Two people with obesity can be struggling for very different reasons. One never quite feels full during a meal, another is hungry again soon after eating, a third burns energy slowly, a fourth eats in response to stress or mood. A research group at the Mayo Clinic has been testing whether lifestyle advice matched to the reason works better than the same advice for everyone. A new analysis in Nutrients looks at what the participants actually ate.
The trial behind it. The intervention comes from the PHENO-Diet trial, a 12-week proof-of-concept study at the Mayo Clinic. It was not randomized. Of 211 adults screened, 165 were assigned according to their enrollment date. One team delivered a standard lifestyle program first (August 2020 to February 2021) and a phenotype-tailored program afterward (March 2021 to February 2022). The tailored program was associated with greater weight loss, which left one question open: did the two groups eat differently?
Who is in this analysis. The original trial had 165 participants, but only 77 of them, about 47%, are in this analysis: those who completed a food recall at baseline and at least one more during the program. Of the 77, 35 were in the tailored group and 42 in the standard group. Not everyone kept reporting, so at week 12 the comparison rests on 56 people with a food record, 27 in the tailored group and 29 in the standard group. That matters for how far the results can be trusted, because the people who stopped reporting may differ from those who continued. Each participant was classified into one of four energy-balance phenotypes: abnormal satiation, reduced postprandial satiety (hunger returning soon after a meal), low energy expenditure, or emotional eating.
How eating was measured. Participants completed three paper food records, each a 24-hour recall: one at the start, one sometime between weeks 4 and 8, and one at week 12. Each record covers a single day, so it is a snapshot of one day's eating, not of a typical week. The figures below are model-adjusted averages across those records, with the headline comparison made at week 12. The records were entered into a nutrition database and checked by a dietitian who did not know which group each person was in. That check covered completeness and consistency, not whether the reported intake was biologically plausible.
What the tailored group reported eating. After adjusting for baseline values, the tailored group reported lower total energy intake at 12 weeks: 1273 ± 182 versus 1894 ± 176 kcal per day (p = 0.018; adjusted averages ± standard error). The gap came from carbohydrate (464 ± 76 versus 818 ± 73 kcal per day, p = 0.002) and fat (479 ± 76 versus 697 ± 73 kcal per day, p = 0.036). Absolute protein intake did not differ, but because the tailored group ate less overall, protein made up a larger share of its calories: 26.8 ± 1.6% versus 19.8 ± 1.5% (p = 0.003). Dairy protein supplied 6.8 ± 0.8% of calories in the tailored group and 3.7 ± 0.7% in the standard group (p = 0.004). Total calories were a pre-specified secondary goal of the parent trial, and the calorie difference held up in the authors' sensitivity checks. The breakdown by carbohydrate, fat, protein and dairy was not pre-specified, and the authors call it exploratory.
Who paid, and who has an interest. Readers should know about two connections. First, four of the authors (Andres Acosta, Michael Camilleri, Matthew Clark and Donald Hensrud) and the Mayo Clinic are named in the paper's declaration as inventors of intellectual property licensed to Phenomix Sciences Inc. By its own description, Phenomix is a company working in precision obesity medicine, and it sells a test that identifies a patient's obesity phenotype. This study evaluates exactly that kind of phenotype-based approach to obesity care. The paper does not say which property is licensed. Another co-author receives research funding from Phenomix Sciences, among other sources, and some authors report consulting or advisory work for pharmaceutical and food companies.
Second, the study was funded by Dairy Management Inc., a U.S. dairy-industry organization, which is worth knowing because one of the highlighted findings concerns dairy protein. The authors state that the funder had no role in the design, analysis, interpretation, writing or decision to publish.
None of this means the results are wrong, but it is context a reader should be able to weigh.
Why this is a signal, not proof. The authors say it directly: the non-randomized design does not allow causal inferences. Because the groups were recruited in two separate periods, anything that changed between them could have contributed. The team also delivered the standard program first, so its counseling may simply have improved by the time the tailored program began, and that cannot be separated from the tailoring itself. Add the small week-12 groups, the fact that more than half of the trial's participants did not contribute dietary data, self-reported one-day recalls, and a follow-up of only 12 weeks. A difference of roughly 620 kcal per day is large, which is one more reason to wait for a randomized trial before reading too much into it.
What it adds. The pattern fits the idea behind phenotype-based advice, which steers people toward strategies suited to their own drivers of overeating rather than a single macronutrient rule. The authors describe a potential role for personalized dietary approaches in obesity management. For now, that is a hypothesis worth testing, not a recommendation to change how you eat.
Medical note. This article is general education, not personal medical advice. Talk to a qualified clinician before changing treatment, medication, or supplement routines.