine, urolithin A and a senolytic, often alongside ordinary supplements like vitamin D and omega-3.

Do longevity supplement stacks work?

No stack has been tested as a stack. Every combination in circulation is assembled from individual compounds, most of which have not demonstrated a benefit alone, and combining unproven things does not produce a proven thing.

Is there any evidence for combining supplements?

Essentially none for these combinations. Trials test single compounds because that is the only way to attribute an effect. A stack is untestable by design: if something changes, nothing tells you which component did it.

What would a defensible stack contain?

Judged strictly on evidence, mostly boring things: creatine if you do resistance training, vitamin D if you are deficient, omega-3 for cardiovascular reasons in some people, and protein sufficiency. That list disappoints everyone and it is what the evidence supports.

Why do stacks keep growing?

Because each compound has a plausible mechanism and no compound has a disconfirming result strong enough to remove it. Absence of proof cuts both ways, and in practice it means nothing ever leaves the stack.

Do the compounds interact with each other?

Nobody knows, which is its own problem. Resveratrol affects clotting, several compounds affect glucose handling, and none of these combinations has been studied for interaction. More components means more unmodeled interactions.

Is a stack safer than a drug?

Usually, and that is the honest appeal. Most of these compounds have mild side effect profiles at ordinary doses. Low risk is a reason not to worry, not a reason to expect a benefit.

What does a typical stack cost?

Several hundred dollars a month once it includes NAD precursors and urolithin A, which are the expensive components. That spending has an opportunity cost against things with better evidence, including a gym membership and better food.

Should I take everything Bryan Johnson takes?

That protocol is one person running hundreds of simultaneous interventions with no control group, which makes it uninterpretable by design. Copying an uninterpretable experiment does not transfer a benefit that was never demonstrated.

Which compound in a typical stack has the best evidence?

Creatine, and it is usually not in the stack because it is cheap and unglamorous. Among the longevity-branded compounds, urolithin A has the most real human trial data, with modest effects and a single dominant sponsor.

Which has the weakest?

Spermidine has the strangest problem: two studies found supplementation barely raises plasma spermidine. Fisetin has the most awkward one: the rigorous multi-site mouse program did not reproduce its lifespan result.

Should stacks be cycled?

No cycling schedule for any of these compounds has been tested. Cycling advice is reasoning by analogy from pharmacology that does not necessarily apply.

How would I know if my stack was working?

You would not. There is no accessible marker for autophagy, senescent cell burden or tissue NAD, and biological age tests have measurement error comparable to the changes people celebrate. The stack is unfalsifiable in practice.

What is the sensible approach?

Decide what you are buying. A small number of compounds as an affordable bet on plausible mechanisms is defensible. A twelve-item stack purchased as a longevity protocol is buying certainty that nothing in the evidence supports.