Intermittent fasting covers several patterns: time-restricted eating, where all food falls in a window of eight to ten hours; alternate-day fasting; and whole-day fasting such as 5:2. The appeal is a rule about when rather than what, which some people find easier to keep.

The claim that fasting does something beyond reducing calories, through metabolic switching, autophagy or circadian alignment, is where the enthusiasm lives. The trials that tested it against plain calorie restriction are the ones that matter, and they have been run.

Any pattern that reduces total intake causes weight loss, and a narrow eating window usually reduces intake without counting. The additional proposals are that extended fasts shift metabolism toward fat oxidation and ketone production, trigger cellular recycling processes, and that eating earlier in the day aligns intake with circadian rhythms in insulin sensitivity. Each is supported in animals and in short human physiology studies. Whether they add to weight loss over months is the question the clinical trials address, and mostly they find that they do not.

Against calorie restriction: no advantage. In a 2022 New England Journal of Medicine trial, 139 adults with obesity were randomized to a year of calorie restriction either within an eight-hour window or spread across the day. Weight loss was 8.0 kg with the window and 6.3 kg without, a difference that was not significant, with no differences in waist, body fat, blood pressure or metabolic risk factors.

Without calorie restriction: little. The TREAT trial randomized 116 adults with overweight to a 16:8 window or three meals a day for twelve weeks with no calorie target. Weight fell by 0.94 kg in the fasting group and 0.68 kg in the control, not significantly different, and the fasting group lost more lean mass.

In metabolic syndrome: a small gain. A 2024 trial in the Annals of Internal Medicine gave 108 adults with metabolic syndrome three months of standard nutritional counseling with or without a personalized eight-to-ten-hour window. HbA1c improved by a tenth of a percentage point more with the window.

Across patterns. A 2025 network meta-analysis in the BMJ of 99 trials in 6,582 adults found every fasting pattern and continuous calorie restriction reduced weight versus eating freely. Compared with continuous restriction, only alternate-day fasting did better, by 1.29 kg, with moderate certainty; time-restricted eating and whole-day fasting did not. In trials of 24 weeks or longer, the differences between fasting and restriction disappeared.

For most adults the risks are hunger, irritability and headache in the first weeks, and the lean mass loss seen in TREAT, which argues for keeping protein and resistance training in place. The groups for whom fasting is a clinical matter are people on insulin or sulfonylureas, who risk hypoglycemia; people with a history of eating disorders, in whom rules about when to eat can become something else; pregnant and breastfeeding women; and older adults at risk of frailty. The trials excluded most of these groups.

Not a product. Fasting apps and programs are sold without any regulatory review of their claims. No clinical guideline recommends intermittent fasting over other forms of calorie restriction, and the diabetes guidelines that mention it do so with the hypoglycemia caution above.