Not by much, and maybe not at all. When calories were matched, keto groups lost about 1.5 kg more on average than higher-carbohydrate groups in short trials, but the gap was unclear once protein was also matched, and tightly controlled studies found no extra body fat loss on keto31.
The question matters because most keto comparisons let people eat freely. Then the diet that happens to cut more calories wins, and you cannot tell whether the carbohydrate cut did the work.
What was studied
Three kinds of evidence answer this question.
A meta-analysis of matched-calorie trials. A 2026 systematic review searched three databases for randomised trials that gave a ketogenic diet and a higher-carbohydrate diet the same planned calories, or the same planned deficit3. A randomised trial assigns people to a diet by chance. A meta-analysis pools the trials into one estimate. The reviewers defined a ketogenic diet as 50 g of carbohydrate a day or less, or 10% of energy or less. The comparison diets drew about 45% to 70% of energy from carbohydrate.
| What | Number |
|---|---|
| Records screened | 3,959 |
| Trials in the review | 7 (6 pooled) |
| Adults pooled | 259 (127 keto, 132 higher-carb) |
| Who | adults 18–65 with overweight or obesity |
| Trial length | 6 to 52 weeks, most 6–12 |
| Trials that also matched protein | 3 |
A metabolic-ward study. In 2016, 17 men with overweight or obesity lived on a metabolic ward, where every meal is measured1. They ate a high-carbohydrate diet for 4 weeks, then a ketogenic diet with the same calories and the same protein for 4 weeks.
A pooled analysis of feeding studies. A 2017 review pooled 32 controlled feeding studies that swapped carbohydrate for fat at equal calories2.
In plain terms: only trials that held calories equal can tell carbohydrate apart from simply eating less.
What the studies found
The pooled difference was small. Across the 6 matched trials, the keto groups lost 1.49 kg more body weight on average (95% confidence interval 0.58 to 2.41 kg more)3. A confidence interval is the range the true value most likely sits in. The trial results were consistent with each other, and the estimate held when any single trial was left out or the trial at high risk of bias was dropped3.
Protein muddies it. In three of the six pooled trials, the keto diet carried more protein than the comparison diet3. Protein can affect fullness and the energy used in digestion (see the thermic effect of food), so it is a second difference hiding inside "keto vs carbs". In the three trials that matched protein, the estimate was 1.79 kg, but the interval ran from 5.69 kg more to 2.11 kg less on keto. The authors wrote that this "neither confirms nor excludes" an effect of cutting carbohydrate itself3.
Keto minus higher-carb diet, body weight change at matched calories · kg
The certainty is low. No trial was at low risk of bias, six had "some concerns" and one was at high risk3. The reviewers graded the evidence low for risk of bias and imprecision. Their own reading is "a small, predominantly short-term difference in body-weight change", not clear superiority3.
In plain terms: a short-term edge on the scale, of uncertain size, that may not be about carbohydrate at all.
Scale weight is not body fat
The meta-analysis did not pool body fat. Its authors caution that early weight reduction on a very low-carbohydrate diet "may partly reflect depletion of glycogen stores and the water associated with them rather than loss of adipose tissue alone"3. Glycogen is the carbohydrate stored in muscle and liver, and it is stored with water. That is a mechanism the authors raise, not a result they measured.
The ward study measured body fat directly with a DXA scan (a low-dose X-ray that separates fat, lean tissue and bone)1. Overall the men were in a deficit of about 300 kcal a day. On the ketogenic diet, energy use rose slightly: about 57 kcal a day in the metabolic chamber and 151 kcal a day by doubly labelled water, a tracer method for free-living energy use. Yet body fat loss slowed on keto, and loss of fat-free mass rose. The authors described the energy change as near the limits of what their equipment could detect.
Change in energy use after switching to keto at the same calories · kcal/day
The pooled feeding studies point the same way. Across 32 of them, with carbohydrate swapped for fat at equal calories, both energy use (by 26 kcal a day) and fat loss (by 16 g a day) were slightly greater on the lower-fat diets2.
In plain terms: when every calorie is counted, cutting carbohydrate did not melt extra fat.
Why keto can still feel faster
A comment under a keto video says "I lost weight way faster on Keto". The trials offer two honest reasons that do not need a special fat-loss effect. First, the early drop on the scale can include glycogen and water. Second, outside a lab, people often eat less on keto without planning to, because many high-calorie foods are carbohydrate-based. That second point is about intake, which these matched trials were designed to hold equal. For how daily weigh-ins mix water and tissue, see trend weight vs scale weight.
What this means for a macro split
These trials describe groups over weeks to months, not what one person will see. In them, what separated the diets at equal calories was small and uncertain; the amount of energy eaten did most of the work. That fits the broader picture in macro splits compared and keto macros explained.
In Zyra Nutrition's calculator, the diet style changes only how a calorie target is split. Protein is set first, from body weight; the keto style then fixes carbohydrate at 30 g a day and fills the remaining calories with fat. The calorie target is the same whichever style is chosen.
The studies at a glance
| Study | Who was studied | Key result | Year |
|---|---|---|---|
| Hall et al., *Am J Clin Nutr*1 | 17 men with overweight or obesity, metabolic ward, 4 + 4 weeks | Same calories and protein: energy use rose slightly on keto, body fat loss slowed | 2016 |
| Hall & Guo, *Gastroenterology*2 | 32 controlled feeding studies | Equal calories: slightly more energy use and fat loss on lower-fat diets | 2017 |
| Żurawski et al., *Nutrients*3 | 6 pooled trials, 259 adults with overweight or obesity | Keto groups lost 1.49 kg more; unclear in the 3 protein-matched trials; low certainty | 2026 |
Limits worth knowing
- Small and short. Six pooled trials and 259 people; most lasted 6 to 12 weeks, and only one ran a year3.
- Planned, not measured. Trials matched the calories people were told to eat; what they actually ate was mostly self-reported3.
- Body fat not synthesised. The meta-analysis pooled weight only, so it cannot say how much of the gap was fat, glycogen or water3.
- The ward study is tiny. Seventeen men, eight weeks, no randomised order of diets1.
- Abstract only. We read the 2016 and 2017 papers as abstracts. We read the 2026 review in full, but not its forest-plot image or online appendix, so no per-trial figures are shown.
- Population. Adults with overweight or obesity; lean people who lift were not studied here.
Latest research
The newest study we opened is the 2026 meta-analysis, published in August 2026 with searches to September 20253. Its conclusion is careful: under comparable calories, a ketogenic diet "may produce a small additional reduction in body weight", but short trials, uncertain intake and protein, no body fat synthesis and low certainty limit confidence in the size, durability and meaning of that edge. The authors call for longer trials that verify intake and track body composition.
The short version
- With calories matched, keto groups lost about 1.5 kg more on average in short trials.
- Where protein was matched too, the difference was unclear.
- Controlled ward and feeding studies found no extra fat loss from cutting carbohydrate.
- Early keto losses may be partly glycogen and water; calories eaten drive most of the result.
Educational overview only — not medical advice.
