Finnish researchers went looking for the people whose genes should make dieting hardest, put them on a diet, and found they did just as well as everyone else. The trial, published in Nature Communications, is a rare thing: a weight-loss study designed from the genome outwards rather than recruited from a clinic waiting list.
How they built it. Starting from 38,621 genotyped Finns, the team calculated a polygenic score for BMI — a single number summing up thousands of common genetic variants associated with body weight — and contacted people at the two extremes: the top 5% and the bottom 5%. A total of 223 overweight or mildly obese adults without diabetes enrolled in the GENEROOS trial (mean age 51.1 years, 74.9% women, mean BMI 30.3 kg/m²) and were randomised to six months of structured dietary coaching or to a control group.
The genes clearly do something. Before anyone changed anything, the top 5% carried a BMI 3.0 kg/m² higher than the bottom 5% — for someone of average height, that is roughly 9 kilos of difference sitting there at baseline, attributable to inherited variation.
And then the diet worked the same for both groups. After six months, the coaching arm dropped 1.51 BMI units (95% CI −1.72 to −1.31) against essentially nothing in the control arm (−0.01, 95% CI −0.22 to 0.19). The test that mattered — whether the polygenic score changed how much the diet delivered — came out flat: an interaction of β = 0.06 (95% CI −1.33 to 1.44), P = 0.94. Not a hint of a difference in either direction.
This is the useful way to think about genetic risk. Your genes appear to set the starting line rather than the slope. High polygenic risk meant a heavier body at the outset; it did not mean the diet worked less well once it started. That distinction matters, because "it's genetic" is one of the most common reasons people give for not bothering — and at least for six months of structured coaching, the data say the effort pays off the same either way.
Where the caution belongs. The researchers are careful about what they have not shown. With 223 participants, a modest interaction effect could still be hiding in the noise; only a much larger trial could rule that out. Six months is short, and weight regain is the usual pattern after that point in almost every diet trial ever run. This was a Finnish population, and polygenic scores travel poorly across ancestries — a score built mainly on European genomes tends to perform worse elsewhere. And the finding applies to a polygenic score, not to single rare mutations that cause severe obesity, which are a different medical situation entirely.
Disclosures. The senior author is the founder of a genetics company, Real World Genetics Oy, and one co-author is an employee of Eli Lilly, which markets weight-loss drugs. Neither tie points in an obvious direction here — if anything, a null result for genetic stratification is the inconvenient one for a company selling genetic insight — but both belong on the table. The trial was registered in advance (NCT06092372).
Medical note. This article provides general education and information about health and related subjects. The information is not medical advice and should not be used to diagnose or treat health problems or conditions. Always consult with a qualified healthcare professional for medical advice and treatment.