Which type of fasting works best? A 99-trial analysis compared them

A 99-trial analysis compared 16:8, 5:2, alternate-day fasting and calorie cutting. All helped with weight loss; differences between types were small.

Published October 2026

Short answer

All the common fasting styles help with weight loss; alternate-day fasting has a small short-term edge, but differences between plans fade within six months. The main safety concern is low blood sugar for people on diabetes medicines. Evidence grade B (Moderate). Network meta-analysis of 99 randomised trials plus two Cochrane reviews; downgraded from A for inconsistency (differences between styles are small and fade in trials of six months or more) and for risk of bias and reliance on indirect comparisons in the network analysis.

Key takeaways

In a 2025 network meta-analysis of 99 trials with 6,582 adults, alternate-day fasting, whole-day fasting (such as 5:2), time-restricted eating (such as 16:8) and daily calorie cutting all beat eating freely, by about 1.7 to 3.4 kg [1].
Alternate-day fasting was the only style that beat daily calorie cutting, by about 1.3 kg — less than the 2 kg the authors set as a meaningful difference [1].
In the 17 trials lasting six months or more, no fasting style was better than another or than daily calorie cutting; only the advantage over unrestricted eating remained [1].
Side effects are usually mild and early: a review of 15 trials found no more tiredness or headaches and no more dropouts than with control diets [2], but people with type 2 diabetes on glucose-lowering medicines can get low blood sugar [3].
A national US survey linked eating windows of 8 hours or less to a third higher death rate, especially in older adults, but surveys cannot prove cause and effect [4].

What the study did

In a 2025 network meta-analysis, researchers pooled 99 randomised trials with 6,582 adults to compare popular types of intermittent fasting: alternate-day fasting, time-restricted eating (such as 16:8) and whole-day fasting (such as 5:2), against daily calorie cutting and unrestricted eating [1]. A network meta-analysis compares treatments even when they were not tested directly against each other, and this one rated the certainty of its findings with the GRADE system [1]. Ninety-three of the trials reported body weight, the main outcome [1].

What it found

Every plan beat eating freely. Weight loss against unrestricted eating was about 3.4 kg with alternate-day fasting, 2.4 kg with whole-day fasting, 2.1 kg with daily calorie cutting and 1.7 kg with time-restricted eating, with moderate to high certainty [1]. Only alternate-day fasting beat daily calorie cutting, by 1.3 kg (95% CI 0.6 to 2.0) — less than the 2 kg the authors set as a meaningful difference [1]. Alternate-day fasting also beat time-restricted eating by 1.7 kg and whole-day fasting by 1.1 kg, while time-restricted eating and whole-day fasting were level with each other [1].

In the 17 trials that lasted six months or longer, the differences between styles disappeared: every plan worked about equally well, and only the comparisons with unrestricted eating stayed significant [1].

Blood markers told a similar story. No plan changed HbA1c or HDL cholesterol, and all lowered fasting glucose and insulin resistance slightly compared with eating freely [1]. Alternate-day fasting lowered total and non-HDL cholesterol more than time-restricted eating [1].

How it fits

Earlier reviews reached similar conclusions. A 2023 network analysis of 24 trials ranked alternate-day fasting first but found fasting overall matched calorie cutting (0.26 kg difference, not significant); it was funded by Nestlé, whose staff were among the authors [5]. A 2024 review of 10 trials found fasting gave about 0.9 kg more weight loss than calorie cutting in the short term, which the authors did not consider clinically significant, and no long-term advantage [6].

The 2026 Cochrane review of 22 trials found intermittent fasting made little to no difference to weight loss compared with regular dietary advice (low certainty), but 3.4 percentage points more than no intervention (moderate certainty) [7]. One year-long trial stands out: in 165 adults with obesity, 4:3 fasting beat daily calorie cutting by 2.9 kg, with both groups in an intensive behavioural programme [8].

Direct comparisons of fasting styles are still rare. In a 4-week trial of 76 adults, alternate-day fasting cut total fat volume by about 700 cm3 more than 16:8 eating, and only alternate-day fasting reduced visceral fat and non-HDL cholesterol; the trial was small and short [9].

Is it harmful?

Most people tolerate these plans well. A meta-analysis of 15 trials with 1,365 adults found no more tiredness, headaches or dropouts than in control groups [2]. Another review found slightly more vomiting (risk ratio 1.11) and irritability (risk ratio 1.22) than controls, mostly early on and resolving without treatment [10].

The clearest risk is low blood sugar for people with type 2 diabetes on glucose-lowering medicines. In a 12-week trial, fasting roughly doubled the rate of low blood sugar events (risk ratio 2.05, 95% CI 1.17 to 3.52), even though participants' medicines were reduced [3].

People with eating-disorder traits should be cautious: a 2025 review found more preoccupation with food and a higher risk of relapse, and possible mood destabilisation or anxiety in vulnerable people, while randomised trials reported few psychiatric problems [11].

Two US national surveys suggest a U-shaped link between eating windows and death rates. In 33,052 adults followed for a median of eight years, eating windows of 8 hours or less were linked to 34% higher all-cause mortality (95% CI 1.07 to 1.67) and more than 50% higher cardiovascular mortality in older adults; the lowest risk was at an 11-12 hour window [4]. A second analysis of 10,561 adults over 60 found the longest nightly fasts (over 12.4 hours) were linked to higher cardiovascular mortality, with the lowest risk at about 11.5 hours [12]. These are observational studies: people who are already ill may eat in shorter windows, so they cannot show cause and effect [4].

Pregnancy data are thin: one trial tested a 10-hour eating window plus exercise before and during pregnancy in 167 people at higher risk of gestational diabetes, found no effect on blood sugar at week 28, and reported no safety outcomes [13].

Limits

The evidence measures weight and blood markers, not disease. No trial in the 2021 Cochrane review reported heart attacks, strokes, heart failure or deaths [14], and only five trials in the 2025 network analysis ran for a year or longer [1]. Short trials may make the differences between diets look bigger than they are: 3-week trials mainly show early weight loss of about 0.5 to 2 kg [1].

Adherence is the weak point. In one 52-week trial, adherence to whole-day fasting fell from 74% at six weeks to 22% at a year [1]. Only three trials have directly compared two fasting styles; many style-versus-style results rest on indirect comparisons, which the authors flag as a source of uncertainty [1].

Funding and ties deserve scrutiny: several authors of the 2025 network meta-analysis declared research funding, speaking fees or consulting from food, sweetener and drug companies [1], and the 2023 comparison review was funded by Nestlé, which employed several authors [5]. The year-long 4:3 trial was funded by the US National Institutes of Health [8].

Go deeper

Intermittent Fasting — Intermittent fasting helps adults with overweight or obesity lose weight: pooled trials show about 3 kg more loss than control groups, and it works about as we…