⏰
Intermittent fasting: no magic, but no demons either
Research review · PubMed
✍️ Автор: Катя
Arguments about intermittent fasting run hot in both directions: fans promise cellular rejuvenation, critics promise heart attacks and baldness. We took the loudest claims from each side and checked every one — 99 randomised trials, year-long experiments, genetics and meta-analyses. Spoiler: reality is duller than either version, and considerably more usable.
🔬
Autophagy: the mechanism is real, the human benefit is not established
Let us start honestly: autophagy exists and works in animals. Genetically enhancing it extends mouse lifespan by roughly **17%**. But **no human study has yet shown** that stimulating it improves health. The best trial available — spermidine, 100 people, one year — returned **nothing**. So the mechanism is real; a clinical benefit for humans has not been demonstrated.
⏱️
"16 hours triggers autophagy" is a mouse number
The study everyone cites was done in **mice**, fasted for 24–48 hours. Mouse metabolism runs roughly **seven times faster** and their glycogen burns down incomparably sooner — 24 hours in a mouse is not 24 hours in you. **No fasting threshold has been established for humans at all.**
📉
Worse: in muscle the markers sometimes fall
This is the critics' strongest point, and it holds. One paper is titled, literally, *Intermittent fasting activates markers of autophagy in mouse liver, **but not in mouse or human muscle***. In 50 women after eight weeks on the protocol, muscle markers **decreased**. The only study that measured genuine autophagic flux in humans phrased its own result cautiously: it "**may** increase".
⚖️
Is it better than an ordinary diet? No
The best synthesis available is a network meta-analysis of **99 trials covering 6,582 people**. Against ordinary calorie restriction: time-restricted eating — **no difference**; the 5:2 pattern — **no difference**; alternate-day fasting — 1.3 kg better, which is close to clinically meaningless. Year-long trials agree: 12 months, 139 people, difference not significant.
💪
"It eats your muscle" is overstated
The viral "65% of the weight lost was muscle" figure comes from a single study where the difference in total lean mass was **not statistically significant**, and the "significant" limb measure fell **within the measurement error of the device itself**. All three year-long trials found no difference against ordinary dieting. **The part that matters:** with resistance training and 1.6 g of protein per kg, muscle is fully preserved — meta-analysis even shows a small gain.
🍽️
"It wrecks your appetite" — the opposite, if the window is early
When bodyweight was held stable, an **early** eating window **lowered** ghrelin by 32 pg/mL and made hunger steadier. Ghrelin does rise during weight loss — but it does so **on any diet**. A meta-analysis of 17 trials found hunger levels **no different** from ordinary calorie restriction.
💔
The viral "+135% cardiovascular mortality": what is actually there
The figure is real and peer-reviewed. But the eating window was derived from **two days of dietary recall** and extrapolated across eight years, and the short-window group contained three times more shift workers, smokers and people with food insecurity. Meanwhile a **larger study** (33,052 people) tells a different story: short windows **+34%**, and windows **longer than 15 hours +25%**. The relationship is **U-shaped**, with an optimum around **11–12 hours**. The data do not say "the longer you eat, the better".
💇
The hair loss story
The mouse mechanism is described solidly. The "human clinical study" is **8 people over 10 days**, where hair growth was a **secondary** measure inside a study about glucose. The effect: hair grew **18% more slowly** — which is not shedding and not baldness. The lead author explicitly asked people not to over-read it: "We do not want to discourage people from intermittent fasting."
⚠️
The caveat that matters most — and it is in none of the viral posts
**Disordered eating.** A study of 2,762 people found a consistent association between intermittent fasting and eating-disorder symptoms, strongest in women. The protocol structurally reproduces the restrict-then-binge cycle. This is not abstract: roughly half of young people report practising some form of it. If you have a history of disordered eating, this format is not for you. **Also not advised:** during pregnancy and breastfeeding, at low bodyweight, for adolescents. With insulin-treated diabetes — only with a doctor and dose adjustment.
✅
What to take from this
Intermittent fasting is **one workable way to create a calorie deficit** — no better and no worse than the others. Its real value is **convenience**: some people find skipping breakfast easier than counting portions. If that is you, do it; absent contraindications it is safe. If it makes you miserable, there is no point — you are not missing any magic. Two practical refinements: an **early window** beats a late one, and **protein plus resistance training** matter far more than the clock.
❓ Часті запитання
Does autophagy switch on after 16 hours of fasting?
No threshold has been established for humans. The figure comes from mice, whose metabolism runs about seven times faster. In human muscle, markers of autophagy sometimes decrease on this protocol.
Is intermittent fasting more effective than an ordinary diet?
No. A network meta-analysis of 99 trials in 6,582 people found no difference from classic calorie restriction. Only alternate-day fasting gave 1.3 kg more, which is barely meaningful clinically.
Will I lose muscle?
Any calorie deficit costs some muscle. Studies have not shown any specific harm from intermittent fasting, and with resistance training plus 1.6 g of protein per kg the mass is fully preserved.
Is it true that it raises mortality by 135%?
That comes from an observational study where the eating window was derived from two days of dietary recall. A larger sample of 33,052 people gives +34%, and also shows that windows longer than 15 hours raise risk by 25%. The optimum is around 11–12 hours.
Who should avoid intermittent fasting?
Anyone with a history of disordered eating, during pregnancy and breastfeeding, at low bodyweight, and adolescents. With insulin-treated diabetes, only under medical supervision with dose adjustment.
📚 Дослідження, згадані в статті (10)
- KEY: network meta-analysis of 99 RCTs, 6,582 people — no difference from ordinary dieting (BMJ, 2025) — джерело →
- One-year RCT: eating window plus deficit vs deficit alone, 139 people (NEJM, 2022) — джерело →
- Autophagy markers rise in mouse liver but not human muscle (Nutrition, 2022) — джерело →
- Enhancing autophagy extends mouse lifespan (Nature, 2018) — джерело →
- Spermidine in humans: one year, 100 people, no effect (JAMA Network Open, 2022) — джерело →
- Early eating window lowers ghrelin and steadies hunger (Obesity, 2019) — джерело →
- U-shaped relationship between eating window and mortality, 33,052 people (Aging Cell, 2025) — джерело →
- Meta-analysis of 17 RCTs: hunger no different from ordinary dieting (Nutrients, 2023) — джерело →
- Association with disordered eating, 2,762 people (Eating Behaviors, 2022) — джерело →
- Hair growth suppression in mice, with a small clinical component (Cell, 2024) — джерело →
Це не медична порада. Маєш діагноз чи обмеження — звернися до лікаря.
Чи була стаття корисною?