Spermidine, Sleep and Autophagy: Part Four: The Trial Now Recruiting

Abstract

Three parts of negative and ambiguous evidence would ordinarily close a topic. Instead, the group that produced the largest null result has come back with a different question, a higher dose and better instruments.

A trial recruiting at University Medicine Greifswald is testing whether 6 mg of spermidine a day changes the electrical architecture of sleep — power spectra and sleep spindles recorded from the scalp — in older adults at risk of cognitive decline.

This part explains why that is a better test than the ones before it, and what would have to be true for it to succeed.

1.The Trial

Registry entry NCT07383311, "Autophagy-Enhancers to Reduce Sleep Disturbances: A Combined Approach", began recruiting on 28 October 2025 and is scheduled to run to December 2028.

It is randomised, parallel-group and double-masked, with 76 participants aged 55 to 70 who have subjective cognitive decline or mild cognitive impairment, alongside healthy controls. The intervention is 6 mg of spermidine daily across three doses, against a placebo of maltodextrin, rice extract, microcrystalline cellulose, citric acid and silicon dioxide.

Six primary outcomes are measured at twelve weeks: EEG mean power spectra, sleep spindle count, sleep spindle power, and computer-based tasks covering sleep-related alertness, visual-spatial memory and verbal memory.

No questionnaire appears among them. That single design choice separates this trial from most of the sleep-supplement literature, including the one our L-theanine series examined, where subjective measures moved and devices recorded nothing.

2.Why Spindles

Sleep spindles are brief bursts of oscillation during non-REM sleep and among the better-characterised physiological correlates of memory consolidation. Counting them is a measurement of a process, not a report of an experience.

The rationale runs from autophagy to neuronal maintenance to sleep architecture to cognitive trajectory. Impaired sleep is a modifiable contributor to cognitive decline, so an intervention improving sleep architecture might slow that decline.

That is a chain with four links, and the preceding parts of this series have tested three of them with mixed results. What this trial adds is a measurement at the end of the chain that cannot be produced by expectation alone.

3.The Dose, and the Objection It Answers

SmartAge delivered 0.9 mg a day, which its authors calculated as raising daily spermidine supply by around ten per cent. A null from an increment that small tells you little about the compound.

Greifswald uses 6 mg — roughly seven times as much, and above the 3.3 mg used in the Austrian dementia trials described in Part Two.

This is what a field correcting itself looks like. A commentary on the null argued that dose was the likely explanation; the follow-up trial raised the dose. Whether 6 mg is enough remains unknown, but the objection has been answered rather than ignored.

4.The Marker That Makes It Interpretable

Among the secondary outcomes is eIF5A hypusination — the molecular modification through which spermidine is thought to act, and the step that would confirm the polyamine pathway was engaged. Blood polyamine concentrations are also measured.

This matters because of Part Two. If oral spermidine is converted to spermine during absorption and blood spermidine does not rise, a null result would be uninterpretable — nobody could say whether the hypothesis failed or the delivery did. Measuring hypusination removes that ambiguity for the blood compartment.

It does not remove it for the brain. The blood–brain barrier problem from Part Two stands: under normal conditions spermidine largely does not cross it, and no blood marker can show what happened above the neck.

So the honest reading is that this trial can distinguish two of the three possible failure modes, which is two more than most.

5.The Group, and Why That Is a Good Sign

Agnes Flöel, the senior author of SmartAge, holds appointments at the German Center for Neurodegenerative Diseases in Greifswald and at University Medicine Greifswald, the sponsor of this trial.

This is the same research programme returning to a compound after publishing its own negative result on it, with a different endpoint, a higher dose and a target-engagement marker. Nobody was obliged to do that; the easy path after a null is to move on and let others cite the exploratory findings.

We have criticised industry-adjacent research repeatedly in this section. It is worth being equally explicit when a group behaves well.

6.How We Would Read the Result

Do target engagement and outcome agree? If hypusination shifts and the spindles do not, the hypothesis has been tested properly and failed. If neither shifts, the trial has established only that 6 mg is still too little. Those are different conclusions and both will appear in the same paper.

Which primary outcome is led with? Six were pre-specified. A report opening on one of the three cognitive tasks rather than on the EEG measures should prompt a look at what the EEG measures did.

And are adverse events reported in full? Part Three describes apoptosis induction in healthy animals at doses that do not translate. A twelve-week human trial at 6 mg cannot address that, but it can report completely.

The timing is the last thing. Completion is scheduled for December 2028. Anything sold on the strength of this trial before then is being sold on a registry entry, which records that somebody is asking a question.

Editorial Comment

MindHeaven® does not use spermidine in any formulation, has no plans to, and has no commercial interest in how this trial turns out. Spermidine has no authorised health claim in the European Union.

Across four parts the evidence reads as follows. The largest randomised trial found nothing at a dose that raised intake by a tenth. The observational literature contradicts itself, partly for reasons of sample handling and absorption chemistry. The mechanism, when verified as working in animals, did not produce benefit and coincided with harm in healthy ones. And a better trial is now running, three years from an answer.

That is not a case for the compound. It is a case for the group testing it, which is a different thing and worth distinguishing.

If your sleep has stopped working, none of this is relevant to you and will not be for years. Persistent sleep problems have identifiable causes — sleep apnoea, thyroid function, depression, medication timing, alcohol — and finding the cause is worth considerably more than any supplement currently on the market for it. That is a conversation for a doctor, not a trial to wait for.