Fasting and Time-Restricted Eating

The hype far outran the human RCTs. Once they caught up, intermittent fasting turned out to be roughly as effective as continuous calorie restriction — no more, no less. Useful for some, but not the longevity breakthrough it was sold as.

Intermittent fasting (IF) has been the most popular dietary "movement" of the past decade. The honest evidence summary: it works for weight loss to roughly the same extent as continuous calorie restriction, with some specific benefits for glycemic control and metabolic flexibility, but it's not a longevity breakthrough.

The four main protocols

Time-restricted eating (TRE)

Eat within a specific daily window (typically 6–10 hours).
Most popular: 16:8 (16-hour fast, 8-hour eating window).
Early TRE (eating window ending ~17:00) is metabolically superior to late TRE (eating window ending after 19:00) — even at matched calories.

Alternate-day fasting (ADF)

Alternating days of normal eating and severely restricted eating (~25% of energy needs).
Often produces ~5–8% weight loss over 8–12 weeks; adherence drops significantly long-term.

Periodic fasting (PF)

Multi-day fasts (24–72 hours, sometimes longer).
Stronger autophagy signals; significant practical and adherence challenges.

Fasting-mimicking diet (FMD)

5-day low-calorie, low-protein, plant-based protocol (Valter Longo's protocol).
Designed to trigger fasting biology while still allowing some food.
Typically 4 cycles per year.

What the evidence actually says

Weight loss

A 2026 Cochrane review of 22 RCTs (n=1,995) found IF did not produce significantly more weight loss than standard dietary advice or continuous calorie restriction.^[1]
The 2025 BMJ network meta-analysis of 99 RCTs (n=6,582) found all IF strategies produce weight loss vs. ad-libitum eating, but the difference vs. continuous restriction is marginal. Alternate-day fasting was the only strategy with a slight statistical advantage (−1.29 kg) over continuous restriction in shorter trials. Differences vanished entirely in trials >24 weeks.^[2]

Glycemic control

Modest improvements in fasting glucose, insulin, and HbA1c.
Strongest in T2D and prediabetic populations. Meta-analyses show IF as adjuvant therapy reduces HbA1c by ~0.5% (oral hypoglycemic users) up to ~2.8% (insulin-dependent — with strict supervision required to avoid hypoglycemia).^[3]
Early TRE specifically shows the most consistent improvements in fasting insulin (−3.32 μIU/mL vs. late TRE).^[4]

Autophagy

Single FMD cycles induce measurable autophagy markers (LC3B-II/I ratio) in human peripheral blood cells, persisting through refeeding.
Ramadan fasting upregulates Beclin-1 and downregulates LC3B / p62 — biological markers of active autophagy.
However, age-related decline in autophagic capacity means older adults respond less robustly than younger adults.

Cardiometabolic risk

Modest improvements in BP, lipids, and inflammatory markers across most protocols.
Less consistent than for weight or glucose.

Longevity (hard outcomes)

No human RCT has tested whether IF extends lifespan.
Animal data (rodents, nematodes, yeast) consistently show lifespan extension under various caloric restriction and fasting protocols, but human translation is uncertain.

Concerning signals

A 2024 American Heart Association abstract reported a 91% increase in cardiovascular mortality in those with eating windows <8 hours in a NHANES analysis. Still observational and contested, but worth knowing.^[5]
Severe TRE (<6 hour eating windows) sustained long-term has not been adequately studied for safety.
Refeeding after extended fasts has been associated in some animal models with paradoxical increases in tumorigenesis — the "refeeding paradox". Human relevance unclear.^[6]

Early vs. late time-restricted eating

This is one of the more robust findings. Multiple meta-analyses converge:

Early TRE (last meal before 17:00, eating window concentrated in morning/midday) is associated with:
- Lower fasting insulin
- Better body composition
- Improved glycemic control
- Better sleep quality (when last meal is well before bedtime)
Late TRE (last meal after 19:00):
- Modest improvements over no restriction
- Less metabolic benefit despite same total fasting hours
- Worse sleep architecture

Why? Insulin sensitivity declines through the day. Glucose tolerance is much better at breakfast than at dinner. Late eating also pushes core body temperature and pre-sleep metabolic activity, degrading sleep.

Practical implication: A breakfast + lunch + light early dinner pattern aligns with circadian biology. "Skip breakfast, big late dinner" may be the worst chronotype for IF.

Resistance training during fasting

The training-fasting interaction matters:

Resistance training requires adequate protein and energy intake — fasted training is fine, but post-training feeding (with protein) supports adaptation.
A 16:8 eating window is fully compatible with hypertrophy — most adults can hit ~150 g protein in an 8-hour window if they plan.
Pre-bed protein (40 g casein or similar) has evidence (Snijders 2015) for overnight muscle protein synthesis. This conflicts with strict early-TRE protocols.

The reasonable compromise: most days do early TRE (eating window 8–10 hours, last meal ~17:00–19:00), but on heavy training days, allow flexibility — including a post-training meal even if it pushes the window slightly later.

Who should not fast

Pregnancy and lactation
Type 1 diabetes (without medical supervision)
History of eating disorders — IF can trigger or reinforce restrictive patterns
Underweight individuals
Children and adolescents
Insulin or sulfonylurea users — extreme hypoglycemia risk; only under medical supervision
Frail older adults — protein intake is more important than fasting; sarcopenia risk

Practical protocols

For most healthy midlife adults, the simplest evidence-based starting point:

Daily early TRE

12–14 hour overnight fast (e.g., last bite at 19:00, first food at 7:00–9:00 the next morning)
Aligns with circadian biology
Effortlessly compatible with normal social patterns
Low downside risk

More aggressive: 16:8 with early window

Eat between 9:00–17:00 or 10:00–18:00
Strong metabolic signal
Compatible with morning training

Periodic deeper fasts

A 24-hour fast (e.g., dinner-to-dinner) once weekly is a moderate intervention
Or 1 FMD cycle (5 days) per year for those drawn to the autophagy framework

What to skip: Aggressive ADF, multi-day water-only fasts without medical supervision, OMAD (one meal a day) sustained long-term — adherence and metabolic side effects are real.