No, not in the trials so far. A 2026 meta-analysis pooled 11 randomised trials in which lifters ate more or less carbohydrate with protein kept the same, and the higher-carbohydrate groups did not gain significantly more muscle1. The certainty is low, and a small benefit is still possible.
That is a narrower answer than "carbs don't matter". This article covers what was compared, what the numbers show, why whole-body scans can flatter carbohydrate, and where carbohydrate does show up: in how much work you can do in a session.
What was studied
Carbohydrate is the starch and sugar in foods like bread, rice, fruit and pasta. Hypertrophy is the scientific word for muscle growth. The question the 2026 review asked was simple: when protein is the same, does more carbohydrate grow more muscle1?
The reviewers searched four databases up to June 2025 and kept only randomised controlled trials in healthy adults aged 18 to 65 who did resistance training for at least 6 weeks1. The key filter: any trial where protein intake differed between groups, by design or by chance, was left out. Protein matched means both groups got about the same protein, so protein cannot explain a difference.
| What | Number |
|---|---|
| Randomised trials | 11 |
| People per trial (average) | 20.6 |
| Training length | 6 to 12 weeks (average 8.5) |
| Who | mostly trained lifters; men only in 7, mixed in 3, women only in 1 |
| Keto diet vs usual diet | 8 trials |
| Carbohydrate added around training | 2 trials |
| One snack swapped for another | 1 trial |
| Muscle measured directly (ultrasound) | 2 trials |
Most trials put a ketogenic diet, one with almost no carbohydrate, against a usual diet1. Two gave a post-workout drink with or without added carbohydrate. The other nine measured whole-body lean mass, meaning everything a scan counts as not fat, with DXA (a low-dose X-ray body scan), bioimpedance or air displacement.
In plain terms: small groups of mostly trained lifters, for about two months, with protein held equal.
What the studies found
The review's yardstick is the standardised mean difference (SMD). It is the gap in muscle gained between the two groups, divided by how much people differed at the start. Zero means no difference. The review's authors used 0.16 as the threshold for a small effect in resistance-training research1.
Pooled across all 11 trials, the SMD was 0.15 with a 95% confidence interval of -0.10 to 0.40, p = 0.231. A confidence interval is the range the true value most likely sits in. This one includes zero, so the result is not statistically significant. The trials agreed with each other (I² = 0, meaning no measurable spread between them beyond chance).
Muscle growth, higher vs lower carbohydrate, per trial · SMD
Three checks gave the same picture1:
- Leave one out. Dropping any single trial kept the pooled effect between 0.08 and 0.21, never significant.
- Same total energy. The 3 trials that also matched calories (isocaloric) gave an SMD of 0.15 (p = 0.60).
- Measured in the muscle. The 2 ultrasound trials pooled to -0.26, leaning slightly toward less carbohydrate. With two trials, no test was run.
Pooled effect on muscle growth, by type of trial · SMD
An earlier review points the other way, at least on the scan. A 2022 meta-analysis of 13 trials in people doing resistance training found keto diets gave a group-average fat-free mass 1.26 kg lower than non-keto diets (95% CI -1.82 to -0.70)4. That review did not require protein or energy to be matched, and fat-free mass includes water, so it answers a looser question.
In plain terms: across these trials, lifters on more carbohydrate did not grow measurably more muscle than lifters on less.
Why whole-body scans can flatter carbohydrate
Muscle stores carbohydrate as glycogen, and glycogen is held with water. The review's authors note that low-carbohydrate diets lower glycogen and its water, which "can transiently decrease muscle volume without affecting contractile tissue"1. A whole-body scan counts that water as lean mass.
One trial showed this directly. After 10 weeks of a keto diet, lifters who brought carbohydrate back in the final week gained an extra 4.8% fat-free mass in that single week, which the authors attribute to glycogen refilling1.
Energy intake is the other confounder. Keto diets tend to lower how much people eat without trying. In the review's own post-hoc check, the lower-carbohydrate groups lost slightly more fat, a trend that was not significant (SMD -0.23, p = 0.09)1. Where carbohydrate looks helpful for lean mass, some of it may be extra food or water, not extra muscle.
The authors' own mechanism line: "Mechanical tension is the primary stimulus for muscle hypertrophy, not acute metabolic stress or substrate availability"1. That is their reading of the wider literature, not something these 11 trials tested.
In plain terms: some of what looks like muscle on a scan is water that follows carbohydrate.
Where carbohydrate does show up: training performance
"Do carbs help me train harder?" is a different question, and the 2026 review did not measure it1.
A 2022 meta-analysis did. It pooled crossover studies, where each person lifts once with carbohydrate and once with a placebo or water: 21 studies, 226 participants2. Carbohydrate raised total training volume in the session, mostly measured as repetitions done to failure, with an SMD of 0.61 (95% CI 0.11 to 1.11), pooled from the 12 studies that reported it. The effect depended on the setting:
- Sessions longer than 45 minutes: SMD 1.02 (0.07 to 1.97). Shorter sessions: 0.23 (-0.21 to 0.67), not significant2.
- After a fast of 8 hours or more: 0.39 (0.06 to 0.72). After a shorter fast: 0.76 (-0.19 to 1.71), not significant, with high spread between studies2.
Training volume in one session, carbohydrate vs placebo or water · SMD
A 2022 systematic review of 49 studies reached a similar line: acutely, higher carbohydrate did not help performance in 13 studies and helped in 6, mainly against fasted control groups and in workouts of more than 10 sets per muscle group3. Over longer diet-and-training studies, performance changes were not influenced by carbohydrate in 15 of 173.
Whether more volume in single sessions adds up to more muscle over months is plausible but was not what these studies measured. For how carbohydrate fits among the other macros, see how many carbs per day and macro splits compared.
In plain terms: carbohydrate can help you do more work in long or fasted sessions; that is not the same as proof it grows more muscle.
What this means in practice
The trials that found no difference kept two things equal: protein and, in the strictest ones, total energy. Their authors' practical line is that "carbohydrate intake may be selected based on personal preferences or criteria other than muscle growth, provided that energy and protein requirements are met"1. They add that a higher-carbohydrate diet may still suit someone who finds it easier to eat enough on it.
So the two numbers that carried the result in these trials were protein (see how much protein per kg) and total energy. Carbohydrate was the fuel mix around them. If you are curious what a very-low-carbohydrate split looks like in numbers, keto macros explained lays it out.
The studies at a glance
| Study | Who was studied | Key result | Year |
|---|---|---|---|
| Ashtary-Larky et al., *Crit Rev Food Sci Nutr*4 | 13 RCTs, 244 people doing resistance training | Keto diets: fat-free mass 1.26 kg lower on average than non-keto (protein and energy not required to match) | 2022 |
| Henselmans et al., *Nutrients*3 | 49 studies, mostly strength trainees | Acute carbohydrate helped performance in 6 of 19 studies, mainly vs fasted controls or over 10 sets per muscle | 2022 |
| King et al., *Sports Med*2 | 21 crossover studies, 226 people | Carbohydrate raised session training volume, SMD 0.61; larger in sessions over 45 min | 2022 |
| Henselmans, Vårvik, Izquierdo, *Sports Med*1 | 11 RCTs, lifters, protein matched, 6 to 12 weeks | No significant effect on muscle growth, SMD 0.15 (-0.10 to 0.40); certainty low | 2026 |
Limits worth knowing
- Small and short. About 21 people per trial and 6 to 12 weeks1. A small effect would be hard to detect.
- Mostly keto vs usual diet. Eight of 11 trials compared near-zero carbohydrate with a usual diet. Moderate changes in carbohydrate were barely tested.
- Mostly young men. Only one trial was women only, and older adults were not studied.
- Energy was often not controlled. Most trials let people eat freely, and even one "isocaloric" trial may have had the higher-carbohydrate group eating more, by the authors' own calculation1.
- Only 2 trials imaged the muscle. The rest used whole-body lean mass, which counts water.
- Certainty low. GRADE (a standard system for grading how far to trust a result) rated it 2 of 4, for imprecision and a moderate risk of bias1.
- Abstract only for the 2022 keto meta-analysis and the 2022 performance review; we read the 2026 review and the King 2022 meta-analysis in full.
- Funding. The 2026 review reports no funding1.
Latest research
The newest study here is the 2026 review itself, published online in February 2026, the first meta-analysis to isolate carbohydrate with protein matched1. Its authors conclude the data are "most consistent with a null effect or a small positive effect", and call for longer trials that image the muscle directly and control energy tightly. We found no newer trial that changes the picture. When one appears, this page will be updated and its date will change.
The short version
- In 11 trials with protein matched, more carbohydrate did not significantly change muscle growth (SMD 0.15).
- The two trials that imaged the muscle leaned slightly the other way; whole-body scans count glycogen water.
- Carbohydrate can raise training volume in long sessions or after long fasts; that is a performance result, not a growth result.
- Certainty is low, so a small benefit is not ruled out.
Educational overview only — not medical advice.
