Gene Variant Tied to Body Fat and High-Fat Diets in Women with Prediabetes

A single-letter change in a gene that helps regulate fat tissue appears to travel with different eating habits and different patterns of body fat in women with prediabetes, according to a cross-sectional study of 202 women in Tabriz, Iran. The variant, known as Val109Asp in the ITLN1 gene, encodes omentin-1, an adipocytokine secreted by adipose tissue that plays a role in the pathogenesis of insulin resistance and diabetes. Researchers led by Roghayeh Molani-Gol and Maryam Rafraf at Tabriz University of Medical Sciences report in the journal Genes & Nutrition that carriers of one version of the variant were less likely to follow a high-fat dietary pattern, while other versions tracked with lower body weight but larger waists. The work was published on 16 May 2026 as an open access article.

Prediabetes is the stage before type 2 diabetes mellitus, a high-risk state characterized by elevated glucose levels that do not yet reach diabetes criteria. The American Diabetes Association defines it by fasting blood sugar levels of 100–125 mg/dL, glycated hemoglobin of 5.7–6.4%, or glucose tolerance test values of 140–199 mg/dL two hours after ingesting a standardized 75 g glucose solution. The prevalence of prediabetes is increasing around the world. In 2021, the global prevalence of impaired fasting glucose was 5.8%, or 298 million people, and impaired glucose tolerance affected 9.1%, or 464 million. By 2045, those figures are expected to reach 6.5%, or 414 million, and 10.0%, or 638 million, respectively. In Iran, 18.2% of males and 15.4% of females have been reported to have prediabetes. Because the complications of type 2 diabetes and the financial burden of managing it are substantial, prevention and control of the disease have become a global concern.

The gene encoding omentin-1 sits on the first chromosome, at position 1q23.3 in the GRCh38 assembly, and carries a single-nucleotide variant, rs2274907, at the fourth exon, where adenine is replaced with thymine. This missense change, described as c.326T>A, results in the replacement of valine with aspartic acid at codon 109 of the omentin-1 protein. Serum omentin-1 levels have been reported to be lower in subjects with obesity, impaired glucose regulation, and type 2 diabetes than in the normal glucose tolerance group. Earlier studies of the same polymorphism in people with type 2 diabetes found a significant association with increased insulin resistance, and work in a Central-European population suggested that the interaction between the variant and body mass index can decrease insulin sensitivity. That earlier study also reported a connection between the polymorphism and daily intake of energy, fat, and protein. No previous study, the authors note, had assessed the association of the ITLN1 polymorphism with dietary patterns as such.

The study recruited women aged 18 to 65 years who were diagnosed with prediabetes according to the ADA fasting glucose criterion and who attended a healthcare center in Tabriz. Participants were selected by convenience sampling. Women with severe obesity, defined as a body mass index of 40 kg/m2 or higher, or wasting, below 18.5 kg/m2, were excluded, as were those who were pregnant or lactating, smoked or had alcohol addiction, had renal failure or cardiovascular, thyroid, liver, or kidney disease, cancer, chronic neurological disease, Cushing’s syndrome, or rheumatoid arthritis, had surgery in the last three months, used vitamin or nutritional supplements currently or in the last three months, followed weight loss diets, or took drugs such as glycemic controllers or lipid-lowering agents. The sample size was calculated with Stata 18 software based on the allele frequency of the variant in an Iranian population, with 80% power, an alpha of 0.05, and about 15% loss to follow-up accounted for, yielding a target of 204 participants, a number that also meets the minimum threshold recommended for robust pattern extraction in exploratory factor analysis.

Each participant completed a 168-item food frequency questionnaire that had been validated previously for the Iranian population, and the responses were analyzed with Nutritionist-4 software to estimate mean daily caloric intake. To identify dietary patterns, the researchers divided the 168 food items into 30 food groups and applied exploratory factor analysis in JASP software, using the minimum residual extraction method with Promax rotation. The number of factors was chosen based on parallel analysis and the change point in the scree plot curve, and food groups with a factor loading of 0.2 or higher were considered contributors to each pattern. The Kaiser–Meyer–Olkin index was 0.57, indicating the data were sufficiently adequate, and Bartlett’s test was significant, showing sufficient correlation between the food groups. The cumulative predictive value of the obtained dietary patterns was 15%. Three dominant patterns emerged: a vegetarian pattern led by vegetables, tomatoes, and nuts; a high-fat pattern characterized by mayonnaise, snacks, and unsaturated oils; and a mixed pattern anchored by refined grains, legumes, and salt.

Genotyping was performed on blood drawn after 8 to 12 hours of fasting using the high-resolution melting-polymerase chain reaction method, in which genotypes are distinguished by differences in the melting curve after DNA extraction and amplification with specific primers. A 10% random sample was re-genotyped for quality control, and the genotype distribution was consistent with Hardy–Weinberg equilibrium. Among the 202 women, 15.8% carried the AA genotype, 52.0% the AT genotype, and 32.2% the TT genotype, with a minor allele frequency of 41.68%. The median age of participants was 49.50 years, 98.5% were married, and 97.0% were housewives. A first-degree family history of diabetes was reported by 51%. The median body mass index was 31.09 kg/m2, and 4% of the women were of normal weight, 32.7% overweight, and 63.4% had obesity.

The central results come from adjusted statistical models. In an ordinal regression adjusted for age, body mass index, and dietary energy intake, participants with the AT genotype had a lower chance of having a high-fat dietary pattern compared with the AA genotype, with an adjusted odds ratio of 0.466 (95% confidence interval 0.219 to 0.989, p = 0.047). In other words, the AA genotype was the version more often linked with the energy-dense pattern rich in mayonnaise, snacks, unsaturated oils, and full-fat dairy products. The analysis of covariance, adjusted for the same confounders, showed that women with the AT genotype had a decreased body weight compared with the AA genotype, with a mean difference of -3.073 kg (95% confidence interval -5.245 to -0.901, p = 0.007), and the TT genotype showed a similar reduction of -3.188 kg (95% confidence interval -5.513 to -0.862, p = 0.006).

Central adiposity told a different story. Women in the TT genotype group had a higher waist circumference than those with the AA genotype, with a mean difference of 2.557 cm (95% confidence interval 0.060 to 5.055, p = 0.045) in the adjusted model. Both the AT and TT genotypes had higher waist-to-height ratios than the AA genotype, with mean differences of 0.028 (95% confidence interval 0.011 to 0.044, p = 0.001) and 0.022 (95% confidence interval 0.005 to 0.040, p = 0.013), respectively. The authors suggest that this pattern of lower body weight but greater central adiposity among certain genotypes may reflect differential fat distribution rather than total adiposity, since visceral adipose tissue is metabolically distinct and more strongly associated with insulin resistance. A previous study in adults with abdominal obesity similarly reported that the TT genotype increased the risk of abdominal obesity, whereas the AA genotype was not associated with the condition. The researchers also note that hormonal differences, particularly estrogen-related fat deposition, may contribute to the divergent pattern observed in these women.

Glycemic factors did not hold up under adjustment. In the crude analysis, participants with the TT genotype had increased serum fasting blood sugar levels compared with the AA genotype, with a mean difference of 3.305 mg/dL (95% confidence interval 0.098 to 6.512, p = 0.043), but this result was non-significant after adjusting for age, body mass index, and dietary energy intake. No significant association between the ITLN1 polymorphism genotypes and glycemic parameters was found in the adjusted model. The authors caution that the rs2274907 variant is a missense substitution, and although direct functional studies are limited, current evidence does not conclusively demonstrate functional alteration, so the variant may act as a marker of adipokine-related metabolic pathways rather than a proven driver. They also acknowledge that glycemic regulation and adiposity are complex, polygenic traits, and that focusing on a single variant limits the ability to capture the broader genetic architecture underlying prediabetes.

The study has clear limits that temper the findings. Its cross-sectional design cannot infer causality, so the results describe associations between genotype, diet, and body measurements rather than effects. Subgroup analyses by genotype may have been underpowered due to allele distribution, and the food frequency questionnaire relies on respondents’ memory, with overestimating or underestimating of food intake possible; adults with obesity commonly underreport their dietary intake. The authors argue that identifying such gene–diet interactions could eventually support personalized dietary recommendations for individuals at high risk, and they call for further well-designed studies in larger and more diverse populations to confirm the findings and to clarify the association of the ITLN1 polymorphism with glycemic indices. For now, the takeaway is narrower: in this group of Iranian women with prediabetes, one variant in an adipokine gene tracked with both eating patterns and the distribution of body fat, while leaving blood sugar measures essentially unexplained.

Subject of Research: Association of the ITLN1 Val109Asp gene polymorphism with dietary patterns, glycemic factors, and anthropometric indices in women with prediabetes

Article Title: Association of the ITLN1 gene polymorphism with dietary patterns, glycemic factors, and anthropometric indices in women with prediabetes

Article References: Molani-Gol, R., Aftabi-Yousefabad, S., Azarsina, G., Shanehbandi, D., Malekian, M., Asghari Jafarabadi, M., & Rafraf, M. (2026). Association of the ITLN1 gene polymorphism with dietary patterns, glycemic factors, and anthropometric indices in women with prediabetes. Genes & Nutrition, 21(1), Article 16. https://doi.org/10.1186/s12263-026-00802-2

Image Credits: AI Generated

DOI: 10.1186/s12263-026-00802-2

Keywords: prediabetes, ITLN1, omentin-1, genetic polymorphism, dietary patterns, central obesity, insulin resistance, nutritional genomics, cross-sectional study, fasting blood sugar, waist circumference, Iran