03/07/2026
Two people can carry the same weight, the same insulin resistance, and the same fasting glucose, and still find that the diet which most improves their metabolic health runs in opposite directions, because what separates them is not how resistant they are but which organ the resistance sits in.
A Dutch team ran what precision nutrition rarely gets: a randomized controlled trial. They took 242 adults with overweight or obesity, measured whether each person's insulin resistance was concentrated in muscle or in the liver using an oral glucose tolerance test, then assigned them for twelve weeks to one of two diets, a high-monounsaturated-fat diet or a low-fat diet high in protein and fiber, arranged so that one group's pairing was the mirror image of the other's. The pre-registered primary outcome, a measure of how well the pancreas compensates for insulin resistance, did not improve in either arrangement.
The arrangement in which the muscle-insulin-resistant people ate the low-fat, high-protein, high-fiber diet and the liver-insulin-resistant people ate the high-monounsaturated-fat diet produced greater improvements in insulin sensitivity, blood glucose control, triglycerides, and C-reactive protein than the opposite arrangement. Same two diets, same twelve weeks, the same degree of insulin resistance at baseline. The only thing that differed was whether each person's diet was matched to the tissue driving their resistance, and that matching alone separated the two groups on markers that track long-term cardiometabolic risk.
Insulin resistance is usually treated as a single condition, but it is not distributed evenly. In some people the defect sits mainly in skeletal muscle, where insulin normally drives glucose out of the blood and into the tissue; in others it sits mainly in the liver, where insulin normally restrains glucose output and fat production. These can be measurably different states, separable by the pattern of glucose and insulin across an oral glucose tolerance test, and recent work shows they carry distinct plasma protein signatures rather than being two names for the same problem. Because muscle and liver handle carbohydrate and fat differently, it is biologically reasonable that the macronutrient mix which helps one is not the mix that helps the other, though this trial demonstrates the pattern more clearly than it explains the precise reason.
Several things keep this a proof of concept:
The pre-registered primary endpoint was null, so the positive results all come from secondary and exploratory analyses, which carry less weight and a higher chance of a false positive. It is a single twelve-week trial in adults with a body mass index between 25 and 40, so it says nothing about lean or older people, or about effects that take longer than three months to surface. Several of the authors are employed by nutrition companies, including a dairy company and companies that sell the kinds of fats and proteins the diets emphasized, which does not invalidate the work but is worth naming. And the phenotyping depends on an oral glucose tolerance test with timed blood draws, which is not something most people or most clinics do, so the path from this finding to an individual's plate is not yet built.
What the evidence supports is that where insulin resistance sits, in muscle or in the liver, appears to change which macronutrient pattern improves a person's cardiometabolic markers, and that treating insulin resistance as one uniform condition may be leaving some of that improvement unclaimed. What it does not support is any specific personalized-nutrition product or test sold on the promise of a diet tailored to your biology, because the primary outcome was null, the effect lives in secondary markers, and no one has yet shown that this matching changes hard outcomes like diabetes or cardiovascular events. The practical reframe is still worth stating: the personalized-nutrition industry is built on the premise that your biology should set your diet, and this is one of the few randomized trials to test a real, measurable version of that premise rather than a genetic or microbiome readout of uncertain meaning. The open question is whether matching diet to insulin-resistance phenotype changes anything a person can feel or that lengthens a life, or whether it is a real but small metabolic signal that will never justify the testing required to act on it.
Trouwborst et al., Cell Metabolism 2023
Abdul-Ghani et al., Diabetes Care 2007
Kemper et al., J Clin Endocrinol Metab 2026