Walnuts in place of part of the starch at every meal: deep belly fat fell from 113 to 97 cm² in 12 weeks. At the finish, the walnut group and the comparison group still could not be told apart

Fat around the middle is not all the same. The fat packed around the organs, called visceral fat, is tied more closely to diabetes and heart disease than fat on the hips and legs. So a diet that moved fat away from the trunk, or stopped it settling there, would be worth having even if the scales barely changed. That is the idea a team at the First Affiliated Hospital of Soochow University set out to test, and their results are in the October 2026 issue of the Asia Pacific Journal of Clinical Nutrition.

They recruited 80 adults aged 18 to 60 with central obesity by the Chinese definition: a waist of at least 90 cm for men or 85 cm for women, and a body mass index between 24 and 32.5. Everyone was put on the same reduced-calorie plan drawn from China's 2021 obesity guidelines: 500 g of vegetables a day, 200 g of fruit, 100 g of staple food at each meal, five protein foods, 30 g of oat bran, three spoons of oil. Half were then told to take 50 g of staple out of each meal and eat walnuts in its place. That turned a diet of roughly 57% carbohydrate and 24% fat into one of about 39% carbohydrate and 42% fat.

Twelve weeks later the walnut group's visceral fat area had dropped from 113 to 97.4 cm², taking the average below the 100 cm² line used in Asia to mark visceral obesity. Their BMI fell from 27.6 to 26.1. In the comparison group visceral fat went from 116 to 109 cm² and BMI from 27.3 to 27.0. The ratio of trunk fat to leg fat, the measure of where fat sits, edged down from 1.80 to 1.74 with walnuts and stayed put, 1.82 to 1.81, without.

The statistics need reading slowly. The test that asks whether one group changed more than the other over time came out in the walnut group's favour on all three measures. But when the two groups were compared directly at week 12, none of the three differences was statistically significant. The paper says so in its results, and then concludes that the walnut diet achieves 'significantly greater and faster fat reduction'. Greater, by the first test. Faster is hard to support: body composition was measured twice, at the start and at the end.

Then there is the calorie puzzle. By their own three-day food diaries, the walnut group ended up eating 1,162 calories a day and the comparison group 1,211, down by about 580 and 680 calories from where they started. On paper the comparison group cut slightly more. Yet its BMI moved by 0.3 while the walnut group's moved by 1.5. Either the walnut diet did something that calories don't explain, or the diaries are not telling the whole story. The authors list self-reporting bias among their limitations, and say they did not strictly hold calories equal between the groups.

Other things to weigh. Sixteen of the 80 people did not finish, and they were not spread evenly: 5 from the walnut group and 11 from the comparison group, so the main results are from 35 against 29. Fat was estimated by bioelectrical impedance, a machine you stand on and grip, not by a scan. The walnuts were free for two weeks and after that the participants paid for them, because the grant ran short. And the gut-hormone results the paper leans on to explain its findings, higher GLP-1 and PYY and lower C-peptide in the walnut group, come from a single blood sample at week 12. Nobody measured them at the start, so there is no way to say the diet changed them.

Our reading: this is a small, honest-looking trial with an oversold last sentence. It cannot separate the walnuts from the low-carb. The walnut group ate 63 g less carbohydrate a day than the comparison group by the end, and any food that pushed the starch off the plate might have done the same. What it does add is modest and useful. Swapping some starch for a measured portion of unsalted nuts, inside a diet that is already cutting calories, did not slow fat loss and may have helped it. The measured portion is the part to keep. Walnuts are dense in calories, and eating them on top of a normal diet is a different experiment from the one run here.