The diet comparison that dominates the internet is not low-carb against low-fat. It is low-carb against eating as much as you like, with low-fat on a shelf somewhere. The trials that actually test the two against each other produce a much flatter document, and it is worth reading it in full.
What the meta-analysis actually compared
Kazeminasab and colleagues, in the European Journal of Clinical Nutrition 2025, set a primary endpoint that almost no diet article uses: flow-mediated dilation. PubMed, Scopus and Web of Science searched through May 2023, weighted mean differences computed for low-carbohydrate diets against low-fat diets, both under hypocaloric conditions. Ten studies reporting flow-mediated dilation, 475 participants.
The design point matters more than anything else in the paper. Both arms were hypocaloric. The question was never whether a calorie deficit works. The question was whether the macronutrient split changes anything once the deficit is already there.
The result, in the order the authors put it
Flow-mediated dilation overall: −1.04 per cent (95% CI −2.28 to 0.20, p = 0.10) — not significant. For very-low-carbohydrate diets specifically, −2.12 per cent (95% CI −3.35 to −0.9, p = 0.001) — significantly reduced.
Read the sign. Reduced flow-mediated dilation means worse endothelial function, which is the opposite of what “healthy fat” marketing implies. It is a small number and a short intervention, and the authors report it without embellishment.
Anthropometric and body composition measures: no change. C-reactive protein: no change. Blood pressure: no change. HDL: no change. Total cholesterol: no change.
Triglycerides: significantly decreased, weighted mean difference −19.94 milligrams per decilitre (95% CI −31.83 to −8.06, p = 0.001). LDL: significantly increased, +20.00 milligrams per decilitre (95% CI 14.09 to 25.90, p = 0.001).
So the honest summary in the authors’ own words: low-carbohydrate diets did not exert superior effects on cardiovascular markers or body composition when compared with low-fat diets in individuals with overweight or obesity, but low-carbohydrate diets reduced triglycerides and increased LDL levels more than low-fat diets.
That is a trade, not a win. Anyone quoting one half of it without the other half is doing the thing this article exists to describe.
The comparison that is missing when a different comparison is present
The 2025 BMJ network meta-analysis of 99 randomised trials with 6,582 adults is in this article for its structure rather than its subject. All intermittent fasting and continuous energy restriction strategies reduced body weight compared with an ad libitum diet. Compared with continuous energy restriction — diet against diet — only alternate day fasting showed a benefit in body weight reduction, mean difference −1.29 kg (95% CI −1.99 to −0.59), moderate certainty.
Which is 1.29 kilograms. In 6,582 people.
Two meta-analyses, different outcomes, same lesson: diet-versus-diet differences are small and often trivial; diet-versus-unrestricted differences are large and belong to the restriction, not to the macronutrients.
When a head-to-head does find something
The counter-example is worth including, because a blog that only reports unfavourable comparisons is as useless as one that only reports favourable ones. Ghaedi and colleagues’ 2019 meta-analysis in Advances in Nutrition covered the Paleolithic diet across eight eligible randomized controlled trials and found reductions in body weight, waist circumference, BMI, body fat percentage, systolic and diastolic blood pressure, and C-reactive protein, with an increase in HDL.
Then the sentence that makes this the right paper: sensitivity analysis revealed that the overall effects on lipid profile, blood pressure and CRP concentrations were significantly influenced by removing some studies, hence the results must be interpreted with caution. The authors state that the evidence is not conclusive and that more well-designed trials are needed.
The headline measurements and the robustness check, in the same abstract. The robustness check is the one to read.
How to use this when you read a diet page
Three questions, in order. Were both arms calorie-restricted — or was the comparison diet against unlimited eating? How many people were randomised? And what happened when the weaker studies were removed?
The first question kills most claims. The second sizes the claim. The third is the only one that tells you whether the claim will still be there next year.
What this evidence does not show
It does not show that low-carb is equivalent to low-fat for every person: 475 participants across ten trials is a real but modest evidence base, and individual responses to a diet are not what a meta-analysis measures. The flow-mediated dilation finding is a vascular endpoint in short interventions and should not be read as a cardiovascular event rate. And nothing here speaks to adherence, which is the variable that decides which diet a person actually ends up on.
This article summarises published meta-analyses of randomized controlled trials. It is not medical advice and it is not a recommendation for any individual.
