Genotype‑Informed Nutrition in Obesity/Prediabetes

The systematic review Personalised Nutrition in Obesity and Prediabetes describes evidence for genotype-guided dietary response in adults with obesity or prediabetes as mixed across loci, endpoints, and study designs. The authors outline a landscape in which some gene–diet interactions appear context-dependent and difficult to replicate, alongside more consistent themes that do not rely on genetic stratification. In their synthesis, broad diet-pattern frameworks and diet-quality constructs sit beside examples in which single variants or polygenic burden are reported to modify associations with adiposity, glycaemic indices, lipids, or inflammatory markers. A recurring contrast is between genotype-agnostic diet-pattern signals and more specific, locus- and polygenic-dependent effect modification.

Across the included human studies, dietary exposures spanned Mediterranean- and DASH-type patterns, macronutrient distribution (including carbohydrate and protein proportions), fat quality, and energy restriction. Outcomes were mapped to anthropometrics, glycaemic measures, lipid traits, and inflammatory/oxidative stress biomarkers in adults described as having obesity, prediabetes, or related cardiometabolic risk. Within that scope, the review describes “baseline diet-quality” targets with cross-study support—particularly restricting saturated fat and preserving carbohydrate quality—and notes in its conclusions an emphasis on replacing saturated fats with monounsaturated and polyunsaturated fats. The authors present these recurring targets as a reproducible substrate onto which more heterogeneous gene–diet interaction findings are layered, forming a genotype-agnostic foundation in their summary.

On aggregate genetic susceptibility, the authors describe a polygenic “risk-amplification” pattern in which associations between diet patterns and metabolic risk differed by polygenic-risk strata. As summarized in the review, healthy dietary patterns were associated with lower risk in higher-risk strata (reported odds ratio about 0.53), whereas unhealthy patterns were associated with higher risk in higher-risk strata (reported odds ratio about 3.69), with associations described as null or attenuated in lower-risk groups. These examples are framed as potential stratification effects rather than uniform diet–risk relationships across genetic backgrounds. In this account, polygenic burden serves as a recurring illustration of effect modification.

Single-variant findings are presented as more granular and often linked to specific dietary thresholds or biomarker directions, with TCF7L2 and APOA2 used as prominent illustrations. For TCF7L2, the authors describe macronutrient “threshold” concepts, including patterns such as protein intake above 18% of energy and carbohydrate intake below 48% of energy, and link these allocations to differences in visceral adiposity and glycaemic indices in the contexts they summarize. For APOA2 −265T>C (rs5082), the review reports genotype-dependent inflammatory and oxidative-stress biomarker associations, including a paradoxical direction in which CC homozygotes showed higher inflammatory/lipid peroxidation markers in association with higher dietary antioxidant capacity, contrasted with more expected anti-inflammatory directions in T-allele carriers in the cited example. Additional loci (including FTO, MC4R, and PPARG) are discussed as further instances of heterogeneous, exposure-dependent signals across outcomes, which the authors portray as locus-specific rather than universally generalizable.

The authors also describe translation constraints that, in their view, bound interpretation across the evidence base: interaction testing that is often underpowered, inconsistent approaches to multiplicity, and uneven ancestry representation across cohorts and trials. They also note transferability concerns for genetic findings, including ancestry-specific signals and the difficulty of applying polygenic scores across populations without local calibration, and they frame locally trained scores as relevant to this limitation. As forward-looking priorities, the review states that future progress requires preregistered, adequately powered genotype-stratified trials and polygenic scores trained and evaluated within the populations in which they would be applied. The synthesis concludes that signals are suggestive in selected settings but constrained by design and population limits.

Key Takeaways:

• The review reports that the most reproducible baseline themes were restricting saturated fat, emphasizing monounsaturated/polyunsaturated fats, and preserving carbohydrate quality, largely independent of genotype.
• Polygenic analyses are summarized as showing a “risk-amplification” association pattern, with diet-pattern links appearing stronger in higher polygenic-risk strata than in lower-risk strata.
• Locus-specific signals (including TCF7L2 and APOA2) are presented alongside ancestry/transferability limitations, with the authors calling for preregistered genotype-stratified trials and locally trained polygenic scores.