Intermittent fasting vs plain calorie restriction: the head-to-head trials

Fasting and plain calorie counting come out about even. Across 99 trials in 6,582 adults, only one fasting schedule beat calorie counting on weight, and only by 1.29 kg: fasting every other day.¹ In the trials that ran 24 weeks or longer, even that edge was gone.¹

The comparison, run 99 times

One 2025 analysis put every version of the question into a single frame. It gathered 99 trials covering 6,582 adults and set four approaches against each other and against eating freely: alternate-day fasting, a daily eating window, whole-day fasting, and plain daily calorie restriction.¹

Every restricted approach beat eating freely. Against daily calorie restriction, only alternate-day fasting showed a weight advantage, 1.29 kg, with a plausible range of 0.59 to 1.99 kg, on evidence the reviewers graded moderate.¹ It also edged out a daily eating window by 1.69 kg (range 0.88 to 2.49) and whole-day fasting by 1.05 kg (range 0.19 to 1.90), gaps the authors themselves call trivial.¹

Alternate-day fasting vs other approaches: weight difference, kg0 = no effect

Vs. daily calorie restriction-1.29 (-1.99 to -0.59)

Vs. a daily eating window-1.69 (-2.49 to -0.88)

Vs. whole-day fasting-1.05 (-1.90 to -0.19)

Then the finding that matters most. Among the trials running 24 weeks or longer, the only advantages left standing were over eating freely.¹ Fasting's edge over counting is a short-trial effect.

That sharpens an older result rather than overturning it. An earlier analysis gathered trials lasting 8 to 24 weeks, all of which cut fast days to a quarter of daily energy needs or less.⁶ Weight loss came out the same either way, a difference of 0.61 kg, which is noise.⁶ Almost nothing else separated the two approaches. Fasting insulin fell slightly more on the intermittent plans, and the authors call the clinical relevance of that uncertain.⁶

The twelve-month head-to-head

The cleanest single trial fixes the calories and varies only the clock. It gave 139 adults with obesity the same prescription, 1,500 to 1,800 kcal a day for men and 1,200 to 1,500 for women, and told half of them to eat it all between 8 a.m. and 4 p.m.³ It ran a full 12 months.

Weight fell 8.0 kg with the eating window and 6.3 kg without one.³ The gap of 1.8 kg was not reliable: the range around it runs from -4.0 to 0.4 and takes in zero.³ Waist size, body fat, lean mass, blood pressure and the metabolic markers all followed the same pattern, and side effects did not differ.³

Eight hours is a demanding window. Twelve months is a long trial. The answer was still no.

An eating window with nothing attached

Take the calorie target away and the effect shrinks further. One trial asked 116 adults to eat only between noon and 8 p.m., or to eat three structured meals a day, for 12 weeks.⁴ Weight fell 0.94 kg in the window group and 0.68 kg in the comparison group, and the difference between them came to 0.26 kg, with a range from -1.30 to 0.78 that spans zero.⁴

Eating window vs same-calorie comparison: weight difference, kg0 = no effect

12-month trial, set calories-1.8 (-4.0 to 0.4)

12-week trial, no calorie target-0.26 (-1.30 to 0.78)

Two details matter more than that headline. Estimated energy intake did not differ between the groups, and eating less is the whole mechanism a window is meant to supply.⁴ And in a 50-person subset, lean mass in the arms and legs came out slightly lower on the fasting schedule, by 0.16 kg for every square meter of height.⁴

Where fasting did win something

Two trials found more than the combined analyses do, and both are informative about why.

In adults with obesity and type 2 diabetes, 75 people were assigned to an 8-hour eating window with no calorie counting, to cutting calories by a quarter, or to a control group, for 6 months.⁵ Weight fell meaningfully in the window group, by 3.56% against controls, and not meaningfully in the calorie group, at 1.78%.⁵ Long-run blood sugar fell in both, by 0.91% and 0.94%, with nothing to choose between them.⁵

The reason sits in the intake data. The window group cut 313 kcal a day and the calorie-counting group cut 197.⁵ People told to watch the clock ate less than people told to count. Neither instruction does anything except by getting someone to eat less, and here the simpler instruction did that better.

A second trial combined both ideas in 209 adults at raised risk of type 2 diabetes, and set them against a group eating 70% of their energy needs every day.⁷ Support ran for 6 months and follow-up out to 18.⁷ The fasting group's blood sugar handling improved more at 6 months, and by 18 months that difference had gone.⁷ They also reported more fatigue than either comparison group.⁷

The adherence argument nobody collected

The strongest case for fasting has always been practical rather than metabolic. The usual argument is that a rule about hours is easier to keep than a running total in your head. That claim is close to untested.

A 2026 Cochrane review brought together 22 studies with 1,995 participants, comparing fasting against ordinary dietary advice, no intervention, or a waiting list.² Against ordinary advice, fasting made little or no difference to the share of body weight lost: 0.33 of a percentage point, on weak evidence.² It made little or no difference to the chance of reaching a 5% reduction, or to quality of life.² Whether it brought more side effects was very uncertain, at a risk ratio of 1.45.²

Then the line that matters most here. None of the included studies reported how satisfied participants were.² The one advantage fasting is supposed to hold over calorie counting is the outcome the trials did not measure.

The diabetes trial caught it in passing. The fasting group stayed inside its eating window on 6.1 days a week, while 68% of the calorie-restriction group met their calorie goals across the 6 months.⁵ That is a single trial in 75 people, and it is most of what exists.

What remains unknown

Duration is the whole gap. The Cochrane studies looked out to 12 months, which its authors say limits what the evidence can be used for.² The large comparison located fasting's advantage only in trials shorter than 24 weeks.¹ The one trial that followed people to 18 months watched its own 6-month difference evaporate.⁷

Body composition is the second gap. One trial found less lean mass in the arms and legs on a fasting schedule, and no analysis here holds enough data to say whether that repeats.⁴

The third is that the differences in play are small enough to be swamped by who sticks around. In the twelve-month comparison, 118 of the 139 people who started made it to the final visit.³ A gap of one or two kilograms between two diets over a year could easily be swamped by that much dropout.