Autophagy and the 16-hour fast: what is real, and what is marketing

What autophagy is, briefly

Autophagy is the process by which cells break down and recycle their own damaged components. It runs continuously in every cell, all the time. It is not a mode you enter.

Yoshinori Ohsumi won the 2016 Nobel Prize in Physiology or Medicine for working out its mechanisms. That work was done in budding yeast, and culminated in identifying the genes responsible. It is foundational biology. It says nothing about how long a person should go without breakfast.

The measurement problem nobody mentions

Here is the fact that reframes this entire topic: until around 2020, there was no way to directly measure autophagic flux in living humans. The researchers who developed the first method said so plainly in print.

Everything published before then measured static markers, most commonly a protein called LC3-II, in a muscle or blood sample. The trouble is that LC3-II is a snapshot of a pool, not a rate. It goes up either because the cell is making more autophagosomes, or because the cell has stopped clearing them. Those are opposite meanings, and telling them apart requires an additional experimental step that not one human fasting study has used.

So when an article tells you autophagy increased by some percentage at some hour, the underlying measurement usually could not distinguish "more recycling" from "recycling has stalled".

What the human studies actually found

There are only a handful, and they are smaller and more equivocal than their reputation.

The most cited is a 2014 study of eight healthy men fasted for 72 hours. LC3B-II rose about 30%, but a second marker, p62, rose 10% as well. The authors themselves wrote that this made interpretation of autophagic flux problematic. The paper most often used to prove human fasting autophagy contains its authors declining to claim it.

The only study that measured across a fast at multiple timepoints took muscle biopsies at 2, 12, 24 and 36 hours. It brackets the supposed 16-hour threshold precisely. It reported no abrupt change around any particular hour, and in untrained participants the markers went down. The authors described the effects as modest.

A 2019 study frequently cited as proof measured a single gene transcript in whole blood in eleven people over four days, and in the same study a competing signal moved in the opposite direction.

The best-measured human study, published in 2025, is the first to use real flux measurement in a decent sample of 121 adults over six months. It found a difference against standard care at six months, but no significant increase from baseline within the fasting group, and the authors describe the finding as exploratory.

Where the 16-hour number probably came from

The most likely ancestor is a 2010 mouse study titled "Short-term fasting induces profound neuronal autophagy". In that paper, short-term meant 24 to 48 hours, in a mouse. A mouse metabolises several times faster than a person, so a 24-hour mouse fast is not a 24-hour human fast in any meaningful sense.

A second problem: the two foundational mouse papers disagree with each other. One found the brain showed no autophagic response even at 48 hours. The other reported profound neuronal autophagy. Popular accounts cite the second and never mention the first.

A direct side-by-side comparison using the same assay in both species found that in mice, every leukocyte population responded to fasting. In humans, only neutrophils did. The mouse response is both larger and more general than the human one.

The assumption that more autophagy is better is also untested

Almost every article on this subject takes for granted that increasing autophagy is desirable. In humans, that has not been established either.

Under severe energy deficit, one study found autophagy activation was the mechanism driving muscle wasting, and the investigators proposed inhibiting it rather than encouraging it. Basal autophagic flux also appears to rise with age rather than fall, which sits awkwardly with the popular framing of autophagy as a youth-restoring process in decline.

And if you want a lever that does measurably affect autophagy in human muscle, the evidence points at exercise intensity more strongly than at diet.

What is genuinely unknown

No measured time course of human autophagic flux across a fast exists. The trial designed to produce one over three days has published a protocol and no results.

No threshold hour has been established, and no study has looked for one.

Human data exist only for skeletal muscle and blood, the two tissues you can reach with a needle. I am not aware of any published human liver or brain autophagy measurement during fasting, and there cannot ethically be one in healthy volunteers. Every claim about fasting and autophagy in the human brain is extrapolated from rodents.

And no trial has linked any measured human autophagy change to a clinical outcome. Not weight, not longevity, not disease risk.

What fasting reliably does, which is the better reason anyway

Here is the part that gets lost in the autophagy argument. The effects that have been measured properly in humans are worth having, and several of them arrive well before hour 16.

Fat release roughly doubles between hours 12 and 24, and about 60% of that entire increase happens inside that window. Around 70% of the fall in insulin also happens in the first day. Growth hormone secretion rises roughly fivefold after two days, driven by more frequent and larger pulses, and that rise measurably limits protein loss: when researchers blocked it during a 40-hour fast, nitrogen loss jumped by half.

Eating earlier in the day appears to help more than eating in a narrow window. In a controlled five-week study of eight men with prediabetes, finishing dinner before 3pm improved insulin sensitivity, beta-cell responsiveness and blood pressure, with weight deliberately held constant. Eight people is a very small trial, but it is an effect that does not depend on losing a pound.

And the plainest benefit of all is behavioural. A shorter eating window helps a lot of people eat less without counting anything, which is exactly why it works when counting has failed. That is the reason most people who succeed at this succeed.

So should you fast?

If autophagy is your only reason, you are choosing a behaviour on the strength of a mechanism nobody has measured the way you are imagining, at a threshold nobody has identified.

But that was never the strongest argument for fasting, just the most repeated one. Fat oxidation rises, insulin falls, growth hormone climbs, and for a great many people a shorter eating window is simply the first approach to food that has ever felt sustainable.

Fast because the structure suits your day and you can keep it up. That reason survives contact with the evidence, which is more than the 16-hour claim manages.