For years, we’ve been having the same conversation with patients:
“Doc, you keep talking about autophagy and cellular cleanup. But how do we actually *know* it’s happening when I fast?”
The honest answer was unsatisfying: “Well, we see it beautifully in mice, and the metabolic markers in humans suggest it’s happening, but...”
That ‘but’ just got a lot smaller.
Researchers at Cedars-Sinai just published something we’ve been waiting for in longevity medicine: the first human trial that directly measures autophagic flux during a fasting intervention. Not metabolic proxies. Not assumptions from rodent studies. Actual human autophagy activation.1
Let me explain why this matters.
The Cellular Housekeeping We’ve Been Banking On
Think of autophagy as your cells’ recycling program. Every day, damaged proteins, dysfunctional mitochondria, and cellular debris accumulate like junk in a garage. Autophagy is the process that tags this junk, packages it up, and breaks it down for parts.
When autophagy works well, your cells stay young and efficient. When it declines — which happens with age, overfeeding, and chronic disease — that cellular junk piles up. The result? Inflammation, metabolic dysfunction, and accelerated aging.
We’ve known from elegant animal research that fasting is one of the most powerful autophagy triggers. But measuring it directly in living humans has been, until now, technically nightmarish. Most studies measured indirect markers like ketones or circulating amino acids and said, “Autophagy is *probably* happening.”
Probably doesn’t cut it when you’re asking patients to go five days on 700 calories.
What the Study Actually Showed
The Cedars-Sinai team enrolled 30 adults in a pilot trial using the Fasting Mimicking Diet (FMD) — specifically Prolon, a 5-day protocol that drops calories to around 700-1,100 per day while maintaining specific macronutrient ratios designed to keep you in a fasted metabolic state.
Here’s what made this study different: they used sophisticated biomarkers to measure autophagic flux directly. This isn’t just “ketones went up” or “mTOR went down.” They tracked the actual turnover of cellular components being degraded through autophagy.
The results? Autophagic activity measurably increased during the FMD period. Not in mice. In humans. In a controlled setting with direct measurements.
Bonus findings that anyone who’s done FMD protocols will recognize: participants also showed improvements in weight (average 5.7 lbs), fasting glucose, insulin sensitivity, and ketone production. But those were the supporting cast. The star of this show was autophagy itself.
Why This Changes the Conversation
I’m not usually one to get excited about pilot studies with 30 people. But this one matters because it validates the mechanistic foundation under an entire category of longevity interventions.
For years, the fasting-for-longevity conversation has rested on three legs: animal data (strong), mechanistic plausibility (strong), and human metabolic improvements (decent). The weakest link was always the question: “But does it actually trigger the cellular repair processes we care about in humans?”
Now we can say yes. Not “probably.” Not “we think so.” Yes.
This doesn’t mean FMD is a magic bullet. The study was small, short-term, and funded by the company that makes Prolon (though conducted by independent academic researchers). We need larger trials, longer follow-up, and head-to-head comparisons with other fasting protocols.
But for the first time, we have human evidence bridging the gap between “this should work” and “this actually does what we think it does at the cellular level.”
The Practical Reality
Here’s what I’m telling patients now: If you’re considering periodic fasting for longevity purposes, this study provides the first direct human evidence supporting the cellular mechanisms we’ve been theorizing about.
The 5-day FMD protocol isn’t easy — my patients (and me personally) report it’s mentally and physically challenging, especially days 2-3. But it’s more doable than water fasting, and now we have human data showing it’s doing what we hoped at the cellular level.
My usual recommendation: try it quarterly if you’re metabolically healthy and looking for a longevity intervention. More often might be appropriate for specific clinical situations, but this isn’t something to do weekly.
The Bottom Line
We finally have direct human evidence that fasting-mimicking diets activate autophagy — the cellular cleanup process linked to longevity and disease prevention. This bridges the critical gap between promising animal research and clinical practice. While we need larger and longer studies, this validates the mechanistic foundation of fasting-based longevity protocols.
For My Fellow Physicians
The methodology here is what excites me. They used LC3-II and p62 measurements — the gold standard autophagy markers — in peripheral blood mononuclear cells. Not perfect (tissue-specific autophagy varies), but far better than the indirect markers we’ve relied on.
Key clinical considerations:
- Screen for contraindications: active eating disorders, underweight BMI, pregnancy, Type 1 diabetes
- Monitor electrolytes if patients are on diuretics or have renal disease
- Consider holding metformin, GLP1s or SGLT2i during the FMD period (hypoglycemia risk)
- The autophagy signal appears strongest on days 3-5, which tracks with the metabolic switch timeline
- This was Prolon specifically, but the principles likely apply to other well-designed FMD protocols
Importantly: this doesn’t prove longevity benefits in humans yet. Autophagy activation is mechanistically linked to longevity pathways, but we’re still connecting dots. Clinical endpoints matter.
Closing Question
Now that we have direct human evidence of autophagy activation, does this change your thinking about periodic fasting protocols? And for those who’ve tried FMD — how did your experience match the science?
References
Espinoza SE, Park S, Connolly G, Qi W, Semwal M, Zhang N, Li Y, Salmon AB, Musi N, Wei M, Hsu W. “Effect of fasting-mimicking diet on markers of autophagy and metabolic health in human subjects.” GeroScience (2025). DOI: 10.1007/s11357-025-02035-4
