Fasting protocols work mostly through calorie restriction. Time-restricted eating shows no independent weight-loss advantage in matched-calorie trials. Pick the protocol you will actually follow. The autophagy-switch narrative is mostly extrapolation from rodent data; the human evidence for a unique metabolic mechanism beyond calorie reduction is thin. None of this means fasting is useless. It means fasting is a structural tool for calorie control, not a metabolic hack.
The trial evidence: Cienfuegos 2020 (n=58) showed that 4-hour and 6-hour TRE windows produced modest weight loss (~3% over 8 weeks) compared to no-intervention controls, with calorie reduction as the apparent mechanism. Lowe 2020 (n=116, the TREAT trial) found no statistical advantage of 16:8 TRE over a 3-meal control, with similar small weight changes in both arms. Across the broader literature, TRE typically produces 2 to 4% body weight loss over 8 to 12 weeks, indistinguishable from any other modest calorie deficit.
5:2 fasting (5 days normal eating, 2 days at ~500 kcal) and alternate-day fasting (ADF) produce larger total deficits and correspondingly larger weight losses, often 5 to 8% over 12 to 24 weeks. They also have higher dropout rates in the trial literature. The signal is consistent: protocol choice should be driven by adherence, not by a metabolic-magic premise.
Where fasting actually earns its margin: people for whom unstructured eating leads to chronic surplus. A defined window or fasting day is cognitively cheaper than calorie counting and produces the same deficit if done consistently. People who eat one large dinner and skip breakfast naturally are already on a 16:8 protocol; formalizing it changes nothing. People who graze across the day and feel anxious about a smaller window will fail it within a month.
The metabolic claims worth tempering: TRE does not uniquely improve glucose control beyond what calorie reduction explains. Autophagy markers do shift on extended fasts, but linking that to clinical outcomes in humans is currently a long inferential chain. None of these caveats argue against fasting; they argue against marketing fasting as something it is not.
- TRE (16:8 or similar): ~2 to 4% weight loss over 8 to 12 weeks, mechanism is calorie reduction.
- 5:2: ~5 to 8% over 12 to 24 weeks; larger deficit, higher dropout.
- ADF: similar to 5:2 in outcomes; tolerability worse for most people.
- Matched-calorie comparisons: TRE shows no independent advantage over 3-meal eating.
- Best fit: people who graze and want a structural rule, or who already eat in a narrow window naturally.
What to actually do
- Pick a protocol on adherence, not on biology. If a 16:8 window feels easy and gets you a 300 to 500 kcal deficit without thinking about it, do that. If you cannot survive without breakfast, do not force a TRE protocol.
- Protect protein and resistance training inside any fasting protocol. The deficit interacts with lean mass the same way it does in any other calorie restriction. 1.6 g/kg/day, 2 to 3 strength sessions/week, regardless of feeding window.
- Stop tracking the autophagy talking point. It is not the lever. The lever is total weekly calories below maintenance. If your fasting protocol delivers that, it works. If it doesn't, no amount of metabolic-flexibility theorizing will rescue it.
Fasting works when it makes calorie control easier. It does not uniquely accelerate fat loss versus matched-calorie eating. For the full trial-by-trial breakdown and the protocol decision tree, see the full article.